United States
Environmental Protection
Agency
Office of Research and
Development
Washington, DC 20460
EPA/625/R-99/002
July 1999
f/EPA
National Conference on
Retrofit Opportunities for
Water Resource Protection in
Urban Environments
Proceedings
Chicago, IL
February 9-12, 1998
fen*
-------
EPA/625/R-99/002
July 1999
National Conference on Retrofit
Opportunities for Water Resource
Protection in Urban Environments
Proceedings
Chicago, IL
February 9-12, 1998
Technology Transfer and Support Division
National Risk Management Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
Notice
Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
-------
Introduction
Water resource managers have been successful in developing approaches for reducing
nonpoint source pollution in newly developing urban areas. Issues become increasingly com-
plex, however, when managers are faced with the challenge of reducing nonpoint source im-
pacts within previously developed urban environments. A diverse assortment of resource man-
agement tools, or "retrofits," is being developed, but their implementation has been hampered
by a lack of technology transfer opportunities. The National Conference on Retrofit Opportu-
nities for Water Resource Protection in Urban Environments was designed to address
these issues and to transfer much-needed information to state and local water resource practi-
tioners.
Held in Chicago, Illinois, on February 9-12,1998, the conference program brought together
an array of progressive scientists and researchers, along with managers of successful local
retrofit projects from across the country. Session topics included retrofit opportunity identifica-
tion, modeling and monitoring approaches for retrofit applications, conservation design strate-
gies, innovative financing approaches, evaluating results and measuring success, newly emerging
technologies, urban revitalization issues, riparian reforestation, and public education and in-
volvement programs.
During the conference, a series of speakers presented papers, 43 of which are reproduced
in these proceedings. The purpose of this document is to present these papers and provide
information to individuals unable to attend. All papers included were peer reviewed. This docu-
ment will be useful to individuals who are interested in information about retrofitting techniques
and approaches to improving protection of urban water resources. A list of the nearly 300 at-
tendees is provided following the papers.
-------
IV
-------
Contents
Securing the Urban Greenfrastructure: Integrating Stormwater Management with
Regional Growth Management
Michael C. Houck
Urban Naturalist,
Audubon Society of Portland and
Natural Resources Working Group, Coalition Fora Livable Future
Portland, Oregon
The Use of Retention Basins to Mitigate Stormwater Impacts to Aquatic Life 6
John R. Maxted
Delaware Department of Natural Resources and Environmental Control
Dover, Delaware
Earl Shaver
Auckland Regional Council,
Auckland, New Zealand
Assessing the Status of Aquatic Life Designated Uses
in Urban and Suburban Watersheds 16
Chris O. Yoder and Robert J. Miltner
Ohio Environmental Protection Agency
Division of Surface Water Monitoring & Assessment Section
Columbus, Ohio
Dale White
Ohio EPA, Division of Surface Water
Information Resources Management Section
Columbus, Ohio
Tampa Bay Environmental Monitoring Program 29
Robert C. Brown
Environmental Management Department
Manatee County, Florida
Retrofit Opportunities for Urban Waters Using Soil Bioengineering 34
Robbin B. Sotir
Robbin B. Sotir & Associates, Inc.
Marietta, Georgia
Restoration of the Waukegan River Through Biotechnical Means 43
Scott Tom kins
Illinois Environmental Protection Agency
Don Roseboom
Illinois State Water Survey
Illinois Department of Natural Resources
Peoria, Illinois
-------
Contents (continued)
Monitoring the Effectiveness of Urban Retrofit BMPs and Stream Restoration 48
John Galli
Metropolitan Washington Council of Governments
District of Columbia
Urban Water Quality Monitoring and Assessment Approaches in Wisconsin 54
Roger Bannerman
Wisconsin Department of Natural Resources
Madison, Wisconsin
Considerations and Approaches for Monitoring the Effectiveness of Urban BMPs 65
Eric W. Strecker, P.E.
Woodward-Clyde
Portland, Oregon
Targets of Opportunity: Alexandria's Urban Retrofit Program 83
Warren Bell, P.E.
City Engineer, City of Alexandria, Virginia
Philip C. Champagne, P.E.
Dewberry & Davis
Fairfax, Virginia
Port Towns Revitalization and Environmental Enhancement - Stormwater Projects
Revitalize Urban Areas 90
S. AM Abbasi
Prince Georges County Dept. Of Environmental Resources
Largo, Maryland
Tollgate Drain -An Innovative Approach to Stormwater Management 98
John LeFevre and Patrick Lindeman
Fishbeck, Thompson, Carr& Huber
Ada, Michigan
A Stormwater Banking Alternative for Highway Projects 100
Robert B. McCleary, P.E.
Delaware Department of Transportation
Dover, Delaware
Financing Retrofit Projects: The Role of Stormwater Utilities 107
Greg Lindsey
Center for Urban Policy and the Environment
Indiana University, Indiana
Amy Doll
Apogee Research/Hagler Bailly, Inc.
Bethesda, Maryland
Credits as Economic Incentives for On-Site Stormwater Management:
Issues and Examples 113
Amy Doll and Paul F. Scodari
Apogee/Hagler Bailly, Inc.
Bethesda, Maryland
Greg Lindsey
Indiana University, School of Public and Environmental Affairs,
Indianapolis, Indiana
VI
-------
Contents (continued)
Conservation Design for Stormwater Management 118
Earl Shaver
Environmental Engineer
Delaware Department of Natural Resources and Environmental Control
(presently a Technical Specialist, Auckland Regional Council)
Auckland, New Zealand
Results of the Site Planning Roundtable 123
Whitney Brown
Center for Watershed Protection
Ellicott City, Maryland
Retrofitting Conservation Designs into the Developed Landscapes of Northeastern
Illinois 127
Dennis W. Dreher
Northeastern Illinois Planning Commission
Chicago, Illinois
Impacts of On-site Sewage Systems and Illicit Discharges on the Rouge River 132
Barry Johnson, P.E., M.S.
Camp Dresser & McKee
Detroit, Michigan
Dean Tuomari
Wayne County Department of Environment
Detroit, Michigan
Raj Sinha
Wayne County Department of Health
Detroit, Michigan
Stormwater Management in an Environmentally-Sensitive Urban Bushland in
Sydney, Australia 136
Dr Stephen Lees
Executive Officer, Upper Parramatta River Catchment Trust
Sydney, New South Wales, Australia
Can a Steel Plant be Clean? 142
Nigel Ironside
Auckland Regional Council, Auckland, New Zealand
Alistair Atherton
Fletcher Challenge Steel Ltd, Auckland, New Zealand
Real World Modelling: A Case Study of the Silver Lake Watershed Project 154
Randell K. Greer, P.E.
Delaware Dept. of Natural Resources and Environmental Control
Dover, Delaware
Water Quality Modeling to Support the Rouge River Restoration 160
Edward H. Kluitenberg, P.E.
Applied Science, Inc.
Detroit, Michigan
Gary W. Mercer, P.E.
Camp, Dresser and McKee
Detroit, Michigan
Vyto Kaunelis
Wayne County Department of Environment
Detroit, Michigan
VII
-------
Contents (continued)
Overview of Urban Retrofit Opportunities in Florida 166
Michael Bateman, Eric H. Livingston, and John Cox
Stormwater Management Program
Florida Department of Environmental Protection
Tallahassee, Florida
Evaluating the Cost Effectiveness of Retrofitting an Urban Flood Control
Detention Basin for Stormwater Treatment 183
Peter Mangarella
Woodward-Clyde
Oakland, California
David Drury
Santa Clara Valley Water District
San Jose, California
Chee Chow Lee
Environmental Technology Institute,
Nanyang Technical University, Singapore
Richard Mattison
Kinnetic Laboratories Inc.
Santa Cruz, California
Retrofitting to Protect Drinking Water Reservoirs from the Impacts of Urban Runoff 189
James D. Benson and Melissa Beristain
New York City Department of Environmental Protection
Valhalla, New York
Empirical Modeling Approaches for Establishing Nutrient Loading Goals for Tampa Bay .. 196
Anthony Janicki and David Wade
Post, Buckley, Schuh & Jernigan, Inc.
St. Petersburg, Florida
Alum Treatment of Stormwater Runoff-An Innovative BMP for Urban Runoff Problems... 205
Harvey H. Harper, Ph.D., P.E. and Jeffrey L. Herr, P.E.
Environmental Research & Design, Inc.
Orlando, Florida
Eric H. Livingston
Florida Department of Environmental Protection
Tallahassee, Florida
An Eight-Step Approach to Implementing Stormwater Retrofitting 212
Richard A. Claytor, Jr. P.E.
Center for Watershed Protection
Ellicott City, Maryland
Identifying Wetland Restoration Opportunities in the Rouge River Watershed 219
Donald L. Tilton
Tilton & Associates, Inc.
Ann Arbor, Michigan
Taking Root: Sowing and Harvesting the Seeds of Public Involvement and Education 223
Josephine Powell and Noel Mullett
Wayne County Department of Environment
Detroit, Michigan
Zachare Ball
Environmental Technology and Consulting, Inc.
VIM
-------
Contents (continued)
The Tollgate Wetland's Educational Experience 227
Patrick E. Lindemann
Ingham County Drain Commissioner
Mason, Michigan
Minneapolis Chain of Lakes Phosphorus Reduction Strategy 230
Jeffrey Lee
Minneapolis Park and Recreation Board
Minneapolis, Minnesota
Restoration in the Sunshine: Retrofitting the Watersheds of Two Urban
Lakes in Florida 237
Craig W. Dye, Keith V. Kolasa, and K. Lizanne Garcia
Southwest Florida Water Management District
Brooksville, Florida
Retrofit Study for the Lower Neshaminy Creek Watershed 244
George Townsend and Mary Beth Corrigan
Tetra Tech, Inc.
Fairfax, Virginia
Terri Bentley
Bucks County Planning Commission
Doylestown, Pennsylvania
David Athey, P.E.
Tetra Tech, Inc.
Christiana, Delaware
The Stormwater Management StormFilter™ 252
James H. Lenhart, P.E. and Bryan 0. Wigginton
Stormwater Management™
Portland, Oregon
Bioretention: An Efficient, Cost Effective Stormwater Management Practice 259
Larry S. Coffman
Prince Georges County, Department of Environmental Resources
Largo, Maryland
Derek A.Winogradoff
Planning Section, Programs and Planning
Largo, Maryland
Innovative Stormwater Treatment in Washington State 264
Stacy Trussler, PE
Northwest Region, Water Quality Program
Washington State Department of Transportation
Seattle, Washington
Bert Bowen
Environmental Affairs Office, Water Quality Program
Washington State Department of Transportation
Olympia, Washington
StormTreat™ Technology for Stormwater Treatment 272
Mark E. Nelson, Director and Scott W. Horsley, President
StormTreat™ Systems, Inc.
Sandwich, Massachusetts
IX
-------
Contents (continued)
Evaluation of Stormceptor® and Multi-Chamber Treatment Train as Urban
Retrofit Strategies 277
Steven R. Greb
Wisconsin Department of Natural Resources
Madison, Wisconsin
Steve Corsi and Robert Waschbusch
US Geological Survey
Madison, Wisconsin
Assessing the Effectiveness of Orlando's BMP Strategies 284
William G. Chamberlin, II
City of Orlando
Orlando, Florida
Evaluating Public Information Programs: Experiences with the Florida
Yards and Neighborhoods Program 287
Billie Lofland
Florida Yards and Neighborhoods Program
University of Florida-Hillsborough County Cooperative Extension Service
Seffner, Florida
Examining the Need for Project Evaluation 291
Thomas E. Davenport
Water Division - Region 5
United States Environmental Protection Agency
Chicago, Illinois
Attendees 296
-------
Acknowledgements
The success of the conference and the preparation of this document are due largely to the
efforts of the presenters as well as the following individuals:
Conference Planning Committee
Robert Kirschner, Conference Coordinator
Northeastern Illinois Planning Commission, Chicago, IL
Thomas Davenport, Project Officer
U.S. Environmental Protection Agency, Region 5, Chicago, IL
Dale Bryson, Camp Dresser & McKee, Naperville, IL
Martin Kelly, Southwest Florida Water Management District, Tampa, FL
Bruce Kirschner, International Joint Commission, Detroit, Ml
Lyn Kirschner, Conservation Technology Information Center, W. Lafayette, IN
Eric Livingston, Florida Department of Environmental Protection, Tallahassee, FL
Richard Mollahan, Illinois Environmental Protection Agency, Springfield, IL
Daniel Murray, U.S. Environmental Protection Agency, Cincinnati, OH
Scott Ristau, Illinois Environmental Protection Agency, Springfield, IL
Peyton Robertson, National Oceanic and Atmospheric Administration, Silver Spring, MD
Roy Schremeck, Michigan Department of Environmental Quality, East Lansing, Ml
Thomas Schueler, Center for Watershed Protection, Ellicott City, MD
Earl Shaver, Delaware Department of Natural Resources and Environmental Control, Dover, DE
William Swietlik, U.S. Environmental Protection Agency, Washington, DC
Peer Reviewers
Donald Brown, U.S. Environmental Protection Agency, Cincinnati, OH
Thomas Davenport, U.S. Environmental Protection Agency, Region 5, Chicago, IL
Brian Hill, U.S. Environmental Protection Agency, Cincinnati, OH
Martin Kelly, Southwest Florida Water Management District, Tampa, FL
Susan Kaynor, Environmental Consultant, Coconut Grove, FL
Frank McCormick, U.S. Environmental Protection Agency, Cincinnati, OH
Jim Meeks, Consultant, Washington, DC
Daniel Murray, U.S. Environmental Protection Agency, Cincinnati, OH
Nancy Phillips, Consultant, Hollis, New Hampshire
Krista Reininga, Woodward-Clyde, Portland, OR
Eric Strecker, Woodward-Clyde, Portland, OR
Technical Editing and Publishing
Alan Everson, Scott Minamyer, Jean Dye, Peggy Heimbrock, and Stephen Wilson of the U.S.
Environmental Protection Agency, Cincinnati, OH
XI
-------
Securing the Urban Greenfrastructure:
Integrating Stormwater Management with Regional Growth Management
Michael C. Houck
Urban Naturalist, Audubon Society of Portland
and
Natural Resources Working Group, Coalition For a Livable Future
Portland, Oregon
Introduction
As a representative of the Coalition for a Livable Fu-
ture1, I was asked to present information on what we are
doing in the Portland, Oregon and Vancouver, Washing-
ton metropolitan region to integrate the "Urban
Greenfrastructure" into ourgrowth management strategies.
I will first describe the context in which we are working to
integrate more progressive stormwater management and
Greenspace (natural area) protection into regional growth
management strategies. Then I will discuss the efforts of
the Coalition for a Livable Future to further integrate that
work into a framework that includes social and environ-
mental growth objectives.
The Portland Park Bureau's 1903 master plan contains
the following admonition to utilize the natural landscape to
address issues of water resource management:
Marked economy may be effected by laying out parks,
while land is cheap, so as to embrace streams that carry
at times more water than can be taken care of....thus,
brooks which would otherwise be put in large underground
conduits at enormous public expense, may be made at-
tractive parkways.
This has a certain Olmstedian ring to it, but it was John
Charles Olmsted not his fatherwho first articulated a policy
of multi-objective stream management some 95 years ago.
While there may be no such thing as "cheap" land any-
more, especially in the cities, realizing Olmsted's vision is
very much within our ability to implement in the urban and
urbanizing environment. That is the path we have set out
on in the 24 cities and three counties of the Portland met-
ropolitan region.
Building and Retrofitting Livable Regions
One of Henry David Thoreau's most quoted statements
is, "In wildness is preservation of the earth." Ironically, some
members of the conservation community, carrying
Thoreau's aphorism into battle, have contributed to the
unfortunate demonization of the city. Some in the conser-
vation community, I believe, have also deified the so-called
"American Dream" of owning a quarter acre, or better yet,
a rural homesite in which to commune with nature, as if
nature cannot be appreciated in an urban setting. Of
course, most of them will then commute to the much-de-
rided city to work. The resultant urban sprawl has con-
sumed vast acreages of prime farm land and productive
forest land; fragmented wildlife habitat; destroyed a sense
of community; created expanding areas of concentrated
poverty in inner cities; and significantly increased the cost
of infrastructure, including stormwater management.
Robert Liberty, Director of 1000 Friends of Oregon, pro-
vided the following data which illustrate the tremendous
consumption of land that is the signature effect of unfet-
tered urban sprawl. Between 1970 and 1990 the Chicago
region's population grew by 4% but its land area increased
by 50%. Kansas City's population grew by 29% during that
same period and its land consumption was 110%.
Michigan's population is projected to grow by 12% between
1990 and 2020 while the urbanized areas in that state will
increase between 63% and 87%. A study commissioned
by the New Jersey legislature concluded that low-density
development consumed 130,000 more acres than a more
compact urban form would have, at an additional cost of
$740 million for roads and $440 million for sewer and wa-
ter infrastructure.
Perhaps Thoreau's adherents would be better served
by a new aphorism, "In livable cities is preservation of the
wild." It will only be through the creation or, where neces-
sary, the re-creation of livable cities that we will success-
fully protect the American landscape and the wilderness.
But we cannot hope to create compact, land-conserving,
urban forms unless we also ensure our cities are places
people want to live, not flee. Without a vibrant, healthy
urban Greenfrastructure (an interconnected system of
streams, wetlands, Greenspaces and greenways), we will
not create, or recreate—retrofit, if you will—livable cities.
Smart Growth and Urban Stormwater
Management
There is a growing national movement toward compact
urban form, although in truth it is not so much a movement
-------
forward as back to a development pattern that is reminis-
cent of our pre-World War II, non-auto-dominated commu-
nities. The weakness of this new Smart Growth movement
is the lack of an explicit nexus between higher-density,
mixed-use, pedestrian-friendly development and redevel-
opment on the one hand, and the protection and long-term
management of the urban Greenfrastructure on the other.
I was recently discussing the Smart Growth movement with
one of its adherents in Washington, D.C. and noticed a
huge, four-foot by six-foot poster on the wall. Amidst the
multi-modal transit schemes, row houses, townhouses and
mixed-use developments was a small, three- by five-inch
area marked "open space." There were no wetlands, no
un-culverted streams, not even a tree in this Smart Growth
scenario.
How do we rectify this? First, we can ensure that the
next version of that poster has not only the progressive
urban planning icons, but also urban waterways with
healthy riparian zones, parks that serve multiple purposes-
including stormwater and floodplain management-and
waterways used by people, fish and wildlife. To promote
this vision we need to form new partnerships between non-
government organizations (NGOs) and the practitioners
of water resource, stormwater and floodplain management.
We also need to build new coalitions among NGOs and
the grassroots citizen groups that can promote the inte-
gration of urban waterway management into the Smart
Growth movement.
The Coalition For a Livable Future has successfully
brought together unlikely partners in the nonprofit commu-
nity to integrate stormwater management into local and
regional land use programs, and to integrate environmen-
tal issues with social and environmental equity concerns.
The Coalition Fora Livable Future (CLF) is a group of more
than 40 nonprofit organizations, working in the Portland-
Vancouver metropolitan region, including: the Urban
League of Portland which represents low-income commu-
nities and people of color; the Community Development
Network, an umbrella organization for the region's afford-
able-housing advocates; Bicycle Transportation Alliance
and other alternative transit advocates; several stream
groups and watershed councils; and three local neighbor-
hood associations. What many would consider more "main-
stream" conservation organizations such as the Audubon
Society of Portland and 1000 Friends of Oregon are also
CLF members.
Coalition Building: Linking Environmental
and Social Concerns to Regional Growth
Management
Robert Liberty, director of the 1000 Friends of Oregon,
provided the catalyst for the formation of the coalition by
bringing Myron Orfield, a Minnesota state legislator, to
Portland. Representative Orfield has studied metropolitan
regions throughout the U. S. and has documented the "hol-
lowing out" of their urban cores. His maps graphically il-
lustrate the economic disparity that develops between
communities as the rapidly growing, sprawling suburbs cap-
ture a larger share of the regional tax base—where de-
mand for social services is lowest—while urban neighbor-
hoods with the highest social needs struggle to meet a
high demand for services, with a dwindling tax base.
The containment and the reversal of these phenomena
was the primary basis for formation of the CLF. While the
Portland metropolitan region does not exhibit all the symp-
toms of urban decline observed throughout the U. S., there
were enough signs that we might be headed down the
same path of metropolitan decay. The result of Orfield's
presentation and subsequent meetings was the writing of
a mission statement and development of core principles
around which diverse partners could join to become a re-
gional coalition. The coalition's mission statement and ob-
jectives were sent to interested organizations and individu-
als with an invitation to join. Every member organization
has been asked to sign an agreement to work not solely
on their individual issues, but to commit to promoting the
entire integrated package of CLF objectives.
CLF's mission is: To protect, restore, and maintain
healthy, equitable, and sustainable communities, both hu-
man and natural, for the benefit of present and future resi-
dents of the greater metropolitan region. The focus of the
coalition is to adapt or change government land use, trans-
portation, housing, public investment, and economic and
environmental policy through advocacy, research, and
public education.
The CLF's objectives are:
1) Protect the region's social and economic health in-
cluding: preventing displacement of low and moder-
ate income residents and people of color; assuring
equitable access to employment and affordable hous-
ing throughout the region; and reversing polarization
of income.
2) Develop a sustainable relationship between human
residents and the region's ecosystems by: changing
patterns of urban expansion to more compact neigh-
borhoods; expanding transportation options; and pro-
tecting, restoring and maintaining healthy watersheds,
fish and wildlife habitat, and Greenspaces both within
and outside the Urban Growth Boundary.
3) Assure fair distribution of tax burdens and govern-
ment investment within the region.
4) Promote a diverse and tolerant society.
5) Increase public understanding of regional growth
management issues; develop effective democratic
discourse; and promote broader citizen participation
in decision-making regarding regional growth issues.
In forming the CLF, we have brought together afford-
able-housing advocates, those working in the jobs-with-
justice arena, and representatives from low-income com-
-------
munities and people of color with the land use and trans-
portation specialists. We are not focusing our attention on
urban stormwater management alone, but on our regional
growth management program, Region 20402, cross-inter-
est education in regional growth issues has been such a
tremendously powerful political tool that affordable-hous-
ing experts testify before local and regional governments
supporting Greenspace protection. By the same token,
elected officials hear about affordable housing and urban
design issues from the Audubon Society of Portland.
Regional Growth Management: The Context
for Coalition Building and Stormwater
Management
In addition to Orfield's catalytic role in the formation of
the CLF, we were fortunate in having a regional planning
program to help focus our energy and develop jointly held
principles and policies. Metro, the only directly elected re-
gional government in the United States, has authority over
the 24 cities and three counties in the Portland metropoli-
tan region. Metro's charter requires it to undertake regional
growth management planning and other issues "of regional
significance." Water resource management is one of the
regionally significant issues that Metro is required to ad-
dress, as is housing, transportation, hazard mitigation and,
with considerable assistance from coalition members,
Greenspaces or natural area acquisition and management.
The CLF supports Metro's work where it is coincident with
our mission and objectives and addresses deficiencies
where necessary. One of the initial deficiencies was weak
stormwater and watershed management policies.
To date, the CLF has succeeded in persuading Metro's
seven-member council to adopt provisions for fair share,
inclusionary zoning for affordable housing (which is, as you
might suspect, a controversial issue among local govern-
ments); low-income community economic revitalization
language in the Regional Framework Plan: and newly
adopted floodplain and water quality management regula-
tions that will be applied consistently throughout the met-
ropolitan region. Additional acquisition of natural areas,
Greenspaces, and implementation of a regional
Greenspaces master plan is also a key element of the
framework plan.
Greenspaces to Stormwater Management;
Securing the Urban Greenfrastructure
One of the first areas of focus for the coalition was par-
ticipation in the development of a regional vision. Metro's
Future Vision Commission developed, among numerous
other recommendations, the following vision forthe region:
Integrate urban, suburban, and rural lands in a water-
shed-wide perspective to ensure reduction in downstream
flooding, reduction in wintertime flows and enhancement
of summer flows, protection of riparian corridors and wet-
lands and restoration of fisheries. Any future development
within the targeted urban reserves must be sensitive to
increased stormwater runoff, erosion, and sources of pollu-
tion and flooding downstream communities. An integrated,
multiobjective floodplain management strategy shall be
developed which recognizes the multiple values of stream
and river corridors including: enhanced water quality, fish
and wildlife habitat, open space, increased property val-
ues, education, flood reduction, aesthetics, and recreation.
An interconnected system of streams, rivers, and wetlands
that are managed on an ecosystem basis and restoration
of currently degraded streams and wetlands are important
elements of this ecosystem approach.
We next took on the task of redefining what the region
viewed as "infrastructure" in our Regional Growth goals
and objectives. We developed an alternative definition, took
it to the regional advisory committee of local elected offi-
cials and the full Metro Council and the following definition
of urban infrastructure was adopted:
Infrastructure: Roads, water systems, sewage systems,
systems for storm drainage, telecommunications and en-
ergy transmission and distribution systems, bridges, trans-
portation facilities, parks, schools and public facilities de-
veloped to support the functioning of the developed por-
tions of the environment. Areas of the undeveloped por-
tions of the environment such as floodplains, riparian and
wetland zones, groundwater recharge and discharge ar-
eas and Greenspaces that provide important functions re-
lated to maintaining the region's air and water quality, re-
duce the need for infrastructure expenses and contribute
to the region's quality of life.
From Greenspace Acquisition to Watershed
Management
Even prior to the formation of the coalition, the Audubon
Society of Portland and several other groups like The
Wetlands Conservancy had worked to create a regionally,
interconnected natural areas system. The Coalition for a
Livable Future identified Metro as the logical government
entity to house a regional natural areas system. Working
with numerous citizen groups and local park providers, the
coalition was able to persuade Metro Council to establish
a Regional Parks and Greenspaces Program at Metro.
Again, coalition-building and partnerships with govern-
ment agencies at every level were key to this successful
grassroots effort. We also had to be creative. We brought
in "outside experts" such as Dr. David Goode, Director of
the London Ecology Unit in England and New Yorker au-
thor, Tony Hiss, who wrote about our efforts in national
publications. We also invited nationally syndicated colum-
nist Neil Pierce to address our newly established coalition
of Greenspace advocates, FAUNA (Friends and Advocates
of Urban Natural Areas). We then organized two field tours
of the East Bay Regional Park District in Alameda and
Contra Costa Counties in the San Francisco Bay area so
local elected officials and park professionals in our region
could see how a regional park system focused on natural
areas can be developed and managed.
In spring of 1989, with funding from the Audubon Soci-
ety of Portland, local neighborhood groups, U. S. Fish and
-------
Wildlife Service, the U. S. Army Corps of Engineers and a
host of other cooperators, Metro commissioned an infra-
red photography project forthe entire Portland-Vancouver
metropolitan region, an area covering 1925 square miles
(55 miles north to south and 35 miles east to west). Dr.
Joe Poracsky, professor of geography at Portland State
University, and his graduate students then digitized this
low-level imagery to produce for the first time in our region
a map of all remaining natural areas in the Portland-
Vancouver metropolitan region3
The result of these efforts was the development of the
political will and broad public support at both the local and
regional level to establish a regional Greenspaces program
at Metro and to pass, with 60% voter approval, a $135.6
million bond measure (from an increase in property taxes)
to acquire up to 6,000 acres of Greenspaces. While this
acquisition program is a very important tool, the acquisi-
tion of 6,000 acres in a region that contains 204 square
miles and measures 38 miles east to west and 26 miles
north to south is inadequate to protect the regional land-
scape.
Regulatory Approach: Region-Wide
Floodplain and Water Quality Management
During the past three years the Coalition's Natural Re-
sources Working Group has focused its efforts in the regu-
latory arena and the development of a region-wide Func-
tional Plan, one element of which-Title 3-addresses flood-
plain and water quality management. Every opinion sur-
vey demonstrates tremendous public support for additional
regulatory approaches to the protection of water quality
and the region's urban waterways. Water quality is viewed
as essential to the maintenance of the region's livability
and long-term economic health. Protecting urban streams
is consistently rated one of the top values in Metro's public
surveys: 60 % of the respondents want to protect urban
streams, even if it means limiting development.
The Portland metropolitan region has 213 miles of 303
(d)-listed streams and rivers (water quality limited). In ad-
dition to these polluted stream miles, 388 miles of streams
have "disappeared" by being culverted, routed underground
or piped under streets and parking lots. An estimated 8,840
household units in the region are in, orclose to, floodplains.
Approximately 1,080 units were built in floodplains since
1992. During the February 1996 flood, 189 homes in the
region were inundated with water. According to the Oregon
Emergency Management Office, the cost of this flood was
about $60 million forthe three counties in Metro's jurisdic-
tion.
To address these issues the Coalition For A Livable Fu-
ture has worked with local stream groups and watershed
councils, and with Metro staff and elected officials at the
local and regional levels to develop a region-wide strategy
to address development in the region's floodplains and the
degradation of water quality in the Willamette River and its
tributaries. One of Metro's most important advisory com-
mittees recently recommended to Metro Council that the
region's cities and counties be required to do the follow-
ing, as one element of the region's integrated Growth Con-
cept:
1. Prohibit new development in the floodplains of the
region's rivers and streams or, at a minimum, require
"balanced cut and fill."
2. Adopt water quality performance standards that fo-
cus on retention of vegetated corridors along all of
the region's streams, rivers and wetlands. Each city
and county will be required to maintain vegetated cor-
ridors which provide shade, stabilize banks, trap soil
and other runoff before it enters the water and mod-
erate stormwater flow. The vegetated corridors will
measure (on each side of the water feature):
15' for seasonal streams that drain between 50 and
100 acres, on slopes of less than 25%
50' for perennial streams or rivers that drain more
than 100 acres, wetlands and year-round springs if
they are in areas where slopes are less than 25%
200' for streams and wetlands where slopes are
more than 25%
3. Adopt Metro's map which delineates all floodplains,
wetlands, stream corridors and steep slopes (over
24%) throughout the region or develop local maps
which "substantially comply" with Metro's maps.
4. Adopt region-wide erosion control for any new devel-
opment (no acreage limitation).
5. Adopt Metro's Model Ordinance or develop a local
ordinance which substantially complies with Metro's
Model Ordinance.
This new regulatory package will be voted on by the full
Metro Council in April of this year (1998). Once adopted
(scheduled for May of 1998) local jurisdictions will have up
to eighteen months to implement the provisions of the
Floodplain and Water Quality Management Functional
Plan. The recent listing of steelhead by the National Ma-
rine Fisheries Service for the lower Willamette River and
the Sandy and Clackamas Rivers, all of which are in Metro's
jurisdiction, has brought the Endangered Species Act to
the Portland metropolitan region in a manner that will as-
sist in the adoption of water resource-oriented growth man-
agement policies.
For example, Oregon's Governor, John Kitzhaber and
agency directors from the departments of Agriculture, Land
Conservation and Development, Division of State Lands,
Water Resources, Fish and Wildlife, Geology and Mineral
Industries, and Environmental Quality and Water Re-
sources submitted a joint statement that Title 3 of the
Coalition's Functional Plan is an important first step in wa-
tershed enhancement...The recent federal endangered
species listing of steelhead in the Columbia and Willamette
River systems elevates the significance of habitat protec-
-------
tion practices at the local government level.. .Weofferthe
following additional recommendations: add a provision for
setback buffers in headwater areas, preferably a minimum
of fifty feet. We encourage Metro's early adoption of strong
Title 3 policies and implementing measures so that
progress can be made soon on the larger work envisioned
by chapters 4 and 5 of Metro's Regional Framework Plan.
A letter of this nature is unprecedented in the Portland met-
ropolitan region.
Next Steps
The Coalition will continue to focus its efforts on the chap-
ters of the Regional Framework Plan that require consis-
tent, region-wide stormwater management: mandated com-
prehensive watershed planning for all the region's water-
sheds within Metro's jurisdiction; development of policies
to reduce landslide hazards; and development of a regional
fish and wildlife habitat protection program that would en-
sure an adequate program in every city and county within
Metro's jurisdiction. We will also work to implement the
Greenspaces Master Plan, which will include an update of
the 1989 infrared Greenspaces inventory, and establish
plans for a regional interconnected Greenspace system
based on maintaining the region's biodiversity and wildlife
corridors.
References
1. More information on the Coalition For a Livable Future
can be obtained at the Coalition's offices at 534 SW
Third Avenue, Suite 300, Portland, OR 97204 (phone:
503-294-2889, email: zack@friends.org).
2. For more information about Metro's Region 2040
growth management planning process contact: Elaine
Wilkerson, Director Growth Management Services,
Metro, 600 NE Grand, Portland, OR 97232.
3. Metro has a Growth Management Hotline, 503-797-
1888 and a website, www.metro-region.org. For more
information about their CIS mapping, contact Metro,
Data Resource Center 503-797-1742 or Metro's
website,
-------
The Use of Retention Basins to Mitigate Stormwater Impacts to Aquatic Life
John R. Maxted
Delaware Department of Natural Resources and Environmental Control
Dover, Delaware
Earl Shaver
Auckland Regional Council
Auckland, New Zealand
Abstract
Physical habitat and biological measurements were
taken in nontidal streams below eight stormwater man-
agement pond facilities (BMPs) during the spring of 1996.
Two of the sites were predominantly in commercial land
use while the remaining six sites were in residential land
use. The results were compared to 33 sites with no
stormwater controls. Three replicate macroinvertebrate
samples were collected in riffle habitats using a kicknet.
Biological quality was determined from six metrics using
100-organism subsamples identified to the species level.
Physical habitat quality was determined from 12 metrics
that defined the condition of the channel, stream bank, and
riparian zone. These biological and physical habitat metrics
were compared with mean values derived from three ref-
erence sites to produce summary index scores for each
site, reported as "percent of reference." The overall
macroinvertebrate community, as measured using a com-
posite of all six biological metrics (Community Index), was
not significantly different between BMP and non-BMP sites.
A similar result was found using a composite of three
metrics that characterized pollution-sensitive organisms
(Sensitive Species Index). The BMPs did not prevent the
almost complete loss of sensitive taxa (e.g., mayflies,
stoneflies, and caddisflies) after development. Further, the
BMPs did not attenuate the impacts of urbanization once
the watershed reached 20% impervious cover. Data are
needed to determine whetherthese controls would attenu-
ate impacts at lower levels of development (5-15% imper-
vious cover). Half of the BMP sites had Habitat Index scores
comparable to the reference condition, indicating mixed
results with regard to the effectiveness of the BMPs in pro-
tecting physical habitat. These results suggest that follow-
ing management actions may be needed: (1) modifications
to traditional urban designs that reduce impervious cover
and preserve natural features (e.g., "conservation design"),
(2) modifications to stormwater retention basin designs
(e.g., expanded capacity, constructed wetlands), and (3)
the restoration and preservation of forest cover along
stream channels, especially along intermittent streams and
first and second order perennial streams.
A data set of this size should not be used to derive de-
finitive conclusions regarding the ability of stormwater con-
trols to protect aquatic life and physical habitat. This study
characterized the condition of only eight sites, and the
stormwater management design criteria varied between
the sites. The lack of comparable studies of other regions
of the U.S., and the use of other measures of stream con-
dition below stormwater controls suggest the need for ad-
ditional research.
Introduction
Overthe last 90 years, the population of the United States
has increased 300%, from 76 million in 1900, to 249 mil-
lion in 1990 (United States Census 1996). This period has
also seen a dramatic shift in the way people live and use
the land. In 1900, the majority of the U.S. population (60%)
lived in rural areas, while in 1990 the majority (75%) lived
in urban areas. This trend continues today although at a
slower rate. But even as the rate levels off, roughly three-
fourths of the estimated 25 million people that will be added
to the population over the next decade will likely live in
urban areas. Delaware's population has experienced a
similar rate of population increase (185,000 to 666,000)
and shift in land use over this period.
This change in demographics and land use has brought
about profound changes in the physical, chemical, and bio-
logical integrity of nontidal streams in Delaware. The ob-
jective of this research was to determine the effectiveness
of stormwater controls, principally retention basins, to pro-
tect stream resources after urbanization. This study focused
on wadeable, nontidal streams and the use of
macroinvertebrates and physical habitat as indicators of
stream ecological health.
-------
Impervious surfaces (roads, parking lots, rooftops, drive-
ways, sidewalks, etc.) increase peak flows during storm
events and reduce base flows during droughts. Urbanized
watersheds with 20-30% impervious cover were found to
have 10-15 times the frequency of small flood events (1-
year recurrence interval) compared to nonurban water-
sheds; large flood events (100-year) doubled in size after
urbanization (Hollis 1975). This change in stream hydrol-
ogy affects the physical structure and stability of stream
channels through accelerated erosion and sediment depo-
sition. The replacement of native riparian vegetation (e.g.,
trees) with lawns, parks, golf courses, and structures (e.g.,
buildings, bridges) along stream channels and floodplains
further impact the geomorphology of urban streams.
Water quality contaminants in stormwater (metals, nu-
trients, organics) further stress aquatic life. Exceedences
of dissolved oxygen (DO) criteria occur in streams and
ponds through nutrient enrichment and the removal of
shade. In a recent survey, unshaded stream channels in
Delaware exceeded the State's acute criteria for DO and
temperature 73% and 38% of the days, respectively, dur-
ing the Summer of 1993 (Maxted et al., 1995). Both physi-
cal and chemical factors associated with urbanization con-
tribute to the overall biological condition of urban streams.
Aquatic organisms, principally fish and
macroinvertebrates, are commonly used to assess the
ecological condition of streams, and several researchers
have used them to assess the impacts of urbanization
(Shaver and Maxted 1995, Jones and Clark 1987, Klein
1979, Limburg and Schmidt 1990, Pedersen and Perkins
1986, Booth and Jackson 1994, Weaver and Carmen 1994,
and Garie and Mclntosh 1986). These studies have re-
cently been summarized (Schueler 1994). What is gener-
ally lacking are studies which use aquatic organisms to
evaluate the effectiveness of stormwater controls.
The water quality impacts of stormwater runoff are fairly
well documented, as is the ability of a variety of stormwater
management facilities to provide water quantity control and
water quality treatment. There has been an inherent as-
sumption that water quality treatment and pollutant cap-
ture directly translate into aquatic life protection. This as-
sumption has never been validated. In addition, stormwater
treatment facilities effectively remove pollutants, but the
level of performance is highly variable and needs to be
expressed in ranges rather than in specific levels of treat-
ment. At great expense, large amounts of data covering
many stormwater events and constituents are needed to
make reasonable statements regarding the performance
of BMPs in removing pollutants (Urbonas, 1995).
Water quality data can also present problems in terms
of data accuracy. Stormwater management facilities often
have multiple inflow points and may receive overland flow
which makes data collection difficult. Monitoring each in-
flow point and the facility outfall increases the potential for
error in data collection and analysis. Coupled with the need
to sample multiple storm events over different seasons and
different years, these factors make it difficult to accurately
assess the performance of the BMP. These factors also
affect the overall cost of monitoring.
What is needed is a simple, long-term approach to sys-
tem assessment which minimizes the cost of data collec-
tion and provides a framework for evaluating the effective-
ness of controls. Presented in this paper is a framework
for assessing the ecological health of aquatic ecosystems
and the performance of stormwater facilities using living
resources, in this case aquatic macroinvertebrates. While
evaluating stream ecological health using aquatic organ-
isms is widespread, the evaluation of stormwater facility
performance using this approach is fairly new. Preliminary
results of the present study have been summarized previ-
ously (Maxted and Shaver, 1997).
Methods
The heavily urbanized piedmont region of northern Dela-
ware was selected for study. Data collected at 33 sites
with no BMPs in the catchment were compared with eight
sites sampled below modern stormwater retention basins
(Figure 1). As of 1984, about half the piedment region (48%)
was in urban land use, 33% was undeveloped, and 19%
was in agriculture. Stormwater controls have only recently
been included as part of new developments in the region.
Therefore, the data collected at the 33 non-BMP sites rep-
resent conditions that existed before the implementation
of regulatory programs for controlling stormwater runoff.
The land use conditions in the watersheds above the 33
non-BMP sites covered the full range of urban land use
from relatively undeveloped watersheds with less than 10%
impervious cover to heavily urbanized watersheds with
greater than 30% impervious cover. Sampling sites were
BMP sites (n=8)
no BMP sites (n=33)
•iff Reference sites (n=3)
Figure 1. Locations of sampling sites within the northern piedmont
region of Delaware.
-------
located 100 meters below the BMP discharge to minimize
the immediate influence of the discharge on the stream
and the influence of construction and maintenance activi-
ties related to the BMP itself (grading, mowing, habitat dis-
turbance, etc.).
The 33 non-BMP sites were sampled in the fall of 1993,
while the eight BMP sites were sampled in the spring of
1996 (between May 2nd and June 6th, 1996). The metrics
used to summarize the biological data were not consid-
ered to be sensitive to seasonal differences between the
fall and the spring, and thus allowed for this comparison.
Macroinvertebrate samples were collected using a 1-
meter2 kick net (750 urn mesh). Each sample was a com-
posite of two collections of a 1-meter2 area of riffle, com-
bined in a sieve bucket (600 urn mesh). Three replicate
collections were made at each site while moving progres-
sively upstream. A single 100-organism subsample was
removed from each sample and identified to the species
level. Six metrics were derived for each sample: taxonomic
richness (TR); richness of the orders of ephemeroptera,
plecoptera, and trichoptera (EPT); % EPT abundance (%
EPT); % Chironomidae (% C); % dominant taxon (% DT);
and the Hilsenhoff Biotic Index (HBI) (Table 1).
Habitat quality assessment included measures of the
channel, stream bank, and riparian zone. Each assess-
ment consisted of the visual characterization of a 100-meter
segment of the stream using the following 12 parameters.
Numerical scores, out of a possible 20 points, were as-
signed to each parameter.
CM - channel modification: the degree of engineering
of the channel shape (e.g., channelized) and the ex-
tent to which it meanders.
BSC - bottom substrate/available cover: the amount
and variety of submerged stable habitat throughout
the stream segment (e.g., riffles, logs, snags, aquatic
plants, root-wads along banks, etc.).
E - embeddedness: the degree to which the substrate
is surrounded or covered by fine sediment.
Table 1.
Biological Metrics Used to Derive Summary Index Scores for
BMP and non-BMP Sites.
Metric Name
Description
Type
taxonomic richness
EPT* richness
% EPT abundance
% dominant taxon
% Chironomidae**
Hilsenhoff (HBI)
total number of unique taxa richness
total number of EPT taxa rich/tolerance
% of sample that are EPTs tolerance/comp
largest % of a single taxon composition
% of sample from this group tolerance
composite tolerance by taxon tolerance
EPT - the orders ephemeroptera (mayflies), plecoptera (stoneflies),
and trichoptera (caddisflies); high richness or relative abundance
indicates high quality.
'Chironomidae - family of midges; high relative abundance indicates
low quality.
RQ - riffle quality: the dominant substrate found in
riffles; cobbles are the most desirable, boulders and
gravels are the least desireable.
FR - frequency of riffles: the abundance of riffle areas
in the stream segment.
SD - sediment deposition: the degree to which new
sediment is deposited in the stream channel as evi-
denced by islands, point bars, and sand and silt cov-
ering stable habitats.
V/D - velocity/depth: the presence of four categories
of flow regime; slow and deep, slow and shallow, fast
and deep, fast and shallow.
BS- bank stability: the proportion (%) of stream banks
that show evidence of recent and active erosion.
BV - bank vegetative type: the dominant vegetation
on the stream bank; trees and shrubs being most de-
sirable, grasses being the least desirable; left and right
banks scored separately and then combined.
S - shading: the percent of the stream surface that is
shaded throughout the day.
RZ - riparian zone width: the width of the riparian zone
showing little or no evidence of human activity; left and
right sides scored separately and then combined.
HCI - Habitat Comparison Index: summary index of
habitat quality; individual parameter scores summed
and divided by a reference value; index values ex-
pressed as "percent of reference."
Three summary index scores, two biological and one
habitat, were derived for each site following procedures
developed by EPA (Plafkin, et al., 1989). Three reference
sites were sampled during the same seasonal period as
the sampling sites and used to derive index scores reported
as "percent of reference." Habitat Index scores were de-
termined by comparing the total habitat score for each BMP
site with the mean total score for the three reference sites.
Community Index (Cl) scores were determined by com-
paring all six biological metric values for each site with the
mean values from the three reference sites. The Cl was
used to define the overall quality of the macroinvertebrate
community. The Sensitive Species Index (SSI) scores were
determined using the three biological metrics (EPT, % EPT,
and HBI) that define the components of the community
that are the most sensitive to organic pollution. Mean Cl
and SSI scores for each site were determined from the Cl
and SSI scores from the three replicate samples.
The biological data were plotted against % impervious
cover estimates determined forthe catchment above each
site. Land use was determined from digitized 1992 land
use data. Percent impervious cover estimates were made
by multiplying the area of each land use category by the %
impervious cover estimate for that category, as published
by the U.S. Department of Agriculture (USDA1986), sum-
-------
ming the values for all the land use categories, and then
dividing the total % impervious area by the total area for
the catchment. The Cl biological index values were plot-
ted against the HCI habitat index values to further charac-
terize the habitat quality of the BMP sites. Mean values
differing by more than one standard deviation were de-
fined as statistically significant.
BMP Site Selection
BMP site selection employed a variety of information
sources including an existing stormwater facility inventory
and discussions with individuals familiar with a number of
facilities. The stormwater facility and stream criteria used
to select BMP sites are listed below. Eight sites met these
criteria and were selected for study; two sites were pre-
dominantly in commercial land use while the remaining six
sites were in residential land use (Table 2). The two com-
mercial sites and one residential site met modern design
standards for peak control of the two-, ten-, and 100-year
storms, and extended detention of the first inch of runoff.
Four residential sites were designed for control of the ten-
year storm only. One residential site was located in the
main channel of Jenny's Run below five separate reten-
tion basins that captured approximately 75% of the the
urban land use in the catchment (Table 2).
Stormwater facility criteria:
• Facilities had to be retention or detention basins, so
that one specific type of BMP (e.g., ponds designed
for stormwater control) could be evaluated.
• To the greatest extent possible, the facility had to meet
current design criteria which included peak rate con-
trol (two-, ten-, 100-year storms) and water quality
performance (24-hour detention for the first inch of
runoff). If a sufficient number of facilities were not found,
due to the recent nature of the State Stormwater Man-
agement Program (effective date July, 1991), older
retention ponds serving a development were consid-
ered.
• Facilities had to be at least two years old. The con-
cern with newer facilities was that there was potential
for construction-related stream impacts. If a new facil-
ity had significant instream impacts, the cause of the
impact might be related to excess runoff and sedimen-
tation during construction ratherthan the performance
of the BMP.
• Impervious cover in the catchment to the facility had
to be at least 20%. This would answer the initial ques-
tion concerning the effectiveness of BMPs in already
urbanized areas. Based on the results of this study,
future studies might address the question of the effec-
tiveness of BMPs at earlier stages of urbanization (e.g.,
5-15% impervious cover).
Receiving stream criteria:
• Discharge from the BMP represented the predominant
flow in the stream.
• Riparian zone had native vegetation (e.g., trees,
wooded, and shaded) and was not directly impacted
by human activities. This criterion might be difficult to
achieve in the heavily urbanized piedmont region of
Delaware.
• The receiving stream had perennial flow. This crite-
rion might be difficult to achieve since the streams
below individual retention basins are first order streams
with fairly small drainage areas.
• Riffles with a cobble substrate were common.
Results
The mean EPT richness (EPT), % EPT abundance (%
EPT), % Chironomidae (% C), and Hilsenhoff Biotic Index
(HBI) metrics were significantly different between the BMP
sites and the reference sites (Table 3). This indicated that
none of the BMP sites prevented a shift in the
macroinvertebrate community from one dominated by pol-
lution-sensitive organisms to one dominated by pollution-
tolerant organisms. Taken together, sites below BMPs had
a low proportion of the pollution-sensitive organisms (14%
EPT) and a high proportion of pollution-tolerant organisms
(54% Chironomidae), while the community at reference
sites was almost exactly the opposite (Table 3). The BMP
sites had half the mean HBI value of the reference sites,
indicating a shift at the species level as well. Similar re-
sults were found using the two summary biological indi-
Table 2. BMP Facility Data
Site
Land Use
% Impervious Cover
Drainage Area
Development
BMP1*
BMP 2"
BMP 3"
BMP 4"
BMPS*
BMP 6*
BMP 7*
BMP 8***
Residential
Commercial
Commercial
Residential
Residential
Residential
Residential
Residential
25
22
65
30
28
31
30
23
88.0 acres
83.0 acres
36.0 acres
32.0 acres
383.0 acres
107.0 acres
157.0 acres
330.0 acres
Corner Katch
Brandywine Com
Core States
Hunt at Louviers
Veranda
Limestone Hills
Chestnut Hills
Jenny's Run
Project design based on peak control of the ten-year storm only.
Project design based on peak control of the two-, ten-, and 100-year strom, in addition to 24-hour extended detention for the first 1" of runoff.
Site in main stream of Jenny's Run; considered stormwater flow from more than one development site.
-------
Table 3. Mean Values for Six Biological Metrics and Two Summary Indices Below Eight BMPs; Taxonomic Richness (TR), EPT Richness (EPT), °A
EPT Abundance (% EPT), % Chironomidae (% C), % Dominant Taxon (% DT), Hilsenhoff Biotic Index (HBI), Community Index (Cl), and
Sensitive Species Index (SSI); Standard Deviation Appears in Parenthesis.
Site
BMP 1
BMP 2
BMPS
BMP 4
BMPS
BMP 6
BMP 7
BMPS
all BMPs
samples
N
3
3
3
3
3
3
1
3
28
TR
35
21
31
26
19
22
29
23
25
(5)
(6)
(3)
(6)
(3)
(7)
(3)
(7)
EPT
7.3 (0.6)
2.0 (1.7)
1.7 (0.6)
6.0 (1.0)
0.0 (0.0)
6.7 (1.1)
5.0 (-)
7.7 (1.5)
4.5 (3.0)
Biological
% EPT %
18 (6)
26 (19)
2 (1)
17 (9)
0 (0)
14 (2)
9 (-)
26 (8)
14 (12)
54
52
71
60
37
38
75
60
54
Metrics
C
(6)
(20)
(10)
(22)
(24)
(36)
(8)
(21)
%DT
15 (5)
29 (7)
27 (11)
27 (14)
51 (26)
47 (22)
22 (-)
20 (5)
30 (17)
HBI
5.2
6.1
5.8
5.3
7.1
6.4
5.4
4.7
5.8
(0.4)
(0.4)
(0.4)
(0.7)
(0.5)
(0.8)
(0.5)
(0.9)
(% of reference)
Summary Indices
Cl SSI
49 (3)
35 (15)
33 (3)
39 (18)
25 (6)
31 (9)
35 (-)
51 (12)
38 (12)
26 (6)
15 (17)
7 (6)
18 (18)
0 (0)
7 (13)
11 (-)
33 (20)
15 (15)
Reference
10
24 (4)
10.3 (1.8)
56 (12)
14 (13)
24 (4)
2.9 (0.6)
100
100
ces. The mean Community Index (Cl) scores for the BMP
sites ranged from 25-51% of the reference condition while
the mean Sensitive Species Index (SSI) scores ranged
from 0-33% (Table 3).
Mean values for the biological metrics and summary in-
dices for the eight BMP sites were compared to 21 non-
BMP sites with similar land use (Table 4). Only non-BMP
sites with greater than 20% impervious cover were con-
sidered in order to provide a similar level of urban devel-
opment between the two groups of sites. There was no
significant difference between the two groups of sites for
the EPT, %EPT, and %C metrics, as well as the Cl and
SSI index values (Table 4). Both the BMP and non-BMP
sites were significantly different from the reference condi-
tion for most biological metrics, indicating that neithergroup
of sites approximated conditions found in undeveloped wa-
tersheds (Table 4). While the BMPs appeared to increase
the relative abundance of EPTs (i.e., % EPT), it had no
effect on either taxonomic richness of EPTs (i.e, number
of unique EPTtaxa) or the HBI; both are good indicators
of pollution tolerance.
A lack of biological improvement with the eight BMP sites
was observed when the data were plotted against % im-
pervious cover. No improvement in biological condition was
observed using either the Community Index (Figure 2) or
the Sensitive Species Index (Figure 3), as compared to
sites without BMPs. BMPs did not prevent the loss of sen-
sitive species found at reference sites. The degree of ur-
banization did not appear to affect biological conditions at
the BMP sites. The one BMP site with 65% impervious
cover had a similar biological condition to the seven sites
with 22-32% impervious cover.
Half of the BMP sites (BMP2, BMPS, BMP6, and BMP?)
had habitat scores less than 90% of reference, indicating
physical habitat impairment (Table 5). These sites exhib-
ited the physical characteristics of urban streams with no
controls, indicating that the BMPs were not effective at
eliminating the impacts of urbanization. The impacts were
most often associated with bank instability and channel
sedimentation. The other half of the sites (BMP1, BMP4,
BMP5, and BMPS) had habitat conditions similar to the
reference sites (i.e., greaterthan 90% of reference). It ap-
peared that some of the BMPs provided stable channel
characteristics, although there was no pattern related to
BMP design type or drainage area. The three sites that
had the highest physical habitat quality (BMP1, BMP4, and
BMPS) represented the full range of both BMP design type
and drainage area. The level of impairment, both physi-
cally and biologically, was also illustrated when biological
quality was plotted against habitat quality (Figure 4).
Discussion
Retention and detention basins designed to control
stormwater did not protect aquatic life from the adverse
Table 4. Comparison of Mean Values for Six Biological Metrics and Two Summary Indices Between Reference Sites and Sites With and Without
Stormwater BMPs; Taxonomic Richness (TR), EPT Richness (EPT), % EPT Abundance (% EPT), % Chironomidae (% C), % Dominant
Taxon (% DT), Hilsenhoff Biotic Index (HBI), Community Index (Cl), and Sensitive Species Index (SSI); Standard Deviation Appears in
Parenthesis.
Site
Reference
BMP
no BMP*
samples
N
10/3
28/8
29/21
TR
24 (4)
15 (7)
20 (5)
EPT
10.3
4.5
4.7
(1.8)
(3.0)
(2.7)
Biological Metrics
% EPT % C %
56 (12)
14 (12)
27 (18)
14 (13)
54 (21)
28 (23)
24
30
27
DT
(4)
(17)
(18)
HBI
2.9 (0.6)
5.8 (0.9)
5.1 (1.2)
(% of reference)
Summary Indices
Cl SSI
100
38 (12)
36 (14)
100
15 (15)
14 (15)
* only urban sites with 20-65% impervious cover included
10
-------
o
O
120 -
100 -
80 -
60 -
40 -
20 -
0 -
o
O 00 O O
o o o
O 8 !
o •• o
04
oo o 7»o o
CF2 7%O O O
• 6 O
O 51" o
O
o
o
10
• BMPs(n = 8)
O no BMPs (n = 33)
20
30 40
Impervious Cover
50
3»
60
70
Figure 2. The effects of urbanization on the macroinvertebrate community; numbers denote BMP sites.
X
-o
c
w
0)
'o
tt)
Q.
CO
Sensitive
120 -i
100 -
80 -
60 -
40 -
20 -
0 -
• BMPs (n = 8)
O no BMPs (n = 33)
O
CD O O
0
O O O
oo 8 o
o •1
00*4 oo
OO <§2 0
Hlg o w
O 5« O OO CO
1 i 1 1 I I i I
0 10 20 30 40 50 60 70
% Impervious Cover
Figure 3. The effects of urbanization on sensitive species of macroinvertebrates; numbers denote BMP sites.
11
-------
Table 5. Habitat Metric Scores Below Eight BMPs; Habitat Comparison Index (HCI) Reported as % of Reference; See Text for Abbreviations.
Site*
CM
BSC
RQ
FR
SD
VD
BS
BV
100-
& 80"
^>*
"ra
a eo H
_g
- 40 -
^U1
20-
•
0
BMPs (n = 8
no BMPs (n
)
= 33)
o
o
00 O©
CD O
O
COO !
o
O
-------
veloped forested watersheds. While one of the three ref-
erence sites had a series of old farm ponds in the water-
shed, two did not. Ponds effectively convert inorganic car-
bon and nutrients to organic matter through photosynthe-
sis. This fine particulate organic matter (FROM) discharged
to the stream during runoff events is a preferred food source
for invertebrates and fish tolerant to organic pollution. The
effect on living resources we observed and quantified be-
low the eight BMP ponds may also be related to the con-
struction of the ponds themselves. The important conclu-
sion remains, however, that ponds constructed to treat
stormwaterwere inadequate to protect aquatic life.
Site selection was more difficult than we first anticipated.
Many of the streams that received a discharge from a re-
tention or detention pond were too small to sample, had a
degraded riparian zone (unshaded), or discharged directly
to a larger tributary. Many of the BMP sites not selected
for study had small drainage areas that would have had
intermittent flow even under pre-development conditions.
Itwas difficult to find a representative sample of stormwater
management BMPs that met modern design criteria due
to the relative short time between the initiation of the
Stormwater Management Program in Delaware and this
monitoring effort. The only sites meeting the state's
stormwater management requirements in terms of when
they were constructed and their design criteria were the
two commercial sites and one residential site. The other
residential sites were selected because retention ponds
are a preferred practice under the new state program.
Construction-related impacts must be expected as a re-
sult of increased stormwater discharges and elevated sedi-
ment loadings. There has to be a period of time after con-
struction, and before measurements should be taken, to
assess the response of a receiving system to site devel-
opment and stormwater management facilities. We can
make recommendations, such as a period of two years
used here, but that recommendation must be considered
"preliminary" and subject to variation around the country
due to differences in climatic and other factors.
The importance of riparian zone protection and restora-
tion cannot be stressed too much. More effective man-
agement of riparian zones must be provided if receiving
systems are to acheive the structure and function of un-
disturbed systems. The greatest difficulty in site selection
was finding BMPs that discharged to streams with undis-
turbed riparian areas. This was not surprising given the
results of a recent statewide survey in which 87% of the
nontidal stream miles were found to have degraded physi-
cal habitat (Delaware DNREC 1994). The environmental
benefits of control efforts (both structural and non-struc-
tural) will be reduced if we fail to restore and protect ripar-
ian habitat.
This study represented a different approach to assess-
ing BMP effectiveness as compared to chemical monitor-
ing. Traditional approaches that focus on chemical con-
taminants determine aquatic life use support based on
pounds of pollutants removed and compliance with chemi-
cal criteria. Our approach looked directly at the aquatic
organisms the controls were designed to protect. Living
resources are the only direct measure of aquatic life con-
dition. All others, including physical habitat, are surrogates
that may underestimate or overestimate the true condition
of living resources.
Some have argued that undeveloped forested water-
sheds should not be used as the reference condition for
evaluating the performance of stormwater controls. They
also assert that fundemental changes in land use in urban
watersheds justifies the establishment of a lower quality
"urban stream standard" for aquatic resources. They feel
that this lower standard is needed because streams in ur-
ban watersheds will never achieve such a high level of
quality even with extensive land use and stormwater con-
trols. Further, they claim that urban land uses impact only
a small percentage of stream resources in most regions.
We reject these arguments for several reasons. First,
such an "urban stream standard" would be nearly impos-
sible to set. How would such a level of "acceptable" or
"achievable" quality be determined? Whatever approach
was selected would undoubtedly be influenced by political
ratherthan scientific factors. Second, conservation design
practices coupled with structural and non-structural con-
trols do not yet exist over extensive areas. We, therefore,
have no way of knowing whether the application of these
controls might achieve a higher level of ecological quality.
Third, if we were to set a lower standard, we would elimi-
nate a principal incentive for challenging and testing the
standard. And lastly, urban areas affect an ever-increas-
ing proportion of nontidal streams, particularly intermittent,
first, and second order streams. Urban sprawl affects
aquatic resources far away from city centers, extending
the proportion of streams affected by urbanization. While
roughly half of the piedmont region of Delaware is in ur-
ban land uses, it adversely affects nearly all of the 270
miles of nontidal streams.
It is too early to panic, as additional studies are needed.
Our results should best be described as preliminary. If fur-
ther studies confirm these results, BMP design criteria will
have to be reconsidered to provide a greater level of pro-
tection to receiving systems. Similar data are needed at
more sites in the piedmont region of the Mid-Atlantic U.S.,
before making definitive conclusions on the effectiveness
of stormwater basins in protecting aquatic life and physi-
cal habitat. Data are also needed for various types of ur-
ban designs (e.g., conservation design), various levels of
impervious cover, and various types of BMP designs, in-
cluding the three presented here and the use of constructed
wetlands. The ultimate question remains to be answered:
Can urban developments that incorporate available con-
trol technologies be cost-effective and marketable, while
protecting living resources?
Urban Retrofit Opportunities
What are the implications of this research with regard to
areas already undergoing various degree of urbanization?
13
-------
First, this research indicates that retention basins are not
sufficient to protect living resources. It is likely that changes
are also needed in the way urban areas are designed and
constructed in the first place. It should be no surprise that
ponds added on to conventional urban developments,
where nearly 100% of the development site is modified for
human uses, did not protect aquatic life. The concept of
"conservation design" used in conjunction with structural
and non-structural controls may be necessary. Conserva-
tion design encompasses a range of alternative design
practices that reduce impervious surfaces (e.g., reduced
roadway width, reduced setbacks), preserve sensitive natu-
ral features (woodlots, wetlands, floodplains), and reduce
collection system infrastructure (grassed swales, smaller
and more numerous retention areas). The State of Dela-
ware, in conjunction with the Brandywine Conservancy,
has recently completed a manual on conservation design
(Delaware DNREC 1997).
Second, protecting intermittent, first, and second order
perennial streams needs to be mentioned since their im-
portance is often overlooked. From a developmental per-
spective, these streams are often filled, piped, rerouted,
or otherwise altered. We then, through regulatory programs,
attempt to protect third and higher order streams and their
associated resource values. It may not be possible to pro-
tect the values of these higher order streams and rivers
unless we first protect the headwater streams at the top of
the watershed.
Third, the design of stormwater retention basins may
need to be modified to enhance their performance, par-
ticularly with regard to dissolved contaminants and tem-
perature effects. It should also be no surprise that ponds
may not be sufficient, by themselves, to attentuate the water
quality, hydologic, and biological effects of urbanization.
In fact, the ponds themselves may be contributing to the
problem. Conservation design in conjuction with extended
detention and constructed wetlands may be necessary to
protect stream ecological health. Further, we may want to
make a requirement that all ponds be "off-line" from wet-
lands, intermittent streams, and first and second order
perennial streams to prevent the direct impact that the
BMPs have on existing aquatic resources.
Fourth, riparian restoration should be implemented im-
mediately, which would provide significant benefits to
stream health even in heavily urbanized areas. Most of
the BMP sites visited during site selection could not be
studied because the streams they discharged to lacked
native riparian vegetation (e.g., trees). They most often
consisted of backyards or parks with grassed areas and
few trees. Through public education and the establishment
of easements, these areas should be preserved and re-
stored as natural wooded riparian corridors. This will be
necessary, eventually, even after the implementation of
structural and non-structural controls, to protect aquatic
resources. Since natural revegetation with trees can take
many years, efforts should be initiated now.
All of these objectives will need to be met if streams in
urban areas are to attain the structure and function of natu-
ral systems. Since prevention is often more effective and
less expensive than treatment, the most critical watersheds
are those in the early stages of urbanization (5-15% im-
pervious cover). The need to focus attention on these wa-
tersheds is important because impacts are often perma-
nent once the urban land use is in place. Additional re-
search and monitoring is especially important in these ar-
eas while we also attempt to retrofit conditions in already
developed watersheds.
Acknowledgments
The authors would like to thank Frank Piorko for provid-
ing assistance in selecting sites. Ellen Dickey provided
assistance in sample collection, data management, and
data analysis. Greg Mitchell and Terry Cole provided as-
sistance in sample collection. The Water Resources Agency
of New Castle County produced the estimates of land use.
References
Booth, D.B. and C.R. Jackson. 1994. "Urbanization of
Aquatic Systems - Degradation Thresholds and the
Limits of Degradation." American Water Resources As-
sociation Summer Symposium, Jackson, WY.
Delaware DNREC. 1994. "Habitat Quality of Nontidal
Streams in Delaware." in 1994 Delaware Watershed
Assessment Report, Vol I, Executive Summary, Ap-
pendix D; Division of Water Resources, Dover, DE.
Delaware Department of Natural Resources and Environ-
mental Control (DNREC). 1997. "Conservation Design
for Stormwater Management." DNREC Sediment and
Stormwater Program, Dover, DE, and The Brandywine
Conservancy, Environmental Management Center,
Chadds Ford, PA.
Garie, H.L. and A. Mclntosh. 1986. "Distribution of Benthic
Macroinvertebrates in Streams Exposed to Urban
Runoff." American Water Resources Association, Wa-
ter Resources Bulletin 22 (3); pp. 447-455.
Jones, R.C. and C.C.Clark. 1987. "Impact of Watershed
Urbanization on Stream Insect Communities." Ameri-
can Water Resources Association, Water Resources
Bulletin 23 (6).
Hollis, G.E. 1975. "The Effect of Urbanization on Floods of
Different Recurrence Intervals." Water Resources Re-
search 11 (3).; pp. 431-435.
Klein, R.D. 1979. "Urbanization and Stream Quality Im-
pairment." American Water Resources Association,
Water Resources Bulletin 15 (4).
Limburg, K.E. and R.E. Schmidt. 1990. "Patterns of Fish
Spawning in Hudson River Tributaries: Response to
an Urban Gradient." Ecology 71 (4), pp.1238-1245.
Maxted, J.R., E.L. Dickey, and G.M. Mitchell. 1995. "The
Water Quality Effects of Channelization in Coastal Plain
14
-------
Streams of Delaware." Delaware DNREC, Division of
Water Resources, Dover, DE; 21 pages.
Maxted, J.R. and E. Shaver. 1997. "The Use of Retention
Basins to Mitigate Stormwater on Aquatic Life." in Ef-
fects of Watershed Development and Management on
Aquatic Ecosystems, edited by L. A. Roesner; Ameri-
can Society of Civil Engineers, New York, NY; pp 494-
512.
Pedersen, E.R. and M.A. Perkins. 1986. "The Use of
Benthic Invertebrate Data for Evaluating Impacts of
Urban Runoff." Hydrobiologia 139, pp.13-22.
Plafkin, J.L., M.T. Barbour, K.D. Porter, S.K. Gross, and
R.M. Hughes. 1989. "Rapid Bioassessment Proto-
cols for Use in Streams and Rivers: Macroinvertebrates
and Fish." USEPA, Office of Water; EPA/444/4-89/001.
Shaver, E. and J. R. Maxted. 1995. "Watershed Protection
Using an Integrated Approach." in Stormwater NPDES
Related Monitoring Needs, edited by H.C. Torno,
American Society of Civil Engineers, New York, NY;
pp.435-459.
Schueler, T.R. 1994. "The Importance of Imperviousness."
in Watershed Protection Techniques 1 (3); edited by
T.R. Schueler, Center for Watershed Protection, Sil-
ver Springs, MD; pp. 100-111.
United States Census. 1996. from personal communica-
tions with Mike McGrath, Delaware Department of
Agriculture, Dover, DE.
United States Department of Agriculture. 1986. "Urban
Hydrology for Small Watersheds." Technical Release
No. 55 (2nd edition); Soil Conservation Service, Wash-
ington, D.C.
Urbonas, Ben R. 1995. "Parameters to Report with BMP
Monitoring Data." in Stormwater NPDES Related Moni-
toring Needs, edited by Harry C. Torno, American So-
ciety of Civil Engineers, New York, NY; pp. 306-328.
Weaver, L.A. and G.C. Carman. 1994. "Urbanization of a
Watershed and Historical Changes in a Stream Fish
Assemblage." Transactions of the American Fisheries
Society 123; pp. 162-172.
15
-------
Assessing the Status of Aquatic Life Designated Uses
in Urban and Suburban Watersheds
Chris O. Yoder and Robert J. Miltner
Ohio Environmental Protection Agency
Division of Surface Water Monitoring & Assessment Section
Columbus, Ohio
and
Dale White
Ohio EPA, Division of Surface Water
Information Resources Management Section
Columbus, Ohio
Introduction
The health and well-being of the aquatic biota in surface
waters is an important barometer of how effectively we are
achieving the goals of the Clean Water Act, namely the
maintenance and restoration of biological integrity and the
basic intent of water quality standards. States designate
water bodies for beneficial uses (termed designated uses)
that along with specific chemical, physical, and biological
criteria, assure the protection and restoration of aquatic
life, recreational, and water supply functions and attributes.
Ohio Environmental Protection Agency (EPA) employs bio-
logical, chemical, and physical monitoring and assessment
techniques to assess the status of these beneficial uses
and to satisfy three major objectives:
1) determine the extent to which use designations as-
signed in the Ohio Water Quality Standards (WQS)
are either attained or not attained;
2) determine if use designations assigned to a given
water body are appropriate and attainable; and,
3) determine if any changes in key ambient biological,
chemical, or physical indicators have taken place over
time.
An integrated biological, chemical, and physical moni-
toring and assessment approach has been used to sup-
port all relevant water quality management activities, in-
cluding urban stormwater issues, within Ohio EPA during
the past 18 years. The details of this process have been
extensively described elsewhere (Ohio EPA1987a,b; Ohio
EPA 1989a,b; Yoder and Rankin 1995, 1998).
Urban Watersheds
Urban watersheds in Ohio exhibit a familiar legacy of
aquatic resource degradation. Few, if any, ecologically
healthy watersheds exist in the older, most extensively
urbanized areas of Ohio (Yoder 1995) and no headwater
streams (i.e., draining <20 mi.2) sampled by Ohio EPAdur-
ing the past 18 years in these areas have exhibited full
attainment of the Warmwater Habitat (WWH) use desig-
nation (Yoder and Rankin 1997).
The activities that have the greatest impacts on aquatic
life in Ohio's urban watersheds include the wholesale al-
teration of watershed hydrology, loss and degradation of
riparian habitat, direct instream habitat degradation via
channelization, culverting, and interceptor sewer line place-
ment, excessive sedimentation resulting from land distur-
bance activities and stream bank erosion (strongly linked
to riparian encroachment), and contributions of excessive
nutrients, oxygen-demanding wastes, and toxic chemical
pollutants via urban runoff, point source discharges (both
permitted and unpermitted), and spills and other releases.
According to the 1996 Ohio Water Resource Inventory
(305[b] report), urban and suburban sources are respon-
sible for aquatic life use impairment in nearly 1000 miles
of Ohio streams and rivers and more than 23,000 acres of
lakes, ponds, and reservoirs (Ohio EPA 1997). These ac-
tivities also threaten existing full use attainment in nearly
160 miles of streams and rivers and may pose a potential
problem in more than 4380 miles of streams and rivers
that have not yet been fully monitored and evaluated. These
are also one of the fastest growing threats as urban and
suburban development extends further into rural water-
sheds.
While much attention has been paid to toxic substances
in urban runoff, evidence suggests that sedimentation is
the most pervasive single cause of impairment associated
with nonpoint sources in Ohio. While sediment deposition
in lotic and lentic environments is a natural process, it be-
comes a problem when the capability of the ecosystem to
16
-------
"assimilate" the sediment load is exceeded. The effects of
sediment on aquatic life are the most severe in the
ecoregions of Ohio where: (1) upland erosion and runoff
are moderate to high, (2) clayey silts that attach to and fill
the interstices between coarse substrates predominate,
and (3) streams and rivers lack the ability to expel the finer
grained sediments from the low-flow channel because of
instream and riparian habitat degradation. Estimates of
gross erosion alone are not consistently correlated with
adverse impacts to aquatic communities, although this is
a frequently used indicator for prioritizing nonpoint source
management efforts (Yoder 1995).
Bioassessment of Urban Watersheds
Ohio EPA uses biological criteria via a bioassessment
approach in the designation and assessment of rivers and
streams. Biological criteria are the principal tool for deter-
mining impairment of designated aquatic life uses and
bioassessments play a central role in the Ohio Nonpoint
Source Assessment (Ohio EPA 1990; 1991), the biennial
Ohio Water Resource Inventory (305b report; Ohio EPA
1997), and watershed-specific assessments of which Ohio
EPA completes from 6-12 each year. Biological criteria rep-
resent a measurable goal against which the effectiveness
of pollution control and other water quality management
efforts can be judged. However, biological assessments
must be accompanied by appropriate chemical/physical
measures, land use characterization, and source informa-
tion necessary to establish linkages between stressors and
the biological responses.
Methods And Analyses
For bioassessments to achieve their maximum effective
use in the assessment of urban streams, a watershed de-
sign to sampling and analysis should be employed. A re-
cent example is the Cuyahoga River basin in northeastern
Ohio and small, wadeable streams of the Columbus met-
ropolitan area (Franklin County) in central Ohio. The former
represents historically and extensively urbanized streams
including a mix of residential, commercial, and industrial
land use, streams draining recent and rapid suburban de-
velopment, and larger streams which are dominated by
point source effluents, principally treated municipal sew-
age. The latter case includes small watersheds affected
mostly by residential urban land use with a wide range of
intensity from older areas to recent and rapidly developed
suburban areas.
Biological and Water Quality Assessments
Fish and macroinvertebrates were sampled respectively,
at 82 and 48 locations, in the Cuyahoga River basin in
1996, and an additional 32 locations were sampled for
macroinvertebrates in 1991. Water samples were collected
up to six times at 40 macroinvertebrate sampling locations
and 63 fish sampling locations, and included standard field
parameters (D.O., temperature, pH, conductivity), nutrient
series (N and P), demand parameters (suspended solids,
BOD, COD), and selected heavy metals. Drainage areas
at Cuyahoga River basin stream sites ranged from approxi-
mately 2 to 700 mi2. Fish communities only were sampled
in the Columbus area, at 80 stream locations with drain-
age areas at all sites less than 35 mi2. No water chemistry
samples were collected. Macroinvertebrate community
performance was evaluated using the Invertebrate Com-
munity Index (ICI; DeShon, 1995). The ICI is a multimetric
index comprising ten attributes of community structure and
composition. The individual metrics were scored against
expectations derived from least-impacted reference sites
(Ohio EPA 1987b, 1989a; DeShon 1995; Yoder and Rankin
1995). Fish communities were sampled using generator-
powered, pulsed D.C. electrofishing units and a standard-
ized methodology (Ohio EPA 1987b, 1989b). Fish com-
munity attributes were collectively measured with the In-
dex of Biotic Integrity (IBI; Karr 1981; Karr et al., 1986)
modified for Ohio streams and rivers (Yoder and Rankin
1995; Ohio EPA 1987b). Habitat was assessed at all fish
sampling locations using the Qualitative Habitat Evalua-
tion Index (QHEI; Rankin 1989,1995). The QHEI is a quali-
tative, visual assessment of the functional aspects of
stream macrohabitats (e.g., amount and type of cover,
substrate quality and condition, riparian quality and width,
siltation, channel morphology, etc.).
Two indicators of urbanization were developed for the
Cuyahoga River basin, housing density and urban land
use cover. Housing density by Census Block Group was
obtained from the 1990 Census of Population (U.S. Bu-
reau of Census,1990). Urban land use cover was derived
from Landsat Thematic Mapper satellite imagery of land
cover classification (September 1994) provided by the Ohio
Department of Natural Resources. The number of hous-
ing units per hectare was calculated for the subwatershed
upstream from each fish and macroinvertebrate sampling
point to the boundary of the watershed. The percent urban
land use for subwatersheds upstream from the fish sam-
pling locations only were similarly calculated for both the
Cuyahoga Basin and Columbus area study areas.
Statistical Analyses
IBI scores were regressed against chemical water qual-
ity parameters, an index of habitat quality (QHEI), and
housing density. ICI scores were regressed against chemi-
cal water quality parameters and housing density. Water
quality parameters were expressed as the average con-
centrations of phosphorus, dissolved oxygen (D.O.),
nitrate+nitrite-nitrogen, ammonia-nitrogen, arsenic, lead,
and cadmium (macroinvertebrates only) based on grab
samples collected 6-8 times during June-October. Lead
was highly intercorrelated with zinc, copper and chromium.
Arsenic and cadmium were intercorrelated at fish sampling
locations. Transformations used to correct departures from
normality are provided in Table 1.
The relationship between different levels of urbaniza-
tion, as indicated by housing density or percent urban land
use (IBI only), and performance of the IBI, ICI, and se-
lected metrics was further quantified using an analysis of
variance model where quartile distributions of housing
density and percent urban land use (e.g., 1st quartile <
17
-------
Table 1. Parameter Estimates from the Regression of IBI on Water Quality Variables, Habitat Quality (QHEI) and Housing Density, and ICI on
Selected Water Quality Variables and Housing Density.
Index of Biotic Integrity (IBI)
N:63
Effect
Constant
L°g10(Ar)
Dissolved Oxygen
Log10(Pb)
1/NH3
QHEI
Log10(TP)
Log10(NOx)
(House/Hectare)
Multiple R: 0.606
Coefficient
23.318
5.123
0.549
3.997
-0.098
0.091
-7.876
-4.484
-7.171
Std Error
11.019
9.740
0.852
5.923
0.107
0.095
4.781
2.053
1.769
Squared multiple R: 0.368
t
2.116
0.526
0.644
0.675
-0.916
0.952
-1 .647
-2.184
-4.053
Adjusted R2:
P(2 Tail)
0.039
0.601
0.522
0.503
0.364
0.346
0.105
0.033
0.000
0.274
Adjusted R2
-0.011
0.006
0.022
-0.011
0.071
0.048
0.063
0.274
25th percentile of housing density, etc.) were used as fac-
tor levels. Metrics of the ICI that were used as dependent
variables included the number of Ephemeroptera,
Plecoptera and Trichoptera (EPT) taxa, the percent com-
position of mayflies, other dipterans/non-insects, and tol-
erant taxa. IBI metrics used included the percent compo-
sition of omnivores, tolerant fishes, sensitive fishes, and
insectivores. IBI scores and metrics from a subset of
samples in the Cuyahoga Basin with drainage areas less
than 100 mi2 were also analyzed according to percent ur-
ban land use in a similar manner to examine for potential
differences due to stream and watershed size. Because
sample sizes varied widely in the subsets, multiple com-
parisons were made using Sheffe's procedure (Neter et
al., 1991). An analysis of covariance model was constructed
for Columbus area streams using quartiles of percent ur-
ban land use as factor levels, QHEI as a covariate, and IBI
scores, percent composition of tolerant fishes, insectivores,
and omnivores, the number of darter and sculpin species,
and number of sensitive species as dependent variables.
Multiple comparisons were made using Tukey's procedure
(Neter etal., 1991).
Because Cuyahoga River basin streams are subject to
a variety of multiple stressors, fish sampling sites were
qualitatively classified by predominant impact type and
regressed against percent urban land use cover (Iog10
transformed) as a comparison to the results derived by
using housing density and to determine the influence of
impact type on the regression function. Impact types were
defined as least impacted, estate (i.e., subwatersheds with
large lot-size residential homes or green space provided
by parks), sites reflecting gross instream habitat alterations
(i.e., channel modifications or impoundment), sites im-
pacted directly by discharges from combined sewer over-
flows (CSOs), sites impacted by wastewatertreatment plant
discharges alone and with CSOs, sites with evidence of
impacts by legacy pollutants, or urbanization only. Regres-
sion coefficients from a subset of least-impacted, estate,
and urban-only sites with drainage areas less than 100
mi2 were compared to the same subset of sites for all drain-
age areas. Results of an ANOVA model using quartile dis-
tribution of percent land use as a factor level effect and IBI
scores as independent variables were compared to those
derived from the housing density model. Housing density,
as an indicator of the degree of urbanization, was further
evaluated by comparison with percent urban land use.
Housing Density and Biological Performance
When paired with chemical water quality data, housing
density explained approximately 27% and 59% of the varia-
tion in IBI and ICI scores in the Cuyahoga River basin (Table
1). Of the waterquality variables tested, only nitrate+nitrite-
nitrogen and ammonia-nitrogen explained a small, but sig-
nificant proportion of the variation in IBI and ICI scores
(=3% and 1%, respectively). For all IBI and ICI scores,
housing density accounted for 31% and 23% of the varia-
tion in scores. Multiple comparisons of factor levels based
on quartile distribution of housing density identified a thresh-
old level of urbanization, coinciding with 2.53 housing units
per hectare, beyond which IBI or ICI scores will increas-
ingly fail to attain the biological criteria for the warmwater
habitat use designation (Figure 1).
Shifts within the macroinvertebrate community were also
associated with a threshold level of urbanization (Figure
2). The number of EPT taxa were significantly higher at
the lowest levels of urbanization. Conversely, the percent
composition of pollution tolerant taxa collected from the
artificial substrate samplers increased sharply at sites ex-
ceeding the twenty-fifth percentile of housing density. Simi-
larly, the percent composition of other dipterans and non-
insects increased with increasing urbanization. The per-
cent composition of mayflies found on the artificial sub-
strates did not change with increasing level of urbaniza-
tion (Figure 2).
Shifts in the compositional metrics of the fish commu-
nity were associated with the degree of urbanization in the
Cuyahoga River basin (Table 2) and included an increase
in the relative abundance of tolerant and omnivorous fish.
The relative abundance of omnivorous fishes, however,
tended to be highest at intermediate levels of urbaniza-
tion, but differences were not statistically significant for the
subset of streams with drainage areas less than 100 mi2.
Insectivorous fishes were least abundant when housing
density exceeded the seventy-fifth percentile threshold.
18
-------
2nd 3rd
Housing Units (Quartile)
4th
60
50 -
40 -
£ 30 -
-------
100
80
c
s
JB
,o
60
40
20
1st 2nd 3rd 4th
Housing Units (Quartile)
IUU
8 80
C
° m
Z 60
S
-8
"5,
b 40
Ji
o 20
-s
o^
n
I
- I
'
I
I
1
!
1
1st 2nd 3rd 4th
Housing Units (Quartile)
D.
UJ
O
CD
J2
Z
35
30
25
20
15
10
5
0
O
8
1st 2nd 3rd 4th
Housing Units (Quartile)
100
80
60
40
20
1st 2nd 3rd 4th
Housing Units (Quartile)
Figure 2. Performance of four Invertebrate Community Index (ICI) metrics in relation to housing density for the Cuyahoga River basin. The level of
urbanization is given by quartiles of housing density per hectare of the subwatershed upstream from sampling locations. Horizontal lines
spanning adjacent box plots indicate similar means. Levels of housing density per hectare corresponding to the 25th, 50th and 75th
percentile are 2.53, 4.45 and 7.26 units/ha, respectively.
Urban Land Use and Biological Performance
The percentage of urban land use cover explained 26.7%
of the variation in IBI scores in the Cuyahoga River basin,
similar to that explained by housing density. When classi-
fied by quartile level of percent urban land use cover, the
mean of IBI scores in the first quartile was significantly
higher than those in the third or fourth quartile (Figure 3).
However, classification by percent urban land use cover
showed a more continuous decrease in mean IBI scores
with an increasing level of urbanization than did housing
density. Multiple comparisons of component IBI metrics
classified by level of urban land use cover showed similar
average responses to increasing urbanization as did clas-
sification by housing density (Table 2). However,
intraquartile variation of the metric responses was greater
among urban land use coverthan for housing density, lead-
ing to fewer significant differences between means and
reflecting the more continuous decrease in mean IBI re-
sponse with respect to percent urban land use cover.
Significant differences in mean IBI scores between the
levels of urban land use were also found for Columbus
area streams (Figure 3). Mean IBI scores from streams
with less than 3% urban land use were significantly higher
than those with greater than 33% urban land use (Figure
3). Shifts in the composition of the fish community associ-
ated with increasing percent urbanization included the loss
of darters, sculpins, and other pollution and habitat sensi-
tive species, decreased abundance of insectivores, and
an increase in the proportion of tolerant fishes (Table 3).
Discussion
Threshold levels of urbanization beyond which biologi-
cal communities are likely to be impaired have previously
been identified in the range of 8% to 20% impervious cover
within a watershed (Schuler 1994). The threshold levels in
our study of approximately 8% and 33% urban land use
cover for the Cuyahoga River basin and Columbus area
streams, as identified by analysis of variance, is in general
agreement with the studies reviewed by Schuler (1994).
20
-------
Table 2. Factor Level Means and Sheffe' Groupings for Selected Fish Community IBI Metrics Sampled in the Cuyahoga River basin in Relation to
Urban Land Use Indicators. Means Sharing a Common Letter are Not Significantly Different. The Asterisks Denote where Significant
Differences Between Groups were Not Detected in Multiple Comparisons for the Percent Tolerant Group from all Sites, and for the
Number of Sensitive Species in Streams Less than 100 Mi2, The Overall F Tests Indicated a Significant (P < 0.05) Linear Relationship.
Urban Indicator
(Quartile)
Number of
Sensitive Species
Percent as
Insectivores
Percent as
Tolerant
Percent as
Omnivores
1st
2nd
3rd
4th
22
21
19
21
A
AB
CB
C
All sites - Housing Units per Hectare
3.0
2.1
1.4
0.4
A
A
A
49.9
39.2
27.4
10.5
A
AB
AB
B
31.9
38.2
48.1
71.4
A
B
B
AB
15.1
28.3
48.4
22.1
All sites - Percent Urban Land Use
1st
2nd
3rd
4th
22
21
19
21
A
A
AB
B
2.5
2.2
1.4
0.7
A
A
B
B
49.5
41.4
18.6
14.6
A*
A
A
A
Drainage Area < 100 mi2 - Percent Urban Land Use
35.9
40.4
54.2
58.2
A
A
A
A
18.7
27.7
38.7
31.0
1st
2nd
3rd
4th
12
11
9
17
A*
A
A
A
2.5
2.0
0.8
0.6
A
A
B
B
44.8
40.8
13.2
10.5
A
A
A
A
46.5
44.3
66.1
69.7
A
A
A
A
22.7
20.4
11.5
24.9
o
'g
In
'o
60
50
40
30
20
12
Unimpacted, Estate and
Urban Only
Upper Cuyahoga Tribs
Breakneck Creek
-Nat. Rec. Area Tribs'
Impact Types
Unimpacted
Estate
Habitat
CSOs
WWTP+/-CSO
Urban
Legacy
"Little Cuyahoga & Tribs.
" Mill Creek, Big Creek, Tinkers Creek
~_ Cuyahoga River (Channelized, Reservoir '
I Release, Impounded, CSO, &WWTP Bypasses)
EE ;^-\A
0.1
1 10
Percent Urban Land Use
100
Figure 3. Distribution of Index of Biotic Integrity (IBI) values plotted by quartiles of percent urban land use cover upstream from sampling locations
for all sites in the Cuyahoga River basin, Cuyahoga basin sites with drainage areas less than 100 mi2, and Columbus area streams. The
shaded areas indicate the applicable biological criterion and the range of insignificant departure.
21
-------
Table 3. Factor Level Means and Tukey Groupings for Selected Components of Fish Communities Sampled in Columbus Area Streams. Means
Sharing a Common Letter are not Significantly Different. Samples Sizes in Ascending Order by Quartile are 20, 18, 20 and 20; Two
Cases in the Second Quartile had Missing Values Due to No Fish.
Percent Urban
by Quartile
1st
2nd
3rd
4th
Number of
Sensitive Species
A
A
B
B
3.3
3.8
0.9
0.6
Number of
Darters/Sculpins
A
AB
CB
C
3.1
2.2
1.1
0.6
Percent as
Insectivores
A
AB
B
B
40.8
32.0
17.0
19.7
Percent as
Tolerant
A
A
B
B
54.6
56.3
82.7
78.3
Percent as
Omnivores
A
A
A
A
13.0
13.9
10.5
5.1
However, the threshold level identified by regression for
the Cuyahoga River basin was influenced by the presence
of other stressors (e.g., CSOs, point sources, legacy pol-
lutants). The elimination of those sites impacted by other
stressors from the regression resulted in an increased
threshold of urbanization (Figure 4). Although other stres-
sors acted as covariates in a sense, these were not ame-
nable to an analysis of covariance because each occurred
in relatively discrete groupings along the continuum of in-
creasing urbanization. Analysis of variance was better able
to identify a threshold level by contrasting discrete ranges
(i.e., quartiles) along the entire range of increasing urban-
ization (Figure 3).
Similar patterns in the effect of increasing urbanization
on biological communities were evident for both the
Cuyahoga River basin and Columbus area streams.
Detectible differences in the number of sensitive fish spe-
cies in Columbus area streams occurred at lower levels of
urbanization than did IBI scores, illustrating the role of sen-
sitive species as sentinels of urban effects. Sensitive fishes
are rare in the Cuyahoga River basin as a whole due to
historic, complex, and widespread anthropogenic stressors,
yielding less response and higher variation associated with
interquartile means compared to the Columbus area
streams. However, the number of EPT taxa, a sensitive
macroinvertebrate guild, similarly acted as sentinels of ur-
banization given that EPT abundance was significantly
reduced at relatively low levels of urbanization. The abun-
dance of mayflies, showing little correlation with the level
of urbanization, did not respond in a manner similar to the
number of EPT taxa. While this may reflect the difference
in collection technique as percent mayflies are based on
the data from artificial substrates, whereas EPT taxa are
based on data collected from natural substrates, it may
also be due to differing sensitivities within the EPT guild.
This result, in combination with the response of the fish
community, implies that substrate degradation is a major
factor which limits aquatic communities at relatively low
levels of urbanization.
The relative abundance of omnivores tended to be high-
est at intermediate levels of urbanization when all sites in
the Cuyahoga Basin were included. This response was
due in part to enrichment by wastewater treatment plant
discharges and CSOs discharging to the Cuyahoga River
mainstem. No differences were detected for the subset of
streams with drainage areas less than 100 mi2, nor in the
Columbus area streams. However, the relative abundance
of insectivores was negatively correlated with increasing
urbanization in both study areas, suggesting a disruption
within the aquatic food web. Conversely, the proportion of
tolerant fishes was positively correlated with increasing
urbanization. The high proportion of tolerant fishes at the
highest levels of urbanization is indicative of both degraded
habitat and water quality, specifically toxicity and organic
enrichment. Collectively, these changes in biological com-
munities suggest a continuous negative response to in-
creasing urbanization starting with the loss of sensitive fish
and macroinvertebrate species at comparatively low lev-
els of urban development (<5% urban land use) due to
substrate degradation, disruption within the aquatic food
web at intermediate levels of development, and a response
to toxicity, organic enrichment, or both at higher levels of
development (>15% urban land use).
Overlaying impact types with percent urban land use
(Figure 4) demonstrates that the negative effects of ur-
banization and associated cofactors (e.g., imperviousness,
polluted runoff, altered hydrology) may be partially offset
by beneficial land use practices. Biological performance
at sites impacted by estate-type residential developments
remained comparatively intact and attained the ecoregion
biocriteria even at relatively high levels of urbanization (up
to 15%). The best performing sites within those watersheds
also had relatively intact stream habitat and we 11-vegetated,
wider riparian buffers. Conversely, sites with increasingly
modified habitats performed poorly and failed to attain the
biocriteria regardless of the degree of urbanization. The
most degraded sites were associated with either poorly
treated sewage, CSOs, and/or a high degree of urbaniza-
tion. These findings agree with those of Steedman (1988)
who demonstrated a co-relationship between riparian zone
quality and land use in terms of how each affected the fish
communities of Toronto area streams. Horner et al. (1997)
found the steepest rates of decline in biological function-
ing (in terms of the B-IBI; Kerans and Karr 1992) to occur
with increases in impervious cover of as little as 1-6% in
streams flowing into Puget Sound, Washington. Excep-
tions occurred where urban land use was mitigated by
extensive riparian protection or other management inter-
ventions, but these factors ceased to be effective above
45% as impervious land cover.
Unlike the Cuyahoga River basin, the Columbus area
streams were not subject to extensive CSO impacts and
22
-------
12 -
0.1
100
Percent Urban Land Use
Figure 4. Index of Biotic Integrity (IBI) scores from sites sampled in the Cuyahoga River basin plotted by stressor group (symbols) agains percent
of urban landuse for sites draining less than 100 mi2. The fitted regression lines are for all points and those lacking stressors other than
urbanization. The shaded areas indicate the applicable biological criterion and the range of insignificant departure.
industrial legacy pollutants were virtually absent. Conse-
quently, the threshold level of urbanization precluding at-
tainment of the biological criteria was higher for the Co-
lumbus area streams (Figure 5), results which are analo-
gous to that for sites influenced by the estate impact type
in the Cuyahoga River basin. In fact there were a few sites
with urban land use as high as 50% which fully attained
the ecoregional biocriterion. This suggests that the type of
urban development strongly influences the attainability of
aquatic life uses within a watershed. Furthermore, factors
such as impermeability and urbanization alone do not au-
tomatically disqualify streams from meeting designated
uses based on biological criteria.
Although housing density and percent urban land use
demonstrated a strong linear relationship (Figure 6), each
urban indicator showed somewhat differing results. The
percent of urban land use indicator, which is a more pre-
cise measure of urbanization and imperviousness, was
negatively correlated with biological community perfor-
mance. By comparison, the housing density indicator
showed a discrete threshold between the lowest quartile
and all others. The principal difference is that high-quality
sites were more frequently associated with the second
quartile of percent urban land use than for housing den-
sity, reflecting good IBI scores from relatively urbanized
subwatersheds containing large residential lot sizes and
more green space. Also, urban land use within successive
quartiles of housing density apparently becomes increas-
ingly mixed as inferred by increasing interquartile varia-
tion in percent urban land use (Figure 6). Higher levels of
housing density coincided with increased industrial, com-
mercial, and transportation related land uses. The differ-
ence in results by urban indicator underscores the impor-
tance of maintaining natural features within a watershed
including instream habitat, vegetated riparian buffers of
adequate width, and green space in addition to minimizing
and controlling chemical impacts from wastewater treat-
ment plants, CSOs, and other sources.
Implications for Use Attainability
Uses designated for specific water bodies are done so
with the expectation that the criteria associated with the
use are reasonably attainable. If CWA goal uses (e.g.,
warmwater habitat in Ohio) are found to be unattainable,
lower uses may be established and assigned on a case-
by-case basis. Federal water quality regulations (40CFR
Part 131.10[g]) generally specify three criteria for setting
designated uses below "fishable/swimmable" standards as
follows: 1) imposition of the criteria for a higher use would
result in widespread, adverse socioeconomic impacts; 2)
the criteria are not attainable due to natural background
conditions; or 3) the criteria are not attainable due to irre-
trievable, anthropogenic impacts.
Compliance with the aquatic life uses defined in the Ohio
WQS are determined primarily by the biological criteria
(OAC 3745-1-07) which are stratified according to desig-
nated use, ecoregion, and stream size. As such this repre-
23
-------
2
a
o
0)
I
"5,
-------
Quality Gradient of Aquatic Life Uses and Narrative
Descriptions of Biological Community Condition
Max.
> k
Index
Value
(IBI, ICI)
Min.
Exceptional
Vtermwater
Habitat (EWH)
"Exceptional"
Vyarmwater
Habitat (WWH)
Modified
Warmwater
Habitat (MWH)
Limited
Resource
Waters
(LRW)
Low
Biological Integrity
-»- High
Figure 6. Relationship between housing density and percent of urban land use cover for subwatersheds upstream from fish sampling locations in
the Cuyahoga River basin. The upper plot shows housing density as a function of percent urban landuse cover. The lower plot shows
distributions of percent urban land use cover within quartile levels of housing density per hectare.
sents a stratified system of uses and criteria that occur
along a gradient of biological integrity as expressed by the
biological indices which comprise the numerical biological
criteria (Figure 7). For most Ohio streams the "default"
expectation is attainment of the warmwater habitat (WWH)
use provided the physical habitat is relatively intact and no
extensive alterations are evident. Obvious anthropogenic
alterations to small urban streams such as culverting, re-
location, bank and channel stabilization with artificial struc-
tures, and extensive channelization are relatively easy to
identify and assess. In such cases, the Limited Resource
Waters (LRW) use designation is assigned which means
that the minimum level of protection (i.e., prevention of le-
thality) afforded by the Ohio WQS applies. The difficulty is
with small urban streams that exhibit adequate habitat (as
defined by the QHEI score), but which fail to attain the
WWH biocriteria. The recent finding that no urban head-
water stream sites in the Ohio EPA database attain the
WWH biocriteria (Yoder and Rankin 1997) only serves to
further the notion that the degree of watershed urbaniza-
tion can preclude the WWH use regardless of the site-
specific habitat quality.
Recently, the imperviousness of the watershed has been
used as an indicator which is correlated with use attain-
ability. If the frequently cited threshold of 25% imperme-
ability is used, streams in watersheds with greater than
this value would be unlikely to ever attain a beneficial use
regardless of site and reach factors. The results of ourstudy
suggest that there is a threshold of watershed urbaniza-
tion beyond which attainment of the WWH use is increas-
ingly unlikely. However, this threshold is different among
watersheds as evidenced by the results from the Cuyahoga
Basin and Columbus area streams. Co-occurring factors
such as pollutant loadings, watershed development his-
tory, chemical stressors, and watershed scale influences
such as the quality of the riparian buffer and the mosaic of
different types of land use also greatly influence the bio-
logical quality in the receiving streams.
While the development of indicators of watershed ur-
banization has merit from a management and decision-
making standpoint, there are simply too many other fac-
tors, some of which are controllable and amenable to
remediation, to use it as a sole determinant for aquatic life
attainability. We suggest that the co-factors in addition to
urban watershed indicators be better developed and tested
using datasets from broader geographic areas and span-
ning the extremes of the urbanization gradient. One goal
should be to develop, if appropriate, an urban stream habi-
tat designation that would fit along the already existing hi-
erarchy of aquatic life use designations in Ohio (Figure 7).
We have indicated on Figure 7 where the biological crite-
ria forthis potential new designation might occur com pa red
to the already existing hierarchy of aquatic life uses in the
Ohio WQS. However, placing it on the existing quality gra-
dient will require substantial calibration and validation with
existing datasets. Having this use would satisfy the desire
to afford streams with the maximum protection practicable,
while recognizing the inherent limitations that urbanization
imposes on stream quality.
In the meantime, simplistic regulatory and management
approaches should be limited, particularly in those water-
25
-------
Cuyahoga Basin Streams
50
m
•c An
-•-•
—
"o
o
<100 sq. mi.
-r
g
T
I 17
I 17
0% 14.7%
1st 2nd 3rd 4th
1st
2nd
3rd
4th
60
50
^ 40 •
Columbus Area Streams
o
In
30
20
12
T
•
. 20
20
3.3% .
1 T
1
1
20 20
. 11.4% 32.5% 1
1st 2nd 3rd 4th
Figure 7. Index of Biotic Integrity (IBI) scores from sites sampled in the Cuyahoga River basin plotted by stressor group (symbols) against percent
of urban land use for sites draining less than 100 mi2. The fitted regression lines are for all points and those lacking stressors other than
urbanization. The shaded areas indicate the applicable biological criterion and the range of insignificant departure.
26
-------
sheds where uncertainty about the attainability of CWA goal
uses (i.e., WWH and higher) exists. For example, initial
approaches such as the nine minimum controls for CSOs
seem reasonable. However, proceeding beyond these re-
quirements with long-term control plans should be done
cautiously and with the aid of sufficiently robust before-
and-after biological and water quality assessments.
The results of our study also point out the benefits of a
regular, sustained, and robust state monitoring and assess-
ment effort (see also Yoder and Rankin 1998). Dealing with
complex water quality management issues such as CSOs,
stormwater, and TMDLs in urban watersheds would be
difficult at best within the confines of the traditional admin-
istrative approach to water quality management. Steed man
(1988) described multimetric biological indices like the IBI
and ICI as being based on simple, definable ecological
relationships which is quantitative as an ordinal, if not lin-
ear, measure and which responds in an intuitively correct
manner to known environmental gradients. Further, when
incorporated with mapping, monitoring, and modeling in-
formation, such an approach has been shown to be valu-
able in determining management and restoration require-
ments forwarmwaterstreams (Steedman 1988; Bennet et
al., 1993). The value added by a robust bioassessment
and tiered use designation framework coupled with suffi-
ciently detailed and accurate CIS information was amply
demonstrated herein.
References
Bennet, M.R., J.W. Kleene, and V.O. Shanholtz. 1993. Total
Maximum Daily Load Nonpoint Source Allocation Pilot
Project. File Report, Dept. of Agricultural Engineering,
Blacksburg, VA. 44 pp.
DeShon, J.D. 1995. Development and application of the in-
vertebrate community index (ICI), pp. 217-243. in W.S.
Davis and T Simon (eds.). Biological Assessment and
Criteria: Tools for Risk-based Planning and Decision
Making. Lewis Publishers, Boca Raton, FL.
Horner, R.R., D.B. Booth, A. Azous, and C.W. May. 1997.
Watershed determinants of ecosystem functioning, pp.
251- 274. In C. Roesner (ed.). Effects of Watershed
Development and Management on Aquatic Ecosystems,
American Society of Civil Engineers, New York, NY.
Karr, J.R. 1981. Assessment of biotic integrity using fish com-
munities. Fisheries 6(6): 21-27.
Karr, J.R., K.D. Fausch, PL. Angermier, PR. Yant, and I.J.
Schlosser. 1986. Assessing biological integrity in run-
ning waters: a method and its rationale. Illinois Natural
History Survey Special Publication 5: 28 pp.
Kerans, B.L., and J.R. Karr, 1992. An evaluation of inverte-
brate attributes and a benthic index of biotic integrity for
Tennessee Valley rivers, Proc. 1991 Midwest Poll. Biol.
Conf., EPA905/R-92/003.
Neter J., W. Wasserman , and M.H. Kutner 1991. Applied
Linear Statistical Models: Regression, Analysis of Vari-
ance, and Experimental Designs. Irwin, Homewood, IL.
Ohio Environmental Protection Agency. 1997. Ohio Water
Resource Inventory, Volume I, Summary, Status, and
Trends. Rankin, E.T., C.O. Yoder, and D.A. Mishne,
(eds.), Ohio EPA Tech. Bull. MAS/1996-10-3-1, Divi-
sion of Surface Water, Columbus, OH. 190 pp.
Ohio Environmental Protection Agency. 1990. Ohio's
nonpoint source pollution assessment. Division of Wa-
ter Quality Planning and Assessment. Columbus, OH.
Ohio Environmental Protection Agency. 1989a. Biological
Criteria for the Protection of Aquatic Life. Volume III:
Standardized Biological Field Sampling and Labora-
tory Methods for Assessing Fish and Macroinvertebrate
Communities, Division of Water Quality Monitoring and
Assessment, Columbus, OH.
Ohio Environmental Protection Agency. 1989b. Addendum
to Biological Criteria for the Protection of Aquatic Life.
Volume II: Users Manual for Biological Field Assess-
ment of Ohio Surface Waters, Division of Water Qual-
ity Planning and Assessment, Surface Water Section,
Columbus, OH.
Ohio Environmental Protection Agency. 1987a. Biological
Criteria for the Protection of Aquatic Life: Volume I.
The Role of Biological Data in Water Quality Assess-
ment. Division of Water Quality Monitoring and Assess-
ment, Surface Water Section, Columbus, OH.
Ohio Environmental Protection Agency. 1987b. Biological
Criteria for the Protection of Aquatic Life: Volume II. Us-
ers Manual for Biological Field Assessment of Ohio Sur-
face Waters, Division of Water Quality Monitoring and
Assessment, Surface Water Section, Columbus, OH.
Rankin, E. T. 1995. The Use of Habitat Assessments in Wa-
ter Resource Management Programs, pp. 181-208. in
W. Davis and T. Simon (eds.). Biological Assessment
and Criteria: Tools for Water Resource Planning and De-
cision Making. Lewis Publishers, Boca Raton, FL.
Rankin, E.T 1989. The Qualitative Habitat Evaluation In-
dex (Qhei), Rationale, Methods, and Application, Ohio
EPA, Division of Water Quality Planning and Assess-
ment, Ecological Assessment Section, Columbus, OH.
Schuler, T.R. 1994. The Importance of Imperviousness.
Watershed Protection Techniques 1: 100 -111.
Steedman, R.J. 1988. Modification and Assessment of an
Index of Biotic Integrity to Quantify Stream Quality in
Southern Ontario. Canadian Journal of Fisheries and
Aquatic Sciences 45: 492 - 501.
U.S. Bureau of Census. 1990. 1990 Census of Popula-
tion. Government Printing Office. Washington, D.C.
Yoder, C.O. 1995. Incorporating Ecological Concepts and
Biological Criteria in the Assessment and Management
of Urban Nonpoint Source Pollution, pp. 183-197. in
D. Murray (ed.). National Conference on Urban Run-
off Management: Enhancing Urban Watershed Man-
27
-------
agement at the Local, County, and State Levels. EPA/
625/R-95/003.
Yoder, C.O. and E.T. Rankin. 1998. The Role of Biological
Indicators in a State Water Quality Management Pro-
cess. J. Env. Mon. Assess. 51(1-2): 61-88.
Yoder, C.O. and E.T. Rankin. 1997. Assessing the Condi-
tion and Status of Aquatic Life Designated Uses in Ur-
ban and Suburban Watersheds, pp. 201-227. in
Roesner, L.A. (ed.). Effects of Watershed Development
and Management on Aquatic Ecosystems, American
Society of Civil Engineers, New York, NY.
Yoder, C.O. and E.T. Rankin. 1995a. Biological Criteria Pro-
gram Development and Implementation in Ohio, pp.
109-144. in W Davis and T Simon (eds.). Biological
Assessment and Criteria: Tools for Water Resource
Planning and Decision Making. Lewis Publishers, Boca
Raton, FL.
28
-------
Tampa Bay Environmental Monitoring Program
Robert C. Brown
Environmental Management Department
Manatee County, Florida
Significant damage to, and loss of natural habitats in
Tampa Bay can be traced to the uncontrolled development
and pollution that started in the 1950s. Although great
strides have been made over the last decade to reverse
this trend, many agreed that a bay management and res-
toration plan, including monitoring programs that would
evaluate bay conditions and progress, were needed.
The nomination and designation of the Tampa Bay Na-
tional Estuary Program (TBNEP) in 1990, provided the plat-
form to assist the community in developing a comprehen-
sive plan to protect and restore the bay.
The process for developing the master plan includes the
following components: identify and rank priority problems;
assess bay conditions and needs; establish specific goals
for the bay; develop management options; prepare the
Comprehensive Conservation and Management Plan
(CCMP); develop an implementing agreement among bay
partners; implement the plan; and monitor progress
(TBNEP 1996).
Methodologies for bay assessment, goal setting and
development of comprehensive monitoring strategies are
described. These methodologies could be useful to others
interested in evaluating environmental protection and res-
toration schemes for natural resources.
Program Organization and Goal Setting
Since there already existed strong local and regional
involvement in bay management, the TBNEP built on this
commitment through the creation of its governing struc-
ture which consisted of the following committees: Policy,
Management, Technical Advisory (TAG) and Community
Advisory (CAC). From the Program's inception, the over-
all management goal was to protect and enhance the bay's
natural resources. In support of this goal, the committees
were required to characterize the natural systems of the
bay and the impacts to these systems and define and imple-
ment actions to address those impacts. A Management
Conference, with participation from all committees and
stakeholders, was convened to identify priority bay issues.
The following priority problems were identified by the Policy
Committee in March 1991 (TBNEP 1996):
•Water quality deterioration/eutrophication
•Reduction/alteration of living resources
•Lack of community awareness
•Increased user conflicts and impacts from various rec-
reational activities, industrial and navigation needs, and
urban development
•Lack of agency coordination and response
•Lack of circulation and flushing
•Hazardous/toxic contamination
Traditionally, monitoring and evaluating water quality
conditions within a watershed are used to measure water-
shed health and productivity. Vigorous bay monitoring and
management activities were being conducted by the time
the TBNEP began in 1991; however, many of these activi-
ties focused on individual components and/or processes
of the bay (Greening 1998). It was necessary to organize
the information, coordinate bay managers' and stakehold-
ers' participation and evaluate how these individual activi-
ties could be integrated to establish bay ecosystem man-
agement.
Quantifiable Restoration and Protection
Goals
In keeping with the overall goal of protecting and restor-
ing the bay's natural resources, the TAG worked to define
species or biological communities which could be used as
"indicators" of functioning bay ecosystems. The significant
loss of submerged aquatic vegetative ("seagrass") habitat
stood out as the premier concern of bay managers, scien-
tists and concerned public. This habitat is crucial for many
invertebrates and fish and provides for sediment stabiliza-
tion (Busby and Virstein 1993). If quantifiable seagrass
restoration goals and management strategies could be
developed and implemented, it would be feasible to de-
velop similar procedures for restoring other targeted habi-
tats and natural resources.
Quantitative targets for the restoration and protection of
seagrass habitat, as well as emergent habitats, were ap-
29
-------
proved at a Management Conference in 1993. The ap-
proach to habitat restoration and protection was as fol-
lows (Janicki et al., 1994):
1. Map the historic living resource distribution during a
benchmark time period.
2. Map the existing distribution of these living resources.
3. Overlay the historical and existing distributions to de-
fine potential restoration and protection targets.
4. Subtract physically altered (non-restorable) areas to
identify restoration targets.
Seagrass Restoration and Protection Goals
Utilizing the approach described above, it was deter-
mined that the benchmark for establishing seagrass pro-
tection and restoration goals would be the period circa
1950. This era was chosen because the area was begin-
ning to experience explosive growth and the major devel-
opment alterations were not yet complete. Additionally,
comparable habitat data were not available before 1950
(NUSCorp. 1986).
Using aerial photography coupled with the Arc/Info CIS
system, it was determined that the extent of seagrass cov-
erage in 1950 (not including areas that were irrevocably
altered by 1990) was estimated to be 40,400 acres (NUS
Corp. 1986). In 1990, Ries (1993) estimated the seagrass
habitat coverage to be approximately 25,200 acres. Hav-
ing already factored the physical losses due to dredge and
fill activities, the remaining losses were most likely caused
by degraded water quality conditions (Janicki et al., 1994).
Recent investigations suggest that the loss of seagrass
meadows can be attributed to lack of sufficient sunlight
because of attenuation by excess phytoplankton, sus-
pended solids and epiphytic algal growth (Morris and
Tomasko 1993; Tomasko 1993; and Stevenson et al.,
1993). Excessive algal concentrations or eutrophic condi-
tions are predominantly caused by excessive nutrient (e.g.,
nitrogen and phosphorous) loading.
Acreage goals for seagrass restoration and protection
were developed by overlaying the 1950, 1990 and non-
restorable acreage data sets. Seagrass areas observed in
1990 were designated as seagrass protection areas. All
areas in which seagrasses were mapped in 1950, but which
did not support seagrass in 1990 and were not classified
as non-restorable, were identified as seagrass restoration
areas (Greening 1998). Based on a review of the data
sources, method evaluation and uncertainty in estimating
the 1950 coverage, the Management Committee agreed
to adopt a minimum seagrass restoration goal of 38,000
acres bay-wide. This goal includes protection of an exist-
ing 25,650 acres and restoration of 12,350 additional acres.
Development of Intermediate Targets
Assessing bay management success via living resource
goals is considerably more difficult than using traditional
water quality criteria because it takes much longer to real-
ize results. It is not too difficult to evaluate annual water
quality trend response to management actions. It has been
demonstrated, however, that seagrass quality and quan-
tity improvements may not be observed for decades after
a management action is implemented (Johansson and Ries
1997). To ensure that correct management actions were
being implemented and bay water quality improvements
would lead to the achievement of the seagrass restoration
and protection goal, it was necessary to establish interme-
diate targets so that more timely evaluations and manage-
ment adjustments could be made if necessary.
In the Tampa Bay area it has been demonstrated that
seagrass health and distribution are adversely affected by
incident sunlight being attenuated within the water column
by elevated suspended solids or phytoplankton concen-
trations (Lewis etal., 1985; Lewis etal., 1991). If seagrass
does not receive adequate light, plant maintenance and
reproduction are inhibited (Janicki et al., 1994).
Forthe purpose of determining the relationship between
nutrient loadings to the bay and adequate water quality to
support the seagrass restoration target, a two-pronged
modeling approach was developed. The first was a series
of empirical regression-based models to estimate exter-
nal nutrient loadings consistent with the proposed seagrass
enhancements (Janicki and Wade 1996), and the second
was a WASP-based box model which provided a process-
oriented examination of relationships between nutrient
loadings, chlorophyll a concentration and light attenuation
(Martin et al., 1996; Morrison et al., 1997).
Both the empirical and mechanistic models produced
similar results, suggesting that acceptable nutrient man-
agement targets could be developed. The critical relation-
ships that were established were external nitrogen (limit-
ing nutrient) loads and resulting chlorophyll a concentra-
tions; chlorophyll a concentrations and density of phy-
toplankton in the water column; and chlorophyll a concen-
trations and light levels at the deep edges of historic
seagrass beds.
Since the estuary is about 1,031 km2 (398 mi2) with vary-
ing land uses, fresh water inflow, nutrient loadings and cir-
culation patterns, it was decided that the best way to man-
age this system was to partition or segment according to
similar conditions. The segmentation scheme defined by
Lewis and Whitman (1985) was adopted to establish the
official management subdivisions of the bay (Figure 1).
Following numerous scientific workshops, the TAG and
Management Committee adopted chlorophyll a targets
necessary to maintain water clarity needed for seagrass
growth for each bay segment. The adopted segment-spe-
cific annual average chlorophyll a targets (8.5 tvg/l for Old
Tampa Bay; 12.3 tvg/l for Hillsborough Bay; 7.4 tvg/l for
Middle Tampa Bay; and 4.6 tvg/l for Lower Tampa Bay) will
be used as indicators for evaluating water quality condi-
tions necessary to meet long-term seagrass restoration
and protection goals.
30
-------
N
Figure 1. Tampa Bay, Florida segmentation scheme.
31
-------
Nitrogen Management Strategy
Based on light conditions observed during a present day
period (1992-1994), it was determined that water quality
conditions were adequate to support the long-term
seagrass restoration goals; therefore, a nitrogen loading
"hold-the-line" strategy was adopted (Janicki and Wade
1996). This means that if the nitrogen loads observed dur-
ing the period 1992-1994 remained constant into the fu-
ture, it would be possible to achieve the seagrass restora-
tion goal. However, it is estimated that by the year 2010,
the watershed will experience a 20% increase in popula-
tion and approximately a 7% increase in annual nitrogen
loading (Zarbock et al., 1996).
In lieu of developing stringent future nitrogen load re-
duction allocations, local governments and agency part-
ners in the TBNEP developed an unprecedented interlocal
agreement (Memorandum of Understanding for the Fed-
eral agencies) pledging the development and implemen-
tation of action plans that will defer or reduce future nitro-
gen loadings, thereby maintaining the "hold-the-line" com-
mitment.
Monitoring and Reporting
The process fordeveloping monitoring strategies forthis
program was as unique as that used in developing the
living resource goals. There were many monitoring activi-
ties ongoing when the TBNEP program was established,
but these activities were localized and designed for spe-
cific needs.
The first task was to evaluate all of the different monitor-
ing programs being conducted for Tampa Bay to deter-
mine whether they would meet the monitoring criteria for
National Estuary Programs (USEPA 1991). Their criteria
include: "measuring the effectiveness of management ac-
tions and programs implemented under the CCMP and
providing essential information that can be used to redi-
rect and refocus the management plan." Additionally, a
1992 monitoring workshop recommended four additional
monitoring objectives (Versar 1992):
• To estimate the areal extent, and temporal trends in
areal extent, of habitat conditions in Tampa Bay not
meeting living resource requirements
• To assess the relative abundance and condition offish
populations of Tampa Bay overtime
• To estimate the areal extent and quality of seagrass,
mangroves, and coastal marshes in Tampa Bay over
time and
• To estimate the areal extent and trends in areal extent
of oligohaline habitat in Tampa Bay and its tributaries
To accomplish most of these monitoring objectives, it
was decided that a probability-based sampling design be
developed that would allow statistically valid, unbiased es-
timates of abundance and areal extent of key indicator spe-
cies on a bay-segment and bay-wide basis. The chosen
design was based on the U.S. Environmental Protection
Agency's (USEPA) Environmental Monitoring and Assess-
ment Program (EMAP) (Versar 1992). Since most of the
existing monitoring activities were biased, fixed station
designs, modifications to these programs were necessary.
In order to prepare and implement the new monitoring
strategies, the local and regional agencies responsible for
sample analyses and data reporting created a coalition
known as the Florida West Coast Regional Ambient Moni-
toring Program ("RAMP"). RAMP participants meet regu-
larly forthe purpose of standardizing methodologies, evalu-
ating quality assurance between laboratories, and coordi-
nating field sampling strategies. These coordinated activi-
ties have 1) allowed the local agencies to develop exper-
tise in areas other than general water quality monitoring
(e.g., benthic and seagrass monitoring); 2) economized
resources by linking bay areas and programs instead of
creating overlap; and 3) allowed utilization of the existing
EMAP probabilistic design to build monitoring programs
required by other regulations (i.e., NPDES stormwater).
Another very important component of the monitoring
strategy is reporting. The monitoring design described has
both short- and long-term targets and goals. In order to
provide bay resource managers timely information, the
TBNEP, with assistance from state, regional and local sci-
entists conducting monitoring and research, will prepare a
biennial Tampa Bay Environmental Monitoring Report. The
information provided in these reports is intended to pro-
vide decision makers timely access to information critical
for successful restoration and protection of Tampa Bay's
living resources.
Conclusions
The restoration and protection strategies designed for
Tampa Bay by local, regional, state and federal participants
epitomize coordinated ecosystem management. The de-
velopment of resource-based targets, as defined by the
environmental requirements of critical living resources, is
difficult but essential for maintaining the health and pro-
ductivity of critical habitats.
The real key to successes experienced in Tampa Bay is
the concerted effort put forth by agency personnel, elected
officials and concerned members of the public in dealing
with difficult, complex issues and making critical manage-
ment decisions. These accomplishments were possible be-
cause participants possessed dedication and commitment
to restoring and protecting the living resources of Tampa
Bay.
References
Busby, D.S. and R.W Virnstein. 1993. SAVI/PAR Execu-
tive Summary. Special Publication SJ93-SP13. Pro-
ceedings and Conclusions of Workshops On: Sub-
merged Aquatic Vegetation Initiative and Photosyn-
thetically Active Radiation. L.J. Morris and D.A.
Tomasko (eds.). St. Johns Water Management Dis-
trict.
32
-------
Greening, H.S. 1998. Tampa Bay Issues and Options: Pro-
tection and Restoration of the Bay's Living Resources.
In preparation.
Janicki, A.J. and D.L. Wade. 1996. Estimating Critical Ni-
trogen Loads for the Tampa Bay Estuary: An Empiri-
cally Based Approach to Setting Management Targets.
Technical Publication #06-96 of the Tampa Bay Na-
tional Estuary Program. Prepared by Coastal Environ-
mental, Inc.
Janicki, A.J., D.L. Wade and D.E. Robison. 1994. Habitat
Protection and Restoration Targets for Tampa Bay.
Technical Publication #07-93 of the Tampa Bay Na-
tional Estuary Program. Prepared by Coastal Environ-
mental, Inc.
Johansson, J.O.R. and T. Ries. 1997. Seagrass in Tampa
Bay: Historic Trends and Future Expectations. Pages
139-150 in Treat, S., (ed.). Proceedings, Tampa Bay
Area Scientific Information Symposium 3.1996 Oct.21 -
23; Clearwater, FL. 396 pp.
Lewis, R.R. Ill and R.L. Whitman, Jr. 1985. A New Geo-
graphic Description of the Boundaries and Subdivisions
of Tampa Bay. Pages 10-18 in S.F. Treat, J.L. Simon,
R.R. Lewis III, and R.L. Whitman, Jr. (eds.). Proceed-
ings, Tampa Bay Area Scientific Information Sympo-
sium.
Lewis, R.R. Ill, M.J. Durako, M.D. Mofflerand R.C. Phillips.
1985. Seagrass Meadows of Tampa Bay- A Review.
Pages 210-246 in S.F. Treat, J.L. Simon, R.R. Lewis
III, and R.L. Whitman, Jr. (eds.). Proceedings, Tampa
Bay Area Scientific Information Symposium.
Lewis, R.R. Ill, K.D. Haddad and J.O.R. Johansson. 1991.
Recent Areal Expansion of Seagrass Meadows in
Tampa Bay, Florida: Real Bay Improvements or
Drought-Induced? Pages 189-192 in S.F. Treat and
PA. Clark (eds.). Proceedings, Tampa Bay Area Sci-
entific Information Symposium 2.
Martin, J.L., P.F. Wang, T. Wool and G. Morrison. 1996. A
Mechanistic Management-Oriented Water Quality
Model for Tampa Bay. Final Report to the Surface Wa-
ter Improvement and Management (SWIM) Depart-
ment, SWFWMD, Tampa, FL.
Morris, L.J. and D.A. Tomasko (eds.) 1993. Special Publi-
cation SJ93-SP13. Proceedings and Conclusions of
Workshops on Submerged Aquatic Vegetation and
Photosynthetically Active Radiation. St. Johns Water
Management District.
Morrison, G., A.J. Janicki, D.L. Wade, J.L. Martin, G. Vargo
and J.O.R. Johansson. 1997. Estimated Nitrogen
Fluxes and Nitrogen-Chlorophyll Relationships in
Tampa Bay, 1985-1994. Pages 249-268 in Treat, S.
(ed.). Proceedings, Tampa Bay Scientific Information
Symposium 3. 1996 Oct. 21-23; Clearwater, FL. 396
pp.
NUS Corporation. 1986. Tampa Bay Estuarine Wetland
Trend Analysis. Final report submitted to the Tampa
Bay Regional Planning Council.
Ries, T. 1993. The Tampa Bay Experience. Pages 19-24
in Morris, L.J. and D.A. Tomasko (eds.). Special Pub-
lication SJ93-SP13. Proceedings and Conclusions of
Workshops on: Submerged Aquatic Vegetation and
Photosynthetically Active Radiation. St. Johns Water
Management District.
Stevenson, J.C., L.W Staverand K.W Staver. 1993. Wa-
ter Quality Associated with Survival of Submerged
Aquatic Vegetation along an Estuarine Gradient. Es-
tuaries. 16:346-361.
Tampa Bay National Estuary Program. 1996. Charting the
Course. The Comprehensive Conservation Manage-
ment Plan for Tampa Bay.
Tomasko, D.A. 1993. Assessment of Seagrass Habitats
and Water Quality in Sarasota Bay. Pages 25-35 in
Morris, L.J. and D.A. Tomasko (eds.). Special Publi-
cation SJ93-SP13. Proceedings and Conclusions of
Workshops on: Submerged Aquatic Vegetation and
Photosynthetically Active Radiation. St. Johns Water
Management District.
USEPA. 1991. National Estuary Program: Monitoring Guid-
ance Document. U.S. Environmental Protection
Agency, Office of Water. EPA 503/8-91/002.
Versar. 1992. Design of Basinwide Monitoring Program for
the Tampa Bay Estuary. Technical Publication #09-92
of the Tampa Bay National Estuary Program. Prepared
by Versar, Inc. and Coastal Environmental Services,
Inc.
Zarbock, H.W, A.J. Janicki, D.L. Wade and R.J. Pribble.
1996. Model-Based Estimates of Total Nitrogen Load-
ing to Tampa Bay. Technical Publication #05-96 of the
Tampa Bay National Estuary Program. Prepared by
Coastal Environmental Services, Inc.
33
-------
Retrofit Opportunities for Urban Waters Using
Soil Bioengineering
Robbin B. Sotir
Robbin B. Sotir & Associates, Inc.
Marietta, Georgia
Abstract
Soil bioengineering retrofits have been used to meet
specific objectives, such as flood control, stormwater man-
agement, bank stabilization, aesthetics and habitat en-
hancement. In response to increasing environmental con-
cerns, soil bioengineering systems using woody vegeta-
tion provide streambank protection in urban waters while
maximizing ecological and water quality benefits in urban
waters. Retrofit opportunities to restore incised and en-
larged channels as well as those that have been relocated
and/or straightened are discussed in this paper.
Case studies are presented to illustrate the use of this
technology on several projects where geotechnical, hydro-
logic/hydraulic, and environmental objectives needed to
be met. These included a flood control stream in an urban
linear park in Charlotte, North Carolina; a relocated stream
in Portland, Oregon; a flood control channel through a resi-
dential neighborhood in Houston, Texas; and an erosion
and flood control stream restoration project in a residential
neighborhood in Wilmington, North Carolina. Information
is presented for evaluating alternative soil bioengineering
streambank protection measures and selecting those that
best achieve the desired goals
Introduction
Nonpoint source impacts on urban streams and adja-
cent lands have become increasingly damaging in many
U.S. metropolitan areas, especially where streams have
been straightened or relocated. Straightening of urban
streams is often done to consolidate land for development,
or to bypass lakes or public parklands. Uncontrolled
stormwater runoff is also a significant contributor to the
erosion of channel beds and banks of urban streams. The
increased frequency and magnitude of these flows and
their associated velocities results in stream damage and
loss of valuable land.
Stream power is proportional to the product of discharge
and slope. Both variables are increased when streams are
straightened during development and subjected to in-
creased discharges. Stormwater detention is a relatively
recent concept, and most of the subdivisions and com-
mercial sites developed during the 1960s and 1970s have
no stormwater detention facilities (Nunnally and Sotir 1997).
Traditional land "repackaging" for convenience and short-
term resale ignore not only the sediment transport that
occurs during the construction phase and long-term chan-
nel stabilization problems; they also ignore water quality,
aesthetic values and a host of environmental benefits in-
cluding aquatic and terrestrial habitat. The opportunity of
retrofit recovery through the use of soil bioengineering is
discussed in this paper using case studies.
Geotechnical, Hydraulic and Hydrologic
Benefits of Soil Bioengineering
Soil bioengineering methods offer a broad range of me-
chanical benefits when installed as retrofits to damaged
urban stream systems (Table 1). Geotechnically, they of-
fer immediate soil reinforcement up to a depth of 12 feet.
The use of brushlayers with natural or synthetic geogrids,
is especially useful where space is constrained, as these
methods may be constructed on very steep slopes (1H to
1V; in some cases as steep as .25H to 1V). Installed veg-
etation offers many hydrologic values in that the embed-
ded branches serve as horizontal drains converting paral-
lel seepage flow to vertical flow, thus offering improved
overall slope stability. Surface protection and reinforcement
is further increased when live branches develop roots and
top growth. The roots tend to consolidate the soil particles
by reinforcing the soil mantle, reducing the possibility of
slips and displacements. The top leaf and branch growth
provides direct bank protection and reduces velocities while
redirecting the flow away from the bank.
Environmental Benefits of Soil
Bioengineering
Soil bioengineering offers a variety of environmentally
sound retrofit opportunities for urban waterways. The main
benefits of different methods have been summarized in
Table 2. They may serve as a useful guide in the selection
of specific vegetative methods. Soil bioengineering sys-
tems that use live siltation construction or create scour
34
-------
Table 1. Soil Bioengineering System Geotechnical and Hydraulic Goals and Benefits
Soil Bioengineering Methods
Goals
and
Benefits
Geotechnical
Hydraulic
*Live
Stakes
fair
fair to
good
Live
Fascine
good to
very good
good to
very good
Live
Siltation
Construc-
tion
n/a
very good
to excellent
Branch-
packing
good
excellent
Brush-
mattress
n/a
very good
Live
Cribwalls
excellent
good to
very good
Vegetated
Geogrid
excellent
good
Live
Boom(s)
n/a
fair to
good
'After established (1 year).
Table 2. Environmental Benefits of Soil Bioengineering Streambank Restoration Systems
Soil Bioengineering Methods
Goals
and
Benefits
Shade and
Vegetated
Geogrid
excellent
Live
Cribwall
excellent
Live
Boom
very good
Live Siltation
Construction
excellent
Brush-
mattress
good to
very good
Live
Fascine
good
Live
Stake
fair
to good
Create or
Preserve Scour good
Riparian fair to
Habitat good
'Recreation very good
very good
fair to
good
very good
excellent
n/a
n/a
n/a
very good to
excellent
good to
very good
n/a
excellent
good to
very good
n/a
good to
very good
good
n/a
fair
to good
fair
'Visual perspective
holes using live booms (dikes composed of woody veg-
etation and soil) are excellent choices as part of a bank
protection system where shade and overhanging cover or
pool habitat is desirable (Sotir 1997c). The habitat for mam-
mals and birds will improve over time in such areas for
nesting, migration and cover.
In addition to the selection and orientation of methods,
the choice of woody vegetation species in soil bioengineer-
ing systems can also have a significant effect on the habi-
tat benefits. Various species of willow are the most com-
monly used woody plants because of their excellent root-
ing capabilities, good overhanging cover and shade, good
nesting habitat for some species of birds, and some cover
for mammals, other species offer better food sources for
land animals. Soil bioengineering designs incorporate
plants that provide the best habitat benefits for target spe-
cies (Sotir 1997c).
Johnson Creek Relocation and Restoration
Johnson Creek is located in a highly urbanized area of
Portland (Oregon), with land uses ranging from heavy in-
dustry to low-density residential. It is a third-order stream
with a 100-year discharge at the project site of about 4,400
cfs. Flood control efforts during the 1930s enlarged, but
did not straighten the stream. A survey of Johnson Creek
revealed that with few exceptions, streambanks are stable,
heavily vegetated, and provide excellent riparian habitat
and overhanging cover for the stream.
The Oregon Department of Transportation (ODOT) pro-
posed relocating a section of Johnson Creek in the Town
of Milwaukie for bridge and highway construction (Figure
1). The relocated section would be about 20% shorterthan
the existing channel with a commensurate increase in gra-
dient. The Johnson Creek Corridor Committee, created
because of concerns over degraded water quality and
aquatic habitat and with an interest in restoring an anadro-
mous fishery, was worried about potential impacts of the
stream relocation. The relocated stream reach is in a highly
visible location, and the riprap channel proposed by ODOT
would present a stark, sterile appearance and cause fur-
ther loss of habitat and aesthetic value.
Robbin B. Sotir & Associates, Inc. (RBSA) was retained
by ODOT to evaluate the proposed channel design for sta-
bility and for potential impacts to aquatic and riparian eco-
systems and to modify the design as needed to address
the concerns voiced by the Johnson Creek Corridor Com-
mittee. The review determined that the proposed trapezoi-
dal channel cross-section shape and gradient were too
uniform and that the floodplain berms were too high. RBSA
35
-------
Figure 1. Johnson Creek after realignment activities.
recommended changes to the channel to improve stabil-
ity, water quality, and habitat value (Sotir and Nunnally
1995). The channel cross-section was altered by lowering
floodplain berms, incorporating a sub-channel to convey
bank overflows, and constructing a low-flow channel to
concentrate flows during the summer months. A pool-riffle
sequence was created by widening the sub-channel and
raising the invert by one foot in cross-over reaches and
lowering the invert by one foot in outside meander sec-
tions.
Streambanks were riprapped to the ordinary high-water
elevation in the outside bends. Banks above were soil
bioengineered, using vegetated geogrids. Siltation con-
structions using live materials were installed on the lowest
floodplain berm adjacent to the sub-channel to provide
cover for waterfowl and overhanging cover for fish. The
upper bank was protected with brushmattress.
The soil bioengineering systems were installed in the
winter of 1993 and spring of 1994. During the early spring
and before the plants had established growth, the site ex-
perienced a 1,750 cfs flood with mean velocities of 6-7-
feet-per-second and maximum velocities estimated in ex-
cess of 10-feet-per-second. The soil bioengineering sys-
tems were secure, and by the end of the growing season
they were providing excellent bank protection and habitat
benefits (Figure 2).
Buffalo Bayou Bank Stabilization and
Aesthetic Improvement
Buffalo Bayou upstream of Sheperd Drive is the only
stream of any size in Houston (Texas) that has not been
channelized for flood controls. The watershed of Buffalo
Bayou is almost totally urbanized, and Addicks Reservoir
was constructed upstream to alleviate flooding. The com-
bination of natural flooding and operation of the flood gates
at Addicks results in abrupt rise and fall of the water level
in the bayou coupled with prolonged periods of both high-
and low-water levels. These hydrologic conditions, com-
bined with sandy and silty soils with little cohesion, have
resulted in widespread erosion and large streambank fail-
ures.
Several soil bioengineering streambank protection
projects were built on Buffalo Bayou between 1990 and
1995 (Nunnally and Sotir 1995; Gray and Sotir 1996). The
1990 sites survived one of the largest floods of record in
1992 without damage. The installation described here was
constructed in 1992-93 following that flood.
The project site, located in an outside bend, is 280 feet
long and its height varies from 25-35 feet. Due to the re-
ceding bank, over 20 feet of land had been lost (Figure 3).
The bank recession was caused by a combination of mass
slope failure and streambank erosion. The instability of the
steepened slope was aggravated by the presence of fine
sands and seepage of 200-2,000 gallons per day from the
bank face. While the main goal for this project was to sta-
bilize the bank and stop the erosion, the client was also
interested in the restoration of the riparian zone, aesthetic
improvements, and the ability to maintain a view to the
Bayou.
To achieve long-term bank stabilization, a foundation of
wrapped concrete rubble was installed in a 7-foot deep
toe trench. A fill slope with a grade of 0.5 H:1Vwas recon-
structed above this foundation. The fill was constructed in
2-foot lifts wrapped with a geogrid. Thick layers of brush
long enough to extend from the undisturbed soil at the back
36
-------
Figure 2. Johnson Creek four years after construction.
Figure 3. Buffalo Bayou erosional failure after a flood event.
37
-------
of the bench and protrude several feet beyond the slope
face were placed between each wrapped soil layer. The
overall constructed height was 42 feet with the upper half
being at 0.251-1:1 V. Because continued seepage would have
substantially reduced the safety factor, it was necessary
to install vertical chimney drain construction to conduct the
water into a gravel trench drain that discharged into the
bayou. Since construction, the site has experienced sev-
eral floods and has remained stable; meanwhile, the in-
stallation is developing into a dense riparian buffer of na-
tive and naturalized species (Figure 4).
Little Sugar Creek Stabilization, Habitat
Restoration & Flood Control
This 4,650-foot section of Little Sugar Creek is in the
Huntington Farms Park area in the City of Charlotte (North
Carolina). This linear park, located along the creek in a
predominately residential neighborhood, is owned by the
City of Charlotte and maintained by Mecklenburg County
Storm Water Services.
Like most other streams in Charlotte, Little Sugar Creek
was channelized to improve drainage in the early 1900s,
and it has been dredged and snagged several times since
then, often leaving the channel without any vegetative
cover. The stream drains much of eastern and central
Charlotte, and the watershed is highly urbanized. The fre-
quent flooding and high peak discharges caused signifi-
cant bank erosion and channel enlargement. The immedi-
ate area had also been used as a constructed landfill in
the past. At several locations, bank erosion had uncov-
ered construction debris burial sites containing tree trunks,
waste construction materials, and miscellaneous organics
(Figure 5).
An interdisciplinary team with expertise in hydrology,
surveying, geotechnical and aquatic science, fluvial geo-
morphology, and soil bioengineering was assembled to
develop a restoration and stabilization project. Goals in-
cluded: bank stability, erosion protection, aquatic habitat
enhancement, water quality and aesthetic improvement,
community education and economic savings.
Design and cost information studies were initiated in April
1996. Erosional bank failures along the creek were evalu-
ated, typed and matched with appropriate solutions. From
this, final plan and specification documents were prepared.
Initially, riprap rock was reduced or completely eliminated
along the toe. This dramatically reduced the project costs.
Soil bioengineering methods such as live fascines and
brushmattress were employed in different configurations
along the banks. Construction was completed in March
1997. Four months after installation, Little Sugar Creek
experienced a flood that exceeded the 100-year event. The
project sustained no damage. While this project is a very
new installation, it has become well-vegetated, offering
enhanced riparian benefits, overhanging cover, aesthetic
improvements and bank stability (Figure 6). The instream
habitat structures (current deflectors and rocks) have also
been performing well, producing a variety of scour hole
cover and resting areas for fish.
Figure 4. Buffalo Bayou five years after construction.
38
-------
Figure 5. Little Sugar Creek erosion failures before construction.
Figure 6. Little Sugar Creek in the first growing season after construction.
39
-------
Long Leaf Hills/Hewletts Creek
Stabilization, Aesthetic and Habitat
Enhancement
This stretch of Long Leaf Creek is located in a residen-
tial neighborhood in Wilmington (North Carolina) known
as Long Leaf Hills Subdivision. Increased stormwater run-
off due to urbanization of the watershed and frequent flood-
ing in the lower section have caused significant bank ero-
sion and channel enlargement. Bank seepage and uncon-
trolled overbank runoff also contributed to bank failure (Fig-
ure 7 and 8). The creek has been used as a dump site for
organic garden debris which kills the bank vegetation and
has worsened erosion. Public meetings focused commu-
nity concern on existing conditions and spurred interest in
stabilization and restoration based on ways that residents
wanted to use and enjoy the creek in the future.
Kimley-Horn & Associates, Inc., the prime consultant,
and Robbin B. Sotir & Associates, Inc. prepared six con-
ceptual alternatives which included a simple intermediate
action for cleanup and stabilization, grass, riprap rock and
concrete liners, box convert, and soil bioengineering. Al-
ternatives were matched against 11 critical issues (Table
3). Soil bioengineering was selected by the neighborhood
as it fulfilled all the criteria. The project is currently in the
final design stages. Construction is scheduled to start in
the fall of 1998 and is expected to be completed by late
winter of 1999. Monitoring will be performed after construc-
tion to evaluate the stabilization and restoration develop-
ment of Long Leaf Hills/Hewletts Creek (Sotir 1997a).
Summary
Urban water restoration and stabilization projects involve
multiple objectives. In addition to controlling erosion in a
cost-effective manner, we are increasingly concerned with
waterquality, habitat, aesthetics, recreational use and other
environmental objectives. Soil bioengineering designs that
employ woody vegetation meet these environmental ob-
jectives better than other types of streambank protection,
especially when integrated with othertechnology. The suc-
cessful retrofit applications of soil bioengineering on ur-
ban waters discussed in this paper indicate that this ap-
proach to stabilization and restoration is successful.
Figure 7. Hills/Hewletts Creek/Long Leaf failure conditions
40
-------
Figure 8. Long Leaf/Hills/Hewletts Creek failure conditions
Table 3. Long Leaf Hills/Hewletts Creek Alternatives and Critical Issues
Critical Issues
ALT. #1
Intermediate
Action
ALT. #2
3:1 Side
Slopes
Grass
Lining
ALT. #3
2:1 Side
Slopes
Riprap
Rock
ALT. #4
2:1 Side
Slopes
Concrete
Lining
ALT. #5
Reinforced
Box
Convert
ALT. #6
Soil
Bioengineering
Stop Erosion & Stabilize
Banks
Clean Out Trash & Debris
Remove Fallen Trees
Safer & Healthier Area
Control Flooding
Timely Project Completion
Environmental Improvement
Aesthetically Enhancing
Meets Bank Stability & Hydraulic
Efficiency
Minimize Property Loss
Financial Feasibility
n/a
n/a
n/a
Adapted from: Kimley-Horn & Associates (Sotir 1997a)
41
-------
References
Beak Consultants, Inc. and Robbin B. Sotir& Associates,
Fall 1996. Tacoma Street Interchange Soil Bioengi-
neering Project - Monitoring and Evaluation of Stabil-
ity and Habitat. Seasonal Monitoring Report Prepared
for Oregon Department of Transportation.
Gray, D.H. and Sotir, RB., 1996. BiotechnicalSoilBioengi-
neering Slope Stabilization: A Practical Guide for Ero-
sion Control. John Wiley & Sons, New York, NY.
Gray, D.H. and Sotir, R.B., 1995. Biotechnical Stabiliza-
tion of Steepened Slopes. Transportation Research
Board, NAS-NRC.
Nunnally, N.R. and Sotir, R.B., 1997. Criteria for Selection
and Placement of Woody Vegetation in Stabilization
Protection, Oxford, MS May pp. 816-821.
Sotir, R.B. and Nunnally, N.R. 1995. Soil Bioengineering
for Stream Restoration. Water Resources Engineer-
ing, Vol. I: 795-799.
Sotir, R.B., 1996. Soil Bioengineering for Slope Protection
and Restoration; 1996 ASAE Annual International
Meeting, Phoenix, AZ, July, Paper No. ASAE 962048.
Sotir, R.B. and Associates, Inc., Sept. 1997a. Reconnais-
sance and Conceptual Design - Long Leaf Drainage
Improvements Project Wilmington, NC. Proposed Soil
Bioengineering Stabilization and Restoration Report
Sotir, R.B., 1997b. Designing Soil Bioengineering
Streambank Protection for Multiple Objectives. Pro-
ceedings, Conference on Management of Landscape
Disturbed by Channel Incision Stabilization Rehabili-
tation Restoration, Oxford, MS, May, pp. 325-350.
Sotir, R.B., 1997c. Urban Watershed Restoration, National
Urban Forestry Conference, Atlanta, GA. Sept. 1997.
USDA/NRCS, Chapter 16: 1997 Streambank and Shore-
line Protection- Engineering Field Handbook.
USDA/NRCS, Chapter 18: 1992 Soil Bioengineering for
Upland Slope Protection and Erosion Reduction - En-
gineering Field Handbook.
42
-------
Restoration of the Waukegan River
Through Biotechnical Means
Scott Tomkins
Illinois Environmental Protection Agency
Don Roseboom
Illinois State Water Survey
Illinois Department of Natural Resources
Peoria, Illinois
Introduction
Many urban Illinois streams have been degraded as a
result of streambank erosion, increased urban runoff, and
increased channelization. Biotechnical stream stabilization
techniques (BSST - Structures added to vegetation) were
implemented on the Waukegan River to reduce the sedi-
ment load discharge to Lake Michigan originating from the
river's eroded streambanks. The Waukegan River is lo-
cated 30 miles northwest of Chicago, Illinois, in the City of
Waukegan, Illinois (Figure 1). Best management practices
(BMPs) were implemented on the Waukegan River in
Washington Park and Powell Park, located in Waukegan.
The Waukegan River Restoration Project was created to
demonstrate whether the biotechnical techniques utilized
Waukegan
Figure 1. V\teukegan River restoration project Section 319 National
Monitoring Project.
were an effective means of resolving streambank erosion.
The project was funded in part, by the United States Envi-
ronmental Protection Agency (U.S. EPA), under the Sec-
tion 319 Nonpoint Pollution Program of the Clean Water
Act.
At the selected severely eroded streambank sites, BSSTs
were a more cost-effective and environmentally sensitive
means of reducing nonpoint source (NPS) pollution than
traditional approaches (i.e., rip rap, concrete lining).
Biotechnical Designs
Installation of the first BSST occurred on the North
Branch of the Waukegan River in Powell Park and in Wash-
ington Park during the fall of 1991. Lunkers and A-Jack
structures were installed in Powell Park, while lunkers with
stone were installed in Washington Park (Figure 2). On
the two lunkers installations, vegetation (willows, dog-
woods, grasses, and other wetlands plants) were placed
into the lower, middle, and upper zones of the lunkers struc-
tures. The structures utilized were chosen to enhance in-
stream habitat and provide a structural base for riparian
revegetation of the bank. Advantages and disadvantages
of using lunkers with vegetation are listed in Table 1.
The next installations of BSSTs were on the South Branch
of the Waukegan River, in the fall of 1994, to control se-
verely eroded streambanks in Washington Park. To ad-
dress these eroded streambanks, lunkers, stone, dog-
woods, willows, and grasses were installed. Other BSSTs
that included coir coconut fiber rolls, willows, and grasses
were implemented to treat specific small streambank ero-
sion sites on the South Branch.
In the winter of 1996, seven low stone weirs (LSWs)
formed by granite boulders were installed to create a se-
ries of pool/riffle sequences to enhance in-stream habitat
on the Waukegan River. These LSWs were constructed to
help resolve a lack of water depth, limited cobble sub-
strates, and limited stream aeration in order to enhance
the aquatic community in the Waukegan River at Wash-
ington Park.
43
-------
Lunker Installation Design
A-Jack Design and Installation
Figure 2. Biotechnical designs.
Table 1. Advantages and Disadvantages of Using Lunkers with
Vegetation.
Advantages
Disadvantages
A. Provides greater public access
to stream.
B. Appearance of a natural stream
functioning in an urban par is
more appealling to the public.
C. Lower cost of installation.
D. Greater fishery benefits by
increasing aquatic habitat
forgamefish.
E. Maintenance operation require-
ments are revegetation, not
construction activities.
During first year, maintenance
and revegetation are critical to
project stability.
Labor for lunker construction
and installation is greater
than riprap bank protection.
Monitoring
The Illinois EPA and the Illinois Department of Natural
Resources are jointly monitoring the effectiveness of the
biotechnical streambank techniques implemented on the
Waukegan River. The U.S. EPA's National Monitoring Pro-
gram (NMP) documents environmental benefits resulting
from the BMPs implemented on the Waukegan River.
On the South Branch of the Waukegan River, protocols
of the NMP were followed to detail the response of the
stream fishery, the macroinvertebrate populations, and the
in-stream physical habitat. The environmental quality of
these three monitoring areas were evaluated utilizing the
Index of Biological Integrity (IBI) for fisheries, the
Macroinvertebrate Biotic Index (MBI) for benthic organisms,
and the Potential Index of Biologic Integrity (PIBI) for in-
stream habitat.
The monitoring plan divided the South Branch of the
Waukegan River stream reach (Figure 3) into an upstream
control (S2) and a downstream bank erosion site (S1) for
biotechnical stabilization and in-stream habitat enhance-
ment. This reach was chosen because no large ravine
system transported urban runoff onto the stream between
S1 and S2.
Between 1994 and 1997, the Waukegan River was moni-
tored three times per year, once each in the spring, sum-
mer, and fall seasons. The monitoring activity documented
44
-------
Washington Park
Water St.
i r
Figure 3. Map showing placement of erosion control techniques in the Waukegan River.
aquatic resources for one year before and one year after
the biotechnical streambank stabilization and in-stream
habitat enhancements were implemented. After the BSST
application, the number of game fish species observed at
S1 increased from four to five (Table 2). Following pool/
riffle construction in 1996, the number of S1 game fish
species increased to nine. Increased numbers of game
fish and pollution intolerant fish species following the addi-
tion of the pool/riffle stream reach resulted in the IBI in-
creasing from 26 to 35 (Table 3). The average number of
fish sampled at S1 increased from 37 to 191 following
lunker habitat enhancement and increased further, to 225,
with the addition of the pool/riffle series. The upstream con-
trol (S2) remained a limited aquatic resource during the
study period, with only 1-2 species present and an IBI of
28 or less during the entire monitoring period (Table 3).
The average number of fish sampled at S2 varied between
16 and 69.
In 1996, the MBI indicated poor water quality atS2, with
a value of 8.3 (Table 3), but better water quality at the S1
pool/riffle site, which remained in the non-limited classifi-
45
-------
Table 2. Comparison of the Fish Species and Abundance for S1 and S2, for 1994 to 1996
1994 1995 1996
S1 S2 S1 S2 S1 S2
Lunkers Riffles
Fish Species and Abundance
Game Fish
Coho
Bluegill
Largemouth bass
Longnose dace
Mottled sculpin
Fathead minnow
Creek chub
Golden shiner
White sucker
Pollutant Intolerant Fish
Black bullhead
Green sunfish
Mosquito fish
Goldfish
Brook stickleback
Ninespine stickleback
Threespine stickleback
No. of species
Abundance offish
1
4
1
1
27
1
1
1
8
37
4
2 64
8
2 17
24
13 20
1
53
3 8
17 191
2
9
12
44
2
4 16
8
2
7 28
3
8
4 2
1
1
3
54 84
4 16
69 225
1
15
2
16
Table 3. Comparison of the Mean Station Values of the Indices for S1
andS2, for 1994 to 1996
1994
S1
S2
1995
S1
S2
Lunkers
IBI
MBI
PIBI
25.82
6.64
41.51
22.18
7.26
41.93
25.33
6.26
41.93
26.00
6.31
41.79
1996
S1
S2
Riffles
34.67
6.99
41.34
28.00
8.26
41.65
cation with a value of 7.0, even with the same stream wa-
ters as S2. The MBI indicates that water quality did not
limit or degrade aquatic resources in 1994 or 1995 (Table
3).
Physical habitat evaluations found deeper pools at the
S1 station, while the S2 site remained very shallow. The
LSWs were designed to transport bedload and scour pools
during high flow events. PIBI scores remained constant
for all three years and for both the S1 and S2 sites, how-
ever, ranging between 41 and 42 (Table 3). The PIBI scores
are predicated on the absence of claypan or silt-mud sub-
strates, the percentage of pools, and stream width. The
S1 and S2 physical habitat had very little or no claypan
substrate initially, which limited the expected change in
PIBI.
A price comparison of the various types of bank stabili-
zation are given in Table 4. The construction and installa-
tion techniques of lunkers make them relatively easy to
use by volunteer citizens' groups. The relative costs of a
rectangular concrete channel design, a riprap channel, a
tri-lock channel, and lunker applications with vegetative
stabilization can be estimated. The cost of a concrete cul-
vert would include more design engineering support to
determine possible offsite flooding effects. The design
channel is 10 ft deep, 25 ft wide, and 300 ft long. The con-
crete channel would have a wall thickness of 10 inches.
This project demonstrated that BSSTs can be effective
for reducing streambank erosion, by enhancing bank sta-
bility, and improving in-stream habitats. Incorporation of
LSWs that created a pool/riffle series added to the in-stream
physical diversity and resulting increased biodiversity. The
project also demonstrated that LSWs are effective in in-
creasing water aeration.
Streambank restoration is only one important step in im-
proving the diversity of fish communities. LSWs provide
additional pool depth and in-stream stone habitat neces-
sary for higher quality fish communities in urban streams.
Table 4. Cost Per Foot of Various Applications
1. Lunker with vegetation $27 per linear ft
2. Riprap with geofabric $52 per linear ft
3. Tri-lock Channel $165 per linear ft
4. Concrete Channel $750 per linear ft
46
-------
References
Bertrand, W.A., R.L. Hite, and D.M. Day. 1996. "Biological
Stream Characterization (BSC): Biological Assessment
of Illinois Stream Quality through 1993." Report by the
Biological Streams Characterization Work Group.
IEPA/BOW/96-058.
Bovee, K.D. 1992. "A Guide to Habitat Analysis Using the
Instream Flow Incremental Methodology." Instream
Flow Information Paper No. 12.
Buchanan, T.J. and W. P. Somers. 1969. Discharge Mea-
surements at Gaging Stations: USGS Techniques, Wa-
ter-Resources Inventory, Book 3, Chapter A8, 65 p.
Fausch, K.D., J.R. Karr and P.R. Yant. 1984. Regional
Application of an Index of Biotic Integrity Based on
Stream Fish Communities. Transactions of the Ameri-
can Fisheries Society 113:39-55.
Gorman, O.T and J.R. Karr. 1978. Habitat Structure and
Stream Fish Communities. Ecology 57(3):507-515.
Hilsenhoff, WL. 1977. Use of Arthropods to Evaluate Wa-
ter Quality of Streams. Technical Bulletin 100: Wis-
consin Department of Natural Resources. Madison, Wl.
Hite, R.L. and WA. Bertrand. 1989. Biological Stream
Characterization (BSC): A Biological Assessment of
Illinois Stream Quality. Special Report #13 of the Illi-
nois State Water Plan Task Force. IEPA/AC/89-275.
Isom, B.C. 1978. "Benthic Macroinvertebrates." In Methods
for the Assessment and Prediction of Mineral Mining
Impacts on Aquatic Communities: A Review and Analy-
sis. U.S. Fish and Wildlife Service. FWS/OBS-78/30.
Keller, E.A. and WN. Melhorn. 1978. Rhythmic Spacing
and Origin of Pools and Riffles. Geological Society of
America Bulletin 89:723-730.
Lane, E.W 1974. Report of the Subcommittee on Sedi-
ment Terminology. Page 14 in WS. Platts et al. Meth-
ods for Evaluating Stream, Riparian and Biotic Condi-
tions. U.S. Department of Agriculture. Forest Service,
Intermountain Forest and Range Experimental Station.
Gen. Tech. Rep. INT-1 38, Ogden, UT.
Platts, W.S., WF. Megahan, and G.W Minshall. 1993.
Methods for Evaluating Stream, Riparian, and Biotic
Conditions. U.S. Department of Agriculture, Forest
Service, Intermountain Forest and Range Experimen-
tal Station. Gen. Tech. Rep. INT-138, Ogden, UT.
Resh, V.H. and J.D. Unzicker. 1975. Water Quality Moni-
toring and Aquatic Organisms: The Importance of Spe-
cies Identification. Journal of Water Pollution Control
47:9-19.
47
-------
Monitoring the Effectiveness of Urban Retrofit
BMPs and Stream Restoration
John Gallh
Metropolitan Washington Council of Governments
District of Columbia
As a part of a larger Anacostia watershed restoration
initiative, efforts have been underway since 1988 to re-
store Upper Sligo Creek. Over the past eight years, more
than $2.0 million dollars have been spent on the restora-
tion of Upper Sligo Creek and its environs. Upper Sligo
Creek is a degraded, third order, urban Piedmont stream
which flows through Montgomery County, Maryland. The
general restoration strategy has featured the comprehen-
sive employment of stormwater retrofits, instream habitat
restoration, riparian reforestation, wetland construction and
restoration, and native fish and amphibian reintroductions.
Extended detention wet pond/marsh systems were em-
ployed on the basis of their ability to reduce pollutant loads
and channel erosion and to create additional wildlife habi-
tat. A prototype parallel pipe storm drain system was addi-
tionally used to divert first-flush stormflows away from an
important feeder stream. Last, a wide variety of instream
habitat enhancement structures such as rootwads, stone
wing deflectors, boulder placement, log drop structures,
etc. were employed. Restoration work was performed in
Phase I and II of the three-phase project which covered the
1990-95 period. Biomonitoring offish and macroinvertebrates
was conducted before, during and after each construction
phase. Physical habitat, hydrological and chemical condi-
tions were monitored in Phase III. The number of estab-
lished fish species residing in Upper Sligo Creek has risen
from three species in 1988 to approximately 12 in 1997.
Monitoring results were used to determine general retrofit
effectiveness, adjust fish stocking strategies, document
recruitment success and critique the overall effort.
Introduction
Attempts to restore the once highly degraded Upper Sligo
Creek stream system exemplify the basic subwatershed
restoration approach being employed throughout much of
the urban, 400 kM2 Anacostia River watershed. The
completion of the Wheaton Branch stormwater manage-
ment (SWM) retrofit facility in June, 1990 marked the be-
1 Project manager and co-investigator with James D. Cummins, Interstate Commis-
sion on the Potomac River Basin and James B. Stribling, Tetra-Tech, Inc.
ginning of a three-phased restoration project. The major
objective of Phase I (1990-91) was to restore Wheaton
Branch, Upper Sligo Creek's largest and most severely
degraded tributary. The centerpiece of this effort was the
three-celled, wet extended detention Wheaton Branch
SWM pond/marsh. This SWM retrofit was designed to pro-
vide both a high level of water quality control and down-
stream channel erosion protection for a 326 ha (805 ac),
55% impervious catchment. Other major components com-
pleted under Phase I included restoration of 300 m of down-
stream aquatic habitat, the creation of two vernal pools for
amphibian breeding habitat, and riparian restoration along
a 350 m stream corridor.
Phase II (1992-94) restoration featured the completion
of the University Boulevard SWM retrofit (a companion,
two-celled wet extended detention pond/marsh). The SWM
facility provides similar water quality and quantity control
for a 162 ha (400 ac), 30% impervious drainage area. In
addition, Phase II included: selective physical aquatic habi-
tat restoration of approximately 7 km of the Upper Sligo
Creek mainstem, construction of a 300 m-long parallel pipe
stormflow diversion system along Flora Lane tributary, cre-
ation of a 0.1 ha marsh, riparian reforestation of 2 ha along
Sligo Creek and the systematic reintroduction of 17 native
fish species into Wheaton Branch, Flora Lane tributary and
the Sligo Creek mainstem. Physical aquatic habitat condi-
tions at 19 sub-project sites were enhanced via the em-
ployment of stone wing deflectors, boulder fields, rootwads,
placed rip-rap, log drops, streambank bioengineering and
cedar-tree brush bundles.
Because of the general lack of adequate surface
stormwater runoff storage sites, physical aquatic habitat
restoration of the Flora/Lane tributary necessitated a flow
diversion approach. The prototype flow-splitting system was
designed to divert stormflow generated from up to 90% of
all one-hour storm events. Peak one-hour discharge from
the 87 ha (216 ac), 50% impervious catchment is approxi-
mately 1.6 m3/s or 55 cfs.
Phase III (1994-95) included biological, physical habi-
tat, hydrological and stream and pond water chemistry
48
-------
evaluations. No major restoration construction work was
performed in Phase III.
Study Design
Between March, 1990 and June, 1995, macroinvertebrate
and fish monitoring was performed at a total of 10 sites to
help assess the success of stormwater retrofit and stream
restoration work on Sligo Creek's aquatic communities. Over
this six-year period, the number of sampling stations grew
from four in Phase I to eight in Phase II and finally, 10 in
Phase III. Of the 10 sites, four were located in the Sligo Creek
mainstem, two in the Flora Lane tributary, two within the re-
stored portion of Wheaton Branch, one in the unrestored
Woodside Park tributary and one in the SWM control com-
parison stream (i.e., Crabbs Branch, located in the neigh-
boring Rock Creek watershed). In addition, a similar head-
waters area of the neighboring, semi-rural Northwest Branch
served as the Piedmont reference stream. Upper Sligo Creek
restoration areas and monitoring station network are shown
in Figure 1.
Stream water quality grab sampling was conducted be-
tween May, 1994 and July, 1995 at the following sites:
Wheaton Branch - W131; Sligo Creek - SL2, SL3, SL4;
Flora Lane tributary FL1 and FL2; and Crabbs Branch -
C131. Paired baseflow and stormflow water samples for
laboratory analysis were collected at WB1 and FL1 be-
tween June, 1994 and July, 1995. Monthly pond water
column sampling of Wheaton Branch Pond No. 3 and the
Crabbs Branch SWM facility was performed between May
and November 1994.2 As part of Phase III, sediment grab
sampling was conducted at six locations: Wheaton Branch
Pond No. 3, WB1, SL2, Sligo Creek mainstern above Flora
Lane tributary, SL4 and FL1. Stream thermal regime char-
acterization via continuous temperature monitoring was
performed between May and November, 1994 at the fol-
lowing locations: SL2, W131, FL2, SL4 and CB1.
Methods
Macroinvertebrate sampling of riffle and pool habitats
was performed using a square foot Surber sampler and
long-handled D-frame net (595 micron mesh opening).
Three Surber samples and a single D-frame sample were
taken from riffle and pool areas, respectively. Specimens
were identified to the lowest practical taxonomic level. Five
metrics were calculated in the study: taxa richness,
Hilsenhoff Biotic Index, EPT, percent contribution of domi-
nant taxon and shredders/total.
Fish sampling was conducted via backpack
electrofishing. Sampling techniques followed procedures
present in Plafkin, et al. (1989) and as described in
2Pond characteristics - Wheaton Branch: D.A. = 326 ha; imperviousness = 55%;
permanent pool surface area 2.4 ha; bottom release design; maximum depth 1.75
m; constructed 1990; SAV absent; 24-36 hr ED control. Crabbs Branch: D.A. =
238 ha; imperviousness = 60%; permanent pool surface area = 3.1 ha; surface
release design; maximum depth 2.60 m; constructed 1983. SAV (Hydrilla) covers
approximately 75% of pond bottom; no formal ED.
Cummins (1989) and (1991). The Zippin (1956) three-pass
depletion method was used for fish population estimation.
In addition, one-pass electrofishing was performed to fur-
ther evaluate fish dispersion, taxa richness and recruit-
ment success in Upper Sligo Creek.
Spot baseflow and stormflow water quality readings were
made in the field using a Horiba U-10, multiprobe water
quality meter and a Hach TDS meter. Paired baseflow and
stormflow samples were collected for WSSC laboratory
analysis from Wheaton Branch and the Flora Lane tribu-
tary. Baseflow samples were collected by immersing a 20-L
polyethylene carboy in an undisturbed pool area. Stormflow
samples were collected using a modified suspended sedi-
ment sampler.
Pond water column samples were collected at estab-
lished representative surface, mid-level and bottom depths
using a 2.0-L Van Dorn sampler. At the Wheaton Branch
pond, one 4-L water sample was collected for laboratory
analysis at each of the following depths: 0.15, 0.61 and
1.22m.
An EPA priority pollutant scan of stream (pool) and pond
sediments was performed by first taking 8-L of fine sedi-
ment with a coring device. Samples from three-to-five lo-
cations at each site were composited and delivered to
Gascoyne Laboratories, Inc. for analysis.
Continuous stream temperature monitoring was accom-
plished through the systematic employment of Ryan
TempMentor recording thermograph thermometers.
Physical aquatic habitat conditions were visually evalu-
ated using both methods described in Barbourand Stribling
(1991) as well as the Rapid Stream Assessment Technique
(Galli, 1996).
Results
Stormwater Pond Influence
Both the Wheaton Branch and Crabbs Branch SWM fa-
cilities exerted a strong influence on downstream hydrol-
ogy, water chemistry, temperature, substrate particle size
and stream bioenergetics. As expected, water quality in
both ponds was typically highest at or near the surface
and declined with increasing depth (Table 1).
During the Phase III study period, Wheaton Branch's 1.1
rn release depth resulted in the periodic discharge of poorly
oxygenated water high in organic materials and fine sedi-
ments (note: the pond's outlet structure was slightly modi-
fied in 1996 resulting in a mid-level release). Previous find-
ings (Environmental Dynametrics, Inc., 1993) strongly sug-
gested that during stormflow conditions this subsurface
release functions as a siphoning device, effectively reduc-
ing the pond's overall pollutant removal efficiency. Of the
stream sites monitored in Phase III, dissolved oxygen (DO)
levels in Wheaton Branch were typically lower. DO con-
centrations there were below 5.0 mg/L on four out of the
30 sampling dates. The study's low stream DO reading
49
-------
Stream Codes
SL = Sligo Creek
WB = Wheaton Branch
FL = Flora Lane Tributary
WP = Woodside Park Tributary
( ,tlniversity Blvd.
Regional SWM Facility
W
Wheaton Brancri
egional SWM »
Facility
Legend:
Phase I Restoration Area
(Wheaton Branch)
Phase II Restoration Area
(Sligo Creek & Flora Lane)
Drainage Divide
Monitoring Station
SWM Facility
1 in.= 1.27km.
Figure 1. Upper Sligo Creek restoration and monitoring station network.
Table 1. Wheaton Branch and Crabbs Branch Pond Water Quality: June, 1994
Location &
Date
Wheaton
Br. Pond
No. 3
(6/22/94)
Crabbs Br.
SWM Pond
(6/23/94)
Air
Depth Temp.
(m/ft) (°C)
0.15/0.5 32.0
0.30/1 .0
0.61/2.0
0.91/3.0
1 .22/4.0
1 .45/4.5
0.15/0.5 32.0
0.30/1 .0
0.91/3.0
1 .22/4.0
1 .83/6.0
2.59/8.5
Water
Temp.
28.5
28.6
28.2
27.5
26.5
26.0
28.4
28.3
27.8
27.0
23.0
25.2
DO
(mg/L)
12.13
12.46
11.79
5.93
1.20
0.23
12.73
12.12
10.44
3.75
0.22
0.11
Field
PH
7.30
7.21
7.12
6.72
6.53
6.21
8.77
8.71
8.51
7.26
6.60
6.57
Cond.
(umhos/cm)
146
146
146
147
210
227
300
299
294
298
423
595
Turb.
(NTU)
16
14
18
17
34
34
11
14
13
14
139
87
Secchi
Depth
(m)
0.50
0.72
(2.87 mg/L) was recorded at both sites WB1 and W132 in
June,1994.
By comparison, Crabbs Branch's surface release design
resulted in the discharge of warmer, yet clearer and more
highly oxygenated water. The larger permanent pool sur-
face area and volume and presence of extensive stands
of Hydrilla verticillata (which cover approximately 75% of
the pond bottom) contributed to Crabbs Branch's appar-
ently better water quality performance.
Wheaton Branch and Flora Lane Tributary
Stormflow Chemistry
Compared to Wheaton Branch, stormflow total sus-
pended solids (TSS), total organic carbon (TOC) and bio-
chemical oxygen demand (BOD) were generally slightly
higher in Flora Lane. Median stormflow TSS, TOC and BOD
concentrations were as follows: Wheaton Branch - TSS
(20 mg/L), TOC (8 mg/L), BOD (9 mg/L); Flora Lane tribu-
tary - TSS (50 mg/L), TOC (10 mg/L), BOD(10mg/L). The
50
-------
median nitrate (NO3) concentration was three times higher
in Flora Lane (1.6 mg/L) than in Wheaton Branch (0.5 mg/
L). Stormflow copper concentration ranges were nearly
identical for both streams. The median stormflow copper
concentration for both Wheaton Branch and Flora Lane
was 20.0 ug/L. This median level was double that recorded
under baseflow conditions. Mean stormflow total hardness
concentrations for Wheaton Branch (80.3 mg/L CaCO3)
and Flora Lane (105.2 mg/L CaCO3) were also consider-
ably lower than under baseflow conditions.
Stream Sediment Chemistry
Results of the EPA priority pollutant scan revealed no
high or unusual concentrations of pollutants in the sampled
stream sediments and were deemed to not pose serious
environmental toxic risks. Not surprisingly, the majority of
contaminants found were associated with road runoff. For
all metals detected, higher concentrations occurred in the
Sligo Creek mainstern below the Flora Lane tributary
confluence than above. For example, lead concentrations
increased from 23 mg/kg above the Flora Lane confluence,
to 50 mg/kg below. This enrichment is likely associated
with the large volume of highway traffic and runoff from
Interstate 495 and Georgia Avenue (MD Rte 97), which
are conveyed via the Flora Lane tributary to Sligo Creek.
1994 Thermal Regime Characterization
Based on continuous water temperature monitoring re-
sults, the thermal regimes of the streams were generally
categorized, per Galli (1990), as follows: 1) Sligo Creek
mainstem - coolwater; 2) Crabbs Branch - coolwater bor-
dering on warm; and 3) Flora Lane tributary - coolwater
bordering on cold. Summer stream temperatures in all but
the Flora Lane tributary regularly exceeded temperature
levels considered optimal (i.e., less than 17-20° C) for many
stonefly, mayfly and caddisfly species (Gaufin and Nebeker,
1973; Ward and Stanford, 1979; Fraley, 1979). Compared
to Wheaton Branch, Crabbs Branch was typically 3-4° C
warmer. This condition remained operative throughout the
temperature monitoring period.
Physical Aquatic Habitat
Major aquatic habitat improvement occurred in Wheaton
Branch following restoration work in April, 1991. Prior to
this date, aquatic habitat at sites W131 and W132 was 49-
56% of reference stream conditions. Following restoration,
aquatic habitat at these two sites increased to 104-108%
of reference. Similar improvements were documented in
both the Flora Lane tributary and Sligo Creek mainstern
upon completion of habitat enhancement work in Febru-
ary, 1994. Marked reductions in embeddedness levels were
recorded throughout. Pre- and post- restoration
embeddedness levels in Flora Lane fell from approximately
85% to 40%.
Macroinvertebrates
From Phase I to Phase II, both the number of individuals
and number of taxa in Wheaton Branch and the Sligo Creek
mainstern downstream of Wheaton Branch increased by
approximately 50%. No discernible change was observed
in the Flora Lane tributary. For the restored stream sites,
the metric percent contribution of dominant taxon ranged
from approximately 67-93% in 1990 spring samples to
approximately 26-78% in 1995 (Table 2).
Fish
Between Phase I and III the number of established fish
species increased as follows: Wheaton Branch - three to
six; Flora Lane tributary - three to six; Sligo Creek mainstern
- three to nine. Follow up, one-pass electrofishing results
in 1996 and 1997 revealed that approximately 12 species
are now established in the Sligo Creek mainstem. By com-
parison, Crabbs Branch and the reference stream support,
12-15 and 16-17 species, respectively. As seen in Figure
2, Index of Biotic Integrity (IBI) scores for restored sites
SI-2, W131, WB2, SI-3, FL1, FI-2 and SI-4 all increased
between Phase I and III (i.e., generally from poor to poor/
fair). During Phase III, Crabbs Branch fish IBIs were con-
sistently in the fair/good category.
Discussion
Monitoring results confirmed that the Upper Sligo Creek
restoration produced several improvements in both bio-
logical and aquatic habitat conditions. These generally in-
cluded: increases in the number of macroinvertebrate in-
dividuals (hence, improved food base for resident fish);
reductions in percent contribution of dominant taxon; an
increase in the number of established fish species from
Table 2. Calculated Macroinvertebrate Metric Values: Spring 1995* (modified from Cummins, et al., 1997)
Monitoring
Site
SL1
SL2
SL3
SL4
WB1
WB2
FL1
FL2
WP1
CB1
Taxa
Richness
7
10
9
10
11
7
2
6
3
10
Hilsenhoff
Biotic Index
7.8
6.5
7.0
6.9
6.8
6.3
8.0
7.3
8.1
7.0
EPT
2
3
3
3
3
3
0
2
0
3
Percent
Dominant
Taxa
63
47
26
30
46
78
67
48
61
43
Shredders
(Total)
0
0.01
0
0.12
0.006
0
0
0.05
0
0.007
51
-------
Phase I
Phase
Phase
-------
Cummins, J.D. 1991. Nineteen Ninety Maryland Anacostia
River Basin Study. Part 2: Fisheries Rapid
Bioassessments. CPRB Report 91-2. Interstate Com-
mission on the Potomac River Basin, Rockville, MD.
Cummins, J.D., J.B. Stribling, and F.J. Galli. 1997. Biologi-
cal Responses to Stream Habitat Rehabilitation in Sligo
Creek Watershed, Maryland. Interstate Commission
on the Potomac River Basin, Rockville, MD.
Environmental Dynametrics, Inc. 1993. Automated Flow
and Water Quality Monitoring Program Final Report
1992. Prepared for Department of Environmental Pro-
tection, Montgomery County, MD.
Fraley, J.J. 1979. Effects of Elevated Stream Temperature
Below a Shallow Reservoir on a Cold Water
Macroinvertebrate Fauna. The Ecology of Regulated
Streams. J.V. Ward and J.A. Stanford (eds.) Plenum
Press, NY.
Galli, J. F. 1996. Appendix A, Final Technical Memoran-
dum: Rapid Stream Assessment Technique (RSAT)
Field Methods. Prepared for Montgomery County De-
partment of Environmental Protection. Metropolitan
Washington Council of Governments, Washington,
D.C.
Galli, F.J. 1990. Thermal Impacts Associated with Urban-
ization and Stormwater Management Best Manage-
ment Practices. Prepared for Maryland Department of
the Environment. Metropolitan Washington Council of
Governments, Washington, D.C.
Gaufin, A.R. and A.V. Nebeker. 1973. Water Quality Re-
quirements of Aquatic Insects. U.S. EPA 660/3-73/004.
Plafkin, J.L., M.T. Barbour, K.D. Porter, S.K. Gross, and
R.M. Hughes. 1989. Rapid Bioassessment Protocols
for Use in Streams and Rivers: Benthic
Macroinvertebrates and Fish. U.S. Environmental Pro-
tection Agency, Office of Water, Washington, D.C. EPA/
440/4-89/001 (previously published as EPA/444/4-89-
00).
Ward, J.V. and J.A. Stanford. 1979. Ecological Factors
Controlling Stream Zoobenthos with Emphasis on
Thermal Modification of Regulated Streams. The Ecol-
ogy of Regulated Streams. J.V. Ward and J.A. Stanford
(eds.) Plenum Press, NY.
53
-------
Urban Water Quality Monitoring and Assessment Approaches in Wisconsin
RogerBannerman
Wisconsin Department of Natural Resources
Madison, Wisconsin
The Wisconsin Department of Natural Resources is
implementing a long-term urban monitoring strategy de-
signed to support Wisconsin's Nonpoint Source and
Stormwater Permit Programs. The purpose of the moni-
toring is to help ensure that management programs will
improve the quality of Wisconsin's urban streams in the
most cost-effective manner. At the core of the strategy are
seven questions that we try to answer as part of preparing
every urban nonpoint source control plan. All of the plans
are enhanced with site-specific monitoring data forthe first
three questions, but the high cost of collecting monitoring
data for the remaining four questions limits the data col-
lection to special urban monitoring projects. Results are
available for 26 special monitoring projects. To improve
our answers to all of the questions, we are planning about
28 new special monitoring projects forthe next five years.
An urban runoff model and stormwater management manu-
als are used to transfer the results of special projects to
other urban watersheds.
One or more types of monitoring data are being collected
to answer each question. Biological data is collected in
every urban watershed to answer the first two questions
and part of question 3. The first three questions are: 1)
what are the designated uses of the streams, 2) what are
the problems in the stream, and 3) what are the pollutants
and/or habitat factors degrading the streams? Special
monitoring projects using biological, chemical and physi-
cal data attempt to answer the last four questions and part
of question 3. The part of question 3 answered by special
monitoring is what, if any, potentially toxic pollutants are
degrading the streams. The last four questions are: 4) what
are the sources of the pollutants, 5) what are the goals for
reducing the pollutants and changing other factors degrad-
ing the stream, 6) what management alternatives will
achieve the goals, and 7) what did the implementation of
practices improve in the streams?
Introduction
Urban runoff has degraded many of Wisconsin's streams
(Masterson and Bannerman, 1994; Simonson and Lyons,
1993). To improve and protect the quality of urban streams
and other resources, Wisconsin's State Legislature cre-
ated a Nonpoint Source Program in 1978. The program is
implemented through "priority watershed projects," which
include the preparation of a priority watershed plan. The
priority watershed plan assesses nonpoint and other
sources of water pollution and identifies best management
practices needed to achieve the designated uses of the
water resources.
The watershed plan guides the implementation of the
best management practices. The Wisconsin Department
of Natural Resources (WDNR) and the Department of Ag-
riculture, Trade, and Consumer Protection administer the
program, while the local units of government implement
the plan. State funds are provided to cost-share the imple-
mentation of the best management practices recom-
mended in the plan. Approximately 10 million is made avail-
able to municipalities every two years. Approximately 6
million is forthe installation of urban practices and the re-
maining 4 million dollars is used for such activities as de-
signing the practices, stormwater management plans, de-
veloping stormwater ordinances, and developing utility dis-
tricts.
A priority watershed plan could not be prepared without
the results from some type of urban monitoring activity.
Results of the monitoring help us make the best manage-
ment decisions to ensure that designated uses of the
streams are achieved forthe least cost possible. Monitor-
ing data is used to strengthen our confidence in such deci-
sions as identification of the pollutant sources and the se-
lection of management alternatives. In response to this
need for data, the WDNR has developed an urban moni-
toring strategy supported by two types of monitoring ac-
tivities.
The purpose of this paper is to describe the monitoring
strategy and some of the results from the monitoring ac-
tivities. Information from the Lincoln Creek Subwatershed
part of the Milwaukee South Priority Watershed Plan is
described as an example of how the monitoring data is
used to prepare the chapters in a plan. Future monitoring
activities proposed forthe strategy are also discussed.
54
-------
Seven Stormwater Management Questions
The urban monitoring strategy is based on answering
seven stormwater management questions (Table 1). Each
question is related to the information needed fora chapter
in the watershed plans. Chapter IV in the priority water-
shed plan for the Milwaukee River South Priority Water-
shed Project is entitled "Water Resources Conditions,
Nonpoint Sources and Water Resource Objectives"
(WDNR,1991). Designated uses and water resource ob-
jectives are usually the same. Results from questions 1 to
5 are needed to prepare this chapter. Answers to question
6 are helpful for making the management recommenda-
tions in Chapter V, which is entitled "Nonpoint Source Con-
trol Needs." The last question matches with the last chap-
ter, Chapter VIM, entitled "Water Quality Evaluation Moni-
toring." Titles and order of the chapters might vary between
plans, but they all cover the same types of information.
These same questions would probably apply to almost any
water resource management effort.
Table 1. Seven Stormwater Management Questions Used To Design
Urban Monitoring Activities.
Question
No.
Questions
1. What are the designated uses of the urban streams?
2. What are the problems in the stream?
3. What are the pollutants and/or habitat factors degrading
the urban streams?
4. What are the sources of the pollutants?
5. What are the goals for reducing pollutant loads and
changing other factors?
6. What management alternatives will achieve goals?
7. What did the implementation of the practices improve in
the urban streams?
All of the questions are important. The quality of the an-
swer to a question depends to some degree on how good
the answer is to the previous question. The order of the
questions is the order in which they are usually answered.
For example, selecting the best management practices
without identifying the sources of the pollutants increases
the risk of wasting money on the wrong practices.
Answers to the seven questions are also helpful in the
implementation of the U.S. Environmental Protection
Agency's Stormwater Permit Program. Priority watershed
plans are available for many metropolitan areas in Wis-
consin. Most municipalities required to have a stormwater
permit are in a priority watershed. So far, some of the per-
mit requirements overlap with the management actions
specified in the watershed plans.
Two Types of Monitoring Activities
Supporting the Urban Monitoring Strategy
Two types of monitoring activities are essential parts of
the urban monitoring strategy. Results from the two types
of monitoring activities are used to answer the seven man-
agement questions for all of the priority watershed projects.
One type of monitoring is done for every priority water-
shed project. Biological data is collected in every stream
to answer questions 1,2, and parts of 3. Information about
the designated uses, the problems in the urban stream,
and the reasons for any problems are very site-specific.
Parts of question 3 assessing the problems caused by
conventional pollutants and a degraded fish habitat are
included in the monitoring done for every stream.
There is no substitute for collecting good biological data
in every watershed project. In most cases, it cost about
$500 to collect the biological data in each urban stream.
One or two sites in every stream are selected for fish, fish
habitat, and macroinvertebrate sampling. The biological
sampling for questions 2 and 3 is always done during the
planning phase of a project, while the answer to question
1 is sometimes determined before the beginning of a wa-
tershed project.
The other type of monitoring provides data to answer
questions 4, 5, 6, 7, and the part of question 3 evaluating
the role of potentially toxic pollutants. These "special ur-
ban monitoring projects" collect data at a few selected sites
that are then extrapolated to other urban areas. Special
urban monitoring projects provide the kind of data needed
by every priority watershed project, but which would be
too expensive and time-consuming to collect for every
project. Concentrations of zinc measured in street runoff
in Madison, for example, is used to estimate zinc loadings
from streets in Milwaukee. Testing the effectiveness of a
best management practice in every priority watershed
project would not only be unnecessary, but would cost over
$100,000 to properly test each device. Results from three
special monitoring projects were used to help answer the
stormwater management questions for the Lincoln Creek
Subwatershed.
Answers to Seven Questions for Lincoln
Creek Subwatershed
Answers to the seven questions for the Lincoln Creek
Subwatershed provide an example of how the information
is presented in many of the priority watershed plans. Spe-
cial urban monitoring project data used for the Lincoln
Creek Subwatershed is also typical of the stormwater data
available to a number of the priority watershed plans pre-
pared before 1993.
Lincoln Creek Subwatershed is the largest urban
Subwatershed in the Milwaukee River South Watershed,
draining 12,600 acres (18.8 sq. miles). Information about
Lincoln Creek was collected as part of preparing the Mil-
waukee River South Priority Watershed Plan (WDNR,
1991). Residential land uses dominate this totally urban
Subwatershed. High density residential areas occupy 35%
of the Subwatershed, while 12% of the Subwatershed is
industrial. Lincoln Creek is almost entirely channelized with
about one-third of the channel being concrete lined.
Lincoln Creek Answers - Questions 1 to 3
As for all the urban streams in the Milwaukee South Pri-
ority Watershed Project, biological sampling was done in
55
-------
Lincoln Creekto answerquestions 1,2, and parts of ques-
tion 3 (Table 2). Electro-shocking offish was done to de-
velop an assessment of the fish community in Lincoln
Creek. The WDNR's stream system habitat rating form (Ball
1982) was used to characterize the habitat in the stream.
Thirteen factors, such as low flow, depth of pools, bank
vegetation protection, lower bank deposition, and charac-
teristics of bottom substrate and cover are ranked to de-
termine the quality of the habitat. Habitat data are very
important because they are a large factor in determining
the potential fish species, composition, abundance, and
age structure. Results from dissolved oxygen, tempera-
ture, pH, and bacteria surveys were combined with the
ranking from the Hilsenhoff Biotic Index (HBI) for
macroinvertebrates (Hilsenhoff 1982) to assess the water
quality of the stream. All of the environmental data is then
used to classify the stream.
Procedures developed by the WDNR are used to clas-
sify the urban streams for fish and other aquatic life (Ball,
1982; WDNR, 1995). Although the procedures are de-
signed to provide a legal classification of a stream, classi-
fications prepared for the priority watershed projects do
not follow all of the required steps and, therefore, carry no
legal authority. The stream use classes are (a) cold water
communities, (b) warm water sport fish communities,
Table 2. Answers to the Seven Stormwater Management Questions
for Lincoln Creek Sub-watershed (from WDNR, 1991)
Table 2.
Question
Continued
Answer to Questions
Question
Answer to Questions
1. What are the designated uses?
2. What are the problems in
the stream?
3. What are pollutants and/or
factors degrading the stream?
4. What are the sources of
pollutants?
Fishery Use'. Below Teutonic Ave.
- warm water sport fish; other
natural reaches - limited forage
fish; and concrete lined reaches -
limited aquatic life
Recreational Use: All reaches -
partial body contact
Fishery Use: Species diversity - 2
(Ref. stream -20)
Macroinvertebrates: Severely
impaired
Recreational Use: Partially
meeting use
Pollutants'. Sediment, potentially
toxic pollutants (eg. lead, zinc,
and copper) in water column and
bottom sediments, bacteria, and
low dissolved oxygen.
Factors'. Poor habitat, flashy
flows, and concrete lining.
Sediment Established urban
area - 29%; construction
sites - 64%; and streambanks
-7%.
Lead: High density residential
-33%; industrial-32%;
Multi-family residential -19%;
and commercial -14%
Safer/a and low dissolved oxygen
-no entry
(continued)
5. What are goals for reducing
pollutant loads and changing
other factors?
6. What management alterna-
tives will achieve the goals?
7. What did the implementation
of the practices improve in the
stream?
Flow Rate and Volume: Reduce
enough to control bank erosion
and scour.
Sediment: A 50% reduction in
sediment
Lead: 40% to meet acute toxicity
standards at outfalls and 50% to
meet chronic toxicity standards in
stream.
Sediment: Construction site
erosion controls designed to
reduce the sediment by 75%
Lead: Wet detention ponds or
their equivalent to control all of
the runoff from critical land uses
(industrial, commercial, freeways,
high density residential, and
multi-family residential).
Flow Rate and Volume: No entry
Develop long-term biological,
chemical, and physical monitor
ing program.
(c) warm water forage fish communities, (d) limited forage
fish communities, and (e) limited aquatic life. Recreational
stream use classifications are also defined. For the pur-
pose of designating fish and aquatic life uses, the biologist
must decide if the factors limiting the ability of a stream to
support certain uses are controllable or uncontrollable. If a
controllable factor, such as urban runoff, is limiting the uses
of a stream, the biologist can assume the urban runoff will
be controlled to some degree when deciding what the po-
tential uses of the stream should be. Although the proce-
dures provide more objectivity to the process of classify-
ing streams, professional judgement usually enters into
the final use class selection.
Not all of question 3 was answered by biological moni-
toring. Although problems caused by excessive sediment,
and sometimes high nutrient loadings, can be identified by
the fish habitat surveys, some grab samples of Lincoln
Creek water were used to identify the presence of poten-
tially toxic pollutants. We recommend raising public aware-
ness about the potential problems with toxic pollutants and
bacteria by collecting grab samples below a storm sewer
outfall during three different runoff events. These samples
should be analyzed for as many of the pollutants found in
stormwater as possible, such as heavy metals and fecal
coliform bacteria.
Lincoln Creek Answers - Questions 4 to 6
Lead is the potentially toxic pollutant that is assumed to
represent all of the other potentially toxic pollutants in the
answers to questions 4 through 6. Lead is also important
to part of the answer for question 3. Extensive stormwater
monitoring of eight storm sewer outfalls in the City of Mil-
waukee identified the types of potentially toxic pollutants
that might be in the stormwaterdischarging to Lincoln Creek
56
-------
(Bannerman, 1983). Results of this special urban monitor-
ing project were available for samples collected at high
density residential, medium density residential, commer-
cial strip, and shopping center monitoring sites. Total re-
coverable lead event mean concentrations exceeded the
acute toxicity standard (hardness of 100mg/l) of 170 ug/l
for 90% of the runoff events sampled at a commercial
landuse site (WDNR, 1989). Although the WDNR does not
currently regulate stormwater discharges using numeric
effluent limitations, acute and chronic toxicity standards
applied to point source discharges for industries and mu-
nicipalities are useful to characterize the potential impor-
tance of different pollutants to the quality of urban streams.
Bottom sediment samples collected for a special urban
monitoring project in the nearby Menomonee River indi-
cated that all urban stream bottom sediments are prob-
ably contaminated with heavy metals (Dong, 1979).
Sources of lead and sediment for the established urban
areas were estimated using an urban runoff model called
Source Loading and Management Model (SLAMM) (Pitt,
1989). SLAMM is widely used in Wisconsin as a planning
tool to better understand sources of stormwater pollutants
and their control. Percent contributions listed in the an-
swerto question 4 are a lot more credible because SLAMM
was first calibrated with the data from the 1983 Milwaukee
stormwater monitoring project. Once the model was cali-
brated it was also used to estimate the pollutant reduction
goals. An average annual lead concentration calculated
with SLAMM for all the outfalls in Lincoln Creek
subwatershed was compared to the acute toxicity criteria
for lead. About a 40% reduction in lead loading was needed
from all the critical landuses in the subwatershed to meet
the acute toxicity standard in the stormwater discharged
from the outfalls. Concentrations measured in the Milwau-
kee River were used to determine the exceedances of the
chronic criteria in the stream.
Two years of samples collected at the inlet and outlet of
a wet detention pond in Madison, was the basis of the an-
swer developed for question 6 (House, 1993). The results
of this special urban monitoring project confirmed that about
a 90% reduction in sediment and about a 60% reduction
in lead could be achieved with wet detention ponds.
Lincoln Creek - Question 7
Lincoln Creek is part of an intense evaluation monitor-
ing effort in Wisconsin. Comprehensive biological, chemi-
cal, and physical monitoring is being done for at least a
ten-year period. Results from Lincoln Creek will be used
to evaluate the benefits of implementing best management
practices in other urban streams. All of the pre-practice
installation monitoring has been completed for Lincoln
Creek.
Special Urban Monitoring Projects
Results from the three special urban monitoring projects
referenced above were available in time to enhance the
answers to the seven stormwater management questions
in the Lincoln Creek Subwatershed. None of these projects
were conducted as part of preparing the Milwaukee South
Priority Watershed Plan. Determination of pollutant reduc-
tion goals would have been more difficult if the monitoring
data had not been available from the 1983 Milwaukee
stormwater monitoring project. Although there was insuffi-
cient data in these three projects to completely defend the
answers to the questions, it was enough to begin an imple-
mentation effort for Lincoln Creek.
Atotal of 26 special urban monitoring projects have been
conducted by the WDNR over the last 17 years. Each one
of the projects was selected to help answer one or more of
the seven stormwater management questions. All but five
of the projects were completed after 1993. This is after the
time priority watershed plans had been prepared for most
of the major metropolitan areas in Wisconsin. More recent
priority watershed projects have used the results of the
later special monitoring projects.
All together, the special monitoring projects cost about
$2.5 million. These costs were shared by the WDNR, EPA,
and local units of government. Between five and eight
projects are completed for each of questions 3, 4, 6, and
7. Our difficulties in selecting goals for reducing pollutant
loads and changing other factors, such as flow volumes,
is reflected in the fact that data is available for only two
projects related to question 5. A report is available for all of
the completed special monitoring projects. Since biologi-
cal monitoring is done in every watershed project for ques-
tions 1 and 2, there are no special projects completed for
these questions.
Role of Toxic Pollutants in Urban Streams -
Question 3
Five special urban monitoring projects are completed
that help characterize the impact of potentially toxic pollut-
ants on the biological integrity of an urban stream. Three
of the projects evaluated the toxicity of stormwater in Lin-
coln Creek. A total of 316 laboratory toxicity tests were
performed on stormwater and baseflow samples with
Ceriodaphnia dubia, Daphnia magna, and Pimephales
promelas. No short term, 48-96-hour acute or 7-day chronic
toxic effects, which could be solely attributed to stormwater
runoff, were identified with the three laboratory test spe-
cies (Ramcheck, 1995). Subsequent toxicity tests were
modified to include longer-term in situ tests. Tests with D.
magna performed in flow-through aquaria showed signifi-
cant increases in mortality for 93% of the tests after 14
days of exposure (Crunkilton, 1996). Longer exposures of
17 to 61 days, with juvenile and adult P. promelas exhib-
ited significant increases in mortality ranging from 30 to
95%. It appears that conventional waste water effluent tox-
icity tests lack the sensitivity to detect the biological deg-
radation observed in Lincoln Creek. The long-term in situ
toxicity tests should be used for future special monitoring
projects evaluating the toxicity of stormwater.
An in vitro bioassay with PLHC-1 (Poeciliopsis lucida)
fish hepatoma cells was used to assess potential toxic
potency of aryl hydrocarbon receptor (AhR) - active com-
57
-------
pounds, collected by semipermeable membrane devices
(SPMDs) exposed to Lincoln Creek water
(Villeneuve,1997). Dialysates from SPMDs exposed to Lin-
coln Creek water caused marked cytochrome P4501A in-
duction in PLHC-1. SPMDs exposed to baseflow had con-
sistently lower potencies than those exposed to high flows.
Emperical evidence suggests that AhR-active polycyclic
aromatic hydrocarbons (PAHs) can account for about 20
to 50% of the potency observed.
Monitoring of several urban streams in Milwaukee
County, showed that the urban streams are highly degraded
(Masterson, 1994). Stormwater discharges are blamed for
high concentrations of pollutants in the water and bottom
sediments, flashy flows, poor habitat, low diversity of
aquatic organisms, and accumulation of pollutants in fish
and crayfish tissue. A reference site was used to deter-
mine the degree of degradation. Water quality data com-
piled from four stormwater monitoring projects showed the
concentrations of many potentially toxic pollutants are high
enough to say that stormwater might be contributing to the
degradation of the urban streams (Bannerman, 1996). All
these findings describe the complexity of developing a
solution to problems caused by stormwater.
Sources of Pollutants - Question 4
Results from six special urban monitoring projects are
available to help determine the sources of stormwater pol-
lutants. All but one of these projects provides data on the
concentrations of pollutants in the runoff from different ur-
ban source areas. New sampling equipment was devel-
oped to collect sheet-flow runoff samples from roofs, park-
ing lots, driveways, streets, industrial yards, and lawns.
Source areas were sampled in residential, industrial, and
commercial landuses. The relative importance of the pol-
lutant load from each source area varies by pollutant and
landuse. Study sites in Madison, Wl, and Marquette, Ml,
showed streets as an important source for most pollutants
and landuses ( Bannerman, 1993; Burnhart, 1993;
Waschbusch, 1998; Steuer, 1997). Lawns are an impor-
tant source of phosphorus for all the study sites, while roofs
contribute a relatively large amount of zinc in commercial
and industrial landuses for the Madison study sites. Park-
ing lots at the Marquette study site are contributing the
largest amount of PAHs.
Not only the results from these projects identify impor-
tant source areas for the study sites, but the data from
these projects is also being used to calibrate SLAMM. This
will increase our confidence in source area loadings de-
termined for future priority watershed plans. Data from
these projects are also helping us identify the activities
responsible for depositing the pollutants on the different
urban surfaces. For example, phosphorus concentrations
in the runoff from streets increased with the greater per-
cent tree canopy overthe street (Waschbusch, 1998). This
information will be used to make the model more sensitive
to the tree canopy variations around a city. To more accu-
rately model the runoff from lawns, runoff parameters were
measured using a rainfall simulator on 20 Madison lawns
(Legg,1996). Rainfall-runoff relations vary substantially
between lawns, while the lawns that have been established
less than three years produce much higher runoff volumes
than older lawns.
An important number in every priority watershed plan is
the comparison of agricultural and urban contributions of
phosphorus and sediment to a stream. Stream phospho-
rus and sediment loads compiled from watersheds around
the state indicated that the phosphorus unit-area loads in
the southeast part of the state are similar for agricultural
and urban landuses (283 and 318 Ibs/sq. mi., respectively)
(Corsi, 1997). Sediment unit-area loads are three times
higher for the urban areas. A simple calculation with these
numbers will be used to determine the importance of con-
trolling urban sources of phosphorus and sediment.
Pollutant Reduction Goals - Question 5
Question 5 is probably the most challenging of the ques-
tions to answer, but it has received the least amount of
attention. An inadequate answerto this question can greatly
lower confidence in the suggested solutions to the
stormwater problems. An interim method for predicting
pollutant reduction goals is to combine the output of
SLAMM with a probabilistic dilution model developed by
the EPA (Corsi, 1995).Atestofthe method in Lincoln Creek
demonstrated a reasonable agreement between the me-
dian measured and predicted event mean suspended sol-
ids concentrations. Pollutant loading reduction goals can
be determined by reducing the pollutant loading output from
SLAMM until the median event mean concentration in the
stream is below the water quality standard. SLAMM loads
are reduced by simply specifying a control in the model
run.
The approach that will eventually replace using a proba-
bilistic dilution model will be based on understanding the
relationships between urban landuse activities and the
conditions in the streams. An investigation of 103 streams
in Wisconsin showed that a high amount of urban land
use in a watershed is strongly associated with poor biotic
integrity and weakly but significantly associated with poor
habitat quality (Wang, 1997). There seemed to be a thresh-
old value of the urbanization between 10 and 20% beyond
which IBI values are consistently low. Performance stan-
dards based on observed threshold values can become
the basis for setting pollutant and water volume reduction
goals for urban streams.
Selection of Best Management Practices -
Question 6
Having eight of the special monitoring projects, the study
of best management practices has received the most at-
tention. Monitoring data is available on the pollutant re-
moval effectiveness for a wet detention pond, a multi-cham-
ber treatment train, a Stormceptor, and street sweeping. A
model was developed to test the removal effectiveness of
infiltration devices. All of these types of practices are be-
ing used in Wisconsin. Two other projects summarized the
58
-------
costs of implementing different best management prac-
tices and the methods for monitoring industrial sites.
More monitoring data appears to be available for wet
detention basins than any other type of practice. They are
probably the most commonly used structural practice in
Wisconsin. Results from monitoring a wet detention basin
in Madison, indicate that a well-designed basin should re-
move about 90% of the solids, 50% of phosphorus, and
60% of heavy metals (House, 1993). The pond's sediment
and associated pollutant removal efficiencies are both in-
fluenced by influent particle size distributions (Greb, 1997).
Concentration data collected at the outlet occasionally
exceeded the acute toxicity standards for zinc and copper.
Toxicity testing on a pilot-scale wet detention basin indi-
cated that toxic reduction goals might not be achieved by
just using basins (Kron, 1998). Mortality for P. Promelas
exposed to the treated Lincoln Creek stormwater was sig-
nificantly reduced in only one of four test periods.
Evaluations of a multi-chamber treatment tank and a
Stormceptor installed at city maintenance yards revealed
very different pollutant removal efficiencies. The multi-
chamber treatment tank achieved levels of control for many
constituents of between 80 and 95%, while the efficien-
cies for the same constituents in the Stormceptor ranged
from 20 to 30% (Greb, 1998). Both devices will easily ret-
rofit into most land uses.
The water quality benefits of using mechanical street
sweepers was evaluated at four paired test sites in Mil-
waukee County. Models developed during the project were
used to determine street sweeping efficiencies for differ-
ent times of the year. Street sweeping is most effective in
the early spring during the heaviest street loads of the year
and in the fall following leaf fall (Bannerman, 1983). Lim-
ited benefits are expected from any intensive sweeping
program the rest of the year. Newer high efficiency sweep-
ers are expected to perform better than the mechanical
sweepers used in this study (Sutherland, 1997).
Although infiltration devices are rarely retrofitted in Wis-
consin, they might be needed to some degree to fully ac-
complish our pollutant and water volume reduction goals.
One concern about using infiltration as a practice is the
potential threat to groundwater quality. A method for deter-
mining the potential mobility of 32 organic and seven inor-
ganic pollutants during the infiltration of stormwater was
developed (Armstrong, 1992). The main variables affect-
ing leaching of stormwater pollutants are soil type selec-
tion, depth to groundwater, and water loading rate. Under
high loading rates, a few meters of soil will probably not
provide adequate protection of the groundwater. Inorganic
pollutants (mostly metals) are less mobile than organic
compounds (pesticides and PAHs). The calculated resi-
dence times per meter of soil for organic chemicals in a
hypothetical infiltration system range from 15 days or less
for "mobile" compounds to over 1,000 years for "very low
mobility" compounds. These calculations assume a high
water loading rate to the infiltration device. Predicted resi-
dence times for inorganic pollutants in a 1.0 meter soil layer
subjected to a water infiltration rate of 60 meters per year
range from less than 1 year for chromium to over 100 years
for lead.
Costs of stormwater practices are sometimes difficult to
estimate because not many of some types of practices
have been installed in Wisconsin, and the costs of the ones
that have been installed are not well documented. Not
knowing the costs of the practices makes it very difficult to
select cost-effective management alternatives for estab-
lished urban areas. Some estimates of the capital and
annual operation and maintenance costs are available for
a number of practices (SEWRPC, 1991).
Different sampling methods are being used to test the
effectiveness of stormwater practices designed to improve
the quality of runoff from industrial sites. Evaluation of the
effectiveness of the industrial practices will be difficult un-
less more is known about how the sampling methods can
affect the interpretation of the data. Five different monitor-
ing methods were tested at five different industrial sites
(Roa-Espinosa, 1994). These five methods were (1) flow
weighted composite, (2) time discrete, (3) time composite,
(4) source area, and (5) first 30 minutes. Assuming that
sampling at the outfall is the most representative sample,
then time composite sampling is the best method. How-
ever, a new type of electronic source area sampler could
make source area sampling a better choice, because the
samples are collected closer to the source of contamina-
tion.
Results from testing the effectiveness of different best
management practices is used to calibrate SLAMM. Prac-
tices not available in SLAMM, such as the multi-chamber
treatment tank, are added to the model. The new effec-
tiveness data will also be used to update the information
about each practice in Wisconsin's stormwater manual
(WDNR, 1994). Average long-term rainfall conditions for
several regions of the state are used to run the model
(Corsi, 1996).
Evaluation Monitoring - Question 7
The ability of the stormwater best management prac-
tices to achieve the designated uses in a stream is being
determined for Lincoln Creek and the Menomonee River
in Milwaukee County. Frequent chemical, biological, and
physical monitoring being done for both streams. Plans
are to continue the monitoring until implementation of the
priority watershed projects is completed. Results from these
two streams will be extrapolated to other urban streams.
At a cost of about $40,000 per year for each stream this
kind of intensive monitoring cannot be accomplished in all
the priority watershed projects. All of the pre-practice in-
stallation monitoring is done for both streams. Results of
the pre-practice installation monitoring have clearly docu-
mented the degradation of the water quality and biology in
both streams (Wang, 1996; Owens, 1997). Several com-
mon statistical techniques have been tested to detect
changes in the water chemistry data (Walker, 1993). The
59
-------
application of non-parametric tests to regression residu-
als for storm load data appears to be the best approach
for estimating minimum detectable change for a known or
estimated "before" condition.
To help quantify the changes in the stream, a version of
the Index of Biotic Integrity (IBI) was developed for warm
water streams in Wisconsin (Lyons, 1992). Guidelines were
developed for evaluating fish habitat in Wisconsin streams
(Simonson, 1994).
Future Urban Monitoring Plans
Results from the completed special monitoring projects
greatly increased the amount of information available to
answer the seven stormwater management questions. In
any new plan, however, we could not totally defend the
answers to the seven stormwater management questions;
if lead is targeted as a pollutant to control, for example, we
still could not totally defend its role in the degraded biol-
ogy or the levels of lead reduction suggested in the plan.
Without good monitoring data, each implementation effort
is to some degree an experiment, whose results will prob-
ably not be known until it is too late to make any major
adjustments to the types of best management practices
implemented.
Having the right kind of monitoring data can also influ-
ence people's acceptance of the solutions offered in a pri-
ority watershed plan. Whenever municipalities, industries,
and others cooperating in the stormwater clean-up effort
have some doubts about the answers to the management
questions, it diminishes the chances of completely imple-
menting the priority watershed plan.
Future Products
Our experience with finding answers to the seven
stormwater management questions gave us some ideas
on the type of additional information we need to improve
our answers. We identified eight products we need to de-
velop using special monitoring projects (Table 3). All the
products are important, but developing biological criteria
and stormwater performance standards for urban streams
would probably give the biggest boost to the credibility of
our answers. We would like to set a goal of having all of
the products over the next five years. Realistically, it will
probably take longer to develop stormwater performance
standards.
Biological criteria are needed to quantify a potential use
of every stream. This should be less subjective than the
stream classification procedures we follow now to deter-
mine the uses of a stream. A set of indices, such as the IBI
and the HBI, would identify the potential use of the urban
streams in every Wisconsin sub-ecoregion. Development
of the criteria would be a five-year effort requiring the col-
lection of data in both rural and urban streams. Some of
the data needed is already in WDNR files.
Closely related to the biological criteria is the identifica-
tion of the pollutants causing a toxic response in urban
Table 3. Products to be Developed using Future Special Urban
Monitoring Projects
Question
No.
Products
1 Biological Criteria for urban streams.
2 None
3 Method to identify which toxic pollutants are important
in each stream and what levels of control are needed.
4 SLAMM calibrated for all source areas and all the
problem pollutants.
5 Stormwater performance standards designed to
achieve biological criteria - standards based on %
connected imperviousness, flow rates and volume, D.O.
levels, buffers, pollutant loadings, and temperature.
6 Method of selecting most cost-effective practices to
achieve performance standards.
SLAMM capable of testing all practices.
Stormwater manual describing suggested management
alternatives for the most commonly occurring land use
mixtures.
7 Location to showcase benefits of stormwater manage-
ment.
streams. Toxic pollutants could be a limiting factor in the
selection of best management practices. Although a toxic
effect has already been identified in one urban stream, it is
not known which pollutants are responsible for the observed
toxicity or what degree of urbanization is required to cause
a toxic response. Using methods already developed, it
would probably take about three years to develop an un-
derstanding of which pollutants are toxic and how the
amount of urbanization effects their toxicity.
SLAMM is calibrated for many of the pollutants washed
off many of the source areas. But more calibration is needed
for the toxic pollutants in runoff from some of the source
areas, especially gas stations, parking lots, and industrial
paved surfaces. A three-year monitoring effort using our
existing source area monitoring methods would produce
the numbers to finish the calibration of SLAMM.
A performance standard is a threshold value for a bio-
logical, chemical, or physical factor that, if achieved, will
help meet biological criteria for a stream. The threshold
values are for the factors that affect the biological integrity
of any stream. At least seven types of performance stan-
dards need to be developed to meet the biological criteria
or improve the biological integrity of a stream. They in-
clude maximum temperatures, minimum dissolved oxygen
levels, minimum and maximum flows, maximum water
volumes, types of riparian vegetative buffers, annual sedi-
ment loading, and the combined annual loading of prob-
lem toxic pollutants. Percent connected imperviousness
is also included as a factor because it is a good way of
combining the effects of all factors without having to un-
derstand the effect of each one. Target values for all of
these factors would be the basis for developing manage-
60
-------
ment alternatives. These target values will probably vary
between sub-ecoregions. The difference between the per-
formance standards for a stream and the existing values
for these factors determines the goals for reducing pollut-
ants or changing other factors.
A multi-variate statistical analysis is going to be done to
determine the importance of each one of these factors. A
great deal is already known about threshold values for
some of the factors, especially flow, temperature, and dis-
solved oxygen (Raleigh, 1986). Most of the monitoring over
the next five years will be designed to better understand
the threshold values for percent connected imperviousness
and pollutant loadings. Work has already started on the
percent connected imperviousness factor with the collec-
tion of biological data from 45 streams with different de-
grees of urbanization. A less expensive method is being
developed to estimate annual pollutant loadings. Data col-
lection has started on a project to calibrate a model de-
signed to predict stream temperature changes during a
runoff event in an urban area.
Another important product is the development of cost-
effectiveness curves for different management alternatives.
Cost-effectiveness would be based on pollutant removal
relative to different annual costs of the alternatives. Usu-
ally, a combination of practices would be included in each
alternative. Curves might vary by land use and/or type of
pollutants being controlled. The curves would identify the
least expensive alternative for the level of control desired.
All the most promising best management practices will
be tested in Wisconsin. A special emphasis will be put on
infiltration and filtration devices. We will also try to docu-
ment the water quality benefits of educating the public on
stormwater management. All the results of these efforts
will be used to calibrate SLAMM and update our stormwater
management manual.
Every environmental management program needs some
place to showcase the benefits of their efforts. This will be
essential to justifying the long-term funding commitments
required by municipalities, industries, and othergroups re-
sponsible for stormwater management. Although evalua-
tion monitoring has already began at Lincoln Creek and
the Menomonee River, at least one more site is needed.
We are looking for a site where a good quality stream in
an urbanizing area could be saved by the proper use of
best management practices.
Future Types of Special Urban Monitoring
Projects
A lot of monitoring is going to be required to produce all
eight products. It is difficult at this time to describe all the
types of special monitoring projects that will be needed to
develop the eight products, but it is useful to suggest a list
of projects. About 28 special monitoring projects should
provide enough information to develop the products we
need (Table 4). Completion of all these projects will re-
quire at least five years or more to complete for a cost of at
Table 4. Future Special Urban Monitoring Projects
Question
No. Monitoring Projects
1 a. Collect biological data at test and reference sites
in all sub-ecoregions.
b. Develop IBI for small warm water streams.
2 None
3 a. Test response of organisms to different toxics in
stormwater.
b. Test response of organisms to serial dilutions of
problem toxics.
c. Test toxic response in streams with different %
connected imperviousness.
4 a. Test electronic sheet flow sampler.
b. Collect runoff from all source areas in three Wl
ecoregions.
c. Measure runoff coefficients for lawns in three Wl
ecoregions.
d. Evaluate relationship between lawn characteristics
and amount of runoff.
e. Measure pollutant concentrations for streets with
different traffic volumes.
f. Measure street phosphorus levels for streets with
different tree canopy.
g. Measure accumulation and washoff functions for
street solids.
h. Measure pollutant loadings during snowmelt.
5 a. Calculate amount of infiltration needed to maintain
normal baseflows.
b. Measure effect of excess runoff volumes on fish
habitat.
c. Collect all types of data in streams with different
degrees of urbanization.
d. Calibrate temperature model.
e. Evaluate importance of flow, habitat, chemistry to
quality of stream.
f. Evaluate importance of different buffer widths to
stream quality.
g. Determine how many grab samples are need to
estimate annual loading.
6 a. Measure effectiveness of two infiltration devices.
b. Measure effectiveness of high efficiency sweepers.
c. Measure effectiveness of two filtration devices.
d. Measure benefits of public education.
e. Summarize cost of building and maintaining all
types of practices.
f. Calibrate selected flow model.
g. Test controls by using historical and new fish data
from urbanizing areas.
7 a. Measure use changes in urban and urbanizing
streams.
least $4.5 million. Work has already started on five of these
projects. Most of the other projects are just at the sugges-
tion stage.
Projects already started include (1) development of IBI
for small warm water streams, (2) determining the rela-
tionship between percent connected imperviousness and
stream quality indicators, (3) calibration of a model to pre-
dict temperature changes in an urban stream during a runoff
event, (4) developing a less expensive method of deter-
61
-------
mining annual pollutant loads, and (5) evaluating benefits
of implementing practices in urban streams.
Summary
A combination of some monitoring in every urban stream
and special urban monitoring projects is helping Wiscon-
sin retrofit urban best management practices that achieve
the designated uses of the streams for as little cost as
possible. All the monitoring activities have been designed
to answer seven stormwater management questions. Each
question relates to a type of information needed to com-
plete one or more chapters in a priority watershed plan
prepared for Wisconsin's Nonpoint Source Program. Bio-
logical monitoring done in every urban stream identifies
the designated uses and the reasons for any of the ob-
served problems. Special urban monitoring projects en-
hance the answers to the first three questions, such as
identifying the potentially toxic pollutants, and provide an-
swers to the last four questions. Results of special projects
to determine the sources of pollutants and the effective-
ness of best management practices are most used results
in the priority watershed plans completed over the last ten
years.
Results are available from 26 special urban monitoring
projects completed over the last 17 years for a cost of about
$2.5 million. Most of these projects were completed after
1993. These results provide excellent answers to parts of
larger problems; others parts of those problems remain
unasked or unanswered. More information is especially
needed to determine goals for each priority watershed
project and to determine the best management alterna-
tives. Although it is important to continue the efforts to ret-
rofit urban areas with what we know, it is also important to
lower the risk in making future management decisions by
conducting additional monitoring projects.
Twenty-eight new special urban monitoring projects re-
sulting in eight essential products are recommended. De-
velopment of biological criteria and performance standards
for urban streams are two of the products essential to the
success of future stormwater management efforts. Work
has already started on five new special monitoring projects.
These projects will help define the use of "percent con-
nected imperviousness" as a performance standard and
provide a method for predicting the changes in stream tem-
perature during a runoff event. Another project will docu-
ment the changes in two urban streams during and after
the implementation of best management practices. Mu-
nicipalities, industries, and othergroups cooperating in the
stormwater management efforts will be able to use the re-
sults of these and other new projects to improve their con-
fidence in management decisions that could cost millions
of dollars.
References
Armstrong, D.E., and Llena, R., 1992. Stormwater infiltra-
tion potential for pollutant removal: Project Report,
Wisconsin Department of Natural Resources,
Madison.WI. p 10-18
Ball, J., 1982. Stream classification guidelines for Wiscon-
sin. Wisconsin Department of Natural Resources Tech-
nical Bulletin. Wl Department of Natural Resources,
Madison, Wl.
Bannerman,R.T, Baun,K., Bohn,M., Hughes,P.E., and
Graczyk,D.A., 1983. Evaluation of urban nonpoint
source pollution management in Milwaukee
County,Wisconsin: Chicago, U.S. Environmental Pro-
tection Agency, available from National Technical In-
formation Service as PB 84-114164m 209p.
Bannerman,R.T., Owens, D.W., Dodds,R.B., and
Hornewer, N.J., 1993. Sources of pollutants in Wis-
consin Stormwater: Water Science Technology, v.28,
no. 3-5, p. 241-259.
Bannerman,R.T, Legg,A.D., and Greb,S.R., 1996. Qual-
ity of Wisconsin Stormwater, 1989-94: U.S. Geologi-
cal Survey Open-File Report 96-458, 26 p.
Burnhart, M., Bannerman, R., Owens, D., and Hornewer,
N.J. 1993. Sources of bacteria in Wisconsin
stormwater. Wl Department of Natural Resources,
Madison, Wl.
Corsi, S.R. and Greb, S.R., 1995. In Press. Modeling
stormwater contributions of contaminants in urban
streams.
Corsi, S.R., 1996. Average annual precipitation data for
five regions in Wisconsin. U.S. Geological Survey Let-
ter Report. Madison, Wl.
Corsi, S.I., Graczyk, D.J., Owens, D.W, and Bannerman,
R.T. 1997. Unit-area loads of suspended sediment,
suspend solids, and total phosphorus from small wa-
tersheds in Wisconsin, U.S. Geological Survey Fact
Sheet FS-195-97.
Crunkilton, R., Kleist,J., Ramcheck, J., DeVita, W, and
Villeneueve, D., 1996. Assessment of the response of
aquatic organism long-term insitu exposures of urban
runoff. In Effects of Watershed Developments on
Aquatic Ecosystems. Proceedings of Engineering
Foundation Conference ed. by L. A. Roesner. Snow-
bird, Utah. American Society of Civil Engineers, N.Y.,
N.Y.
Dong, A., 1979. Dispersibility of soils and elemental com-
position of soils, sediments and dust and dirtfromthe
Menomonee River Watershed - Volume 6. The IJC
Menomonee River Watershed Study. EPA-905/479-
029-F, U.S. Environmental Protection Agency,
Chicago,II.
Greb, S.R., and Bannerman, R.T, 1997. Influence of par-
ticle size on detention pond effectiveness. Water En-
vironment Research, Volume 69, Number 6
62
-------
Greg, S.R., Corsi, S., and Wascbusch, R. 1998. (In Press.)
Evaluation of Stormceptor and Multi-Chamber Treat-
ment Train as urban retrofit strategies.
Hilsenhoff, W.L., 1982. Using a biotic index to evaluate
water quality in streams. Technical Bulletin No. 132.
Wl Department of Natural Resources, Madison,Wl.
House, L.B., Waschbusch, R.J., and Hughes, P.E., 1993.
Water quality of an urban wet detention pond, 1987-
88. U.S. Geological Survey Open-File Report 93-172.
Kron, D., 1998. Assessment of a wet detention basin for
toxicological remediation of urban runoff. Thesis in
progress, College of Natural Resources, Stevens Point,
Wl.
Legg, A.D., Bannerman, R.T., and Panuska, J., 1996. Varia-
tion in the relation of rainfall to runofffrom residential
lawns in Madison, Wisconsin, July and August 1995.
U.S. Geological Survey, Water- Resources Investiga-
tions Report 96-4194.
Lyons, J., 1992. Using the index of biotic integrity (IBI) to
measure environmental quality in warmwater streams
of Wisconsin. Gen. Tech. Rep. NC-149, U.S. Forest
Service, North Central Forest Experiment Station, St.
Paul, MN.
Masterson, J.P., and Bannerman R.T., 1994. Impacts of
storm water run off on urban streams in Milwaukee Co.
Wl. pp 123-133, In Proceedings of the American Wa-
ter Resources Association, National Symposium on
Water Quality ed. by G.L. Pederson. Chicago,IL,
American Water Resources Association, Herndon,VA.
Owens, D.W , Corsi, S. R., and Rappold, K.F., 1997. Evalu-
ation of nonpoint-source contamination, Wisconsin:
selected topics for water year 1995: U.S. Geological
Survey Open-File Report 96-661 A. 41 p
Pitt, R., and Voorhees, J., 1989. Source load and manage-
ment model-an urban nonpoint source water quality
model: Wl Department of Natural Resources, v. l-lll,
PUBL-WR-218-89, various pages.
Raleigh, R.F., Zuckerman, L.D., and Nelson, P.C. 1986.
Habitat suitability index models and instream flow suit-
ability curves: brown trout. Biological report 82(10.124)
Fish and Wildlife Service, U.S. Dept. of Interior, Wash-
ington, DC.
Ramcheck, J.M., 1995. Toxicity evaluation of urban
stormwater runoff in Lincoln Creek, Milwaukee, Wis-
consin. Masters Thesis, University of Wisconsin,
Stevens Point, Wl. 87p
Roa-Espinosa, A., and Bannerman R.T., 1994. Monitoring
BMP effectiveness at industrial sites, pp 467-486, In
Stormwater NPDES Related Monitoring Needs, Pro-
ceedings of an Engineering Foundation Conference
ed. by H.C. Torno. Mount Crested Butte,CO., Ameri-
can Society of Civil Engineers, New York,NY.
Simonson,T, and Lyons, G., 1993. Evaluation monitoring
offish and habitat during priority watershed projects;
a progress report for 1992. Wl Department of Natural
Resources. 45 pages.
Simonson,T.D., Lyons, J., and Kanehl, P.O., 1994. Guide-
lines for evaluating fish habitat in Wisconsin Streams.
Gen. Tech. Rep. NC-164, U.S. Forest Service, North
Central Forest Experiment Station, St.Paul, MN.
Southeastern Wisconsin Regional Planning Commission
(SEWRPC). 1991. Cost of urban nonpoint source wa-
ter pollution control measures. Technical report num-
ber 31, SEWRPC. Waukesha, Wl.
Steuer, J., Selbig, W, Hornewer, N., and Prey, J., 1997.
Sources of contamination in an urban basin in
Marquette, Michgan and an analysis of concentrations,
loads, and data quality. U.S. Geological Survey. Wa-
ter-Resources Investigations Report 97-4242.
Sutherland, R.C. and Jelen, S.L., 1997. Contrary to con-
ventional wisdom: street sweeping can be an effective
BMP. Published in Advances in Modeling the Manage-
ment of Stormwater Impact, Ed. by Dr. William James,
CHI Publication, Ann Arbor Press, Volume 5.
Villeneueve, D.V, Crunkilton, R.C., and DeVita, W.M.,
1997. ARYL hydrocarbon receptor-mediated toxic po-
tency of dissolved lipophlic organic contaminants col-
lected from Lincoln Creek, Milwaukee, Wl, USA, to
PLHC-1 (PoeciliopsisLucida)f\sh hepatoma cells. En-
vironmental Toxicology and Chemistry, vol 16, No.%,
pp 977-984.
Walker, J.F., 1993. Techniques for detecting effects of ur-
ban and rural landuse practices on stream water chem-
istry in selecting watersheds in Texas, MN, and IL. U.S.
Geological Survey Open-File Report 93-130.
Wang, L., Lyons, J., and Kanehl, P., 1996. Evaluation of
the Wisconsin priority watershed program for improv-
ing stream habitat and fish communities. Progress re-
port for 1996. Wl Department of Natural Resources,
Madison, Wl.
Wang, L., Lyons, J., Kanehl, P., and Gatti, R., 1997, Influ-
ences of watershed land use on habitat quality and
biotic integrity in Wl streams: Fisheries, v. 22, no. 6,
June 1997, p 6-12.
Waschbusch, R.J., Selbig, W.R., and Bannerman, R.T.,
(1998). Sources of phosphorus in stormwater from two
urban residential basins in Madison, Wl, 1994-95.
Wisconsin Department of Natural Resources. 1989. Sur-
face water quality criteria for toxic substances. Chap-
ter NR 105, Register No. 398.
Wisconsin Department of Natural Resources (WDNR).
1991. A non-point source control plan for the Milwau-
kee River South Branch Priority Watershed Project.
PUBL-WR-245-91. Madison,WI.
63
-------
Wisconsin Department of Natural Resources (WDNR). Wisconsin Department of Natural Resources (WDNR).
1994. The Wisconsin stormwater manual part one: 1995. Water quality standards for Wisconsin surface
overview. Publication Number: WR-349-94. Wiscon- waters. Wisconsin Administrative Code Chapter NR
sin Department of Natural Resources, Madison, Wl. 102. Register No. 477, Madison, Wl.
64
-------
Considerations and Approaches for
Monitoring the Effectiveness of Urban BMPs
Eric W. Strecker, RE.
Woodward-Clyde
Portland, Oregon
The purposes of this paper are to 1) describe some of
the problems with typical Best Management Practice (BMP)
monitoring and effectiveness reporting and to 2) suggest
the utilization of consistent stormwater monitoring tech-
niques. This will allow the data collected on the effective-
ness of individual best management practices (BMPs), in-
cluding retrofit BMPs, to be useful fora particular site, and
to also be useful for comparing studies of similar and dif-
ferent types of BMPs in other locations. Many BMP effec-
tiveness studies in the past have provided only limited data
useful for assessing BMP design and selection on a wide
scale. This paper overviews some of the problems of past
BMP effectiveness studies from the perspective of com-
parability between studies. It suggests some of the ways
that data could be collected to make it more useful for as-
sessing factors (such as settling characteristics of inflow
solids and physical features of the BMP) that might have
led to the performance levels achieved. Finally, it also dis-
cusses other considerations that affect data transferabil-
ity, such as effectiveness estimations, statistical testing,
etc.
Introduction
Many studies have been completed which have as-
sessed the ability of stormwater treatment BMPs (e.g., wet
ponds, grass swales, stormwater wetlands, sand filters,
dry detention, etc.) to reduce pollutant concentrations and
loadings. However, in attempting to summarize the infor-
mation gathered from these individual BMP evaluations it
is very apparent that inconsistent study methods and re-
porting make wider scale assessments difficult. For ex-
ample, individual studies often included the analysis of dif-
ferent constituents and utilized different methods for data
collection and analysis. These differences alone contrib-
ute significantly to the range of BMP effectiveness reported.
This makes assessing what other factors may have con-
tributed to the variation in performance almost impossible.
In one review of the use of wetlands for stormwater pol-
lution control (Strecker et al., 1992), a summary of the lit-
erature on performance of wetland systems and the fac-
tors that may have led to the reported pollutant removals
was prepared. The literature was inconsistent with respect
to the constituents analyzed and the methods used to
gather and analyze data. A number of pieces of informa-
tion, if collected and recorded, would have improved the
ability to evaluate the effectiveness of stormwater wetlands
as BMPs and facilitated the transfer of that knowledge into
better design practices. Urbonas (1994 and 1995) and
Strecker (1994) summarized the information that should
be recorded about the physical, climatic, and geological
parameters which likely affect the performance of a BMP,
and considerations regarding sampling and analysis meth-
ods. This paper presents 1) a suggested list of constitu-
ents for analysis along with recommendations for report-
ing data, 2) methods of reporting pollutant removal effi-
ciencies, 3) a brief discussion of statistical approaches to
selecting the number of samples needed, 4) methods for
including detection limit data, 5) sample collection consid-
erations, and 6) the need for dry weather assessments.
BMP Performance Study Inconsistencies
Studies of BMP effectiveness have utilized significantly
different:
• Sample collection techniques (e.g., from sample col-
lection types (grab, composite, etc.), flow measure-
ment techniques, to how the sample was composited,
etc.);
• Constituents, including: chemical species, methods
(detection limits), form (e.g., dissolved vs. total, vs. total
recoverable, etc.), and treatment potential;
• Data reporting on tributary watershed and BMP de-
sign characteristics (e.g., tributary area or watershed
attributes such as percent impervious, land use cat-
egories, rainfall statistics, etc.);
• Effectiveness estimation (at least fourtechniques have
been utilized to assess effectiveness which can cause
significant differences in pollutant removal reporting,
with the same set of data), and potential alternatives
to reporting just concentration/loading reductions; and
• Statistical validation of results (typical lack of statisti-
cal tests to determine if the reported removal efficiency
65
-------
can in fact be shown to be statistically different than
zero).
Any of the above topics would require an in-depth dis-
cussion beyond the scope of this paper to fully explain.
Therefore, this paper will present a brief overview of each
of these and some potential solutions to improving how
data is collected. EPA together with ASCE is currently de-
veloping a set of protocols and a database on BMP perfor-
mance studies with the purpose of improving the consis-
tency of BMP monitoring information. This project includes:
• Developing Protocols for BMP Monitoring
• Conducting an evaluation of existing information to
assist the EPA Wet Weather FACA and contribute to
EPA's Stormwater Toolbox (as identified in Draft Phase
II Stormwater Regulation Preamble)
• Developing a data base on BMP performance studies
The overall goal is to improve the BMP effectiveness
information base to:
• Develop information to improve designs
• Improve performance information
The data base specifies a chosen set of reporting infor-
mation, but does not tell how to develop such information.
For example, it does not specify what a flow-weighted com-
posite sample is and how it should be collected. The next
step beyond the EPA protocols and data base effort should
be a guidance document on monitoring data collection strat-
egies and techniques to improve their consistency and
transferability. It should be recognized that with the devel-
opment of the database and the protocols, it will be a num-
ber of years (5 to 10) before significant new studies on
BMPs are conducted utilizing the protocols to allow for a
more rigorous evaluation of BMP selection and design fac-
tors.
Sample Collection Techniques
The differences among sample collection techniques
alone is enough to make comparing different studies ques-
tionable. These include differences among how flows are
measured to how samples are composited to formulate an
"event mean concentration." Some studies have utilized
grab samples, and the results of these studies in evaluat-
ing BMP performance are limited. Typically studies will in-
clude the collection of flow-weighted composite samples
(either automated or hand collected). These studies involve
various techniques (often not reported very well) for mea-
suring flows. The flow measurements themselves are sub-
ject to a large variation.
The Federal Highway Administration is currently conduct-
ing a study of monitoring techniques for characterizing
Stormwater runoff hydrology and water quality from high-
ways. The study, being completed by Woodward-Clyde,
included a component conducted by the USGS
(Waschbusch and Owens, 1998) which addressed the
potential differences in flow measurement techniques in a
pipe system in Madison, Wl. An in-depth dye-dilution
method was utilized to calibrate a Palmer-Bowlus flume
with a bubbler pressure measurement. The study evalu-
ated 23 flow measurement techniques including commer-
cially available packages and individual component sys-
tems.
Figure 1 is a summary of the results of flow measure-
ments, showing the average percent differences from the
calibrated flume. These data summarize 50 storm events
which were measured over a 6-month period. As the fig-
ure demonstrates, the error in flow measurements is eas-
ily on the order of plus or minus 25% over a range of storms.
The flow measurements for individual storms varied even
more. If samples are composited based upon flows (either
using automated or using grab samples), they are subject
to an error in collection times (for automated systems) or
in composited amounts (grab sample composited) and
therefore could result in errors in estimates of event mean
concentrations (especially for constituents which vary over
the course of a storm event). It should be strongly noted
that these results are for one site only and should not be
interpreted as indicative of how any particular system iden-
tified might perform at another site. It is imperative that
researchers thoroughly evaluate potential flow measure-
ment alternatives and implement the method that will re-
sult in the best information possible.
Another aspect of the study addressed how many
samples should be collected to compile a "flow-weighted"
composite sample. Figure 2 demonstrates the large vari-
ability in sampler bottle configurations. These configura-
tions often drive researchers into selecting the number of
"grab" composite samples to collect. For example, in the
NPDES monitoring for Texas (Brush, etal., 1994), the cho-
sen strategy was to collect one sample into each bottle of
the 8-bottle configuration (this was successful if it rained
sufficiently). In the Portland and Eugene NPDES Sampling
(WCC 1993a and WCC 1993b), an attempt was made to
collect 24 "grab" samples during the course of an event.
Figure 3 shows a typical storm event from the Portland
program and specifically the points at which a sample was
collected. From the variability in flows observed, one can
surmise the pollutant concentrations were also fluctuating
extensively (later confirmed by within-storm sampling).
Having only eight samples during this event may not have
accurately characterized the event mean concentration
(EMC). Collecting three times the samples to "construct" a
flow-weighted sample would appearto reduce the chances
of anomalies (variability) during a storm event influencing
the overall estimate of the average concentration. Early
results from our FHWA study indicate that one should at-
tempt to collect at least 12 to 16 individual samples to form
a composite sample.
The study also has evaluated the potential effects of
sample lift (e.g. pumping up from underground or from
stream bottoms) and has found that the newer samplers
66
-------
D)
.g
to
a:
sm
el
ota
ace
K
D)C
C O
11
|g-
2iu
•g oi
H
•C3IH
^
IX \—\
End-of-Pipe Ultrasonic
lx — 1— 1
_L2 — 1
=H
[H
I® 1
Ultrasonic Approach Rated
Bubbler Throat Rated
Ultrasonic Throat Rated
Bubbler Approach Theoretical
Ultrasonic Approach Theoretical
Bubbler Throat Theoretical
Ultrasonic Throat Theoretical
| 1 X h
Down-stream Ultrasonic
Down-stream Bubbler
Marsh McBirney 250 Ultrasonic
(Corrected Velocity)
Marsh McBirney 250 Bubbler
Marsh McBirney 250 Ultrasonic
Marsh McBirney 250 Bubbler
BVT Ultrasonic
BVT Bubbler
ADS
ISCO 4250
Sigma End-of-Pipe 950
Marsh McBirney Flow-tote
Manufacturer Corrected
Corrected ADS
Manufacturer Corrected
Sigma End-of-Pipe 950
-75%
-50%
-25%
0%
25%
50%
75%
100%
125%
Figure 1. Boxplot of the percent differences between total storm volumes computed using various flow estimation methods and the total storm
volume of the bubbler approach rated discharge (bold line atO%). (Waschbusch and Owens, 1998).
67
-------
3 Gal.
Polyethylene Container
2.5 Gal.
Glass Container
(24) 575 ml
Polyethylene Bottles
4 Gal.
Polyethylene Container
5.5 Gal.
Polyethylene Container
(8) 2.3 Liter
Polyethylene Bottles
(8) 1.9 Liter
Glass Bottles
(24) 1 Liter
Polyethylene Bottles
(24) 350 ml
Glass Bottles
Figure 2. Typical automated sampling bottle configuration options.
68
-------
Flow and Rainfall Results for Station C1
Storm 4, December 5-7, 1991
(Data begin at 8:00 Thursday, 12/5)
-, 0.14
- 0.12
Flow Rate
Rainfall
Sample Taken
A
Sample
Not Taken
O
Time (hrs)
Figure 3. Typical hydrograph indicating measured rainfall, runoff, and water sample collection times from automated flow and water quality
sampling for the Portland NPDES Stormwater Monitoring Program.
(with stronger pumps) do not appear to cause any separa-
tion of suspended solids as they are lifted up to 20 feet. At
the end of the study, a guidance document on sampling of
highway runoff will be developed. These are just some of
the numerous differences in sampling methods that could
lead to differences in results between BMP studies.
Constituents Assessed
A very wide variety of pollutants have been analyzed in
both BMP studies and characterizations studies. The EPA
protocols study has developed a recommended set of con-
stituents for BMP testing programs. These were developed
from the review of previous studies and an understanding
of costs and likelihood of providing meaningful results. Be-
low is a discussion of how these constituents were se-
lected (adapted from Strecker, 1994).
Since NURP and priorto the Phase I Stormwater NPDES
monitoring programs, there have been a number of stud-
ies which continued to assess pollutant concentrations in
Stormwater runoff. These included the Federal Highway
Administration's highway runoff program (Driscoll et al.,
1990) and some selected studies done in a few locations.
These studies typically were not consistent with the stan-
dard NURP protocols. Based upon the 1987 amendments
to the Clean Water Act, EPA required operators of munici-
pal separate storm drainage systems that served popula-
tions of over 100,000 to collect flow-weighted composites
at a minimum of five stations to characterize residential,
commercial, and industrial runoff quality. Only a few addi-
tional parameters have been identified as "problems" in
Stormwater, based upon these post-NURP studies (this
despite the improved analytical methods that have become
available for conducting laboratory analyses). In addition,
NURP focused primarily on residential and commercial land
uses, while NPDES testing included industrial land uses
which were suspected of having more pollutants present.
However, there has not been a comprehensive review
by EPA or others of the newly collected Stormwater infor-
mation to assess the results of requiring the analysis of
over 130 constituents, including priority pollutants. This type
of review is needed. EPAs requirements included moni-
toring three storms at selected stations. This number of
storms is only useful for identifying potential problem pol-
lutants. Statistically, these are not enough data to perform
69
-------
a meaningful regional or other factor analyses of urban
stormwater concentrations, although they could provide
useful information on rates of detection. This analysis would
be helpful in selecting constituents for BMP monitoring.
The choice of constituents to include as "standard pol-
lutants" is a subjective one. As an example, some would
argue that cost should be a primary consideration; others
would say that it should not. In making the recommended
list of monitoring constituents, the following characteris-
tics were considered:
• The pollutant is prevalent in typical urban stormwater
at concentrations that could cause water quality im-
pairment.
• The analytical test can be related back to potential
water quality impairment.
• Sampling methods forthe pollutant are straightforward
and reliable for a moderately careful investigator.
• Analysis of the pollutant is economical on a widespread
basis.
• The pollutant is one for which treatment is a viable
option.
Not all of the pollutants recommended fully meet all of
the factors listed above; however, the factors were consid-
ered in the recommendations. When developing a list of
pollutant analyses for an individual BMP evaluation, it is
important to consider the upstream land use activities. The
parameters recommended below are present and of con-
cern in "typical" urban stormwater.
The Nationwide Urban Runoff Program (NURP) (EPA,
1983), which included monitoring of land use runoff and
BMP performance at over 28 cities nationwide, adopted
consistent data collection methods and analytical param-
eters. Results from the NURP program could be used to
evaluate similarities and differences in pollutant concen-
trations in urban stormwater from different and similar land
uses, and could be used to explain what might be causing
these differences. The following pollutants were adopted
by NURP as "standard pollutants characterizing urban run-
off':
TSS
BOD
COD
TP
SP
TKN
NO +
CU
PB
ZN
NO,
Total suspended solids
Biochemical oxygen demand
Chemical oxygen demand
Total Phosphorus
Soluble phosphorus
Total Kjeldahl nitrogen (as N)
Nitrate + nitrite (as N)
Copper
Lead
Zinc
Oil and grease was not included because of the diffi-
culty in obtaining representative samples. On a less con-
sistent basis, NURP also monitored for pollutants includ-
ing other metals, dissolved metals, semi-volatile organics,
volatile organics, pesticides, and herbicides.
Presented below is a brief discussion, by group, of the
pollutants that are recommended to be included in a base
list, then several that may occasionally be recommended.
Total Suspended Solids (TSS). The term "suspended
solids" is descriptive of the organic and inorganic particu-
late matter which is of a size and type that allows the par-
ticles to stay suspended in water. The solids load in a
waterbody is influenced by a number of factors including
but not limited to: particle sizes, stream flows, climate,
geology, and vegetation of each drainage system. The
conditions under which suspended solids are considered
a pollutant is a matter of definition. In general, suspended
solids are considered a pollutant when they significantly
exceed natural concentrations and have a detrimental ef-
fect on water quality and/or beneficial uses of the water
body.
Suspended sediments are often used as a surrogate for
other contaminants which bind or adsorb easily with fine
particulate matter, including heavy metals. Although TSS
is often highly correlated with other parameters, it is gen-
erally not a strong enough correlation to eliminate the need
to address other parameters specifically. Figure 4 shows
the relationship between TSS and zinc for pooled
stormwater runoff monitoring data from all ten stations
monitored in Portland, Oregon for the NPDES program
(WCC, 1993a) and from the seven stations that were from
piped systems. Although the relationship is statistically sig-
nificant (R2 of .38 for piped stations), it does not explain a
significant amount of the variability. Similar results were
found for almost all other parameters. It should be noted
that for individual stations, the relationships between TSS
and many pollutants were sometimes much higher, but this
would mean that one would have to monitor enough times
to establish the relationship. Therefore, TSS does not ap-
pear to be a good predictor of other pollutants, without
significant data collected from each station. However, TSS
is one good indicator of pollutant removal efficiency (e.g.,
because of the tendency for many pollutants to be associ-
ated with fine particulates) and should be included in any
evaluation of BMP performance.
Many BMPs rely on sedimentation as the primary pollut-
ant removal mechanism. It is recommended that samples
also be analyzed for some measure of the expected set-
tling rate (treatment potential) of TSS. The performance of
a BMP that relies on sedimentation and even filtering can
be greatly affected by the particle sizes and densities
present in the influent. If the influent TSS is characterized
by very small particle sizes, and therefore slow settling
velocities, it will be much more difficult to treat. The settle-
ability of influent solids has not been adequately addressed
in performance comparisons, and may be one of the sig-
nificant reasons that measured performance varies so
highly from similar BMP to BMP.
For consideration, the particle size distribution in street
dirt found in Sartor and Boyd (1972), as shown in Table 1,
70
-------
Regression Plot: Zinc (Zn) vsTSS
... i .... i .... i .... i ... . i .... i
1 -
N
-6
O
o
12 34 567
TSS
Y = -4.147 + .477 * X; R*2 = . 187
Pooled Data
From All
Stations
Regression Plot
N
-.5-
-1 -
-1.5 -
-2 -
-2.5 -
-3 -
-
-3.5 -
°5 0 ° °
0 0° %°
_ o
o ° o
O03 _ O
o °
(S) O ®
Ooo «.
-------
Table 1. Particle Size Fractions of Street Dirt from Selected Locations.
Size
Ranges
>4800 n
2000 - 4800 n
840 - 2000 n
246 - 840 \i
104- 246 n
43 - 1 04 n
30 - 43 n
14-30n
4- 14 |i
>4 |i
Sand %,
43 - 3800 \i
Silt %,
4-43|i
Clay %,
<4 |i
Milwaukee
12.0%
12.1
40.8
20.4
5.5
1.3
4.2
2.0
1.2
0.5
92.1
7.4
0.5
Bucyrus
-%
10.1
7.3
20.9
15.5
20.3
13.3
7.9
4.7
-
74.1
25.9
-
Baltimore
17.4%
4.6
6.0
22.3
20.3
11.5
10.1
4.4
2.6
0.9
82.1
17.1
0.9
Atlanta
-%
14.8
6.6
30.9
29.5
10.1
5.1
1.8
0.9
0.3
91.9
7.8
0.3
Tulsa
-%
37.1
9.4
16.7
17.1
12.0
3.7
3.0
0.9
0.1
92.3
7.6
0.1
Note: |i = microns
Source: Sartor and Boyd, 1972
assessing oxygen demand are not straightforward indica-
tors of potential problems.
Biochemical Oxygen Demand (BOD). The 5-day BOD
test provides an indirect measure of the quantity of bio-
logically degradable organic matter in water in terms of
the amount of oxygen required by microorganisms to oxi-
dize it to carbon dioxide and water. The BOD test is quite
variable. A number of factors can affect results, including
the quality of the seed culture utilized in the test. The BOD
test can also be inhibited by toxicants in the sample, which
may react differently once the runoff mixes with the receiv-
ing water. The levels of BOD that are normally found in
urban stormwater are near detection limits for the BOD
test. Therefore, they are subject to wide variation. There-
fore BOD has not been recommended as a parameter.
Instead, TOC (Total Organic Carbon) has been identified
as a more consistent measure of available organic mate-
rial, which could be contributing to oxygen demand.
Chemical Oxygen Demand (COD). The COD test pro-
vides a more rapid and consistent measure of oxygen de-
mand than BOD tests. The consumption of oxygen from
an introduced strongly oxidizing chemical agent is mea-
sured by this test. As a result, it typically measures appre-
ciably higher levels of oxygen demand than will be pro-
duced by biological decomposition because it oxidizes
some organic compounds that are not biodegradable, and
may also react with inorganic compounds as well. In ur-
ban stormwater, for example, COD levels are typically found
to be about 8 to 10 times greater than BOD levels. COD
measures a "maximum possible," but not probable, oxy-
gen demand.
Nutrients. Nutrients are necessary for the growth and
support of biota in natural watersystems. Excessive quanti-
ties can result in the over-stimulation of biological growth
and the creation of objectionable water quality conditions
(eutrophication). Some forms of nutrients can also be toxic
(e.g., ammonia). In general, the most important nutrient
factors causing an acceleration in algal production are ni-
trogen compounds and phosphorus.
Nitrogen. Nonpoint sources of nitrogen include lawn fer-
tilizers, leachate from waste disposal in dumps or sanitary
landfills, atmospheric fallout, nitrite discharges from auto-
mobile exhausts and other combustion processes, natural
sources such as mineralization of soil organic matter, and
farm-site fertilizers and animal wastes. Many water treat-
ment methods have no significant effect on nitrate removal
from water (Dunne and Leopold, 1978).
Three forms of nitrogen have been analyzed extensively
in stormwater runoff water quality studies. These are ni-
trite plus nitrate (NO2 + NO3), ammonia nitrogen (NH3),
and total Kjeldahl nitrogen (TKN). The latter, named after
the analytical test procedure, provides a measure of am-
monia and organic nitrogen forms that are present. The
first (NO2 + NO3) provides a measure of the inorganic ni-
trogen. There is usually very little nitrite in stormwater. Ni-
trate (NO3) is very mobile and is usually difficult to treat
utilizing stormwater BMPs. Ammonia nitrogen can be toxic
to aquatic life. It can be assessed fortoxicity to aquatic life
with data on pH and temperature. The inorganic (NO2 +
NO3) and ammonia nitrogen are recommended. All forms
should be reported as mass of nitrogen (N).
Phosphorus. Phosphorus is used by algae and higher
aquatic plants and may be stored in excess of use within
plant cells. With decomposition of plant cells, some phos-
phorus may be released immediately through bacterial
action for recycling within the biotic community, while the
remainder may be deposited with sediments.
72
-------
Phosphorus enters waterways from many of the same
sources as nitrogen. Domestic sewage contains significant
concentrations of phosphorus which are contributed by
detergents and human wastes. Primary and secondary
treatment processes normally remove only about 20 to 30%
of this element from sewage (Dunne and Leopold, 1978).
Fertilizers and the erosion of soils rich in phosphorus can
also be a potential source.
Three forms of phosphorus have been somewhat rou-
tinely analyzed in stormwater runoff studies. These include
total phosphorus (TP), soluble phosphorus (SP), and ortho-
phosphate (OP). Ortho-phosphate indicates the phospho-
rus that is most immediately biologically available. Soluble
phosphorus includes both the ortho-phosphate and a frac-
tion of the organic phosphorus. Most all of the SP is usu-
ally OP, however. Total phosphorus includes phosphorus
in the forms that may not be as readily biologically avail-
able plus the forms discussed above. TP and OP are rec-
ommended for inclusion in a monitoring program, as they
characterize both the total and bioavailable forms of phos-
phorus. All forms should be reported as mass of phospho-
rus (P).
Metals. Heavy metals such as copper, lead, and zinc
are naturally released in very small quantities by the weath-
ering of exposed soils and mineral deposits, corroding
metal surfaces, decomposing paints, and certain corrosion-
control compounds. Heavy metals tend to have compara-
tively low solubilities and are often mobilized by forming
soluble complexes with humic materials or by becoming
attached to clay particles. Heavy metals have been con-
sistently identified as the most significant toxics found in
urban stormwater and often exceed water quality criteria
for aquatic life.
These metals are present in the biosphere as trace ele-
ments and are micronutrients necessary for plant and ani-
mal growth. Heavy metals are of concern because elevated
concentration levels of soluble forms in natural water bod-
ies can produce toxic effects in biota. Sources include do-
mestic and industrial point-source discharges, urban
stormwater runoff, and direct atmospheric deposition. In
this paper, copper (Cu), lead (Pb), zinc (Zn), and cadmium
(Cd) have been recommended for inclusion in a monitor-
ing program because stormwater runoff water quality stud-
ies conducted at many urban locations have indicated that
these metals are almost always present, and are at con-
centrations which tend to be elevated, relative to other
heavy metals. They also can be used as surrogates for
other heavy metals, as they tend to display the range of
transport characteristics for heavy metals. However, other
heavy metals should be analyzed if there are known
sources of significant quantities of these metals in influent
flows.
It is recommended that both the total and dissolved form
of each be analyzed. Based upon EPA's recommendation,
the dissolved fraction should be compared to water quality
criteria, with modifications to the criteria as noted in EPA
(1993b). To compare data to criteria, hardness should be
measured for each sample. Too often, metals data are
compared to criteria using an average hardness value not
directly associated with the monitoring, and not associ-
ated with storm events. In the Williamatte Valley of Or-
egon, stormwater sampling has shown that hardness val-
ues during storm events are quite low, which results in low
criteria values.
Total concentrations are valuable in assessing the over-
all reduction of the heavy metal in both soluble and par-
ticulate forms. There is a concern about the long-term
bioavailability of these metals in sediments and sediment
standards are beginning to be developed and implemented.
When conducting these tests, it is recommended that
low detection limits be achieved. For copper, lead, and zinc,
the detection limit should be 1 u,g/l and for cadmium 0.2
u,g/l. This will minimize problems with analyses that include
below detection limit data, which can severely impact per-
formance evaluations. Special "clean" procedures will be
necessary to achieve low detection limits, both in the labo-
ratory and in the field.
Too often, BMP effectiveness for metals is estimated
based upon data that is very near or below detection. This
is troublesome when both the inflow and outflow concen-
trations are at or near detection, and effectiveness is based
upon a storm-by-storm comparison of loads or concentra-
tions. It is recommended that if both the influent and efflu-
ent concentration are within five times the method detec-
tion limit, the pollutant data pair not be considered in the
effectiveness analysis if a storm-by-storm method is used.
If statistical characterizations of the inflow and the outflow
concentrations are utilized to assess effectiveness and
some of the data are below detection, appropriate tech-
niques should be utilized. Driscoll et al. (1990) describes a
method to address detection limit data. The setting of be-
low-detection values to 0 or 1/2 the detection limit or the
detection limit, will typically lead to an underestimation of
the mean.
Oil and Grease. Oil and grease is a prevalent constitu-
ent in urban runoff and often exceeds discharge limits set
by states (such as 10 mg/l in Oregon for industrial
stormwater permits). In a study of oil and grease concen-
trations in urban runoff in Richmond, California, Stenstrom
et al. (1984) found that oil and grease concentrations in
runoff from commercial properties and parking lots are
about three times higher than from residential and open
areas. The NURP program did not address oil and grease
as a standard constituent. Accurately measuring oil and
grease is very difficult due to its affinity for coating sam-
pling bottles and sampling tubes and its highly non-uni-
form distribution in the water column (except in the most
turbulent situations). Other tests include total petroleum
hydrocarbons, which measure the petroleum based frac-
tion of oil and grease. Other sources of oil and grease in-
clude animal and vegetable. For BMPs which are designed
to address oil and grease, it is suggested that some mul-
tiple, within a storm, grab sample analyses would be ap-
propriate. For most BMPs, it is recommended that the
73
-------
parameter be optional. If completed, the TPH evaluation
is recommended as the most appropriate measure to gauge
effectiveness of a BMP at reducing man-induced sources
of petroleum oil and greases.
Pesticides/Herbicides. Pesticides and herbicides are
regularly detected in urban runoff. However, the number
of constituents usually detected is low and most often at
levels below available criteria. In Portland, Oregon (WCC,
1993a) the frequency of detection of pesticides herbicides
was less than 1% of all the pesticides and herbicides tested.
However, the city has noted locations where pesticide con-
centrations in sediments are high. This could indicate that
the problem might be due to misuse or dumping, rather
than a general stormwater problem. Although it is possible
that pesticides accumulate in sediments from low concen-
trations in stormwater, some regional assessments of the
effectiveness of source control measures (education, iden-
tification and elimination of dumping problems) are needed.
TheAlameda County, CA monitoring program (Cooke and
Lee, 1993) and other studies have recently identified that
the pesticide Diazinon may be a primary cause of toxicity
at very low concentrations (below 8140 method detection
limits) to cerodaphrin dubia in receiving streams in the south
bay area of San Francisco. More research is needed to
further define the level of this problem in relation to the
actual instream biota, rather than test organisms. At this
time, I would not recommend including the pesticide in a
standard list, but research studies on the magnitude of the
problem and the effectiveness of BMPs on these pesti-
cides should be performed. Due to the low values at which
these constituents can cause problems, it would be very
difficult to assess BMP performance on a wide-scale ba-
sis. For example, it may be more appropriate to eliminate
or control the use of Diazinon rather than research BMP
effectiveness on concentrations that are below 1 ppb.
Volatile and Semi-Volatile Organics. These pollutants
have not generally been detected at a high frequency and
in quantities that exceed available criteria [with the excep-
tion of Polynuclear Aromatic Hydrocarbons (PAHs), which
are discussed separately]. In the recent City of Portland
and Eugene sampling programs (WCC, 1993a and 1993b)
detection rates were less than 2% of all the tested con-
stituents and below all available criteria. These parameters
are not recommended for general analysis unless a BMP
effectiveness study is being conducted in an industrial area
suspected or known to have elevated levels of organics.
Polvnuclear Aromatic Hydrocarbons. The carcinogenic
properties of PAHs have generated increased interest in
the study of their sources, transport, fate, and aquatic tox-
icity. Major sources include the combustion of fossil fuels,
uncombusted petroleum products (fuels, etc.), and natu-
ral and man-caused fires. PAHs have recently been ana-
lyzed utilizing detection levels that are significantly below
those achieved utilizing the standard semi-volatile organic
scans (WCC, 1993a and 1993b; Cooke and Lee, 1993).
These tests (GC-MS methods at the nanogram per liter
level) have shown that PAHs in stormwater are above hu-
man consumption criteria by significant amounts (up to over
100 times). However, these tests are specialized (only a
few laboratories provide this level of analysis) and expen-
sive (about $500 to $600 per analysis). In addition, there
are no criteria for aquatic life, and toxicity identification
evaluations performed in the San Francisco Bay Area have
not identified PAHs as the source of toxicity in either de-
veloped land-use runoff or in stream stations. For these
reasons, PAHs are not recommended for the standard list
of constituents to be monitored. However, because of their
carcinogenic nature and their tendency to bioaccumulate,
new studies may identify potential long-term aquatic life
impacts that may require reevaluation of this recommen-
dation.
Data Reporting
Practical and technical data reporting considerations,
including consistent formatting of data, the clear indication
of QA/QC results, standard comparisons to water quality
criteria, reporting of tributary watershed characteristics, and
BMP design information would facilitate data usefulness.
The last two items are considered critical for evaluation of
what contributed to BMP effectiveness in one location over
another.
Data Formatting. It is recommended that all constituent
concentration data be reported as event mean concentra-
tions (EMCs). Table 2 is an example format for reporting
storm event EMCs. It indicates the date of the storm, the
EMC value for each sampling point, the data that are esti-
mates based upon QA/QC evaluations, method used for
analysis, and detection limit achieved. Also included are
summary statistics of the EMCs. These statistics should
be based on use of the lognormal distribution. The NURP
and FHWA studies (EPA, 1983; Driscoll etal., 1983) iden-
tified the lognormal distribution as suitable for characteriz-
ing EMC distributions. An example of the variability in data
is shown in Figure 5. The figure shows a log-probability
plot for total copper collected at a commercial land use
station. The event mean concentrations ranged from 6 to
70u,g/I.This high degree of variability is why proper statis-
tical techniques should be employed to evaluate whether
a measured difference between BMP before/after or in-
put/output is truly different.
The inclusion of outlet data as a part of any paper or
report will allow comparisons of typical outlet concentra-
tions and may allow the determination of the lowest or av-
erage expected concentration from a particular type of BMP.
For example, it may be that wet ponds may only be able to
treat to some minimum concentration range at the outlet
and the "effectiveness" is greatly impacted by the inlet con-
centrations.
Quality Assurance/Quality Control (QA/QC). All monitor-
ing studies should include a QA/QC program. The results
of the QA/QC program should be reported in monitoring
study reports and summarized in papers. It is especially
important to discuss when data are characterized as esti-
mates due to QA/QC results and when detection limits were
affected. Too often this information is not included.
74
-------
Table 2. Example Data Reporting Table from Eugene NPDES Monitoring Summary Report (Woodward-Clyde Consultants, 1993b)
Chromium (mg/L) Method EPA 7191 Receiving Water Quality Criteria*
Storm Event Date Sample R-1 C-1 1-1 I-2 M-1 M-2 Detection Limit
#1
#2
#3
#4
#5
#6
#7
#8
Copper
Storm Event
#1
#2
#3
#4
#5
#6
#7
#8
9/23/92
1 2/5/92
12/16/92
1/19/93
3/1 4/93
Median
COV
Mean
(mg/L)
Date Sample
9/23/92
1 2/5/92
12/16/92
1/19/93
3/1 4/93
Median
COV
Mean
0.034
0.005
0.004
0.004
0.003
0.006
1.27
0.010
R-1
0.081
0.004
0.011
0.009
<0.030
0.012
2.03
0.030
0.016
0.004
0.012
0.006
0.008
0.70
0.010
C-1
0.130
0.016
0.046
0.027
0.040
1.11
0.060
0.031
0.005
0.008
0.019
0.020
0.014
0.86
0.018
Method
1-1
0.071
0.01
0.037
0.076
0.034
0.037
0.97
0.051
0.008
0.003
0.009
0.011
0.017
0.008
0.71
0.010
EPA 6010
I -2
0.016
0.01
0.03
0.034
0.025
0.021
0.54
0.024
0.001
0.008
0.004
0.003
-
-
M-1
0.009
0.027
0.020
0.017
-
-
0.003
0.003
0.004
0.007
0.004
0.42
0.004
M-2
0.019
0.009
0.012
<0.030
0.013
-
-
0.007
0.007
0.007
0.007
0.007
Receiving Water Quality Criteria**
Detection Limit
0.001
0.001
0.004
0.003
0.004
Results expressed as mg/L (ppm) unless otherwise noted. COV is the Coefficient of Variation. ** Criteria are hardness dependent.
"nd" means none detected at or above the detection limit listed. If no value is shown, the lab analysis was not performed.
Summary statistics are based on the assumption that the samples of EMCs are lognormally distributed.
Italicized values are considered estimates due to QA/QC review but are included in the calculations.
100
10
o
99.9
Individual Station Variability
Santa Monica Pier- Commercial
N = 9
Median = 19
COV = 1.10
Mean = 28
PPCC = 0.968
Skew = 0.330
Kurtosis = 1.707
95 90 80 7060504030 20 10
Percent Equal or Greater
Figure 5. Example log probability plot of storm event mean concentrations from data collected by the Los Angeles County Stormwater Monitoring
Program at the Santa Monica Pier (Santa Monica, CA) Commercial Land Use Station.
75
-------
Comparisons to Water Quality Criteria. Another method
to gauge effectiveness could be to monitor how the BMP
effects the number of times that criteria are exceeded in
both the inflow and the outflow, to assess how the BMP
reduces (ordoes not reduce) the frequency of storm events
where water quality criteria are exceeded. For heavy met-
als analyses, it is recommended that hardness be collected
for all storms monitored and that comparisons to criteria
be made utilizing the dissolved fraction with the computed
aquatic criteria as modified by EPA (1993b). Figure 6 pre-
sents an example presentation of metals exceedances for
data collected in Portland, OR (WCC 1993a). These data
could be compared to BMP data for exceedances to de-
termine whether or not a BMP was actually reducing po-
tential toxicity.
Watershed BMP Design Parameters. Urbonas (1995)
described information that should be collected regarding
the physical, climatic, and geologic parameters, which in-
clude watershed and BMP design characteristics that could
likely affect the performance of a BMP. Table 3 (Strecker
and Urbonas, 1995) presents a summary of these param-
eters. More detailed and updated lists will be published
upon completion of the EPA study referenced earlier.
Estimation of Pollutant Removal
Effectiveness
BMP pollutant removal effectiveness estimations are not
straightforward and a wide variety of methods have been
employed. Martin and Smoot (1986) discussed the follow-
ing three types of methods to compute efficiencies:
• The first method employs an efficiency ratio (ER), which
is defined in terms of the average event mean con-
centration (EMC) of pollutants from inflows and out-
flows, thus:
ER = 1 - Average outlet EMC
Average inlet EMC
• The second method is based on the summation of
loads (SOL) of pollutants removed during the moni-
tored storms, thus:
SOL = 1 - Sum of outlet loads
Sum of inlet loads
• The third method of determining efficiency, developed
by Martin and Smoot (1986), defines the ratio as the
slope of a simple linear regression of inlet loads and
outlet loads of pollutants. The equation forthe regres-
sion of loads (ROL) efficiency is thus:
Loads in = B Loads out
where B equals the slope of the regression line, with
the intercept constrained at zero.
The ER and SOL methods assume that monitored storms
include samples representative of all storms that occur.
The SOL method assumes that enough samples were col-
lected so that any significant input loads or output loads
were not missed. They are different in that one gauges
effectiveness in terms of concentration reduction, while the
other gauges effectiveness in terms of load of pollutant
removed. The ROL method assumes that the treatment
efficiency is the same for all storms, which is likely not the
case.
Open
Residential
Commercial
Transportation
Industrial
Cadium
Copper
Lead
Zinc
Figure 6. Frequency of water quality criteria exceedances of Oregon urban stormwater data collected for the Municipal Stormwater NPDES
Programs.
76
-------
Table 3. Parameters to Report with Water Quality Data for Various BMPs
Retention Extended Wetland Grass/Swale Sand/Leaf Oil & Infiltration
(Wet) Detention Pond Wetland Compost Sand Trap and
Parameter
Type
Tributary
Watershed
Parameter
(1)
Tributary watershed area
Pond
(2)
•
Basin
(3)
•
Basin
(4)
•
Channel
(5)
•
Filter
(6)
•
(Vault)
(7)
•
Percolation
(8)
•
General
Hydrology
Water
General
Facility
Wet Pool
Total tributary watershed impervious
percentage
Percent of impervious area hyd.
connected
Gutter, sewer, swale, ditches in watershed?
Land use types (res, comm, ind. open) and
acreages
Average storm runoff volume
50th percentile storm runoff volume
Coefficient of variation of runoff volumes
Average daily base flow volume
Average runoff interevent time
50th percentile interevent time
Coefficient of variation of interevent times
Average storm duration
50th percentile storm duration
Coefficient of variation of storm durations
2-year flood peak velocity
Depth high groundwater of impermeable
layer
Water temperature
Alkalinity, hardness and pH
Sediment setting velocity distribution,
when available
Facility on- or off-line?
If off-line, amount of flow bypassed annually
Type and frequency of maintenance
Inlet and outlet dimensions and details
Solar radiation, when available
Volume of permanent pool
Permanent pool surface area
Littoral zone surface area
Length of permanent pool
(continued)
77
-------
Table 3. Continued
Parameter
Type
Parameter
(1)
Retention Extended Wetland Grass/Swale Sand/Leaf Oil & Infiltration
(Wet) Detention Pond Wetland Compost Sand Trap and
Pond Basin Basin Channel Filter (Vault) Percolation
(2) (3) (4) (5) (6) (7) (8)
Detention
Volume
Pre-
Treatment
Wetland
Plant
Detention (or surcharge) volume
Detention basin's surface area
Length of detention basin
Brimfull emptying time
Half-brimfull emptying time
Bottom stage volume
Bottom stage surface area
Forebay volume
Forebay length
Other BMPs upstream?
Wetland type, rock filter present?
Percent of wetland surface at P03 and
P06 depths
Meadow wetland surface area
Plant species and age of facility
Adapted from Urbonas (1995)
Some researchers have suggested that one should uti-
lize an efficiency measure based upon storm pollutant loads
into and out of the BMP on a storm-by-storm basis. This
would weight the effectiveness considering that all storms
are "equal" in computing the average removal. However, it
is readily apparent that all storm volumes and their associ-
ated concentrations are not equal. Similarly one could uti-
lize concentrations on a storm-by-storm basis.
One factor that complicates the estimation of effective-
ness is that, for wet ponds and wetlands (and other BMPs
where there is a permanent pool), comparing effectiveness
on a storm-by-storm basis neglects the fact that the out-
flow being measured may have a limited or no relationship
to the inflow. In analysis of rain gauges utilizing SYNOP
(Driscoll, et al., 1989), if a basin sized to have a perma-
nent pool equal to the average storm, about 60 to 70% of
the storms would be less than this volume. In many cases,
the flows leaving may have little or no contribution to flows
entering the pond. Therefore, storm-to-storm comparisons
are probably not valid. In cases like this, it is probably more
appropriate to utilize statistical characterizations of the in-
flow and outflow concentrations to evaluate effectiveness
or, if enough samples are collected (i.e., almost all storms
monitored), to utilize total loads into and out of the BMP.
Using the same set of data, Table 4 compares three of
the methods including percent removal by storm with a
statistical characterization of inflow/outflow concentration
and a simple comparison of total loads in and out for the
sampled storms. As one can see, the removals estimated
differ by up to 19 percentage points. In this record, there
are several storm events where inflow concentrations were
relatively low and therefore the system was not "effective".
Based upon these factors, it is recommended that the
statistical characterization of inflows vs. outflows be uti-
lized (ER). This enhances the ability to conduct statistical
tests of whether the reported differences are greater than
zero. If enough data on storms are collected (e.g. continu-
ous samples over an extended period), the total loads in
and out (SOL) is probably an acceptable method also.
BMP Evaluations - Statistical
Considerations
As noted in many studies of urban runoff, the variability
in runoff concentrations from event to event is large. If one
were to attempt to statistically characterize a BMP influent
concentration (and outflow), the more data the better, Fig-
ure 7 is a schematic of how more data can improve (re-
duce confidence interval of) results. As mentioned above,
there are a number of types of BMP evaluations that can
be conducted. First, the standard evaluation of a single
BMP, testing input and output; second, the evaluation of
multiple BMPs within a basin (before/after or control ba-
78
-------
Table 4. Example Wetland TSS Removal
Volume of
Storm (ft3) Inflow = Outflow
1 445,300
2 649,800
3 456,100
4 348,111
5 730,261
Med
Cov
Mean
Concentration
In Out
(mg/L)
352
30
99
433
115
139
1.48
249
Cone
24
25
83
141
63
65
.86
85
66%
Load
In Out <
(Ibs)
9780
1220
2820
9410
5240
28,470
Loads
670
1010
2360
3060
2870
A
9,970 V
G
65%
% Removal
by storm
93%
17
16
67
45
48%
90%
Confidence
Interval
T
rea:
den>
i
Mean\ Decreased 90%
Confidence Interval
Figure 7. Expected change of 90% confidence interval of station mean with additional data.
sin); and finally a third, the evaluation of a BMP with mul-
tiple inlets (where it might be very difficult (expensive) to
evaluate the BMP utilizing input/output). All methods should
require that a rigorous statistical approach be applied in
selecting the number of samples to be collected to assure
detection of a given level of change.
As an example of the number of samples required to
detect a "true" difference, Table 5 presents an analysis of
two of the Portland NPDES monitoring stations (WCC,
1993a) where 10 flow-weighted composited samples were
collected. The Fanno Creek station is a large (about 1,200
acres) residential catchment, while the M1 station is a
smaller (about 100 acres) mixed land use station. An analy-
sis of a variance-based test was utilized with the existing
data to determine how many samples are estimated to be
needed to detect a 5%, 20%, and 50% change in the mean
concentration at the station. The test was performed con-
sidering an 80% probability that the difference will be found
to be significant, with a 5% level of significance (Sokal and
Rohlf, 1969). This analysis does not consider potential
seasonal effects on the collection of data as a factor. Even
so, quite a large number of samples would be required to
detect a 5% to 20% difference in concentrations. Figure 8
shows a map of the US plotting the average number of
storms per year (over 0.1") as determined by EPA (Driscoll
et al., 1989) occur. One can see that in many locations, it
would take a number of years of sampling all storm events
to be able to detect small differences.
There are numerous examples in the literature where
small differences (2 to 5%) are reported based upon much
fewer samples than indicated by this analysis. This high-
lights the need to be more rigorous with regard to statisti-
cal testing of reported effectiveness estimates. To detect
larger changes, the number of samples becomes reason-
able. The mixed land use catchment in Portland is cur-
rently being studied forthe effectiveness of the implemen-
tation of a number of source controls and other controls
that do not lend themselves to input/output testing. Ex-
amples include maintenance changes (catch basin clean-
ing, street sweeping), education (business and residences),
tree planting, etc. Post-BMP monitoring will be conducted
along with qualitative evaluations.
As an example that demonstrates how one could evalu-
ate whether one catchment is different than another, Fig-
ure 9 presents results of analysis of stormwater monitor-
ing data collected in Oregon. The figure presents a statis-
tical characterization of land use data, demonstrating that
for Total Copper, the open and residential land use sta-
tions are statistically different from all other land uses as
well as from each other. A similar analysis technique should
be employed for all before and after tests, as well as "con-
trol" tests.
79
-------
Table 5. Analysis of Sample Sizes Needed to Statistically Detect Changes in Mean Pollutant Concentrations from 2 Stations in Portland, OR.
Number of Samples Required to Detect the
Indicated % Reduction in Site Mean
Concentration*
Monitoring Site
Parameter
5%
*80% certain of detecting the indicated % reduction in mean of the EMCs.
20%
50%
R1 - Fanno Creek
Residential
M1 -NE 122nd
Columbia
Slough Mixed Use
TSS
Copper
Phosphorus
TSS
Copper
Phosphorus
202
442
244
61
226
105
14
29
16
5
15
8
4
6
4
2
4
3
20
40
IB
10
10
78
20
30 40
Figure 8. Annual average number of storms, (storms/year)
Other Considerations
There is a need to conduct dry weather analyses be-
tween storms on BMPs with dry weather flows; it may be
that pollutants captured during storms are slowly released
during dry weather discharges.
Biological and downstream physical habitat assessments
such as aquatic invertebrate sampling and habitat classifi-
cation should be explored as an alternative to merely uti-
lizing chemical measures of effectiveness (see Maxted,
these proceedings); long-term trends in receiving water
quality, coupled with biological assessments, would likely
be a much better gauge of the success of the implementa-
tion of BMPs, especially on an area-wide basis.
Summary and Recommendations
There is a great need for consistency in the constituents
and methods utilized for assessing BMP effectiveness. This
80
-------
Ln (Total Cu) by Land-use Types
Commercial Indus-Open Indus-Piped
Open
Residential Transport.
overlap
marks
group I
mean
50% "^
quantile.
x'
overlap
marks
*
"* JUfflrf
-------
Sartor, J.D., and Boyd, G.B. 1972. Water Pollution Aspects
of Street Surface Contaminants. EPA-R2-72-081, U.S.
Environmental Protection Agency, Washington D.C.
Sokal, R.R. and F. James Rohlf. 1969. Biometry: The Prin-
ciples and Practice of Statistics in Biological Research.
W H. Freeman and Company. San Francisco, CA.
Stenstrom, M.K., Silverman, G.S. and T.A. Bursztynsky.
1984. "Oil and Grease in Urban Stormwaters," Jour-
nal of Environmental Engineering Division, ASCE. Vol.
110, No.1, pp. 58-72.
Strecker, E. 1994. Constituents and Methods for Assess-
ing BMPs. Proceedings of the Engineering Founda-
tion Conference on Stormwater Related Monitoring
Needs. ASCE. Aug. 7-12, Crested Butte, CO.
Strecker, E.W., Kersnar, J.M., Driscoll, E.D., and R.R.
Horner. 1992. The Use of Wetlands for Controlling
Stormwater Pollution. The Terrene Institute. Washing-
ton, D.C.
Strecker, E. and B. Urbonas. 1995. Monitoring of Best
Management Practices. Proceedings of the 22nd An-
nual Water Reources Planning and Management Di-
vision Conference, ASCE, NY. pp48-51.
Strecker, E., M. lannelli, and B. Wu. 1997. Analysis of Or-
egon Urban Runoff Water Quality Monitoring Data
Collected from 1990 to 1996. Prepared by Woodward-
Clyde for the Association of Clean Water Agencies.
U.S. Environmental Protection Agency. 1983. Final Report
on the National Urban Runoff Program. Water Plan-
ning Division, U.S. EPA. Prepared by Woodward-Clyde
Consultants.
U.S. Environmental Protection Agency. 1989. Analysis of
Storm Events Characteristics for Selected Rainfall
Gauges throughout the United States. Draft Report to
the U.S. Environmental Protection Agency, Office of
Water, Nonpoint Source Division, 43 pp.
U.S. Environmental Protection Agency. 1993a. Manual of
CSO Control. EPA-6251R-93-0007.
U.S. Environmental Protection Agency. 1993b. Memoran-
dum. Office of Water Policy and Technical Guidance
on Interpretation and Implementation of Aquatic Life
Metals Criteria.
Urbonas, B.R. 1994. Parameters to Report with BMP Moni-
toring Data. Proceedings of the Engineering Founda-
tion Conference on Stormwater Monitoring Related
Monitoring Needs. ASCE. August 7-12, Crested Butte,
CO.
Urbonas, B.R. 1995. "Recommended Parameters to Re-
port with BMP Monitoring Data. J. Water Resources
Planning and Management, ASCE. 121 (1), 23-34.
Waschbusch, R. and D. Owens. 1998. Comparison of Flow
Estimation Methodologies in Storm Sewers. Prepared
by USGS for the Federal Highway Administration.
January 16.
Woodward-Clyde Consultants. 1993a. Final Data Report:
Data from Storm Monitored between May 1991 and
January 1993. Submitted to Bureau of Environmental
Services, City of Portland, OR.
Woodward-Clyde Consultants. 1993b. Draft Data Report:
Data from Five Storms Monitored between Septem-
ber 1992 and March 1993. Submitted to Department
of Public Works, City of Eugene, OR.
82
-------
Targets of Opportunity: Alexandria's Urban
Retrofit Program
Warren Bell, RE.
City Engineer, City of Alexandria, Virginia
Philip C. Champagne, RE.
Dewberry & Davis
Fairfax, Virginia
During 1992 preparations for a stormwater quality pro-
gram as part of the Alexandria (Virginia) Chesapeake Bay
Preservation Ordinance, the city engineering staff made a
survey to identify opportunities for future urban BMP retro-
fitting. The objective of the current "Targets of Opportu-
nity" program is to enhance the minimum requirements of
the Virginia Chesapeake Bay Preservation Act (CBPA) pro-
gram by providing treatment of stormwater runoff from built-
up areas not directly addressed by the CBPA, in order to
further reduce pollutants reaching the bay and its tributar-
ies.
Since the inception of the Targets of Opportunity Pro-
gram, almost 1,000 acres of urban BMP retrofits have been
installed within Alexandria. A substantial part of Alexandria's
urban retrofits have been voluntarily designed and con-
structed by developers of adjacent downhill properties.
While comprising only 3.3% of the urbanized area within
the Potomac and Shenandoah basin, the city has already
contributed almost 23% of the total urban retrofit coverage
proposed in the Shenandoah and Potomac River Basins
Tributary Nutrient Reduction Strategy (Commonwealth of
Virginia, 1996.) Estimated nutrient reductions already ex-
ceed the urban retrofit phosphorus and nitrogen reduction
targets contained in the Shenandoah Strategy, and are
within 3% of meeting the nitrogen reduction target.
Characteristics of Alexandria
Alexandria is a city of approximately 115,000 citizens
located on the Potomac River in Northern Virginia between
Fairfax County and Arlington County. Founded in 1749 as
a deep river seaport, the city is currently some 15.75 square
miles in size. The population is diverse and relatively afflu-
ent, some 56,000 households with a median income of
approximately $53,000. Directly across the Potomac from
Washington, D.C., Alexandria's largest employer remains
the federal government. However, extensive development
during the past decade has resulted in a thriving commer-
cial sector, with divisional, regional, and multinational head-
quarters for operations ranging from research and devel-
opment to high technology, associations, and professional
services now located in the city. With a 1990 population
density of 7,281 people per square mile. Alexandria is the
most densely developed city in Virginia and the eleventh
most densely populated city in the United States. Approxi-
mately 41% of the total city area is covered with impervi-
ous surfaces.
Alexandria is a city bounded by and laced with streams.
The eastern boundary is 5.6 miles of Potomac River shore-
line. The northern boundary with Arlington County includes
1.9 miles of Four Mile Run, 100-year flood channel recon-
structed by the U.S. Army Corps of Engineers in the 1970s
following extensive flooding during Hurricane Agnes. Asimi-
lar 1.7 miles of channel conveying Cameron Run borders
the Capital Beltway, the southern border with Fairfax
County. Approximately 8.4 miles of small tributary streams
flow north or south into the boundary channels, approxi-
mately 20% to Four Mile Run, 20% directly into the Potomac
River, and 60% into Cameron Run or its major tributaries,
Backlick Run and Holmes Run. Almost all of Alexandria's
streams except the Potomac River are severely degraded
urban streams. Protection and partial restoration of these
streams and the Chesapeake Bay into which they flow is
the focus of the city's nonpoint source programs.
Governing Clean Water Programs
Alexandria's clean water programs are governed by sev-
eral federal and state authorities. Initially, the city was clas-
sified as a medium-sized city under the National Pollution
Discharge Elimination System (NPDES) Permit require-
ments for separate stormwater systems, requiring an MS4
permit. (The city was later reclassified as a small city based
on population reductions allowed for regions served by
permitted combined sewer systems). Virginia has also
adopted a number of programs in support of the federal
and multi-state Chesapeake Bay Program. The Virginia
CBPA includes a mandatory program requiring provisions
for stormwater quality on development projects within the
bay watershed.
The Virginia Stormwater Management Act is a discre-
tionary program which allows stormwater quality to be re-
quired of development projects throughout the common-
wealth. The Virginia Erosion and Sediment Control Act re-
quires stormwater quality measures during construction
83
-------
on all but very small development projects. The Chesa-
peake Bay Tributary Nutrient Strategies require nutrient
reductions from all point and nonpoint sources to achieve
40% reductions in the Bay. These reductions to get nutri-
ent levels below the 1985 baseline have been targeted for
the year 2000 by the bay program signatories. Only the
Tributary Strategies address the question of urban BMP
(Best Management Practices) retrofits.
Tributary Targets
The Shenandoah and Potomac River Basins Tributary
Nutrient Reduction Strategy (the Strategy) defines urban
BMP retrofits as "Modifying existing stormwater facilities
to enhance water quality and/or retrofitting stormwater
drainage systems to add water quality components in al-
ready developed areas to slow runoff, remove sediment
and nutrients, and provide a basis for restoring eroded
stream channels." The strategy sets a target of 4,356 acres
of urban retrofit within the entire basin, of which 1,156 acres
was in addition to that existing at the time of the printing.
The Strategy lists total urbanized watershed as 454 square
miles (290,400 acres), including Alexandria's 15.75 square
miles (10,080 acres), approximately 3.5% of the water-
shed.
Targets of Opportunity Program
Stormwater Quality Program Adoption
Alexandria's initial stormwater quality program was its
erosion and sediment control ordinance, which was
adopted in the 1970s. In 1990, the city began an intensive
effort to prepare an application for an NPDES permit for its
stormwater program. Concurrently, city staff enacted a
Chesapeake Bay Preservation Ordinance (City of Alexan-
dria, 1991) which contains provisions from both the Chesa-
peake Bay Preservation Act and the Stormwater Manage-
ment Act. City staff identified several sites where urban
retrofits appeared possible. Staff members responsible for
reviewing development proposals were directed to discuss
the possibility of including urban retrofit as part of proposed
developments. The objective of this program was to en-
hance the mandatory requirements of the Chesapeake Bay
Preservation program with additional treatment of
stormwater runoff from built-up areas not directly addressed
by that act, to further reduce pollutants reaching the bay
and its tributaries. The program was already in place and
functioning when the Virginia Potomac Basin Tributary
Nutrient Reduction Strategy was developed and adopted
in 1995-1996 (Commonwealth of Virginia, 1996).
Elements of the Targets of Opportunity
Program
There are four basic elements to the Targets of Oppor-
tunity Program. The first is knowledge of the watersheds
within the jurisdiction. The Alexandria staff used aerial pho-
tographs, topographic maps, and the sewer outfall map
prepared for the Part I NPDES Stormwater submission.
Discussions with storm and sanitary sewer maintenance
personnel with many years of experience in the city were
also very valuable.
The second element is the identification of potential op-
portunities for urban retrofits. Sites in the watersheds of
streams which receive large numbers of stormwater outfalls
are especially desirable to maximize the effectiveness of
retrofit BMPs. Alexandria also focused on areas with ex-
isting ponds and detention basins which could be adapted
in the future for service as either regional retention basins
(wet ponds) or extended detention basins (dry ponds).
The third element is one of the most crucial: early explo-
ration of urban retrofit options with owners/developers.
Alexandria's zoning ordinance requires a pre-submission
conference with the city staff for all significant construction
projects. This conference usually occurs prior to finaliza-
tion of the stormwater concept plan for the respective site,
allowing the staff an opportunity to discuss retrofit options
with the development team. For smaller projects, the staff
almost always becomes aware of proposed development
well before formal submission through informal contacts
with the engineering community.
Once contact is established, the fourth element comes
into play: creating "win-win" situations for both the devel-
opers and the public. In some cases, developers may find
it less expensive to treat the entire flow of existing storm
sewers transiting a site rather than construct a separate
"off-line" system to collect and treat stormwater runoff. Al-
ternatively, construction (at the developer's expense) of
regional facilities on public land may be more economi-
cally beneficial to a developer than construction on-site
BMPs. Fostering a spirit of cooperation between the par-
ties rather than an adversarial regulator/regulated rela-
tionship is crucial to obtaining results in a program of this
nature.
Status of the Targets of Opportunity
Program
Seven significant urban stormwater BMP retrofit projects
have been approved underthe Targets of Opportunity Pro-
gram. A discussion of each of the projects illustrates how
the various program elements were implemented.
Winkler Run Regional Retention Facility
The Mark Winkler Corporation, which owns a large de-
velopment tract in western Alexandria, proposed to con-
struct a combined stormwater detention/water quality pond
system to stop a severe erosion problem and to provide
detention and water quality for future buildout of their prop-
erty. Noting that the watershed draining through the site
included significant built-up areas, the staff discussed with
the developer's engineer the possibility of sizing the pond
to provide water quality for the entire watershed at full
buildout under city zoning. The developer agreed and of
the total watershed of 221 acres, 126.7 acres was urban
retrofit. Estimated impervious cover on this acreage, which
includes 26 acres of Interstate 395 and the heavily trav-
eled Seminary Road and Beauregard Street, apartment
houses, and hotel and office complexes, is 82.3%. The
Winkler Run Pond System was built to a design from the
84
-------
Washington Council of Governments manual, Controlling
Urban Runoff, (Schueler, 1987) and is rated to provide 45%
phosphorous removal. The system of two ponds in series
was constructed in 1992. Based on information in the
manual, city staff estimate total nitrogen removal at ap-
proximately 30% (see the Attachment for nutrient estimat-
ing methodology). At these ratings, the total yearly reduc-
tions from the urban retrofit areas are 397 pounds per year
of phosphorus and 1,960 pounds per year of nitrogen. The
system was constructed solely at the developer's expense.
Lake Cook Regional Retention Facility
The next significant urban retrofit involved a project to
restore a viable habitat for aquatic life in Lake Cook, a
recreational lake owned by the city and operated by the
Northern Virginia Regional Park Authority (NVRPA). In
1993, NVRPA requested the city to restore sufficient depth
to maintain for recreational fish in Lake Cook in the
Cameron Run Regional Park. Originally four feet deep,
the three-acre lake had silted until less than two feet of
depth remained. During an earlier review of the outfall
map, the engineering staff had noted that the lake was fed
by Strawberry Run, a stream receiving the outfalls from
over 30 storm sewers having a diameter of 36 inches or
greater. The staff determined that if the lake were deep-
ened to an average of six feet, the pool could serve as a
regional wet pond BMP for approximately the 385 acres of
fully developed watershed draining into the lake. A sedi-
ment forebay was also added to trap sediments at the up-
stream end of the lake, protecting the fish habitat and eas-
ing future maintenance.
The lake was deepened during the winter of 1993-1994.
Only approximately 20% of the total project cost of $75,000
(funded by Alexandria general revenues) was attributable
to the stormwater quality features. Based on the BMP phos-
phorus removal ratings currently proposed for inclusion
within the Virginia Stormwater Management Regulations,
Virginia Chesapeake Bay Protection Regulations, and the
WASHCOG nitrogen removal estimates for such ponds,
the staff estimates that the lake now removes approximately
926 pounds of phosphorus and 4,222 pounds of nitrogen
per year from stormwater runoff entering Cameron Run.
Cameron Lake Regional Retention Facility
In the early 1990s, the Department of Defense decided
to close the 164.5-acre Cameron Station Army Base in
Alexandria and sell the land for private development. While
developing future zoning of the property, city staff noted
that two connected lakes on the station which would be-
come a city park acted as stormwater detention ponds for
approximately 60 acres of runoff. Total surface area of the
lakes was almost 5 acres, and their permanent pools could
be deepened to serve as a regional retention pond for a
much larger area, including some of the most intensely
developed areas of the city. Due to early coordination with
the Army, the city was able to insert a condition that re-
quired all future development on the site to drain through
the lakes. The condition also required city storm sewers
which transited the base to be rerouted through the lakes.
The purchaser of the base readily participated in this pub-
lic-private partnership, recognizing that it would provide a
win-win relationship.
By allowing the developer to retrofit the existing lakes,
greater densities of development were created, which more
than compensated for the cost of upgrading the retention
facility. Additionally, if the developer had held the density
of development as originally planned, a series of sand fil-
ters would have been needed to provide water quality for
the 97.5 acres of development at a cost considerably higher
than the retrofitting of the existing lakes. Recognizing that
the regional facility would provide a win-win situation, the
developer agreed to the incorporation of several state-of-
the-art features into the retrofit.
Work began with draining the lake and removing approxi-
mately 20,000 cubic yards of material to create an appro-
priate permanent pool. The existing outlet structure, con-
sisting of little more than a concrete flume with a wire trash
rack, was removed and replaced with an upflow anaerobic
trickling filter. Additional features include a sediment fore-
bay which can be isolated from the permanent pool during
maintenance and an oil skimmer to retain floating hydro-
carbons, trash, etc., from reaching the main basin of the
lake. Facilities to monitor flow rates through the pond and
chemical composition of the flows were also provided.
Constructed during the summer of 1997, the regional re-
tention facility is treating runoff from 246.83 acres, 187
acres of which did not previously drain through the lakes.
The entire drainage shed except for the 97.5 acres of de-
velopment property is urban retrofit. Based on Virginia and
WASHCOG BMP ratings, the staff estimates that, when
full buildout of Cameron Station is completed, the facility
will remove approximately 709 pounds of phosphorus per
year and 3,235 pounds of nitrogen from stormwater runoff
entering Backlick Run.
Park Center Regional Extended Detention
Facility
Since 1992, the city staff has been recommending the
conversion of a large 100-year storm detention facility in
the western part of the city into a stormwater quality BMP.
Early plans to convert it to a regional retention facility col-
lapsed when one of the adjacent property owners objected
to the presence of a permanent pool. However, in the fall
of 1996, the owner of the basin submitted a development
plan to construct new office towers adjacent to the basin.
Ratherthan require construction of new BMPs forthe 8.12-
acre development, the city proposed to the developer's
engineer that the basin be converted into a regional ex-
tended detention facility to serve a 245-acre watershed.
An adequate dam and riser structure able to accommo-
date a 100-year storm was already in place; the conver-
sion involved only modifications to the riser structure to
provide a reduced orifice at the bottom of the dam and
new overflow openings at the top of the new BMP deten-
tion pool. The entire expense of this construction, which is
currently in progress, is being borne by the developer. Using
the Virginia BMP rating of 35% for phosphorus reduction,
85
-------
the staff estimates that the 237 acres of urban retrofit is
removing approximately 307 pounds of phosphorous from
Lucky Run, a tributary of Four-Mile Run and the Potomac
River. The staff is not currently convinced that significant
nitrogen removal occurs in extended detention ponds.
Slater's Village Regional Extended
Detention Facility
When a development with 145 townhouses and 128
condominium units was proposed in 1996 for part of the
old Potomac Rail Yard at the northern portion of Alexan-
dria, the developer's engineer observed that reinforced
concrete storm sewer serving existing developments tra-
versed the site. It was determined that it would be less
expensive for the developer to construct an extended de-
tention facility to treat all of the runoff conveyed by this
sewer than to construct a completely separate storm sewer
system and BMP to serve only the new townhouses. He
therefore proposed to drain the new development directly
into the storm sewer and treat the total runoff downstream
of the townhouses. This provided an additional 38 acres
of pure urban retrofit to the new extended detention facil-
ity. Using Virginia BMP ratings, the staff estimates that this
urban retrofit removes an additional 75 pounds of phos-
phorus from runoff flowing directly into the Potomac River.
The BMP was constructed as an erosion control basin in
the summer of 1997 and will be converted to a full BMP
upon completion of the construction project.
Episcopal Seminary Regional Retention
Facility
The Episcopal Theological Seminary, a large landholder
in central Alexandria, recently decided to construct a state-
of-the-art stormwater retention facility to use as a teaching
tool for environmental classes at the private high school
on site. The pond, was also designed as a stormwater ret-
rofit pond to serve existing development on the property to
provide stormwater quality for any future expansion of fa-
cilities. The pond serves a 51-acre watershed with an ulti-
mate runoff factor of 0.44. The city staff considers this pri-
vate pond to be totally urban retrofit. Initial estimates sug-
gest that this BMP will remove 128.5 pounds per year of
phosphorus and 586 pounds per year of nitrogen currently
reaching Cameron Run and the Potomac River.
Potomac Retail Center Urban Retrofit
When design began on a 60-acre shopping center to
occupy part of a former rail yard in the northern part of
Alexandria, the developer's engineerwas required to deal
with the runoff from 9.9 acres near U.S. Route 1 including
adjacent properties which already drained through ditches
in the rail yard. Rather than provide a separate convey-
ance for this off-site water, the engineer proposed to route
it through a large retention pond being built to treat devel-
opment runoff. The retention pond, which included up-to-
date features such as a sediment forebay, was sized ac-
cordingly. The city staff estimates that this urban retrofit is
removing approximately 31.8 pounds/year of phosphorus
and 188.7 pounds per year of nitrogen.
Program Impact on the Chesapeake Bay
The seven completed urban BMP retrofit projects de-
scribed above have provided a total of 996.8 acres of ur-
ban retrofit since 1992, 23% of the Potomac/Shenandoah
basin total target and 82.9% of the increased coverage
target. Total phosphorus removal from these projects is
estimated at 2,544.8 pounds per year and total annual re-
moval of nitrogen is estimated at 10,193.0 pounds. The
annual phosphorus removal represents 220% of the Nutri-
ent Strategy total basin target and 839% of the increased
coverage target. The annual nitrogen reductions represent
97% of the total basin target and 368% of the increased
coverage target.
In December of 1997, the City of Alexandria was awarded
a Community Innovation Award by the Chesapeake Bay
Local Government Advisory Committee for "its contribu-
tion and commitment to the protection and restoration of
streams, rivers, and the Chesapeake Bay through the
implementation of its Stormwater Urban BMP Retrofit Pro-
gram - Targets of Opportunity."
Transferability of Program
Any jurisdiction having a formal stormwater quality pro-
gram, such as the Virginia Stormwater Management Pro-
gram or Chesapeake Bay Preservation Program, could
institute an urban retrofit program similar to Alexandria's.
Detailed engineering studies need not be made to begin a
program, nor are sophisticated tools such as CIS systems
a necessity. A review of aerial photographs of the jurisdic-
tion by staff engineers and storm sewer personnel is usu-
ally sufficient to identify the "targets of opportunity" for fu-
ture urban retrofit upon development or redevelopment in
the watersheds.
References
Bell, Warren, Stokes, Lucky, Gavan, Lawrence J., and
Nguyen, Trong Ngu (1995) Assessment of the Pollut-
ant Removal Efficiencies of Delaware Sand Filter
BMPs. Department of Transportation and Environmen-
tal Services, City of Alexandria, VA.
Commonwealth of Virginia Chesapeake Bay Local Assis-
tance Department (1989) Local Assistance Manual,
Richmond, VA.
Commonwealth of Virginia Department of Conservation
and Recreation (1990) Virginia Stormwater Manage-
ment Regulations, Division of Soil and Water Conser-
vation, Richmond, VA.
Commonwealth of Virginia Department of Conservation
and Recreation (1992) Virginia Erosion and Sediment
Control Handbook, Third Addition, Division of Soil and
Water Conservation, Richmond, VA.
Commonwealth of Virginia Departments of Environmental
Quality, Conservation and Recreation, and Chesa-
86
-------
peake Bay Local Assistance (1996) Virginia's
Shenandoah and Potomac Basins Tributary Nutrient
Reduction Strategy, Richmond, VA.
City of Alexandria, (1991) Chesapeake Bay Preservation
Ordiance, Alexandria, VA
Northern Virginia Planning District Commission (1992)
Northern Virginia BMP Handbook, Annandale, VA.
Schueler, Thomas R. (1987) Controlling Urban Runoff,
Metroploitan Washington Council of Governments,
Department of Environmental Programs, Washington,
D.C.
Schueler, Thomas R., Kumble, Peter A., and Hearaty,
Maureen A. (1992) A Current Assessment of Urban
Best Management Practices. Department of Environ-
mental Programs, Metropolitan Washington Council of
Governments, Washington, D.C.
U.S. Department of Commerce, (1991) Census of Popula-
tion and Housing, 1990: Summary Tape File 1C on
CD-Rom, Bureau of the Census, Washington, D.C.
87
-------
Attachment 1
Nutrient Removal Estimates Methodology
The method to calculate loadings recommended by the
Virginia Chesapeake Bay Local Assistance Department
(CBLAD) and adopted by Alexandria is the Simple Method
derived by Thomas R. Schuelerin the Metropolitan Wash-
ington Council of Governments (COG) handbook, "Con-
trolling Urban Runoff. "The Simple Method is described as
follows:
viousness. The Simple Method uses the following formula
to compute Rv:
L=PxPxR xCxAx2.72/12
where,
L = phosphorus loadings (pounds/year-lb/yr).
P = average annual rainfall depth (inches) = 40
inches per year for Alexandria.
R = unitless correction factor for storms that
produce no runoff = 0.9.
Rv = runoff coefficient = expresses the
fraction of rainfall converted to runoff.
C = flow-weighted mean pollutant concentration
(m i 11 ig ra ms/l ite r- mg/l).
A = area of development site (acres).
2.72 and 12 are conversion constants.
Further reducing the Alexandria constants in the formula
yields:
L = 8.16xRvxCxA
The runoff coefficient describes the fraction of rainfall
converted to runoff. While dependent on soil type, topog-
raphy and cover, it is most influenced by watershed imper-
R = 0.05 + 0.009 (I)
where
I = the % of site imperviousness in whole
numbers.
For watersheds with greaterthan 20% impervious cover,
CBLAD recommends using a flow-weighted mean concen-
trations of phosphorus of 1.08 mg/l.
Asix-month BMP monitoring project in Alexandria in 1994
established an actual flow-weighted mean concentration
of nitrogen in stormwater runoff of 8.0 mg/l.
The new Virginia Stormwater Management Regulations
(also to be used by CBLAD) rating for retention ponds with
2.0 inches of runoff from impervious surfaces in the per-
manent pool is 65% forTP.
Based on various studies reviewed, including
WASHCOG's Controlling Urban Runoff, city staff estimates
TN removal for such ponds at 40%. Pending further analy-
sis of monitoring studies, city staff is not currently assert-
ing any TN removal from extended detention facilities.
Retention ponds with permanent pools of less than 2.0
inches of runoff were rated using data from the Northern
Virginia BMP Handbook and Controlling Urban Runoff.
88
-------
Attachment 2
Total Phosphorus Reduction Calculations
Project
Winkler Run Pond
Lake Cook Retrofit
Cameron Lakes Retrofit
Park Center Basin Retrofit
Episcopal Center Pond
Slater's Village ED Retrofit
Potomac Yard Shopping Center
Total
Urban
Retrofit
Acres
126.7
385.0
149.3
236.9
51.0
38.0
9.9
996.8
%
Imper-
vious
82.3
41.0
86.6
41.0
43.3
65.3
75.5
-
Runoff
Factor
RV
0.79
0.42
0.83
0.42
0.44
0.64
0.73
-
Alexandria
Constant
8.16
8.16
8.16
8.16
8.16
8.16
8.16
-
C
TP
(mg/l)
1.08
1.08
1.08
1.08
1.08
1.08
1.08
-
Annual
LOAD
TP-lb.)
882.1
1425.0
1092.1
876.9
197.8
197.8
63.7
-
BMP
Type
Ret. Pond
Ret. Pond
Ret. Pond
ED Pond
Ret. Pond
ED Pond
Ret. Pond
-
BMP
Effic'y
(%)
45
65
65
35
50
35
50
-
Annual
Load
Reduced
(Ibs)
397.0
926.3
709.8
306.9
98.9
75.0
31.8
2554.8
Total Nitrogen Reduction Calculations
Project
Winkler Run Pond
Lake Cook Retrofit
Cameron Lakes Retrofit
Park Center Basin Retrofit
Episcopal Center Pond
Slater's Village ED Retrofit
Potomac Yard Shopping Center
Total
Urban
Retrofit
Acres
126.7
385.0
149.3
236.9
51.0
38.0
9.9
996.8
%
Imper-
vious
82.3
41.0
86.6
41.0
43.3
65.3
75.5
-
Runoff
Factor
RV
0.79
0.42
0.83
0.42
0.44
0.64
0.73
-
Alexandria
Constant
8.16
8.16
8.16
8.16
8.16
8.16
8.16
-
C
TN
(mg/l)
8.0
8.0
8.0
8.0
8.0
8.0
8.0
-
Annual
LOAD
TP-lb.)
6534.1
10,558.8
8089.4
9897.5
1464.9
1587.6
471.8
-
BMP
Type
Ret. Pond
Ret. Pond
Ret. Pond
ED Pond
Ret. Pond
ED Pond
Ret. Pond
-
BMP
Effic'y
(%)
30
40
40
_
40
-
40
-
Annual
Load
Reduced
(Ibs)
1960.2
4222.3
3235.8
_
586.0
-
188.7
10,193.0
89
-------
Port Towns Revitalization and Environmental Enhancement
Stormwater Projects Revitalize Urban Areas
S. AliAbbasi
Prince Georges County Dept. Of Environmental Resources
Largo, Maryland
Background
In 1993, Prince Georges County, Maryland (Figure 1)
embarked upon an ambitious neighborhood revitalization
program that targets communities inside the National Capi-
tal Beltway. The primary purpose of this effort is to revive
older communities as attractive places to live and work.
By concentrating resources and developing local institu-
tions, state and county officials hope to stem disinvest-
ment and abandonment of these communities.
Prince Georges County has a population of 800,000 and
a median household income of $60,540. It covers 488
square miles and indues 28 incorporated areas. Its eco-
nomic wealth is tied to both Washington, DC, and Balti-
more. Business is largely clustered in the northern part of
the County. The county's southern area is still largely ru-
ral.
The county is bisected, geographically and economically,
by the National Capital Beltway. Although the majority of
the population is located in the municipalities inside the
Beltway, higher income levels are concentrated outside
the Beltway (Tatar, et al., 1995).
The centerpiece of the County's revitalization effort is
the renewal of communities known as the Port Towns, lo-
cated in the heart of Prince Georges County. What started
as a thriving tobacco port and trading point along the
Anacostia River in Colonial days is now a predominantly
blue-collar neighborhood. Built mostly in the 1940s and
1950s, the Port Towns have fallen into disrepair and have
struggled to attract new residents and businesses (Pierre,
1997).
The Port Towns revitalization initiative has generated pub-
lic interest and support due to this community's keen sense
of identity and heritage. With millions of dollars already ear-
marked for various projects, the Port Towns community is
now one of the most prominent revitalization areas in Mary-
land.
Stormwater Revitalization Projects
What is unique about the Port Towns is that innovative
Stormwater retrofit projects are helping to pave the way to
economic renewal. Construction of bioretention1
streetscaping, shallow marsh wetlands, stream rehabilita-
tion, and river restoration projects are being used to revi-
talize the towns. As a new urban revitalization tool, these
Stormwater projects are intended to mitigate adverse en-
vironmental impacts in older urban areas, improve land-
scaping, enhance community pride, and create a whole-
some community image that invites private investment.
Port Towns
The Towns of Bladensburg, Colmar Manor, and Cottage
City are collectively known as the Anacostia "Port Towns"
(Figure 2) industrial and commercial activity. The Port
Towns residential areas include a wide range of generally
pleasant housing along with industrial and commercial
activity. Economic decline in the Port Towns began due
to relocation of retail shops to newer outlying malls, as
well as constraints imposed by older infrastructure.
Facing similar development issues, common economies,
and proximity to US Route 1, the three towns agreed to
coordinate their revitalization efforts. To guide the revital-
ization effort, the Port Towns developed a comprehensive
Vision and Action Plan in 1995. Although an extensive park
or "greenway" along the banks of the Anacostia River
makes an important contribution to the character of this
community, the Port Towns generally lack adequate trees,
streetscaping, and Stormwater management controls (Legg
Mason et al., 1995).
A New Revitalization Tool - Stormwater
Retrofit Projects
For years, the County has considered the restoration of
the Anacostia River an important part of its capital pro-
gram. Although the County viewed the Stormwater retrofit
of the Anacostia watershed as an environmental goal, the
Port Towns saw it as an opportunity to serve an economic
1 Bioretention BMPs are Stormwater retention facilities designed to mimic forested
systems that naturally control hydrology through infiltration and evapotranspira-
tion (Prince George's County Low Impact Development Manual, 1997).
90
-------
MONTGOMERY
COUNTY
'Xf ••-•t\-->t;•_!, j"'' a''"'/ e **! »> ;k i -'A' r«? .••'""l-—<' '" T- \*.
V- v.v^spLA---*v fei T'./ %**« ft
.^* ', / C H A R L i S
MARYLAND
Figure 1. Prince Georges County, MD (white area on map).
91
-------
ANACOST A RIVER
PORT TOWNS AREA
Figure 2. Aerial photograph of the Anacostia Port Towns area (outlined in white).
goal as well. Accordingly, when the Port Towns developed
a revitalization vision for their community, they included
the environmental restoration of the river as an important
objective.
Reconstruction of river wetlands and pond retrofit,
stream restoration, and bioretention streetscaping projects
are now viewed not only as a way to clean the river but
also as an opportunity to renew the Port Towns image.
With this in mind, county and state agencies began to re-
direct stormwater projects to the Port Towns (and other
revitalization communities). The County and the Port
Towns hope to use an environmental restoration theme
as a catalyst for future growth and investment. Stormwater
projects are at the forefront of this new urban revitaliza-
tion paradigm. Stormwater revitalization project goals are:
• Mitigate adverse environmental impacts
• Beautification
• Improve community image
• Enhance community pride
• Attract private investment
The realization that communities can help revive their
economies by cleaning up their environment is not new.
However, the Port Towns are probably among the first com-
munities to envision that stormwater retrofit projects play
a leading role in urban revitalization. The County's
Stormwater Management Tax District and state and fed-
eral funding sources are paying for construction of inno-
vative bioretention streetscaping, shallow marsh wetlands,
and river restoration projects.
Stormwater retrofit projects can be a new tool in the
repertoire of revitalization projects that help to renew older
urban areas (Table 1). The basic premise is that the Port
Towns stormwater retrofit and other public projects2 will
help create a more attractive place to live and work. Even-
tually, it is hoped that the private sector will fulfill long-
term investment needs in the Port Towns.
Proposed Projects
Key Port Towns projects funded fully or partially by
stormwater funds are as follows:
Port Towns Waterfront Restoration- This project involves
both the rehabilitation of a river marina as a center to pre-
serve the County's rich Colonial history and reconstruc-
tion of about 80 acres of wetlands as a waterfront park
2 Other proposed projects include new town centers, a new railroad bridge,
Brownfield projects, an ecoindustrial park, and road improvements.
92
-------
Table 1. Revitalization Projects
Infrastructure
New town center
New CSX railroad bridge
Marina redevelopment
Road improvements
$3M
$21M
$6M
$6M
Environmental
"Brownfields" cleanup
Eco-industrial park
$1.5M
$1M
Stormwater projects
Bioretention streetscaping, pond retrofits, stream $6M
restoration, river wetlands and drainage rehabilitation
along the Anacostie River (Figure 3). The plan proposes
to increase natural wildlife habitat areas, enhance water
quality, and increase economic tourism opportunities. The
planned Historic Waterfront Park will be the focal point of
community activities, providing opportunities for residents
and visitors to gather and participate in a variety of out-
door recreational cultural, and environmental activities
along the Anacostia River. Estimated costs include $6 mil-
lion for the marina and $5 million for wetlands construc-
tion.
Ponds and Stream - Existing dry ponds are being con-
verted to shallow marsh systems (Figure 4). These ponds
are designed not only to improve water quality; they also
include walkways, benches, and carefully designed land-
scaping to enhance the community's environmental, aes-
thetic, and recreational experiences. Eroded stream corri-
dors are being restored to create greenways. Estimated
costs are $1 to $2 million.
Urban Streetscape - In the early stages, this effort has
involved the construction of several pilot bioretention
streetscape projects (Figure 5) in each town. In the future,
all major road corridors will be gradually reconstructed with
landscaping that improves appearance and treats urban
stormwater runoff. Streetscape improvements are intended
to convey a sense of physical connection between the three
towns and mitigate the adverse effects of pollutants found
in urban stormwater runoff. Bioretention areas and other
water quality BMPs will be incorporated into the streetscape
projects. Estimated costs are $4 million.
Econursery- A self-sustaining nursery facility will be cre-
ated for the community to grow native trees and shrubs
and to produce seeds to maintain bioretention and rain
garden improvements. The facility will also serve as a lo-
cal science education center, a community garden, and a
composting yard. Estimated cost is $300,000.
Ecoindustrial Park3 - Environmental and infrastructure
enhancements are planned for the Port Towns and sur-
rounding industrial areas. The project will include water
quality enhancements, waste minimization, and
streetscape improvements. Estimated costs are $1 million.
(Abbasi, 1997)
Project Funding
Funds for the Port Towns stormwater projects are pro-
vided through federal and state grants, a Stormwater Man-
agement Enterprise Fund, and the sale of stormwater rev-
enue bonds. The debt service on bonds is paid from a tax
levied on all assessable property in the County's
Stormwater Management District. The current tax is 13.5
cents for every $100 of assessed property value. Under
this special tax district, a 5-year, $12 million Environmen-
tal Revitalization Program has been adopted to fund revi-
talization projects. This amounts to about 20% to 30% of
the 1999 annual stormwater capital improvement budget.
Due to the comprehensive nature of the revitalization
initiative and strong community involvement, the Port Towns
have drawn considerable interest from federal and state
agencies (Table 2). Grants are expected to fund up to 75%
of the project capital costs. For example, 50% of the river
wetland project funding is expected to come from the Mary-
land Department of the Environment and the US Army
Corps of Engineers. Up to 75% of the cost of the
streetscape improvements is expected to be paid by the
Maryland State Highway Administration. Funding for the
eco-industrial park is expected to come from a variety of
federal, state,and private sources.
Municipalities have agreed to provide for the operation
and maintenance (O&M) of pilot projects. In the future, O&M
costs may be shared by the municipalities and a proposed
Commercial District Management Authority funded by busi-
nesses and property owners.
Site Feasibility
An inventory of several stormwater projects, divided into
short- and long-term objectives, has been developed.
Projects siting is based on several goals, including urban
design enhancements, water quality improvements, prop-
erty acquisition, permitting, and cost. Short-term pilot
projects are selected on the basis of achieving quick re-
sults. Quick results lead to political and public support for
revitalization and also develop institutional capacity to un-
dertake more complex and ambitious projects (Johnson,
1997). Collectively, these environmental projects are de-
signed to improve existing infrastructure and community
appearance while helping to attract private sector invest-
ment.
Project selection begins with the guidelines of the Port
Towns Vision and Action Plan. Stormwater projects are
sited in commercial districts, transportation corridors, and
3 An ecoindustrial park is generally characterized by closely cooperating manufac-
turing and service businesses that improve their environmental and economic
performance. Industries coordinate their activities to enhance the efficient use of
raw materials, minimize waste and associated disposal costs, and conserve en-
ergy and water resources. This resource efficiency results in benefits to the
industires, while the surrounding community gains environmental performance
and job creation benefits.
93
-------
Figure 3. Drawing of the proposed historic Bladensburg Waterfront Park.
'
Figure 4. Dry ponds converted to shallow marsh systems improve water quality and enhance the aesthetic beauty of the Port Towns community.
94
-------
Figure 5. Bioretention streetscaping (planted areas to the right of the street in the photo) help to treat stormwater runoff and improve the
appearance of the community.
Table 2. Project Sponsors
Prince Georges County Government
Maryland National Capital Parks and Planning Commission
Port Town Community Development Corporation
Town governments
Maryland Department of the Environment
Maryland Department of Transportation
US Army Corps of Engineers
Commercial Management District
Community
natural resource areas that are important renewal areas.
Public exposure and acceptance of the initial pilot projects
are important to enhancing visual and spatial impacts. The
availability of land and easements is normally confirmed
before starting the design phase. To lower project costs
and to help knit a public-private partnership, publicly owned
or "gratis" private easements are sought first. Permits for
bioretention streetscape projects are generally easily ac-
quired. Permits for reconstruction of wetlands in existing
floodways are more complex. Extensive hydraulic model-
ing of the floodway channels is submitted to the federal
and state governments for approval. Approval by the fund-
ing agencies is critical before the final site selection is com-
pleted.
Project Team
To guide the planning and implementation of these
projects, project teams consisting of citizens and staff from
various county and municipal agencies are formed. Mem-
bers are also recruited from universities, nonprofit organi-
zations, and consultants. An informal inter-agency project
support group (Figure 6) leads the overall project. This
organizational arrangement promotes greater integration
of agency functions, stakeholder participation, and com-
munity based initiatives.
Environmental, engineering, urban planning, and project
management staff are all needed for the project. Support
group members provide different functions. For example,
the Department of Environmental Resources provides tech-
nical support in the areas of engineering, project manage-
ment, and funding coordination; the Maryand National Capi-
tal Park and Planning Commission provides community
planning and liaison; and town officials, a community de-
velopment corporation, and various citizen groups repre-
sent public and private interests, respectively.
Costs
To date, three small pilot bioretention streetscape projects
have been completed in the Port Towns. Drainage areas
for all of the bioretention projects were less than one acre
in size and 100% impervious. The average cost of these
bioretention retrofit projects (Table 3) amounts to $44,000
per acre of impervious drainage area.
Two shallow marsh extended-detention ponds have also
been constructed in the Port Towns at an average cost of
95
-------
Eco-Tourism River
Wetlands Committee
Historic Bladensburg
Waterfront Restoration
Committee
I
Project
Support
Group
Eco-lndustrial Park
Committee
Urban Streetscape
Committee
NOTE:
All committee members, project managers, and team
members are a partnership, comprising members
from the County, state, federal, consultants, business,
and industrial community.
Figure 6. Port Towns revitalization organizational structure.
Table 3. Stormwater Construction Cost
Project Name
Drainage
Area (acre)
% Imp.
Total Cost
Cost/acre
Imp. Area
Bioretention Pilot Projects - Commercial Land use
Chesley .94
Colmar Manor .65
Bladensburg .66
Shallow Marsh Pond - Residential Land use
Bladensburg 230.00
Cottage City 38.50
100%
100%
100%
55%
69%
$27 K
$44K
$23 K
$395 K
$91 K
$29 K
$67 K
$35 K
$3.1 K
$3.4K
$3,250 per acre of impervious area. Costs for bioretention
retrofit projects in existing urban areas are relatively high
due to complexities related to limited space, intense traffic
controls, and presence of existing utilities. Although aver-
age costs for similar BMPs in the County over the last 6 to
8 years exceed those of projects in the Port Towns, such
costs should be comparable in the longer term.
The cost of river wetlands reconstruction is expected to
be much higherthan the norm forthe County, due to exist-
ing flood levees. An assessment of all the Port Town retro-
fit opportunities and their associated costs is being pre-
pared.
Conclusion
The Port Towns is one of the first communities to envi-
sion that stormwater retrofit projects can play a leading
role in urban revitalization. Stormwater retrofit projects
comprise a new approach to publicly funded infrastructure
rehabilitation projects that help renew old urban areas. Due
to the comprehensive nature and strong community in-
volvement of the Port Towns revitalization initiative, the
96
-------
projects have drawn considerable funding support from
federal and state agencies.
Pilot stormwater projects were built within a 12-month
period to demonstrate tangible results and achieve politi-
cal and public support for revitalization projects. These
projects demonstrate that stormwater projects constitute
an effective tool to retrofit and improve the appearance
and image of existing urban communities. Planning and
design of more ambitious and complex projects are already
underway, including reconstructing wetlands in the
Anacostia floodway, streetscaping major transportation
corridors, and cleaning up the industrial park.
References
Abbasi, S. A., "Environmental Projects Lure Business to the
Port Towns, Maryland,"American Institute of Architects
Miami Conference on Economic and Environmental
Balance -the 21 st Century Outlook, November 1997.
Giannini-Spohn, S., "Eco-industrial Parks: One Strategy for
Sustainable Growth," Development, pp 3-5, January
1997.
Johnson, J., "Downtown Revitalization in Five Communi-
ties," Preservation, p M109, May/June 1997.
Legg Mason Realty Group, Inc., Feindesign Associates,
Inc. and Neighborhood Revitalization Division, Prince
Georges County Planning Department, "MarketAnaly-
sis and Revitalization Action Program and Strategy,
Bladensburg, Maryland," pp i-vi, 1995.
Maryland Department of Business & Economic Develop-
ment, "Prince Georges County - Brief Economic
Facts," 1995-96.
Maryland-National Capital Park and Planning Commission,
"Port Towns Revitalization Action Plan - Bladensburg,
Colmar Manor, and Cottage City," Spring 1996.
Prince Georges County Department of Environmental Re-
sources, Low Impact Development Design Manual,
November 1997.
Prince Georges County's Department of Environmental
Resources and Maryland-National Park and Planning
Commission, "Environmental Restoration -The Cor-
nerstone of Economic Revitalization," Limited circula-
tion, September 1993.
Pierre, R. E., "Maryland Port Towns' Past Serving as An-
chor for Communities' Future," Washington Post, July
6, 1997.
Tatar, D. D., Boroughs, A. K., and Cousins, M. J., "Prince
Georges County Maryland Community Investment Op-
portunities," Federal Bank of Richmond, Virginia, Feb-
ruary 1995.
97
-------
Tollgate Drain -An Innovative Approach to
Stormwater Management
John LeFevre and Patrick Lindeman
Fishbeck, Thompson, Carr & Huber
Ada, Michigan
The Tollgate Drainage District is a dedicated drainage
district that was established, in the late 1800s to be uti-
lized by Lansing Township, the City of Lansing, and Ingham
County. The Tollgate Drain has used the City of Lansing's
combined sewer system as its outlet since the early days
of the City's sewer system. As part of a recent 30-year
plan developed by the City of Lansing to control combined
sewer overflows to the Red Cedar and Grand Rivers, the
Tollgate Drainage District, underthe direction of the Ingham
County Drain Commissioner's Office, was mandated to
implement a combined sewer separation project. Now the
City of Lansing sends sanitary water to their wastewater
treatment plant, while the stormwater is diverted to a wet-
land park detention basin and a series of detention ponds
incorporated into the reconstruction of the local Groesbeck
Municipal Golf Course.
Objectives
The Ingham County Drain Commission recognized the
importance of redeveloping property forthe dual purposes
of the neighboring Fairview Park and for stormwater man-
agement. The primary goals of the Tollgate Drain project
were to eliminate combined sewer overflows through sewer
separation and develop a wetland ecosystem that improved
storm water quality, while also meeting the aesthetic needs
of Fairview Park and the Groesbeck Municipal Golf Course.
The Project
In accomplishing these objectives, the Tollgate Drain
project created a wetland to act as a natural filtration sys-
tem for stormwater runoff. The wetland helps maintain
water quality by removing nutrients and sediments In the
water. It involves the development of a stormwater sepa-
ration, retention, and recharge system. The end result is a
state-of-the-art urban wetland management system that
uses innovative and cost-effective methods of water man-
agement.
The 210-acre Groesbeck neighborhood had a one-pipe
sewer system which was built in the 1950s. This system
was recently incorporated into a new two-pipe combined
sewer system. One pipeline transports household waste
to the C4 of Lansing's wastewater treatment plant while
the other pipe, containing stormwater, is diverted. A tradi-
tional method of stormwater disposal is to drain it into the
nearest river. Using an innovative method, Tollgate Drain
diverts stormwater to the lowland area of Fairview Park,
where it is naturally cleansed of non-point source pollut-
ants and then recharged into the air and the ground. The
water is also used to irrigate the Groesbeck Municipal Golf
Course.
Most stormwater systems incorporate little if any non-
point source pollution abatement. Stormwater picks up road
oil, organic debris, fertilizers, salt and other forms of pollu-
tion as it makes its way through the stormwater system to
the rivers and their tributaries. For this reason, the Toll-
gate Drain project is unique. Unlike other stormwater sys-
tems, it does not outlet to a river and it has nonpoint source
pollution abatement properties.
Seven-Steps of the Tollgate Drain
Overall, the project consisted of seven key elements:
Step 1: Develop a catch basin maintenance plan.
Step 2: Create a filter chamber to act as a secondary
cleaning chamber.
Step 3: The wetland design.
Once the stormwater reaches the wetland, it runs
through mechanical oil skimmers and sediment
traps which remove petroleum products and ex-
cess sand and mud. Peat-sand and limestone are
used as filters in the system. Their high phospho-
rous (P), biological oxygen demand (SOD), pH
(acidity), and pathogen removal capabilities,
coupled with simple design, low maintenance, and
affordability make them an attractive method. The
wetland provides a variety of functions such as
flood control/water storage and filtration of pollut-
ants, and it creates eleven acres of wildlife habi-
tat. From the Fairview Park wetland the water trav-
els through a pipe underWood Streetto Groesbeck
Municipal Golf Course where it flows into additional
wetland detention ponds. This evaporates some
98
-------
of the water and allows the suspended sediments
to further settle out. At this point, the golf course
has the option to use this water for irrigation.
Step 4: The holding ponds on the golf course.
Step 5: The ultimate discharge to the City sewer at a
restricted rate.
Step 6: A proactive public outreach program within the
drainage district to inform and educate the dis-
trict on their role in this project
Step 7: A public outreach program with a broader per-
spective for the community at large.
Educational Aspects
The Drain Commissioner's Office ran an extensive pub-
lic outreach and education program for this project. All resi-
dents received a survey and a door-to-door visit to dis-
cuss sump pump connections and elimination of illegal
storm water cross-connections to the sanitary sewer sys-
tem.
An on-site office staffed with Ingham County Drain Com-
missioner representatives was available throughout the
construction phase of the project with a hot line so that
residents' concerns and questions were dealt with imme-
diately. A door-to-door follow-up survey was conducted to
obtain feedback and continue the urban storm water edu-
cation.
The overall maintenance of the project depends on how
the residents of the district take care of it. Residents are
encouraged to participate in the success of the project by
tailoring their daily activities to decrease the amount of
pollution, and therefore decrease the maintenance costs.
Dumping oil, pet waste, cigarette butts, or other garbage,
and blowing grass clippings and other yard waste into the
streets increases the number of times the catch basins
will have to be cleaned out. The use of fertilizers, pesti-
cides, and herbicides on lawns brings pollutants that can
also increase the number of times filters in the system will
have to be cleaned out and replaced.
Project Challenges
One of the main obstacles to overcome in the project
was the "ownership" of Fairview Park. Technically, the park
was owned by the State of Michigan, located in Lansing
Township, and maintained by the City of Lansing as a Lan-
sing Park. A lengthy battle overthe use of the land caused
uncertainty among the residents and between the differ-
ent governmental agencies. Before the drainage district
could proceed with design plans for the project these par-
ties had to come to an agreement. This was the most diffi-
cult phase of the entire project and today, all parties are in
agreement and cooperating fully with the Drain
Commissioner's office to make the project a success.
A design challenge was to determine a cost-effective
outlet for the storm sewer discharges so the sewer sepa-
ration could be completed. The district is surrounded by a
developed City on three sides and Groesbeck Municipal
Golf Course on the west. A conventional piped storm sewer
outlet to the Grand River would have had to extend over
one mile through a densely-developed residential area. This
option cost was in excess of $15 million. Three other routes
were examined ranging in costs from $15 million to $20
million. This project, chosen instead of those options, cost
$6.2 million.
Innovations/Benefits
• The key to the savings is putting nature to work. The
storm water is pumped to what was once a little-used
11-acre nearby park.
• An oil and grit chamber was used to trap any oil washed
into the storm collection system.
• Contaminants settle out of the water into ponds. As
the water moves through connecting channels, lime-
stone rocks buffer the acid it contains. Fast-flowing
streams increase oxygen and encourage the growth
of pollution-eating microbes. A peat bog filters out fer-
tilizers and pesticides.
• The water enters a wetland where it evaporates or is
recycled for more treatment. Additional water goes into
water hazards at the Groesbeck Municipal Golf Course
and is used for the golf course irrigation system.
• Numerous trees, plants, and grasses were planted in
the system to trap contaminants. The species were
selected to maximize evapotranspiration.
• The final selected system design was $6.2 million,
about one-half the cost of the other options. Tollgate
Drain not only saved millions of dollars, but it is a pro-
totype for environmentally sound water management
practices.
• The system not only keeps basements from flooding,
but cleans the water so it can be used for irrigation at
the nearby Groesbeck Municipal Golf Course.
• Storm water is managed on-site, ratherthan being ex-
ported.
Conclusions
In the future, direct river discharges will more than likely
have to be rebuilt to accommodate non-point source pol-
lution abatement, similar to this project, before discharg-
ing into the river. In this sense, the Tollgate Drain is ahead
of its time. The Tollgate Drain involves the creation of a
new wetland ecosystem designed to naturally clean and
recharge the neighborhood's storm water.
Aside from the physical challenges, the success of this
project is dependent upon the cooperation of the City of
Lansing, Lansing Township, and the State of Michigan. But,
most importantly, the future success of this project lies in
the hands of the residents within the Tollgate Drainage
District.
99
-------
A Stormwater Banking Alternative for Highway Projects
Robert B. McCleary, RE.
Delaware Department of Transportation
Dover, Delaware
Purpose
The purpose of this paper is to make others aware of
the approach the Delaware Department of Transportation
(DelDOT) is taking toward stormwater quality management.
This is a discussion paper that looks at the costs and sav-
ings of the stormwater banking approach adopted by
DelDOT and provides useful information regarding pro-
gram implementation for anyone considering initiating a
similar approach.
Introduction
In 1996, in response to impacts from water quality con-
trol laws at both the state and federal levels, a memoran-
dum of agreement (MOA) was drawn between DelDOT
and the state's stormwater regulatory agency, the Dela-
ware Department of Natural Resources and Environmen-
tal Control (DNREC)1. This MOA establishes criteria
whereby DelDOT can consider a regional alternative to
the on-site approaches set forth by statute. It allows
DelDOT to mitigate the water quality impacts associated
with highway projects elsewhere in a watershed if on-site
options are not practicable. The agreement is for water
quality control only. Increases in peak flow rates associ-
ated with highway development must still be controlled on-
site. While the approach deviates from the on-site approach
stipulated in state regulations, both parties to the agree-
ment believe it is consistent with state and federal water
quality goals.
The MOA is often referred to as the stormwater "bank-
ing" agreement. The "banking" term is used because imple-
mentation of water quality control best management prac-
tices (BMP) is tracked using a system of credits and deb-
its. Credits and debits are accrued by watershed. Water-
shed delineations can be nebulous, but in Delaware
DNREC officially delineated and defined 41 watersheds2.
The MOA is consistent with this delineation. Each water-
shed represents a separate bank account. Water quality
credits and debits from multiple highway projects may be
applied to each watershed. That is, some projects may be
built without providing water quality control by accruing a
debit to the watershed while other projects that do build
BMPs accrue credits.
This balancing of water quality credits and debits allow
a more flexible implementation of water quality BMPs for
each watershed. Rather than siting BMPs within the limits
of specific projects, DelDOT may look anywhere in a re-
gional watershed. It is believed this approach will direct
the limited funding available for BMP implementation to-
ward locations most conducive to water quality treatment.
In this way, Delaware can realize a more cost-effective and
environmentally beneficial infrastructure of water quality
treatment measures.
Presently, DelDOT is the only agency within Delaware
to use such a system of stormwater banking. So balanc-
ing or trading of water quality credits is only between
DelDOT projects. It is hoped that other agencies will de-
velop similar agreements so that trading of water quality
credits may be conducted between multiple users. This
could include both private and public entities.
Local Factors That Enable MOA
Development
Forthose contemplating development of a similar agree-
ment, it is useful to understand the regulatory climate in
Delaware that allowed development of the MOA.
Program Delegation
EPA delegated the National Pollutant Discharge Elimi-
nation System (NPDES) permit program to the state
DNREC. It also delegated the Coastal Zone Management
Program (CZM) to the state. Being a delegated agency of
EPA for both the NPDES and CZM programs gives the
State of Delaware some latitude in implementing its sur-
face water quality control programs.
Also, DelDOT's situation is unique among state DOT'S
in that DelDOT is delegated by DNREC to administer its
own stormwater management program. Embodied in the
state stormwater management law is a provision that al-
lows delegation of program functions to other state agen-
cies that can demonstrate the technical and financial abili-
ties to implement this program. DelDOT sought and re-
ceived program delegation in 1991. This gave DelDOT the
ability to design, review, and permit its own projects for
stormwater management.
100
-------
Having control of its own stormwater permitting program
affords DelDOT some freedom in program implementa-
tion. In fact, DNREC encourages each delegated agency
to implement policies and procedures that address local
needs and initiatives. This MOA stems from that philoso-
phy and was developed to offer an alternate means of
achieving the state water quality control goals in a way
that considers the limitations of highway projects.
Need For MOA
To understand the need for this MOA one needs to un-
derstand the traditional site-specific approach toward
stormwater management stipulated in the state regulations
and its impact on DelDOT.
Under the state law, DelDOT is required to implement
stormwater management controls on every project involv-
ing land disturbances of 5,000 square feet or more.
Stormwater ponds and other control measures are required
for drainage areas measuring only fractions of an acre.
Highway projects, being long and linear, cut across mul-
tiple watersheds, sub-watersheds, catchments, and sub-
catchments - requiring multiple stormwater ponds on ev-
ery project. DelDOT has found this site-specific require-
ment leads to a proliferation of small expensive stormwater
management ponds.
One project in particular illustrates this fact. State Route
1, North of Smyrna, Delaware, consists of a 6 mile stretch
of 4-lane dual divided highway on a new alignment3. The
preliminary project plan submittal proposed 43 ponds to
manage the runoff from every drainage area affected by
the project. Each drainage area was on the order of 1-2
acres. Laterthis number was reduced to 13 ponds by com-
bining the runoff from multiple drainage areas. But this is
still a large number, especially since it only addresses the
needs for one project. When the whole transportation sys-
tem was considered, it became evident to DelDOT that
the site-specific approach to stormwater quality control
would lead to an unsupportable expansion of public infra-
structure.
Somewhat worse situations arose on widening projects
where multiple stormwater ponds had to be fit into previ-
ously developed landscapes. On projects such as the wid-
ening of Naamans Road, DelDOT actually purchased
homes to make room for stormwater ponds4. The cost in-
cluded the fair market value of the homes, relocation of
the residents, demolition of the existing structures, and
construction of the ponds. In one case, the cost of a pond
to treat less than a 2-acre drainage area exceeded
$300,000.
Because of the SR-1 and Naamans Road experiences,
in the fall of 1994 DelDOT took an inventory of all
stormwater management facilities planned for construc-
tion over the next 6 years (FY95-2000). An examination of
project plans indicated that at least 114 new stormwater
management ponds were in some stage of design or con-
struction. Of those facilities, a representative sample of 37
ponds examined to determine some average conditions
that could be applied for cost estimating purposes:
• The average land area required per
pond 0.75 acres
The average volume required per
pond
3,000 cubic yards
• The average depth of the ponds 4 feet
• The average cost of real estate $25,000/acre
From these parameters, stormwater pond construction
was estimated by applying the mean cost of DelDOT con-
struction pay items. The cost per pond came to $85,225.
This included real estate acquisition, design, and construc-
tion costs. The projected construction cost for all 114 ponds
was estimated at $9,715,650.
The inventory and cost estimate also included other types
of stormwater management practices planned for imple-
mentation on DelDOT projects forthis time period. Included
were infiltration trenches, biofiltration swales, and sand fil-
ters. Their costs were estimated to be $497,460. The total
construction cost of all stormwater management practices
for the 6-year period (FY1995-FY2000) was estimated at
$10,213,109.
Compared to the $463,349,000 Capital Transportation
Improvement Program Budget forthis period, the cost of
stormwater quality control implementation amounted to only
2.2% of the budget. But while 2.2% may seem small, it still
amounts to a substantial investment in public infrastruc-
ture. And certain recent projects raised doubts as to
whether the limited funds available were being used in the
most effective manner.
For example, the Lancaster Pike widening project in-
cluded 816 linear feet of sand filters to treat the runoff from
about two acres of roadway pavement5. Of the two acres,
only about one acre was new pavement. The cost for the
filters was $326,400. It seemed exorbitant to managers at
DelDOT, but at the time of design no alternative existed.
The project's steeply sloped and high-cost real estate sur-
roundings were not conducive to less-expensive options
and the on-site requirements in the regulations obligated
DelDOT to provide water quality control forthis drainage
area. This project caused DelDOT managers and mem-
bers of the public to question whetherthe high cost of water
quality control was worth the seemingly miniscule envi-
ronmental benefit which was difficult to measure. Because
of this and several similar situations on other projects,
DelDOT and DNREC collaborated to devise a better way.
The Stormwater Banking Concept
The quality of water in a stream depends on many fac-
tors - just one of which is stormwater runoff. When consid-
ering non-point sources of pollution, it is commonly thought
that all land surfaces contribute some degree of pollution.
And the relative amounts of pollution will vary naturally
101
-------
within watersheds from one sub-area to another. In this
context, the quality of stormwater runoff from any single
sub-area is not a significant determinant of stream water
quality. Pollution levels then can be increased or decreased
from one area to another with no adverse affect to stream
water quality. As long as the cumulative impact to the
stream stays the same, or is reduced, stream water qual-
ity will be preserved or even enhanced. The MOAis based
on this concept of balancing the levels of pollution from
multiple sub-watersheds.
Specifically, the MOA allows DelDOT to provide
stormwater quality.controls at an a tentative location in the
event the implementation of similar controls at a specific
project site is not practicable. The measures installed at
alternate locations must provide stormwater quality treat-
ment for an equivalent amount of highway runoff as that
going untreated at the project site. In choosing alternate
locations, preference is given to sites within the same wa-
tershed as the project. Projects requiring water quantity
control must still address it on-site. Ways of providing wa-
ter quantity control without building a pond are discussed
in appendix 'A' of the MOA.
This concept of balancing stormwatertreatment from one
area to another, literally treating some areas while letting
other areas go untreated, is often referred to as stormwater
"banking". And, as is the case with Delaware's MOA, the
concept normally uses an accounting system of credits
and debits to track the overall level of water quality control
implementation in each watershed - hence the term "bank-
ing".
Stormwater "banking" offers an alternative to the site-
specific approach by helping to facilitate a regional planned
approach to stormwater quality management. Regional
planning for water quality control involves prioritizing the
various water quality treatment needs in each watershed
and targeting implementation of control measures in the
locations they will do the most good. In theory, this should
minimize the overall number of stormwater management
measures and maximize their cumulative effectiveness. In
this way, taxpayers should receive the greatest return on
their investment in public infrastructure designed to treat
highway runoff- both in terms of initial construction costs
and long term maintenance.
This stormwater "banking" concept is not particularly new
or original. Similar approaches are frequently employed
on projects all over the country. But it is seldom well docu-
mented and is often viewed as bending the rules. This MOA
formalizes the criteria by which DelDOT will determine
compliance when it is not practicable to manage stormwater
quality "on-site". And it validates the approach as an ac-
ceptable alternative.
MOA Triggered by Variance
It should be noted that DNREC was reluctant to depart
completely from the requirement to manage stormwater
quality on-site. In Section 2.2 of the MOA, it was DNREC's
desire to limit the use of stormwater "banking" by allowing
its use only after first exhausting all on-site alternatives. In
its final form, the MOA is reserved for projects located in
areas that pose difficult site constraints or which other-
wise offer little opportunity to implement permanent water
quality control measures on-site. The terms of the MOA
may be invoked only through the granting of variance in
accordance with Section 3.3 of the Delaware Sediment
and Stormwater Regulations. Variances may be granted
only after demonstrating that exceptional circumstances
exist at a project site which would result in unnecessary
hardship and not fulfill the intent of the regulations.
Site Selection
Selection of appropriate sites for stormwater banking is
accomplished through guidance provided by DNREC wa-
tershed managers. The MOA encourages a collaborative
effort in selecting a site.
Section 1.2 of the MOA defines 41 regional watersheds
which is consistent with the delineation established by state
and federal water resource managers (see the-draft Dela-
ware Wetland Banking Agreement)2. It is preferable under
the MOA to mitigate the waterquality impacts from a project
within the same watershed. However, if the committee of
resource managers established under Section 3.4 deter-
mines it is appropriate to mitigate outside the watershed,
then this option may also be considered.
Types of Water Quality Control Alternatives
Allowed
Section 3.3 of the MOA lists the alternative types of wa-
ter quality treatment methods available to DelDOT. The
goal of this section is to encourage the use of alternative
water quality control methods that best meet the water
quality control needs for each watershed. The available
options focus on protecting key natural areas such as
streams and wetlands. Other banking agreements re-
viewed emphasized providing only stormwater manage-
ment ponds or infiltration measures6. DelDOT and DNREC
felt it was important to encourage wetland creation, resto-
ration, and enhancement as a water quality improvement
measure. The list includes: source controls, removal of
existing pavement, reforestation of cut woodlands, replace-
ment of riparian vegetation, retrofitting existing stormwater
ponds, removal of illicit connections. The list itself is not
meant to be all-inclusive. All reasonable water quality im-
provement techniques will be considered under this MOA
provided they help meet the water quality goals for the
watershed being considered.
One alternative that was considered, but later ruled out
was conservation/preservation easements. This option
would have allowed DelDOT to purchase the development
rights to lands deemed worthy of protection such as up-
land forests which are extremely important from a water
quality perspective. Unfortunately, DNREC felt this option
did not mitigate increased pollution. As they saw it, preser-
vation easements only maintain the status quo. DNREC
102
-------
argued that if DelDOT were allowed to increase pollution
at one location in a watershed, then an equivalent amount
needed to be reduced elsewhere. Preservation easements
did not meet that test.
Accounting System
Section 3.5 establishes the accounting procedures for
water quality debits and credits. They are accrued by wa-
tershed and each watershed can be thought of as a sepa-
rate account. Currently, DelDOT is tracking the number of
credits and debits using Microsoft Excel software.
Counting credits and debits seems like it ought to be a
fairly easy thing, but it becomes very complicated if cer-
tain factors are considered such as whetherthe land treated
is impervious, farmland, subdivision, or forest. Questions
arise such as, is it fair to give equal credit to a measure
that treats runoff from fallow fields as one treating road-
way runoff. Other complicating factors include whether the
treatment measure is in the same watershed as the project.
Should it be given equal credit? Maryland's agreement
attempts to consider these factors. DelDOT decided this
approach was just too cumbersome for our purposes. We
limited the credits to the actual acreage of impervious sur-
face treated. Even with this simplification, a supplemental
worksheet was prepared at the request of project design-
ers struggling with the accounting of water quality credits
and debits.
Modification and Termination of MOA
It was important that both parties have the ability to alter
the agreement. Since the regional concept had not been
tried in Delaware previously, neither party was quite sure
how well the concept could be implemented with this simple
agreement. It is expected that the MOA will need updating
from time to time as our understanding of the regional
stormwater management approach matures. Therefore,
Sections 3.4 and 3.7 of the MOA allow for modifications
upon written agreement of both parties.
Both parties acknowledge the agreement relies heavily
on a mutual understanding of each agency's needs and
limitations. It will succeed as long as conditions exist which
foster a cooperative spirit. This could change over time
because of political or personnel changes which might re-
sult in philosophical differences. Should the relationship
degrade, the MOA may be terminated upon written notifi-
cation by either party in accordance with Section 3.8.
Funding of Stormwater Banking Projects
The MOA itself does not stipulate the way DelDOT will
fund stormwater banking projects. Funding opportunities
will vary over time so there seems no reason to create
binding arrangements in the agreement. The agreement
does, however, establish a funding time frame under Sec-
tion 3.5. This Section obligates DelDOT to fund a banking
project to mitigate for previous debits within three years of
first using the MOA.
There is no expectation that federal or state grants will
help fund this program. Rather, funding will be part of
DelDOT's operating and capital improvement budgets.
However, DelDOT is investigating other funding options.
Current funding options being used or considered include
the following alternatives:
1. Banking projects funded as component of highway
contract
2. Percentage of contract cost held in escrow from mul-
tiple projects or programs
3. Public-private partnerships
4. Public-public partnerships
Under item (1), if banking can be accomplished within
the limits of an existing project, the costs can be made
part of that project. Normally, this is done when an oppor-
tunity exists to manage the runoff from more land than what
is required under the project, such as adjacent existing
highway. In this way, water quality credits are accrued for
the watershed in question. Future projects then may be
built without stormwater quality controls by taking debits
against the credits accrued by earlier jobs.
Under item (2), DelDOT will hold a certain percentage
of program funds aside for stormwater banking implemen-
tation. For instance, this is being done with DelDOT's Pave-
ment Management Program which funds pavement over-
lays, shoulder paving, and minor (V-31) lane widening
projects. In FY99, 1% of program funding ($320,000.) is
being set aside to address water quality concerns arising
from this program. The amount set aside is based on an
estimate of additional acreage of impervious surfaces cre-
ated under the program. Mitigation efforts will be focused
in one or two high-priority watersheds to balance the im-
pacts of many projects from multiple watersheds around
the state. Under this scenario, highway projects may be
started prior to actually having a banking project initiated.
So long as the funding is available, the impacts from ear-
lier projects can be mitigated within the 3-year time limit.
Item (3) has been discussed but no agreements have
been reached as of this writing. However, it is envisioned
that a private developer or group of developers could part-
ner with DelDOT to build one or more regional facilities
that manage the runoff from both private and public land.
There are multiple ways to fashion a partnership under
this scenario. The main bargaining chips include land, de-
sign services, construction, and future maintenance.
The public-public partnership under item (4) presents
itself in locations where multiple public agencies share real
estate, but maintain separate operating budgets. DelDOT
has identified several locations where other state agen-
cies, local governments, and school districts may partner
with DelDOT to share the costs of building and maintain-
ing a stormwater banking facility. No agreement has been
signed to date, but several are in draft stages.
Both options (3) and (4) rely on equitable distribution of
costs. DelDOT is settling on a formula of distributing costs
103
-------
based on percentage of land contributing runoff. The cost
to each partner is the total cost multiplied by each part-
ners respective acreage of land as a percentage of total
acreage contributing runoff to the facility. This formula is
used to divide construction costs. It can also be used to
determine each partner's annual share of maintenance
costs.
Property Acquisition Concerns
It remains to be seen what authority DelDOT will be able
to exercise in acquiring property for stormwater projects
when it has to mitigate for highway jobs located a consid-
erable distance away. There was concern that DelDOT
would never be able to settle on a site because property
owners would argue that we could always locate it some-
where else. In his legal review, the Deputy Attorney Gen-
eral (DAG) felt that DelDOT would enjoy all the same au-
thority we have now. That is, if we needed property for
stormwater management purposes we could obtain it ei-
ther voluntarily or through invoking the state's right of im-
minent domain. The DAG's opinion was that as long as we
can show that the sites we pick are the most practical and
feasible locations we would be justified in the taking. He
did not think we would need to prove the chosen sites are
the only feasible locations7.
Cost Comparison on Porter Road Project
Phase I of the Porter Road widening project serves as a
good example to illustrate the potential savings of the re-
gional approach allowed under the MOA. The project be-
gins at the intersection of Route 896 and extends 2.225
miles East to the intersection of route 72. It involves the
widening of the existing 18-ft roadway to a variable width
of 48 - 60 ft. The project has outfalls to three of the 41
watersheds defined in the MOA. Within those watersheds,
13 sub-areas were identified as requiring separate
stormwater management measures. The initial design was
submitted with 13 stormwater management ponds to con-
trol the increased peak rates of runoff and non-point source
pollution associated with the roadway project for each of
those 13 subareas. The proposed measures would treat
the runoff from only those areas within the immediate
project limits. Their design was typical of small-scale
stormwater ponds, lacking in aesthetic appeal and mar-
ginal in the overall water quality benefit to downstream
areas. The estimated cost of this site-specific approach
was in excess of $1 million.
Later it was determined that the MOA criteria would of-
fer a better alternative forthis project. The design proposed
building wetlands instead of stormwater management
ponds. The revised design included the following compo-
nents:
• Wetland creation in the headwaters to Belltown Run
to prevent downstream flooding and improve water
quality,
• Restoration of 1 acre of previously filled wetland at the
Porter Road Belltown Run crossing restoring flood plain
storage and stream habitat,
• Retrofitting of an existing county owned stormwater
pond to incorporate water quality control components
(i.e. with extended detention), and
• Enhancement of an existing degraded wetland at the
Porter Road and Route 72 intersection which involves
the eradication of phragmites and creates wetlands in
the upland areas adjacent to a narrow band of exist-
ing wetlands.
In addition to treating the runoff from the Porter Road
project, an additional 25 acres of existing roadway was
afforded water quality treatment that accrued as credits in
the Christina Riverwatershed. These credits are being used
to balance the water quality impacts from the Salem Church
Road project in the same watershed. The cost of this ap-
proach is estimated at $1.2 million. However, substantial
savings will be realized when future projects make use of
the credits afforded by the Porter Road Project.
This comparison illustrates the potential economic sav-
ings that can be generated through use of the MOA but
DelDOT believes the measures installed under the MOA
are also more effective from an environmental standpoint.
The larger scale of these facilities allows more innovation
in design resulting in many secondary benefits in terms of
wildlife habitat, aesthetics, and public acceptance.
Consistency with Federal Surface Water
Quality Control Programs
It was DelDOT and DNREC's intent to ensure the MOA
was consistent with other water quality control programs
at the state and federal levels. Section 2.1 of the MOA
makes very general statements regarding this consistency
merely to confirm that in fact these programs were consid-
ered. A more in-depth discussion of how the MOA meets
the water quality requirements is provided below.
TMDL Program
Section 303(d) of the Clean Water Act requires states to
develop a list of water bodies that need additional pollu-
tion reduction beyond that provided by the application of
existing conventional controls. The law requires states to
identify all waters needing water quality improvement.
Those portions of streams not meeting designated use
standards are termed "Water Quality Limited".
Water quality limited waters require the application of
Total Maximum Daily Loads (TMDLs) to determine the
allowable stress for each stream. A TMDL is the level of
pollution or pollutant load below which a water body will
meet water quality standards and thereby allow designated
use goals, such as drinking, water supply, swimming, fish-
ing, or shellfish harvesting to be achieved.
The TMDL approach to watershed management recog-
nizes that streams have a certain capacity to carry pollu-
tion without any discernible impact to the designated use
of a stream. It recognizes that restoration of stream water
quality may require a balancing of pollutant loading from
104
-------
multiple sources in a watershed. The MOA is consistent
with this concept and may to some extent help facilitate a
system of trading water quality credits between multiple
users within a watershed.
NPDES Stormwater Permit Program
Section 402(p) of the Clean Water Act establishes per-
mit requirements for certain municipal and industrial
stormwater discharges. New Castle County, Delaware was
identified under the Phase I NPDES Stormwater Permit
program as requiring a permit for the discharge of
stormwater from the municipal separate storm sewer sys-
tem (MS4). Also, all construction activity disturbing more
than 5 acres of land was identified as an industrial activity
requiring a stormwater discharge permit. This legislation
affected all storm drains owned and operated by DelDOT
in New Castle County and also affected DelDOT construc-
tion activity statewide.
The statute mandated that owner/operators of storm
sewer systems implement regional stormwater manage-
ment plans utilizing a watershed approach. Stormwater
banking ran be a component of such a plan.
The concept of banking stormwater quality improvement
credits is consistent with the federal statutory requirement
of implementing controls to reduce the discharge of pollut-
ants from municipal separate storm sewer systems to the
maximum extent practicable. The operative phrase in the
statute is, "reduce... to the maximum extent practicable".
Neither the law nor the regulations requires the discharge
of pollutants associated with stormwater runoff to be elimi-
nated or reduced at all cost. While the implementation of
stormwater quality controls on each and every transporta-
tion project may be a desirable goal, it is recognized such
a goal may not be realistic, cost effective, or practicable.
DelDOT believes it will be able to demonstrate compli-
ance with the legal intent of the statute because the bank-
ing approach is based on the water quality control needs
of the overall watershed. This is especially true if DelDOT
can show consistency with the TMDL for each stream sec-
tion. However, It may be possible for a citizen to lodge a
complaint under the statute if there is a measurable in-
crease in pollution at a specific site where water quality
controls were determined to be impracticable to implement
These types of complaints would likely come when a wa-
ter quality impact is readily noticeable by the general pub-
lic, such as where a storm drain discharges trash, debris,
sediment and the like from a roadway onto adjacent prop-
erty. This would most likely be the case on new alignments
if control measures were not implemented. Improvements
to existing alignments would not be as likely to generate
these types of complaints because roadway type pollution
would already be present. Adding another lane under a
widening project is not likely to change the character of
the pollutants to a great degree, although the total mass of
the various pollutants may increase slightly.
Forthese reasons the agreement emphasizes the imple-
mentation of water quality controls onsite for major road-
way improvement projects, such as new alignments. It
encourages utilization of the banking concept only on more
minor types of projects, such as intersection improvements
and lane widening projects where it is typically more diffi-
cult to incorporate stormwater management measures.
Also, a maximum debit limit of 5 acres is allowed to accu-
mulate statewide before it must be mitigated by implemen-
tation of a water quality control project. The 5-acre limit
was chosen to coincide with the 5-acre limit established in
the NPDES stormwater regulations for construction activ-
ity. The law requires implementation of water quality con-
trols when 5 or more acres of ground is being disturbed by
construction. Under the agreement, DelDOT may do sev-
eral small projects each disturbing a fraction of the 5 acres.
But once the aggregate of all watersheds exceeds the 5-
acre limit, DelDOT must undertake a project to mitigate for
those cumulative impacts.
Coastal Non-Point Pollution Control
Program
Section 6217 of the Federal Coastal Zone Act Reautho-
rization Amendments (CZARA) of 1990 mandated that each
state in the coastal zone initiate a coastal non-point pollu-
tion control program. The intent of the law was to encour-
age EPA, NOAA, and the states to place special and ex-
peditious attention on protecting the nations coastal water
from urban sources of nonpoint pollution.
EPA excluded from coverage under Section 6217 all
stormwater discharges covered by Phase I of the NPIDES
stormwater permit program. That is, any stormwater run-
off that ultimately is regulated under an NPIDES permit
will not be subject to the requirements of Section 6217 of
the CZARA once the permit is issued. For instance, dis-
charges of stormwater from construction activities disturb-
ing more than 5 acres of land and New Castle County's
municipal separate storm sewer system were excluded
from the Coastal Non-point Pollution Control Programs.
That still left several sources of pollution that needed to
be addressed under the Coastal Zone Act. Specific areas
affecting DelDOT included requirements to control runoff
from existing roadways and bridges and runoff from con-
struction sites that result in the disturbance of less than 5
acres of land.
The notion of building stormwater treatment measures
as the only item of work was not commensurate with the
mission of DelDOT which is to build transportation sys-
tems, not water quality treatment systems.
Fortunately, the Section 6217(g) Guidance encouraged
a whole watershed planning approach in implementing
stormwater management measures. Again, the MOA on
stormwater banking is consistent with this concept and may
to some extent help facilitate the process. From DelDOT's
perspective, the MOA helps DelDOT justify the expendi-
ture of transportation funds on water quality control mea-
sures for existing roadways and bridges. As discussed
above, the implementation of stormwater management
105
-------
control measures in accordance with the terms of the MOA
can be less expensive than a site-specific approach. There
is an economic incentive then for DelDOT to undertake
projects solely forthe purposes of treating stormwater runoff
because of the savings accrued to future roadway projects.
Section 404 Weffands Permitting Program
The MOA allows the implementation of many alternative
types of surface water quality control measures, one of
which is creating wetlands. Wetland creation in areas des-
ignated as uplands will provide stormwater quality treat-
ment in accordance with the stormwater regulations. With
the US Army Corps' concurrence, it may also qualify for
wetland mitigation credits required for highway projects.
In searching for a site to build a regional stormwater
management facility, if is often the case that a stream or
wetland is identified as the only feasible location. How-
ever, it is not usually possible to acquire permits to locate
stormwater management measures in existing wetlands,
nor does the MOA encourage this activity. Under certain
circumstances, however, the regulatory agency may be-
lieve work in a wetland is beneficial to the resource such
as by restoring a previously filled wetland. For instance,
DelDOT has restored previously filled wetlands for mitiga-
tion credits on several projects. If these restored wetlands
rely on surface runoff from roadways to provide the hy-
drology needed to support the wetland, then stormwater
quality credits may also accrue under the MOA for the
watershed in question.
Conclusion
The on-site approach to implementing water quality treat-
ment measures encourages a proliferation of small, ex-
pensive, and maintenance intensive practices on DelDOT
highway projects that may not offer the best solutions
needed forthe watersheds in question. Stormwater bank-
ing offers one possible alternative because it allows imple-
mentation of treatment measures anywhere in a water-
shed so they may be targeted toward the areas they are
needed most.
Delaware's MOA allows a broad array of treatment meth-
ods, such as wetland creation, reforestation, and elimina-
tion of existing pavement. This is intended to encourage
innovation in meeting the water quality requirements and
protect key natural areas. The MOA also offers a uniform
procedure for tracking water quality credits and debits ac-
crued in each watershed. DelDOT is finding the stormwater
banking approach to be more flexible and cost effective
than the traditional on-site approach.
References:
1. Memorandum of Agreement on Stormwater Quality
Management. March 1996. Delaware Department of
Transportation and the Delaware Department of Natu-
ral Resources and Environmental control.
2. Delaware Wetland Banking Agreement, November
1994. Delaware Department of Transportation, Dover,
DE.
3. SR-1, North of Smyrna to Townsend. 1994. Prelimi-
nary Plan Submittal. Delaware Department of Trans-
portation Contract 91-110-14.
4. Naamans Road, West of Marsh Road to Merribrook
Road. 1994. Delaware Department of Transportation
Contract 93-102-01.
5. Lancaster Pike (SR 48), West of Centerville Road to
SR 141. 1994. Delaware Department of Transporta-
tion Contract 92-118-01.
6. Memorandum of Agreement on Stormwater Quality
Management Banking. May 1992. Maryland State
Highway Administration and the Maryland Department
of the Environment.
7. Frederick Schranck, Esq., 1995. Personal Com
munication. Delaware Deputy Attorney General. Dela-
ware Department of Transportation, Dover, DE.
106
-------
Financing Retrofit Projects:
The Role of Stormwater Utilities
Greg Lindsey
Center for Urban Policy and the Environment
Indiana University, Indiana
Amy Doll
Apogee Research/Hagler Bailly, Inc.
Bethesda, Maryland
To achieve national water quality objectives, we must
retrofit existing stormwater infrastructure to manage pollu-
tion in runoff. Many public works and water resources pro-
fessionals have suggested that stormwater utilities are an
important, if not essential, funding source for retrofit
projects. We examine in this paper the role of stormwater
utilities in financing retrofit projects and programs. Based
on a broad assessment of the need for funding and a brief
overview of the evolution of stormwater utilities, we con-
clude that stormwater utilities are perhaps the best institu-
tional approach to financing retrofit programs, but that they
are not a panacea. The major issues in implementation of
effective retrofit programs will be economic and therefore
political. Stormwater managers can help constrain and fo-
cus political debate through careful analysis.
How Much Do Retrofits Cost? What is the
Need For Funding?
The answers to questions about the costs of retrofits
and programs to control the quality of stormwater runoff
depend on many different factors. These factors include
the characteristics of runoff quantity and quality, the size
of the watershed where projects are being planned, the
severity of water quality problems, the water quality objec-
tives, and the types of best management practices that
are being proposed. A short, safe answer that recognizes
variability among places is that programs will be expen-
sive, ranging from tens of thousands of dollars in relatively
small places to achieve modest objectives, to tens or hun-
dreds of million dollars in larger cities with moderate to
severe problems.
We provide here several brief examples of the potential
magnitude of costs of retrofit programs. Our examples are
by no means exhaustive and they are not necessarily rep-
resentative. We have chosen published estimates or used
cases with which we are familiar simply to demonstrate
that experts believe costs will be significant and contro-
versial. These examples should be sufficient to convince
skeptics or individuals who have not thought systemati-
cally about the economic aspects of retrofit programs, that
a critical task in implementation of a program is identifica-
tion of sources of revenue.
Our examples include estimates of the costs of programs
at the national level, for a watershed, for a large city, and
for a small town (Table 1). In a project for the American
Public Works Association, James M. Montgomery (1992)
estimated the capital and operation and maintenance
(O&M) costs for large and medium cities to comply with
EPA's stormwater rule. Capital estimates ranged from $147
million to more than $400 billion, depending on assump-
tions about the level of treatment for runoff. Estimates of
O&M costs ranged from $1.2 billion to more than half a
trillion dollars, again depending on assumptions about level
of treatment. Reasonable questions can be raised about
these estimates. Some experts suggest that they are too
high because advanced treatment never has been con-
templated for stormwater. Other critics contend that these
estimates were made primarily for political purposes and
to support opponents to the then-proposed federal
stormwater rule who argued that costs were prohibitive.
Regardless, they are suitable for our purposes. They dem-
onstrate clearly the need for financing and they show that
the costs of programs will be controversial.
More recently, EPA modeled the Phase I Storm Water
needs to inform Congress of the costs of programs to con-
trol pollutants in urban runoff. Approximately 266 Phase I
stormwater permits that regulate about 850 municipalities
will be issued. The Phase I needs estimates were prepared
to determine the stomwater management costs that might
be eligible understate revolving fund (SRF) loan programs.
The SRF-eligible costs include costs for developing and
implementing municipal management programs, including
capital costs for structural controls and BMPs. The total
modeled costs are $7.4 billion. These costs do not include
O&M costs, costs of land acquisition, permitting costs, costs
of developer-financed BMPs; or several other categories
of costs. These estimates, which were subject to peer re-
view prior to their release, also are significant. These esti-
107
-------
Table 1. Selected Costs for Stormwater Programs
Estimated Costs for BMPs in Regulated Municipalities (Montgomery 1992)
Capital
O&M
1. Source controls
2. Increased maintenance + 1
3. Construction of moderate controls + 2
4. Construction of detention basins + 3
5. Advanced treatment plants + 4
Modeled SRF-eligible costs
Core (source controls)
Segment (planned, new areas)
Segment (existing areas)
Total
Capital
Household
$147,100,000
$147,000,000
$83,139,500,000
$91,130,900,000
$406,734,900,000
Estimated Costs for Phase I Storm Water Programs (EPA 1997)
$7,400,000,000
Costs for Pollutant Reduction in the Menomonee River Watershed,
Milwaukee, Wisconsin (WDNR 1992)
$3,400,000
$11,700000
$94-$184,000:00
$110-$200,000,00
Estimated Rehabilitation Costs in Indianapolis, Indiana
$283,000,000
$54
$1,155,000,000
$32,607,800,000
$86,223,700,000
$90,097,500,000
$542,036,700,000
mates also have political dimensions; they were prepared
to inform Congressional debate over funding for water
quality programs.
In general, better cost estimates can be made for smaller
geographic areas because site specific factors can be taken
into consideration and fewer general assumptions need
be made. The Wisconsin Department of Natural Resources
(1992) has estimated the costs to achieve pollutant reduc-
tion objectives for the Menomonee River in Milwaukee,
Wisconsin. The Menomonee River watershed is 136
square miles, is 60% urban, and contains 18 municipali-
ties and parts of four counties. To meet ambient water qual-
ity standards, programs are needed to reduce sediment
by 50%, phosphorus by 50%-70%, and lead by 35%-70%.
The corresponding cost estimates for "segment" controls
for existing areas of development range from $94 million
to $184 million.
In many of the nation's larger cities, stormwater infra-
structure has fallen into disrepair, and significant invest-
ments will be required simply to meet generally accepted
engineering standards for stormwater conveyance and
flood control, let alone implementation of BMPs to meet
water quality objectives. In Indianapolis, Indiana, for ex-
ample, a mayor's blue-ribbon panel estimated the costs to
rehabilitate stormwater infrastructure at $283 million. The
infrastructure includes 1750 miles of storm sewers, more
than 1000 outfalls, more than 50 miles of levees, and a
number of regional detention ponds. The panel did not
estimate costs for programs to manage pollution in runoff.
In smaller towns, individual projects that in larger cities
would be considered routine can pose significant burdens.
In Vincennes, Indiana, a city with a population less than
20,000 and a median household income two-thirds of the
state median, the city is responsible for pumping water
from a drainage ditch over levees into the Wabash River
whenever water in the ditch reaches specified elevations.
The pumps are more than 50 years old and are in poor
repair. The city estimates annual costs for City Ditch to be
approximately $50,000, but no existing sources of revenue
are available to pay for rehabilitation and related O&M
costs.
In sum, while the estimated costs of retrofits vary tre-
mendously with the scale and scope of a program, invari-
ably new sources of revenues will be required to pay for
new programs. The costs of programs are debated by of-
ficials who have responsibility for implementation of them.
Stormwater utilities have emerged from these debates as
the option of choice to fund new programs.
How Can Retrofit Programs Be Funded?
What Are Stormwater Utilities?
Most jurisdictions historically have paid for investments
in stormwater infrastructure with revenues from property
taxes and other general revenues. Many, if not most juris-
dictions, now rely on a variety of sources to finance com-
prehensive stormwater programs. Table 2 is an abbrevi-
ated list of sources of revenues available to pay for differ-
ent elements of stormwater programs. One key observa-
tion from this list is that the sources of revenues most im-
portant for retrofit programs are property taxes and
stormwater user charges.
Stormwater user charges or fees are charges based on
some indicator or proxy for the actual volume of stormwater
runoff that leaves a property. The most common type of
charge is based on the amount, or square footage, of im-
108
-------
Table 2. A Functional Approach to Stormwater Financing
BMP Option
•Watershed planning
• Source controls
- Enforce ordinances
- Development regulation
•Maintenance (e.g., street
sweeping)
• Capital projects
- new development
- retrofit existing areas
- general revenues (property, income
taxes)
- stormwater user charges
- general revenues, stormwater user
charges
- plan review & inspection fees
-general revenues
- stormwater user charges
- developer exactions, fees-in-lieu
- bonds, sinking funds
- general revenues, stormwater user
charges
pervious area on a parcel. Other bases for stormwater
charges include the area and proportion of impervious
cover on a parcel, the intensity of development, and the
type of land use. In some instances, an estimate of the
actual volume of runoff or some estimate of the concen-
tration of pollutants in runoff may be used as the basis of
charges. Examples of rate structures are shown in Table 3.
Stormwater charges usually are administered by a
stormwater utility, an administrative unit or institution es-
tablished within or across jurisdictions for the purpose of
managing runoff and related problems. Revenues collected
by utilities are placed in separate enterprise funds or ac-
counts and can be used only for stormwater related ex-
penditures. The first stormwater utilities were established
in the mid-1970s, primarily to provide sources of revenue
for maintenance of stormwater infrastructure. Since the
1970s, the number of utilities has grown tremendously,
fueled in part by the efforts of stormwater managers des-
perate for funds to do their jobs.
Since the 1980s, as part of efforts to develop new sources
of revenues for stormwater programs, a number of sur-
veys of stormwater utilities have been completed. Table 4
is a summary of some of the results of these surveys. Im-
portant observations include:
•Average annual charges for residential property own-
ers range from $15 to $130.
•Average annual charges have increased overtime.
• Stormwater charges are the source of most revenues
for most stormwater utilities.
• The proportion of charges from different types of prop-
erty varies considerably.
• Total revenues from charges are significant and in-
creasing.
What Are the Advantages and
Disadvantages of Property Taxes and
Stormwater User Charges? Why Has the
Number of Stormwater Utilities Increased?
Stormwater utilities and user charges offer a number of
advantages over property taxes, the main alternative, al-
though taxes are preferable by some criteria (Table 5). It is
useful to consider the drawbacks of charges first.
Stormwater user charges are more difficult and costly to
implement than are taxes because institutions and proce-
dures to levy and collect taxes are already in place. User
charges are not deductible from federal and state income
taxes, and they are not elastic. Property taxes, on the other
hand, are deductible, and revenues from them increase
as property values appreciate without explicit decisions by
officials to increase rates or levies. Revenues from user
charges increase only if officials vote to increase rates.
Despite these disadvantages, reliance on stormwater
user charges is increasing, partly because user charges
are perceived as a more stable source of revenues. As
noted above, revenues from charges are placed in enter-
prise funds and can be used only for stormwater related
expenditures. Funding from general revenue sources like
property taxes is never secure because of fierce competi-
tion among political leaders and program managers for
scarce dollars. Under property tax systems, stormwater
managers often cannot count on budget allocations, do
not have as much control over their budgets, and cannot
plan as well.
Perhaps the most important reason that the number of
user charge systems is increasing is that property owners
believe charges are fairer. Impervious area - the basis for
most stormwater charges - can be measured and is a rea-
sonably objective measure. The idea that property owners
pay in proportion to the measured amount of hard surface
on their property seems fair. Property values, conversely,
are unrelated to the problem of runoff and perceived as
highly subjective. Many surveys suggest that property taxes
are the least popular form of tax.
A final reason that charges are preferable to taxes is
that they provide incentives for property owners to reduce
the amount of impervious area on their property and thereby
reduce volumes of runoff. Depending on how credits
against charges are structured, they also can provide in-
centives for on-site management
Local officials routinely consider these tradeoffs when
evaluating sources of funds for new programs like retrofit
programs. Because perception of fairness is such an im-
portant factor in public finance, it is useful to elaborate on
the issue of equity.
Who Pays More Under Property Tax and
User Charge Systems?
Although charges typically are perceived as fairer than
property taxes, this does not necessarily mean that any
109
-------
Table 3. Utility Rate Structures in Austin, Cincinnati, and Ft. Collins
Austin, Texas Cincinnati, Ohio
Ft. Collins, Colorado
Land Use Rate
Categories Factors Rate Categories
•Undeveloped .10 -Class A
Residential (<
10,000 sq.ft.)
• Residential .40 -Class B
Residential (>
10,000 sq.ft.)
• Nonresidential .80 -Class C
- Commercial
- Industrial
- Multi-family
- Transportation
- Institutional
-Agriculture
-Park
- Undeveloped
Area Range
Numbers
1
2
3
4
Intensity of
Development
Factors
.25
.20
.85
.75
.60
.50
.40
.08
.05
.00
Area (sq. ft.)
0-2000
2001 -4000
4001 -6000
6001 -8000
Basic
Category Runoff
Development Coefficient
•Very Light .00-.30
•Light .31 -.50
•Moderate .51 -.70
•Heavy .71 -.90
•Very Heavy .91-1.0
Runoff Coefficient (C)
C = Percent Impervious Area x 0,95
+ Percent Pervious Area x 0.20
+ Percent Semipervious x 0.50
where
-impervious means roof, concrete, etc.
-pervious means lawn, open space, etc.
-semipervious means gravel, etc.
Rate
Factor
.25
.40
.611
.80
.95
Table 4. Overview of Selected Stormwater Utility Surveys, 1988-1996 (Ungan 1997)
Date
1988
Survey
Stormwater
Management
Range of
Population
Served
20,000-
684,565
Range of
SFR*
Charges
$1 .25-
$3.63
Range of
Total Utility
Revenues
(000)
$263-
$8200
Range of
Total
Revenues
from
Charges
(000)
$425-
$8200
Range of
Charge
Revenues
as % of
Total
Revenues
78%-1 00%
Range of
SFR
Charges as
% of all
Charges
24%-62%
Adminstration,
Maryland Department
of the Environment
(MDE)
(Lindsey, 1988)
1990 MDE NA
(Update of 1988
Survey)
(Lindsey, 1990)
1991 The Florida NA
Department of
Environmental
Regulation
(1991)
1992 Black & Veatch 11,000-
Communications 329,227
(1992)
1992 Apogee Research Inc. 4,300-
(1992) 535,000
1993 Apogee Research Inc. NA
(1994)
$1.07-
$7.45
$1.00-
$4.50
$0.24-
$9.06
$1 -$4.50
$0.24-
$9.08
$75-
10,471
$118-
$6850
NA
NA
NA
$75-
$10,471
$118-
$6850
NA
$75-
$18,316
NA
82%-100%
19%-100%.
62%-100%
8%-100%
NA
15%-78%
NA
NA
NA
NA
(continued)
110
-------
Table 4. Continued
Date
1995
1995
Survey
Delaware Survey
(1995)
Florida Association of
Storm Water Utilities
(1995)
Range of
Population
Served
6000-
2,000,000
6000-
2,000,000
Range of
SFR*
Charges
$0.50-
$7.16
$0.50-
$7.43
Range of
Total Utility
Revenues
(000)
S19.7-
$21,600
S19.7-
$21 ,600
Range of
Total
Revenues
from
Charges
(000)
NA
NA
Range of
Charge
Revenues
as % of
Total
Revenues
NA
NA
Range of
SFR
Charges as
% of all
Charges
NA
NA
1996 Raftelis
March (Water and
Wastewater Survey)
(1996)
1996
July
1988-
1996
Indiana University,
Center for Urban
Policy and the
Environment
(Ungan,1997)
Min:
Max::
NA
11,141-
487,779
4300
3,489,779
$0.15-
$10.46
$0.24
$10.98
$0.15
$10.98
NA
$53-
$28,000
$53,000
$28,000,000
NA
$1.8-
$28,000
$1880
$28,000:000
NA
NA
0.7%-92%
1%
100%
0.7%
92%
Table 5. Advantages and Disadvantages of Taxes and Charges
Criteria Charges Taxes
• Cost of implementation
• Ease of implementation
• Deductible by property owner
• Elasticity of revenues
• Stability of revenues
• Fairness
- user (polluter) pays
- ability to pay
- Incentives for on-site controls
particular property owner will be better off under a charge
system than a system of property taxes. It is useful, there-
fore, to examine the relative burden on property owners
under the two systems. Analyses of the relative burden
typically show that, to generate a fixed sum of revenues,
residential property owners pay less under a user charge
system than under a property tax system. Non-residential
property owners like owners of commercial and industrial
properties typically pay less under a property tax system.
For example, to generate $500,000 in Roseville, Minne-
sota, residential property owners would bear 51% of the
burden under a property tax system but only 28% of the
burden under a user charge system (Table 6). Similar re-
sults have been reported in most jurisdictions where utili-
ties have been considered or established. The main rea-
son is clear: non-residential properties are highly impervi-
ous, while residential properties are only moderately im-
pervious, depending on their density. Another reason for
the difference in burden is that tax-exempt property own-
ers like churches, hospitals, and school pay charges. For
residential property owners, the benefit is partially offset
by the fact that charges are not deductible. Nevertheless,
they typically are better off under charge systems.
What Are Obstacles To Implementing User
Charge Systems?
Stormwater utilities are an attractive source of funds for
retrofit programs, and the number of utilities has grown
constantly over the past 20 years. Nevertheless, there are
a number of obstacles that limit their use. We believe that
the main obstacles are economic and therefore political.
Many people are opposed to all new taxes, regardless of
whether the taxes are perceived as fair. Hence, anytime a
utility is proposed, property owners will debate the merits
of the proposal, and political debate will occur. Two recent
cases from Indiana illustrate this point well.
In Vincennes, the Mayor sought new sources of funding
to pay for pumps in City Ditch. The Vincennes City Council
adopted an ordinance that established a mechanism for
allocating charges among property owners in the City Ditch
watershed based on parcel-level estimates of runoff vol-
umes. The Council did not, however, pass a companion
ordinance to establish a volume-charge. The Mayor lost
the next election, and efforts to establish the charge sys-
tem have foundered.
In Indianapolis, background studies for creating a utility
were completed in the 1980s, but no action to establish a
utility was taken. In 1997, following endorsement by the
Chamber of Commerce, a member of the City-County
Council proposed a new utility. The Mayor, who had been
111
-------
Table 6. Distribution of Property Taxes and User Charges in Roseville, Minnesota
(Honchell, 1986)
Total
Utility Charges
$519,000
100%
Property Taxes
User/Land Use Category
1 . Residential
2. Cemeteries/Golf Courses
3. Parks
4. Schools
5. Apartments/Churches
6. Commercial
Total
Revenues
$148,000
$4,000
$10,000
$11,000
$44,000
$302,000
Percent of
Revenues
28.5%
0.8%
1.9%
2.1%
8.5%
58.2%
Total
Revenues
$260,000
$10,000
$46,000
$203,000
Percent of
Revenues
50.1%
1.9%
8.9%
39.1%
$519,000
100%
elected on a pledge of no new taxes, did not endorse the
utility, but did not publicly oppose it. Many citizens and some
taxpayer groups opposed the proposal, as did some indi-
vidual members of the Chamber of Commerce. Votes to
establish the utility have been delayed because the pro-
posal lacks the necessary number of votes.
These cases are instructive because they demonstrate
that proposed new utilities will be controversial even when
stormwater problems are long-standing and well known
and the proposals are backed by political leaders. In many
communities, political leaders are unwilling to endure the
high cost of advocating new charges or taxes. Advocates
for retrofit programs necessary to achieve water quality
objectives must convince political leaders that the benefits
of retrofit programs exceed the costs.
Stormwater managers can inform debates through care-
ful analysis. For example, the perceived equity of a pro-
posed system can be enhanced through careful design of
the rate structure, including features such as credits for
on-site controls. Experience of local jurisdictions that have
successfully established utilities demonstrates that there
is not a single, correct approach. Innovative applications
of basic concepts can help provide funds for retrofit pro-
grams.
References
Apogee Research, Inc., (1991). Storm Water Utilities: Key
Components and Issues. Prepared forUSEPA, Wash-
ington DC. Bethesda, MD.
EPA (1997). 7996 Clean Water Needs Survey Report to
Congress. EPA832-R-97-003. Washington DC.
Gebhardt, Alicia and Greg Lindsey (1993). NPDES Re-
quirements for Municipal Separate Storm Sewer Sys-
tems: Costs and Concerns. Public Works, Vol. 124,
No. 1, p. 40-42.
Honchell, Charles. (1986). Creating a Storm Drainage Util-
ity. APWA Reporter, p. 10-11.
Lindsey,Greg (1991). An Update to A Survey of Stormwater
Utilities. Maryland Department of the Environment, Sedi-
ment and Stormwater Administration, Baltimore, MD.
Lindsey, Greg. (1990). Charges for Urban Runoff. Issues in
Implementation. Water Resources Bulletin, Vol. 26, no.
1,pp. 112-125.
Montgomery, James M. (1992). A Study of Nationwide Costs
to Implement Municipal Stormwater Best Management
Practices. Prepared for American Public Works Asso-
ciation, Southern California Chapter, Santa Ana, CA.
Ungan, Nur. (1997). A Survey of Stormwater Utilities. Envi-
ronmental Planning Quarterly. For the Environmental,
Natural Resources and Energy Division, American Plan-
ning Association, by the Center for Urban Policy and the
Environment, Indianapolis, IN. p. 3-7.
Water Environment Federation (1990). Organizing a Self-
Sustaining Utility for Stormwater Services. Alexandria,
VA.
Water Environment Federation (1993). Implementing Suc-
cessful Stormwater Utility. Alexandria, VA.
Wisconsin Department of Natural Resources (WDNR) et al.
(1992). A Nonpoint Source Control Plan for the
Menomonee River Priority Watershed Project. WDNR,
Bureau of Water Resources Management, Nonpoint
Source and Land Management Section, Madison, Pub-
lication WR-24492.
112
-------
Credits as Economic Incentives for On-Site Stormwater Management:
Issues and Examples
Amy Doll and Paul F. Scodari
Apogee/Hagler Bailly, Inc.
Bethesda, Maryland
Greg Lindsey
Indiana University, School of Public and Environmental Affairs,
Indianapolis, Indiana
Stormwater utilities provide an institutional mechanism
for incentives such as credits or reduced user charges in
the implementation of onsite Stormwater management.
Such incentives create greater flexibility by allowing each
user to chose the least-cost option—paying the Stormwater
utility charge or implementing onsite Stormwater manage-
ment. This paper provides examples of Stormwater utili-
ties with credits for onsite storm water management, in-
cluding credits for peak runoff controls, implementation of
water quality best management practices, and proper
maintenance of onsite Stormwater facilities. Also discussed
are credits as economic incentives to encourage preven-
tion or reduction of Stormwater runoff problems. As eco-
nomic incentives, credits must be sufficient to induce
changes in behavior; however, their impact on total utility
revenues must be examined carefully.
Introduction
A Stormwater utility is a public utility established to pro-
vide Stormwater management services. Stormwater utili-
ties, like other utilities, rely on dedicated user charges re-
lated to the level of service provided. These user charges
are usually based on the amount of impervious area on a
property (i.e., a proxy for the estimated amount of runoff
discharged from a property). Stormwater utility charges
typically are paid by property owners and managed in a
separate enterprise fund, which is dedicated to financing
local Stormwater management services. Most Stormwater
utilities are administered under public works departments
or local departments of utilities that also provide wastewa-
ter or water services.
Experience with Stormwater utilities has shown that they
are capable of generating substantial revenues for local
Stormwater management programs at relatively nominal
charges. Typical monthly charges for residential users
range from $2 to around $6 per month. Nonresidential prop-
erty owners typically pay more because their property is
generally larger and developed more intensively.
Stormwater utilities offer three major advantages over
financing local Stormwater programs from the general fund
through property tax revenues. A Stormwater utility:
• Provides a dedicated, and stable source of funds for
all facets of Stormwater management programs (pol-
lution prevention, capital investments, and operation
and maintenance);
• Raises funds through charges based on a user's con-
tribution to local Stormwater runoff problems an ap-
proach often seen as more equitable to rate payers or
the public; and
• Provides an institutional mechanism to incorporate in-
centives (e.g., reduced charges) for implementation
of onsite Stormwater management.
Overview of Credits as Incentives for Onsite
Stormwater Management
The impetus for establishing credits in a Stormwater util-
ity rate structure is that a utility may achieve greater flex-
ibility in protecting water quality and aquatic habitat in ur-
ban watersheds at a lower overall cost to the community.
This greater flexibility can also help a utility lower the total
costs of Stormwater management forthe community. A util-
ity could also reward those users that go beyond minimum
requirements in the local Stormwater management code,
if a credit approach is structured accordingly.
Credits are usually made available only to nonresiden-
tial property owners. For utilities where charges to resi-
dential properties account for a significant proportion of
total revenues, there is less potential forthe efficiency gains
possible through lowering the total costs of Stormwater
management.
From an economic perspective, the extent to which a
credit will increase the efficiency of a Stormwater program
113
-------
depends partly on the conditions in which it applies. For
example, if individuals who develop property are not given
the option either to build stormwater management facili-
ties and receive a credit or to pay charges and avoid build-
ing facilities, then some of the incentive effect is lost. In
cases where retrofitting is desired, whether or not a credit
will induce property owners to build new stormwater man-
agement facilities where none exist or retrofit existing fa-
cilities to reduce stormwater charges depends on the size
of the charge and the magnitude of the credit.
Examples of Credit (Fee Reduction)
Approaches
A recent survey of stormwater utilities (NAFSMA, 1996)
asked utilities whether they included incentives, such as
reduced user charges, for commercial and industrial prop-
erties that implement onsite stormwater management. Of
the 38 utilities that responded, 71% (27 utilities) had no
fee reduction. Of the remainder of (11 utilities), two major
types of fee reduction approaches were reported: 16% (6
utilities) had fee reduction for peak runoff controls, and
8% (3 utilities) had fee reduction for implementation of wa-
ter quality best management practices or proper mainte-
nance of onsite stormwater facilities. An earlier report on
stormwater utilities (USEPA, 1992) found over 20 utilities
with various types of credits as incentives for onsite
stormwater management.
Some stormwater utilities offer credits for onsite
stormwater detention/retention facilities in new develop-
ments. Credits can also provide incentives for onsite
stormwater detention/retention through retrofitting older dry
detention basins to extended detention basins or control-
ling peak flows through rooftop or underground storage
tanks. Examples of credit approaches for selected utilities
are highlighted below and summarized in Table 1.
Gainesville, Florida
The City of Gainesville's Stormwater Management Util-
ity provides reduced monthly fees for nonresidential prop-
erties with privately maintained, onsite stormwater man-
agement retention systems. The maximum allowable credit
is 100% of the utility's "base" fee, which is based on the
amount of impervious area and one-half of pervious park-
ing areas. The percentage of fee credit is determined by
the volume of onsite retention provided (detention volume
is not considered since that stormwater is discharged). The
required volume is determined by the 25-year, 24-hour
storm. Most credits range from 15 to 35%.
Orlando, Florida
In the City of Orlando, the stormwater utility provides a
lower rate for commercial and multi-family residential prop-
erties with onsite stormwater management facilities. Such
properties with approved onsite retention or detention get
a credit on the rate charged per ERU (equivalent residen-
tial unit). The typical rate is $66.00 per ERU. The lower
rate for properties with approved onsite stormwater facili-
ties is $38.28 per ERU. Overall, this provides a 42% credit
on the stormwater utility fee.
Wichita, Kansas
The City of Wichita's Stormwater Utility offers credits only
for properties with 50 or more equivalent residential units.
Two credits on the drainage fee are available. First, up to
40% credit on the fee is available for detention that equals
or exceeds the city's new development standards (based
on 100-year design storm). Second, an 80% credit on the
fee is available for retention (no runoff from site). No cred-
its are being given because the stringent standards are
difficult to achieve.
Louisville & Jefferson County Metropolitan
Sewer District, Kentucky
Credits are provided primarily for commercial properties
with onsite detention for control of peak flows in the Louis-
ville/Jefferson County Metropolitan Sewer District (MSD).
A range of credits is available depending on how the de-
tention basin functions. Basins must be sized for the 2-
year, 10-year, and 100-year storms and also limit dis-
charges to the pre-development rate of runoff. Credits are
available for each type of storm, with an 82% maximum
credit if all criteria are met. MSD is currently evaluating
how to incorporate stormwater quality measures into its
credit approach.
St. Paul, Minnesota
The City of St. Paul provides a rate of discharge credit
for nonresidential properties on its storm sewer system
charge. For nonresidential properties, this charge is based
on actual parcel acreage and a standardized peak runoff
rate determined forselected land use classifications. Where
the peak stormwater runoff rate is limited by onsite facili-
ties such as detention ponds owned and maintained by
the property owner, up to a 25% credit is available. A10%
credit is provided for parcels that provide onsite storage
for the 5-year design storm that also limit its discharge to a
maximum of 1.64 cubic feet per second per acre. An addi-
tional 15% credit is provided for parcels that provide onsite
storage for the 100-year design storm that also limit its
discharge to a maximum of 1.64 cubic feet per second per
acre. Both new developments and redevelopment are eli-
gible for apply for credit. Existing nonresidential properties
can retrofit to provide onsite storage for the 5-year design
storm and get the 10% credit. Most credits were provided
in the first few years after the credit approach was estab-
lished. Currently, around 3-4 credits are approved annu-
ally. In St. Paul, the credit approach increased the political
acceptability of the storm sewer system charge.
Charlotte, North Carolina
The City of Charlotte provides one or more credits for
commercial, industrial, institutional, and multi-family resi-
dential properties and residential homeowner associations
that mitigate the impacts of runoff on the stormwater sys-
tem. Eligibility for one or more credits to the service rate
114
-------
Table 1. Summary of Credit Options
Utility Eligible Users
Wichita, KS
St. Paul, MN
Charlotte, NC
Durham, NC
Cincinnati, OH
Tulsa, OK
Austin, TX
Bellevue, WA
King County, WA
Indianapolis, IN
Basis for Credit
Design Storm
Maximum Credit
Typical Credit
Gainesville, FL
Orlando, FL
Nonresidential
Commercial and multi-
family residential
Volume of onsite
Onsite retention or
detention
25-year, 24-hour storm
NA
1 00% of base fee
42%
1 5-35%
42%
Properties > 50 ERUs
Two credits: volume 1) 100-year storm
of detention or retention 2) Complete retention
Louisville-Jefferson Commercial properties
County, KY
Nonresidential
properties
Commercial, industrial,
institutional, multi-
family residential;
homeowner association
Nonresidential
properties
Onsite detention of
peak flows
Onsite detention of
peak flows; acreage,
peak flows
1) peak discharge
2) total runoff volume
3)annual pollutant
loading reduction
Pollution credits for
Water quality and
quantity controls
Commercial properties Onsite retention
2-year, 10-year, 100-
year storms; pre-
development runoff
5-year, 100-year
storms; release limited
to 1.64 cfs/acre
1) 10-year, 6-hour
2) 2-year, 6-hour
3) reduction in loading
State standards for
facility design; esti-
mated pollutant
removal efficiency
Limit discharge to pre-
development runoff
Privately maintained
facilities
Commercial properties
All properties
50% greater detention;
maintenance costs of
onsite facilities
Onsite detention,
inspection
Onsite detention;
intensity of development
Commercial properties Private maintenance
Nonresidential
properties
Discharge to specified Tier Two:2-, 10-, 25-,
streams; onsite 50-, 100-year events
retention or detention
watershed size
1) 40%
2) 80%
82%
10% (5-year storm)
25% (100-year storm)
1)50%
2) 25%
3) 25%
Up to 100%
25%
50%
60%
50%
Reduction of one rate
(intensity of
development) class
Reduction of one rate
class
Tier One: 25%; <$50
Tier Two: 35%; <$250
Currently no
applications
Varies with
degree of
control
Varies with
degree of
control
Varies with
degree of
control
Few applications
Credit never
used
Varies
50%
Varies
Varies
(proposed)
charge is proportional to the extent those stormwater man-
agement measures address the impacts of peak discharge,
total runoff volume, and annual pollutant loading from the
site. Portions of the service rate charge are available for
credit as follows: up to 50% for reducing peak discharge
from a 10-year, 6-hour storm; up to 25% for reducing total
runoff volume from a 2-year, 6-hour storm; and up to 25%
for annual pollutant loading reduction. Each credit allowed
against the service charge is conditional on continued com-
pliance with the Charlotte Mecklenburg Land Development
Standards Manual and may be rescinded for noncompli-
ance with those standards. If 100% credit is given, the af-
fected property will receive no stormwater service charges.
Durham, North Carolina
The City of Durham provides up to a 25% pollution credit
on the stormwater utility fee for selected structural
stormwater contra Is on nonresidential properties. Currently,
the maximum pollution credit goes to standard basin de-
signs that are identified as achieving maximum pollutant
removal efficiency in state performance standards. For
other structural controls in the state's standards, the city's
pollution credit will be linearly variable, with no credit given
fora removal efficiency of 0% of total suspended solids to
a 25% credit for a removal efficiency of 85% of total sus-
pended solids. The city recently approved sand filters in
addition to the approved onsite basin designs, but no pol-
lution credits are established yet for sand filters. Durham
receives few applications for credits.
Cincinnati, Ohio
The City of Cincinnati's Stormwater Management Utility
offers a credit for commercial properties that install onsite
115
-------
retention that goes beyond normal building requirements
(i.e., limit discharge to pre-development level of runoff).
Such properties can apply for a credit of up to 50% on the
utility's storm drainage service charge. This credit has never
been used in Cincinnati.
Tulsa, Oklahoma
Under the City of Tulsa's stormwater drainage system
service charge, credits are provided for private mainte-
nance of approved onsite detention or retention facilities.
An approved onsite facility must provide at least 50% more
detention than required by the city. The amount of credit
varies based upon the estimated maintenance costs if the
city were providing the maintenance. The maximum credit
is 60% of a property's annual stormwater charge. This
maximum was established at 60% because around 60%
of the stormwater utility budget in Tulsa goes to mainte-
nance. Upon inspection, if an onsite facility is not perform-
ing adequately, then the property owner must pay the typi-
cal stormwater drainage service charge.
Austin, Texas
The City of Austin's Drainage Utility provides a 50% credit
on the drainage fee for commercial property owners that
construct and maintain approved onsite detention facili-
ties. The city inspects these onsite facilities annually to
ensure proper maintenance.
Bellevue, Washington
The City of Bellevue Storm and Surface Water Utility
provides a credit on its storm and surface water drainage
service charge for approved onsite detention facilities. This
credit has worked well to get approved detention facilities
built on large residential and commercial plats. Bellevue's
utility rate structure classifies each property according to
its percentage of developed property (from undeveloped
land to very heavy development). A reduction of one inten-
sity of development classification is provided for installa-
tion and maintenance of approved onsite detention facili-
ties. This reduces the rate (based on the intensity of de-
velopment classification) and the storm and surface water
drainage service charge for such properties.
King County, Washington
Under the new King County Surface Water Drainage
Design Manual, any development of parcels with over 5,000
square feet of impervious area must provide onsite deten-
tion/retention. For commercial properties, King County pro-
vides a credit through a reduction of one rate classification
for the utility fee for private maintenance of an approved
onsite detention/retention facility. The facility must be built
to code and meet King County maintenance standards.
Issues in Establishing Credits for Onsite
Stormwater Management
Like stormwater utility charges, there is no "correct"
method for establishing credits. Each utility must consider
local stormwater management goals in deciding whether
to incorporate such incentives into their utility rate struc-
ture. The amount of impervious area on a property is usu-
ally the basis for stormwater utility charges. The quantity
of stormwater runoff is generally the rationale behind charg-
ing property owners for stormwater management services
(e.g., a user-pay approach). The adverse environmental
impacts of urban runoff are related to both stormwater
quality and quantity. To date, few stormwater utilities have
attempted to incorporate measures of the quality of runoff
as a basis for utility charges. Additionally, few utilities in-
corporate site characteristics other than impervious area
(e.g., slope and soil characteristics) that also influence the
adverse impacts of runoff. These factors may be impor-
tant in setting charges and credits to induce the expected
behavior of choosing the least-cost option. On the other
hand, if stormwater quantity (as measured by the amount
of impervious area) is closely correlated with adverse im-
pacts of runoff related to both stormwater quantity and
quality, the amount of impervious area may be a sufficient
basis for setting charges that create the desired incentives.
Although credits must be sufficient to induce changes in
behavior, their impact on total utility revenues must be ex-
amined carefully. An approach that gave large credits for
onsite stormwater management could significantly reduce
revenues for a local stormwater management program.
Each community should evaluate whether charges and
credits proposed for its utility are likely to promote onsite
stormwater management and whether mechanisms are in
place to ensure that onsite stormwater management
achieves the desired environmental results.
Finally, public acceptability and political support is im-
portant to establishing a utility rate structure, whether or
not it includes a credit approach. The nature of local gov-
ernment is that key players in utility design and implemen-
tation are seldom the key players in local politics. In de-
signing a credit approach, a utility can attempt to minimize
controversy by developing education and involvement pro-
grams for informing and gaining the support of local gov-
ernment officials and the public.
Case Study of Issues Associated with
Proposed Credits in Indianapolis, Indiana
The City of Indianapolis is currently attempting to de-
sign a credit approach for its proposed stormwater utility.
Considerable controversy has arisen over the proposed
utility and a credit system is under consideration in part to
help overcome general opposition to new charges or taxes.
Through a credit system, utility planners and local elected
officials are attempting to make the proposed stormwater
utility charges more equitable and acceptable politically.
The credit system in the most recent draft ordinance (Pro-
posal No. 657, 1997) is a relatively complex approach to
provide a reduction in stormwater user fees for nonresi-
dential properties based on 1) certain qualifying conditions
(location in relation to a major waterway), 2) activities that
mitigate the impact of increased stormwater runoff from a
property on a continuing basis, or 3) activities that reduce
the city's cost of providing stormwater management ser-
116
-------
vices to a property. The draft ordinance outlines a two-
tiered credit that is based on watershed area as well as
the size of the onsite detention/retention basin. The city
will also develop a proposed Storm Water Credit Manual
for use in reviewing and acting upon applications for credit.
A credit application fee is also authorized in the draft ordi-
nance. Efforts to establish a credit system for onsite de-
tention/retention have addressed concerns of property
owners and generally increased the perceived fairness of
the proposed rate structure, and it is clear that the pro-
posed utility could not be adopted without some type of
credits. Inclusion of a credit system, however, has not been
sufficient to ensure adoption of the stormwater utility and
overcome other objections.
Conclusion
Economists have long advocated pollution charges as
an approach to achieve greater flexibility and efficiency in
pollution control. If such charges are set to reflect the en-
vironmental damage actually caused by polluted dis-
charges, economic theory suggests they can create in-
centives for each user to choose the least-cost option—
paying a pollution charge or implementing pollution con-
trol requirements. Making credits available on stormwater
utility charges for implementation of onsite stormwater
management can create comparable incentives for users
and potential efficiency gains by lowering the total costs of
a stormwater management program. Additional research
is needed to evaluate the efficiency and equity issues as-
sociated with credits and stormwater utility charges. Until
the economic and data issues in establishing a credit ap-
proach are better understood, communities considering a
credit approach should examine the experience of those
utilities that have implemented credits to evaluate whether
such approaches are appropriate for local stormwater
management goals and problems.
References
Doll, Amy, John Cameron, and Rick Albani. "Stormwater
Utilities: A User Pay Approach to Stormwater Manage-
ment," Partners for Smart Growth Conference, spon-
sored by the Urban Land Institute and U.S. Environ-
mental Protection Agency, December 2-4,1997, Balti-
more, MD.
Doll, Amy, et al., "Storm Water Management: Financing
Local Programs with a Utility Approach," Finance Alert
(Summer 1992), pp. 2, 89.
Lindsey, Greg. "Charges for Urban Runoff: Issues in Imple-
mentation." Water Resources Bulletin (February 1990),
pp. 117-125.
National Association of Flood and Stormwater Manage-
ment Agencies. "1996 NAFSMA Survey of Local
Stormwater Utilities."Washington, DC: National Asso-
ciation of Flood and Stormwater Management Agen-
cies (1996).
U.S. Environmental Protection Agency, Office of Policy,
Planning and Evaluation. "Storm Water Utilities: Inno-
vative Financing for Storm Water Management." Un-
published report prepared by Apogee Research, Inc.
(March 1992).
117
-------
Conservation Design for Stormwater Management
Earl Shaver, Environmental Engineer
Delaware Department of Natural Resources and Environmental Control
(presently a Technical Specialist, Auckland Regional Council)
Auckland, New Zealand
Background
The State of Delaware has developed a manual to pro-
vide guidance for site design which incorporates conser-
vation into land development (DDNREC and BC, 1997).
The intent is to provide an incentive for land developers to
retain and incorporate existing natural site features into
the site development process and thereby reduce or elimi-
nate the need for structural stormwater management con-
trols. Other benefits are certainly realized through Con-
servation Design, such as more closely approximating the
predevelopment water budget, protection of habitat, and
reduced overall impact to the receiving system. Site fea-
tures discussed in the manual include:
• Wetlands
• Floodplains
• Forested areas
• Meadows
• Riparian buffers
• Soils
• Other natural features
Design procedures are provided which allow site design-
ers to incorporate practices inherently known to be good,
but which have not had the detailed design guidance that
ensures plan approval. That guidance is provided in the
manual for a variety of situations. The design approach is
flexible enough to allow for various conservation practices
to be combined on one site and to quantify the benefits of
that combination.
It must be emphasised that structural controls will still
be essential on many sites. A heavily wooded site having
a significant portion of the tree canopy removed will still
have a significant increase in stormwater runoff, even with
aggressive conservation planning. The practices detailed
in the manual are provided as additional tools in the
stormwater management toolbox. They supplement struc-
tural control practices and may, in some situations, elimi-
nate or reduce the need for structural practices while pro-
viding attractive site amenities.
Limitations of Structural Stormwater
Management
Most stormwater management programs place a heavy
reliance on implementation of structural stormwater man-
agement facilities. These facilities include ponds, both wet
and dry; infiltration; filtration; and other variations of them
all. The implementation of these facilities is necessary for
their water quantity and water quality benefits and is ex-
pected to remain integral to program implementation, but
there should not be an overreliance on them. These prac-
tices, in and of themselves, cannot eliminate adverse im-
pacts of urban development. In addition, there are a num-
ber of limitations to structural facilities.
A stormwater management program relying solely on
structural practices has a number of weaknesses.The ex-
istence of these weaknesses has been recognized forsome
time, but there has been little information available on al-
ternative approaches that would justify their inclusion in a
stormwater management program. The following items
present some of the weaknesses.
• Lack of flexibility in site design
• Altered site hydrology
• Expense
• Loss of site area
• Potential increased impacts to site and watershed
natural resources
• Configuration of development
• Connection of impervious areas
• Disregard of site resource conservation benefits
• Maintenance obligations
118
-------
Conservation Design Approaches
Conservation Design approaches reflect a totally differ-
ent philosophy towards site design which integrates
stormwater management into the very core of site design,
as opposed to an afterthought. These approaches can in-
clude an almost endless universe of practices, strategies,
planning, and common sense. The manual doesn't include
all potential components, but provides guidance and infor-
mation on many that are currently recognized where data
exists or can be generated to substantiate their benefits
from a water budget perspective.
It is important to develop a conservation ethic which treats
stormwater runoff as a "resource" ratherthan a "byproduct"
of development. As such, there are a number of key site
design components to consider:
• Reducing impervious surfaces
• Constructing biofiltration practices
• Creating natural areas
• Leaving areas undisturbed
• Clustering development
Conservation approaches are discussed throughout the
manual, but some are briefly discussed here to provide an
initial awareness of the range of options that will be dis-
cussed later in greater detail. Examples of conservation
approaches include the following.
Reducing Impervious Surfaces
Impervious surfaces (roads, roofs, sidewalks) prevent
the passage of water through the surface into the ground.
Water must then be transported across the surface to a
point of discharge. Reducing the total amount of impervi-
ousness is the single most important conservation tool
available. Residential subdivisions can reduce the width
of roadways, or design the roadways to limit the total length
needed to service individual properties. Roof downdrains
should not be directly connected to streets when providing
splash blocks, but should discharge the water away from
impervious surfaces (sidewalks, streets) to allow for a
greater amount of water to infiltrate into the ground.
Just as important in limiting impervious surfaces and
separating roof drains from direct connection to streets is
the need for education of homeowners regarding their re-
sponsibility to ensure continued function of these practices.
Homeowners often change the orientation of downspouts
or otherwise redirect lot drainage to impervious surfaces
which undoes a lot of conservation benefits. Community
education and involvement is integral to effective program
implementation.
Constructing Biofiltration Practices
The use of vegetative swales and buffer strips can pro-
vide a significant water quality benefit in addition to reduc-
ing the total volume of stormwater runoff. The primary pro-
cesses involved in their performance are filtering of pollut-
ants contained in stormwater run off, and infiltration of run-
off into the ground.
Even where curbs are needed to restrain traffic move-
ment to paved surfaces, curb cuts or openings can be
placed to allow water to pass off of the paved surface into
a biofiltration facility. This would allow for both public works
and stormwater objectives to be attained.
Creating Natural Areas
In many site development situations, the predevelopment
condition may be farmfield or other disturbed condition.
Creation of a meadow as open space would have signifi-
cant stormwater management benefits for both water quan-
tity and water quality. The area, if well designed and con-
structed, could become an attractive amenity to a commu-
nity and enhance the value of the properties.
Leaving Areas Undisturbed
Many sites have existing resources which, in addition to
other values, have stormwater management benefits.
These natural systems include forested areas, wetlands,
and other areas of natural value such as meadows.
Forested areas provide for rainfall interception by leaf
canopy. In addition, an organic "duff area" develops on the
woodland floor which acts very much as a sponge to cap-
ture the water and prevent overland flow. In addition, trees
use and store nutrients for long periods of time. Trees also
moderate temperatures during the summer and provide
wildlife habitat, thus providing other environmental ben-
efits.
Wetlands are valuable resources and provide numer-
ous benefits including flood control, low streamflow aug-
mentation, erosion control, water quality, and habitat. They
are very productive ecosystems whose maintenance would
have significant water quantity and quality benefits. Where
they exist on a land development site, they could become
an important element in site design.
Cluster Development
How a site is developed and to what degree the entire
site is utilized will have a significant impact on stormwater
runoff from the site. Conventional land development en-
courages sprawl, while innovative approaches to land de-
velopment can provide significant stormwater benefits.
Cluster development encourages smaller lots on a portion
of a site, allowing the same site density, but leaving more
site area in open space. Clustering designs residential
neighborhoods more compactly, with smaller lots for nar-
rower single-family homes, found in traditional villages and
small towns. Cluster development can provide for protec-
tion of site natural areas, while at the same time reducing
total site imperviousness by reducing the areal extent of
roads.
119
-------
The Conservation Design Procedure in
Detail
Conservation Design can be thought of as a series of
questions which must be asked as it is applied to each
site. If site designers rigorously address all of these ques-
tions, the Conservation Design procedure will have been
accomplished, and the "answers" will be successfully iden-
tified for each site. The overriding objective is to achieve a
new way of thinking about site design.
The procedure has been kept simple by intention. It is
grounded in effective and complete site analysis, and an
upfront commitment by the site designers to inventory and
evaluate the various "systems" which define the site and
which pose problems as well as opportunities for site de-
velopment. The more clever the development "tinkering"
can be, the more successful Conservation Design can be-
come. Extra effort up front pays important dividends in the
long run. Conservation Design requires a major departure
from the conventional mindset of stormwater disposal -
which is a reactive end-of-the-line process forcibly imposed
onto a development program. Conservation Design is pro-
active in the best sense of the word, based on understand-
ing natural system opportunities which enable us to inte-
grate essential stormwater quality and quantity manage-
ment objectives into the devlopment design from the very
beginning.
Rather than provide a lengthy discussion of conserva-
tion design procedures, this paper provides a checklist of
items or "questions" which should be considered in con-
servation design. Those questions are listed as follows:
1. Site Analysis Background Factors: How do Back-
ground Site Factors Affect the Conservation Design
Process?
Hydrologic issues:
Is the site tidally dominated?
Does the site flow to special waterbodies with spe-
cial water quality needs?
Are there known downstream flooding problems?
The site is located in what watershed?
Does the site discharge into 1st, 2nd, 3rd order
streams?
Is the site in the upper, middle, or lower part of the
watershed?
2. Site Analysis Site Factors Inventory: What Site Physi-
cal Factors Affect Conservation Design?
Site size and shape:
Does site size limit Conservation Design?
Does site shape or other factors limit Conserva
tion Design?
Natural features:
What is the basic site hydrology?
Perennial streams?
Intermittent swales?
Describe site soils
Describe site vegetation
Describe site critical features:
Do wetlands exist?
Are there floodplains?
Are there riparian areas?
Are natural drainageways present which are
not perennial streams perse?
Are there special habitat areas?
Do special geological formations exist (i.e.,
carbonate)?
Do steep slopes exist?
Are there high watertable, bedrock, other limi-
tations?
Built/developed features:
Does the site have centralized sewer?
Does the site have centralized water?
3. Site Factors Analysis: What Site Factors are Con-
straints and Opportunities in terms of Conservation
Design?
Site Constraints:
Where should building and roads be avoided?
In terms of vegetation?
In terms of soils?
Are any areas off limits for all forms of disturbance?
Site Opportunities:
Where does most recharge occur?
In terms of vegetation?
In terms of soils?
4. Building Program: How do Building Program Factors
Enter into the Conservation Design Procedure?
Can the proposed building program be reduced in
terms of total number of units?
Can the type of units be modified (e.g., from single-
family to townhouse)?
What is existing site zoning?
Are zoning options allowed?
Have building setbacks been made to be flexible?
Have innovative development concepts such as
zero lot line or clustering been considered?
What does the comprehensive plan indicate for the
site and adjacent areas?
What are the adjacent land uses?
Other Management/Regulatory issues:
What municipal/county requirements exist for
stormwater?
Will some aspects of Conservation Design require
waivers?
What other municipal/county requirements exist
for land development?
Will some aspects of Conservation Design require
waivers?
5. Lot Configuration and Design: How Can Lot Config-
uration and Overall Site Design Prevent Stormwater
Generation?
120
-------
Have lots been reduced in size to the maximum de-
gree?
Have lots/uses been clustered/concentrated to the
maximum degree?
Have lots been configured to avoid critical areas?
Have lots been configured to take advantage of ef-
fective Conservation Design mitigative practices?
6. Impervious Coverage: Have Impervious Surfaces
Been Reduced as Much as Possible?
Have road lengths and widths been reduced to the
maximum degree?
Have driveway widths and lengths been minimized
to the maximum degree?
Have parking ratios and parking sizes been reduced
to the maximum extent?
Has potential for shared parking been examined fully?
Have cul-de-sacs and turnarounds been designed to
minimize imperviousness?
Have sidewalks been designed for single-side move-
ment?
Can porous surfaces be used for overflow parking,
low impact shoulders, other applications?
7. Minimum Disturbance/Maintenance: Has Disturbance
of Site Vegetation and Soils Been Minimized?
Has maximum total site area, including both soil and
vegetation, been protected from clearing and any
other type of development disturbance?
Are zones of open space maximized?
Do these open space zones make sense internally,
externally?
In terms of individual lots, has maximum lot area, in-
cluding both soil and vegetation, been protected from
clearing and other development-related disturbance?
Do structures correspond to site features such as
slope, in terms of type of structure, placement on lot,
elevation, and so forth?
Have revegetation opportunities been maximized
throughout the site?
Have revegetation opportunities been maximized in
critical areas such as riparian buffer zones?
8. Use of Mitigative Conservation Design Practices:
Which Practices are Most Effective and How Can
Their Positive Effects be Maximized?
Are vegetated swales with check dams being used?
Are vegetated filter strips with level spreading devices
being used?
Are berms and other terraforming technique being
used in conjunction with zones of natural vegetation?
9. The Conceptual Stormwater Management Plan: How
Can All Preventive Approaches and Mitigative Tech-
niques be Integrated into an Optimal Conservation
Design Plan?
How has the stormwater plan been integrated into
the overall site design?
Has prevention been maximized through Conserva-
tion Design Approaches?
Has mitigation been maximized through Conserva-
tion Design Practices?
What other benefits are achieved through Conserva-
tion Design (i.e., open space, enhanced marketabil-
ity, cost reduction, habitat protection, stream water
temperature, biota impacts, other stream impacts?)
10. Stormwater Calculations: How Has Conservation
Design Affected Stormwater Calculations? What Con-
ventional Stormwater Techniques are Necessary to
Manage Any Residual Stormwater Need not Mitigated
by Conservation Design?
How has impervious cover been reduced?
What are the implications for Curve Numbers?
How have total runoff volumes been affected?
Has time of concentration been maximized?
How has peak discharge rate been affected?
How has recharge volume been affected?
11. Selection of Additional Stormwater Controls: If Con-
servation Design has not Fully Met all Stormwater
Requirements, What Additional Requirements Must
be Provided?
Watershed Wide Approaches
While not a focus of the manual, watershed-wide con-
siderations are important and should be the context from
which many resource-based land development decisions
are made. The manual strongly supports watershed-based
approaches to land use decisions. This context is impor-
tant from a number of perspectives.
• Watershed approaches allow for a recognition and
consideration of where growth distribution should oc-
cur.
• Consideration of land use from a watershed perspec-
tive allows for a greater awareness of the cumulative
impacts of watershed development.lmpervious sur-
faces are important to consider if downstream areas
are to be protected.
• A comprehensive approach to resource protection can
be developed and implemented based on consider-
ation of watershed specific issues such as steep
slopes, high watertable, the need for aquifer recharge,
etc.
• A watershed approach allows for developers and the
general public to understand the basis by which land
use decisions were made in a rational format which
can be easily understood.
• Land use decisions based on watershed-wide analy-
ses provide the local government with a basis for mak-
ing land use decisions that can be defended.
As desirable as watershed-wide approaches are, it must
be recognized that significant resources and costs may be
121
-------
needed to accomplish those efforts. Depending on the
goals of the effort, significant data needs may exist.
Conclusion
Over the past 20 years, stormwater management has
evolved from water quantity control, to water quality con-
trol, to attempting to address stream ecology. What has
become apparent is that traditional end-of-pipe controls
such as ponds do not provide the level of protection nec-
essary to protect in-stream resources. We have gone full
circle in again having to consider water quantity, but this
time not just to reduce downstream flooding concerns. The
total volume of water running off the land, in addition to
riparian buffer protection, becomes critically important.
Site features, as mentioned in the abstract, must be con-
sidered integral to site development. Too often we have
totally reconstructed a landscape for an individual's eco-
nomic benefit only. We must recognize the economic and
resource impacts that occur downstream from sites being
developed.
There are ways to develop sites and protect or enhance
existing resource values. It is not rocket science. If we as
a society consider conversion of land to urban use as a
desirable societal product, we must do more than accept
the adverse impacts that those site activities cause. We
can minimize adverse downstream impacts if greater
weight is given to existing site resources. In many situa-
tions, as shown in case studies, land can be developed,
less expensively using greater protection of existing site
resources, than when using a conventional approach.
The choice is ours.
References
Delaware Department of Natural Resources and Environ-
mental Control and the Brandywine Conservancy, Con-
servation Design for Stormwater Management, Sep-
tember, 1997
122
-------
Results of the Site Planning Roundtable
Whitney Brown
Center for Watershed Protection
Ellicott City, Maryland
The Site Planning Roundtable, originally convened by
the Center for Watershed Protection in October 1996,
brought together representatives from various national
planning organizations, development and environmental
communities, and local government. The goal was to pro-
vide the technical, professional, and real-world validation
required to promote environmentally sensitive, locally rel-
evant, and economically viable development. In line with
this goal, the Roundtable has developed a set of 22 Model
Development Principles that aid local planners and zoning
officials in identifying how existing ordinances can be modi-
fied to reduce impervious cover, provide effective
stormwatertreatment, and conserve natural areas. These
principles are not national design standards. Instead, they
identify areas where existing subdivision codes can be
changed to better protect streams, lakes and wetlands at
the local level.
Conventional zoning standards outline minimum lot ar-
eas, setbacks, frontages, and road widths, often resulting
in significant impervious cover in the form of wide streets,
expansive parking lots, and large-lot subdivisions. Plan-
ners, landscape architects, and developers can utilize a
wide range of innovative site planning techniques to re-
duce imperviousness at the site level. In some cases, full
utilization of these techniques requires changes to outdated
zoning regulations or inflexible subdivision codes. The
Model Development Principles focus on changing these
regulations and codes. Each principle presents a simpli-
fied design objective; techniques for achieving the objec-
tive should be based on local conditions.
Residential Streets
An important objective of the Site Planning Roundtable
effort was to identify practical and cost-effective strategies
to overcome barriers to implementation of the Model De-
velopment Principles. One particular area of concern
emerged: residential road width. Most local governments
model their residential street design standards upon state
and/or federal highway criteria, although the traffic capac-
ity and function of residential streets differ considerably
from that of highways. Consequently, residential street
widths tend to be wide rather than narrow. Efforts to re-
duce road widths are often met with strong opposition on
a variety of fronts. Local planners and engineers are re-
luctant to modify standards due to safety concerns. Public
works officials wish to maintain adequate access for emer-
gency, service, and maintenance officials. Residents voice
concerns about impacts to parking.
The following discussion will present alternative design
standards to reduce imperviousness and demonstrate how
many of the impediments to narrow streets are already
being overcome with careful site design.
Perceptions and Realities: Parking Demand
Why are residential streets wide? Parking is a major fac-
tor. On-street parking on both sides of the street can in-
crease site imperviousness by approximately 25% (Sykes,
1989). Limiting parking to one side of the street or the use
of queuing lanes can significantly reduce this impervious-
ness. The reduction of on-street parking is often cited as
an impediment to narrow streets. This impediment can be
overcome. In Portland, Oregon, parking is accommodated
through the use of "queuing streets" which are 20' or 26'
wide (Figure 1).
Perceptions and Realities: Safety
The potential for increased vehicle-pedestrian accidents
is an often cited reason for prohibiting narrow streets. Many
studies, however, indicate that narrow streets may actu-
ally be safer than wider streets. The Federal Highway Ad-
ministration (1996) noted that narrow widths tend to re-
duce the speed at which drivers travel, providing greater
driver reaction time. Further, in a study of over 5000 pe-
destrian and bicycle crashes, a narrow road was a factor
in only two cases (FHWA, 1996). Unsafe driving speed,
on the other hand, contributed to 225 accidents.
Case Study: Longmont, Colorado
The City of Longmont, Colorado, is experiencing rapid
growth. The quality and type of new development has be-
come an important issue as more development and non-
conventional site designs are proposed. Part of this dis-
cussion involves acceptable residential street design. Swift
and Associates (1998) examined over 20,000 police re-
ports to determine the relationship between street design
123
-------
26' Roadway (2 Queuing Lanes)
< T > « 12> >< 7' >
Queuing
Lane
Moving
Lane
o
Queuing
Lane
20' Roadway (1 Queuing Lane)
« 13' » < 7' >
It
Moving
Lane
S3
D
I
Queuing
Lane
Figure 1. Queuing Lanes in Portland, Oregon.
and safety. The study focused specifically on residential
streets with maximum average daily traffic (ADTs) of 2,500.
Accidents attributable to poor road conditions or substance
abuse were excluded from the study. The results of the
Longmont study indicate that in general, narrow, curved
streets can be safely used in residential developments.
Specifically, streets between 22 to 30 feet in width were
found to be the safest (see Figure 2).
Perceptions and Realities: Adequate
Access
The conventional wisdom is that very wide streets are
needed to provide adequate access for emergency, ser-
vice and maintenance vehicles. But the facts do not sup-
port this concern:
• Trash trucks require only a 10.5' travel lane (Waste
Management of Montgomery Count, 1997), with a stan-
dard truck width of approximately 9' (BFI of Montgom-
ery County, 1997).
• Half-ton mail trucks, smaller than many privately owned
vehicles, are generally used in residential neighbor-
hoods. Hand delivery of mail is also an option (US Post
Office, 1997).
• School buses are typically nine feet wide from mirror
to mirror. Many jurisdictions require only a 12' driving
lane for bus access.
• Snowplows, mounted on pick-up trucks, with 8' width,
are common. Some companies manufacture alterna-
tive plows on small "Bobcat" type machines (Frink
America, Incorporated 1997).
• A number of local fire codes permit roadway widths as
narrow as 18' (Table 1).
Narrow Streets
Reduced pavement widths can significantly reduce the
impervious impact of residential developments. Site de-
signers should consult with public works, emergency ser-
vice, and residents to confirm that the community's needs
are met. Adequate access, parking, and safety can be
ensured through careful site design.
In addition to environmental benefits, significant construc-
tion cost savings can be achieved by building narrower
streets. Pavement construction costs are approximately
$15 per square yard. Suppose, for example, that a local
jurisdiction currently requires all residential streets with one
parking lane to be a minimum of 28 feet wide. The jurisdic-
tion then adopts a new standard: 18 feet wide queuing
streets. This new standard would reduce the overall im-
perviousness associated with a 300-foot road by 35% and
construction costs by $5,000. Additional economic ben-
efits include reduced clearing and grading costs and re-
duced long-term pavement maintenance costs.
Acceptance of the narrow streets design requires imple-
mentation as a flexible, locally adapted strategy. There-
fore, the Model Development Principles must be consis-
tent with the larger community goals (both economic and
environmental) that are put forth in comprehensive growth
management, resource protection, and watershed man-
agement plans. Finally, the Site Planning Roundtable en-
courages local, state, and federal agencies to provide the
technical support, financial incentive, and regulatory flex-
ibility needed to promote and implement the Model Devel-
opment Principles, and to fundamentally change the way
development takes place.
124
-------
12
ID-
4--
2--
I ' ' I •—'
I'M
20 22 24 30 32 34 36 38 40 42 44 46 48 50
Street Width (ft)
Figure 2. Relationship between street width and accidents in Longmont, Colorado, based on Swift and Associates (1998).
Table 1. Street Width Requirements for Fire Vehicles
Width
Source
Comments
18-20'
24' (on-street parking)
16' (no on-street parking)
18' minimum
24' (no parking)
30' (parking on one side)
36' (parking on both sides)
20'
US Fire Administration (Cochran,1997)
Baltimore (MD) County Fire
Virginia State Fire Marshal
Prince Georges County (MD) Department
of Environmental Resources
Prince Georges County (MD)
Fire Department
Represents typical "fire lane"
width
Road width
Road width
Road width
Road width
References
BFI of Montgomery County. 1997. Personal Communica-
tion. Rockville, MD.
Cochran, John L, 1997. Personal Communication, U.S.
Fire Administration, Emmitsburg, MD.
Federal Highway Administration. 1996. Pedestrian and Bi-
cycle Crash Types of the Early 1990's. U.S. Depart-
ment of Transportation, Federal Highway Administra-
tion, Washington, DC. FHWA-RD-95-163.
FrinkAmerica Incorporated. September 1997. Personal
Communication. Clayton, NY.
Swift and Associates. 1998. Residential Street Typology
and Injury Accident Frequency. Swift and Associ-
ates. Longmont, CO.
Sykes, Robert D., 1989. Protecting Water Quality in
Urban Areas: Best Management Practices for Min-
nesota. Minnestoa Pollution Contro Agency.
125
-------
Post Office, Fleet Maintenance Center. 1997. Personal Waste Management of Montgomery County. 1997. Per-
Communication. McLean, VA. sonal Communication. Rockville, MD.
126
-------
Retrofitting Conservation Designs into the Developed
Landscapes of Northeastern Illinois
Dennis W. Dreher
Northeastern Illinois Planning Commission
Chicago, Illinois
Abstract
There is a small but growing trend to retrofit developed
landscapes in northeastern Illinois with more environmen-
tally friendly designs. The motivation for these activities
varies from site to site. Commonly, those initiating retrofits
are hoping to reduce landscape maintenance costs, fix
erosion problems, improve water quality, enhance aesthetic
conditions, and/or reduce flooding.
Four types of retrofitting have been identified. Retrofit-
ting conventional turf landscapes with native prairie/wild-
flower vegetation is one of the more visible and exciting
trends. Notably, prairie landscaping has become a desir-
able option for several high-visibility corporate, school, and
government campuses. Stream channel retrofitting also
has become common. Many of the recent stream projects
have been stimulated by demonstration projects, funded
through Section 319 of the Clean Water Act, to reduce bank
erosion. Detention retrofitting has been done on a more
limited scale. Detention basin retrofitting has been initi-
ated to improve stormwater runoff control, reduce shore-
line erosion, improve aesthetics, and limit excessive
Canada goose populations. Conversion of storm sewers
to open drainageways has been performed on a very lim-
ited basis to improve treatment of runoff pollutants.
While water quality has not been the principal impetus
for many of the recent retrofit projects, the water quality
benefits can be substantial. As a consequence, watershed
managers are increasingly recommending large-scale ret-
rofitting to enhance the beneficial uses of urban and sub-
urban waterbodies and are selling the retrofit concept on
aesthetic and cost-saving grounds.
Motivation for Retrofitting
From a water resources perspective, there is a growing
realization that the developed landscape of northeastern
Illinois has some serious design and performance flaws.
One of the most obvious reminders of this fact is the fre-
quent and increasing incidence of damaging floods. Where
there was once considerable ignorance regarding the
causes of flooding, the average resident can now readily
relate local and regional flooding to new roads, parking
lots, and subdivisions and their contributions of increasing
stormwater runoff.
Similarly, but to a lesser degree, there is an increasing
awareness that conventional urban development designs
have led to impaired water quality and degraded recre-
ational uses of waterbodies. Contributing factors include
polluted stormwater runoff and the outright destruction of
wetlands and riparian corridors.
Increasingly, concerns also are being raised about the
sustainability of traditional development. In particular, in-
dividuals are questioning the costs of maintaining both the
structures and the landscapes that dominate the develop-
ments of the recent past. In particular, questions are being
raised about the continued reliance on turf grass as the
dominant landscaping material for commercial, office, and
residential developments. There are concerns about the
expense and environmental impacts of a maintenance
approach that relies on frequent mowing, irrigation, and
the extensive use of chemicals for fertilization and pest
control.
Fortunately, there is also a growing awareness and ap-
preciation of an alternative urban design ethic that incor-
porates "natural" elements into developed properties. This
alternative ethic is based on both ecologic and aesthetic
considerations. Central to this ethic is the belief that it is
both possible and desirable to commingle natural areas
and materials with developed landscapes. Elements of this
ethic include use of native plants for landscaping, preserv-
ing or restoring natural buffers at the edges of develop-
ments, and incorporating native ecosystems — prairies,
wetlands, and woodlands — into development site designs.
Cumulatively, this growing awareness of the shortcom-
ings of conventional design and the potential benefits of
naturalistic approaches provide the basis and motivation
for retrofitting elements of the developed landscape.
Types of Retrofitting
Retrofitting, as described in this paper, includes a range
of activities. In other contexts, some of these activities might
127
-------
be termed rehabilitation, restoration, or renovation. Regard-
less of terminology, retrofitting is assumed to involve a
substantive, long-term change to an existing facility or land-
scape resulting in demonstrable improvements to water
quality and aquatic ecosystems.
Fourtypes of retrofitting projects have been implemented
in northeastern Illinois and are documented in this paper.
They are:
• converting conventional turf landscapes to native veg-
etation,
• restoring eroding and/or channelized stream and riv-
ers,
• retrofitting stormwater detention basins, and
• converting, or "daylighting," storm sewers to open
drain age ways.
Converting Turf to Natural Landscapes
For decades, exotic turf grass has been the dominant
landscaping material for almost all new development in
the region. Considering that the pre-development land-
scape on most sites is cropland, and considering the rela-
tive paucity of remnant native landscapes such as prai-
ries, savannas, and woodlands, this landscaping philoso-
phy is not surprising. Recently, however, there has been
growing and enthusiastic support for natural landscaping,
an alternative approach that utilizes native plants that are
adapted to the local climate and soil (Northeastern Illinois
Planning Commission, 1997).
Natural landscaping applies to an array of landscaping
techniques that incorporate native vegetation, particularly
prairie, wetland, and woodland plants. Natural landscap-
ing also includes natural drainage techniques, such as
swales and vegetated filter strips, instead of storm sewers
and artificial drainage channels.
Benefits
The benefits of natural landscaping are the most broad-
ranging of any of the retrofitting techniques. In addition to
water quality benefits, they include flood reduction, habitat
enhancement, improved air quality (climatological benefits)
aesthetic enhancement, and cost-savings.
Water quality benefits are derived in two ways. First,
unlike conventional landscapes of turf grass and ornamen-
tal plants, native plants do not usually require chemical
additives after their initial establishment. Fewer applica-
tions mean greatly reduced runoff of fertilizers, pesticides,
and herbicides. Second, natural landscapes, particularly
with the deep root zones of many native plants, can effec-
tively soak up, filter, and transform contaminated
stormwater run off from roadways and parking lots, greatly
reducing the pollutant loads discharged to the "receiving
stream".
Flood reduction occurs due to the greater infiltration ca-
pacity provided by deep-rooted native plants. In contrast
to the four-to-six-inch root zones of turf grass, the dense
root systems of native prairie plants commonly extend sev-
eral feet into the soil, creating passageways for the rapid
infiltration of precipitation and runoff. The dense root sys-
tems also enhance evapotranspiration.
Habitat enhancements provided by the diversity of plants
found in a natural landscape, in contrast to the conven-
tional near-monotypic stand of turf grass. Native wildflow-
ers host numerous birds and insects whereas turf will not.
Natural landscapes are particularly valuable adjacent to
lakes and streams where they provide habitat for aquatic
insects and amphibians that spend time both in the water
and in terrestrial environments.
Improved air quality results, in part, from reduced usage
of lawn maintenance equipment that discharges hydrocar-
bons and nitrogen oxides into the environment. Native
plants also enhance air quality by filtering out particulates
and converting carbon dioxide to oxygen. In a related man-
ner, natural landscapes (particularly trees) provide clima-
tological benefits via shading and wind breaks, thereby
moderating temperatures and enhancing human comfort.
Natural landscaping also can provide aesthetic enhance-
ment. Natural landscapes provide a great variety of tex-
tures, colors, and shapes that vary seasonally. They also
attract a variety of wildlife, particularly birds and butter-
flies, enhancing their visual appeal.
The cost-savings in maintenance costs of natural land-
scapes can be dramatic. Natural landscapes need little of
no fertilizer or pesticide, as already noted. They do not
require regular irrigation, as does turf. They also need little
or no mowing. The preferred long-term maintenance ap-
proach for many naturally landscaped sites is prescribed
burning, performed every one-to-three years, much as it
was done by Native Americans.
Local Examples
Natural landscaping has been retrofitted onto numerous
sites throughout northeastern Illinois. Most of the retrofit-
ting has occurred on residential lots. In terms of commu-
nity impact, some of the most striking retrofits have been
on large office campuses and public properties.
Residential landscape conversions commonly involve the
replacement of turf grass or annual flowers with perennial
wildflowers and prairie grasses. Most conversions occur
gradually as residents discover and appreciate the advan-
tages of natural landscaping. In some cases, virtually the
entire lawn is converted, although in many cases signifi-
cant buffers of conventional landscaping are retained to
minimize potential conflicts with neighbors. As public edu-
cation and acceptance increase many communities that
formerly prohibited tall grasses and ungroomed land-
scapes, are now growing more flexible toward natural land-
scapes.
Commercial and office campus sites provide some of
the most impressive examples of natural landscape con-
128
-------
versions owing to their high visibility and large expanses
of land. Notable examples include the AT&T corporate cam-
puses in suburban Lisle and Naperville and the Prairie
Lakes commercial redevelopment in Homewood. Natural
landscape conversions on these sites have been motivated
by a combination of factors, including corporate image,
and the influence of employees.
Public properties, notably schools, parks, government
centers, and roadways, are increasingly popular targets
for landscape retrofitting. Schools provide logical retrofit
opportunities considering their typically large expanses of
high-maintenance turf and the potential educational op-
portunities. Wheaton Warrenville South High School, for
example, retrofitted 2.5 acres of turf into dry and wet prai-
rie vegetation. Students have been involved in planning,
planting, and management of the restored areas.
Restoring Stream and Rivers
The streams and rivers of northeastern Illinois reflect a
history of abuse and neglect. Over 40% of the stream miles
have been channelized or severely modified to provide
agricultural drainage or accommodate urban development.
Uncontrolled development of upstream watersheds has led
to severe flooding, streambank erosion, and water quality
degradation. Hence, there is a great need and opportunity
for retrofitting. In the context of this paper, retrofitting is
used to describe stabilization or limited rehabilitation of
the physical characteristics of the channel and its riparian
zone. Retrofitting includes stabilization of eroding banks,
enhancement of instream habitat, and restoration of the
near-stream riparian zone (Dreher 1998).
Benefits
Potential benefits include bank stabilization, improved
water quality, improved habitat, and enhanced aesthetics.
Bank stabilization can be achieved by a number of tech-
niques. The preferred approach would incorporate the use
of soil bioengineering techniques that are largely based
on natural materials and vegetation. Effective bank stabili-
zation reduces the loss of riparian land, protects stream-
side infrastructure (such as bridges and buildings), and
reduces sediment load.
Water quality improvement can be accomplished by a
number of retrofit techniques. One way to improve water
quality is to restore the natural pollutant filtering capability
of the riparian zone and floodplain. This is most readily
accomplished by re-planting streambanks and riparian
buffers with native vegetation, particularly indigenous wet-
land, prairie, and woodland plants that were common prior
to settlement. Another way to improve water quality is to
stabilize stream temperatures. Often, degraded streams
suffer from over-heated conditions during the summer due
to a loss of shading in combination with overly-wide, shal-
low channels. Establishing native vegetation, particularly
along sensitive headwater streams, can result in substan-
tial improvements.
Stream habitat is improved by restoring critical elements
such as meanders, pools, riffles, and natural substrate to
a degraded channel. Some habitat improvements can be
readily accomplished as part of other stream rehabilitation
projects. For example, the addition of rock substrate at
appropriate locations can accomplish both riffle enhance-
ment and stream bed stabilization. However, restoration
of meanders to a straightened stream channel typically
would be performed as an independent restoration project.
Aesthetic enhancement is accomplished in most resto-
ration projects, whether intended or not. Replanting of na-
tive vegetation to stabilize an eroded streambank, for ex-
ample, also results in a visual improvement. Intentional
enhancement of a degraded stream may be in the best
interest of some land developers, particularly of residen-
tial properties, to improve the marketability of a project.
For example, re-meandering and replanting a channelized
stream can convert an ugly ditch into an attractive stream
in the eyes of home buyers.
Local Examples
The Northeastern Illinois region has been the fortunate
recipient of funding to implement several stream restora-
tion demonstration projects. These projects, funded prin-
cipally through Section 319 of the Clean Water Act, have
provided highly visible models for others to emulate.
Two projects in which the Northeastern Illinois Planning
Commission has been involved are restorations of the
Skokie River and Flint Creek, located in suburban areas
north and northwest of Chicago, respectively (Price, 1997).
Both projects successfully demonstrated the use of "soil
bioengineering" techniques for streambank stabilization
and both restored significant areas of riparian buffer. The
projects also attempted to restore instream aquatic habi-
tat, although on a limited basis.
Several parks and golf courses have implemented
stream restorations to beautify their grounds and to re-
duce the loss of recreational lands to excessive streambank
erosion. Such restoration projects have utilized public edu-
cation to overcome the historical bias that favors mani-
cured landscapes over ungroomed "natural areas."
Several large residential developers have implemented,
or initiated planning for, significant stream restoration
projects. These have been done with the intention of en-
hancing the visual appeal of the developments and/or
accommodating development on sites constrained by flood-
plain locations. Whatever the motivation, aquatic habitat,
water quality, and hydrologic functions stand to benefit
substantially. A notable example is the Fox Mill develop-
ment in west suburban Kane County. This project resulted
in the re-meandering of the ditched headwaters of Mill
Creek and the conversion of a large riparian buffer to na-
tive wetland and prairie.
Citizen organizations have been active in stream resto-
ration. These include loosely organized volunteer groups
129
-------
as well as land trusts that actually own and manage land.
Restoration activities have ranged from streambank stabi-
lization projects supported by volunteers to more-exten-
sive restorations of riparian buffers and wetlands. Notable
examples include the restoration of a 1000-foot buffer along
Flint Creek by the Citizens for Conservation and the re-
building of 200 feet of river edge and restoration of a ripar-
ian buffer along the Middle Fork of the North Branch Chi-
cago River by the Lake Forest Open Lands Association
(Price, 1997).
Retrofitting Stormwater Detention Basins
Northeastern Illinois communities have required
stormwater detention for new development since 1970.
Currently, the vast majority of municipalities and counties
have detention ordinances. While local ordinances are
some of the most restrictive in the nation with respect to
flood prevention, most have not incorporated water quality
designs until the recent past. For example, many older
basins are simple dry bottom designs, often with paved
low-flow channels, that provide little pollutant removal ben-
efit. Thus, there are substantial opportunities to retrofit older
basins to enhance their effectiveness. Retrofitting projects
can range from simple repairs or alterations to major reha-
bilitation, depending on the project objectives and the ex-
isting conditions of the basin.
Benefits
Detention basin retrofitting can be targeted to a range of
objectives. These include improved pollutant removal, im-
proved flow control, reduced maintenance, and enhanced
aesthetics.
Improved pollutant removal can be achieved in most
older dry-bottom basins. One retrofitting technique is to
revegetate basin bottoms with wetland plants in place of
turf. If paved low-flow channels exist, they can be replaced
with vegetated swales. Pollutant removal also can be en-
hanced by excavating settling basins at the inlets and/or
outlet of the existing basin. Settling basins can greatly en-
hance the removal of suspended solids and attached pol-
lutants. Outlet structures also can be modified to increase
detention times for small-to-moderate-sized storms,
thereby enhancing pollutant settling. Finally, pollutant re-
moval can be enhanced in basins where inlets and outlets
are located in close proximity, causing short-circuiting. This
is accomplished by lengthening flow paths through the
construction of low berms.
Improved flow control can be readily achieved in some
basins by modifying the outlet structure. This is particu-
larly beneficial in older basins that were designed to con-
trol only the 100-year discharge. The outlets of such ba-
sins can be retrofit with restrictor plates or berms, or re-
placed with completely new structures, to provide control
of smaller storm flows, such as the 2-year event. Such
control is important in stabilizing downstream flows to re-
duce the potential for streambank erosion. However, it
should be recognized that increasing the control of smaller
storms will result in less storage availability for larger flood
events.
Reduced maintenance is an objective in many older
basins. One way to reduce maintenance is to replace turf
grass on basin bottoms and side slopes with low-mainte-
nance native vegetation. Depending on wetness conditions,
either wetland plants or prairie grasses and wildflowers
can be used. Native vegetation requires only occasional
mowing or prescribed burning. Maintenance needs can be
reduced in some basins by excavating settling basins at
basin inlets. Properly sized basins can concentrate the
settling of most particulate matter at the inlets, thereby fa-
cilitating long-term sediment removal from the basin.
Enhanced aesthetics is a concern in many older basins.
Problems range from eroding shorelines to excessive popu-
lations of Canada geese. Shoreline erosion in wet-bottom
basins can be controlled with the introduction of buffers of
water-tolerant native vegetation into shoreline zones.
Where erosion is severe, installation of soil bioengineer-
ing measures, as previously described for stream restora-
tion, can be effective. Introduction of shoreline buffers of
taller native plants also can be an effective control for
Canada geese. Indications are that geese are not com-
fortable moving through tall vegetation and are, therefore,
more likely to seek conventionally landscaped basins. They
also prefer short turf grass as a food source.
Local Examples
In contrast to natural landscape conversions and stream
restorations, there has not been a widespread retrofitting
of detention basins in northeastern Illinois. Perhaps the
most likely explanation is that detention basin owners gen-
erally are unaware of the performance deficiencies of older
basins, particularly their inability to effectively remove
stormwater pollutants. Without regulatory incentives for
such retrofitting, little has occurred. Detention basins own-
ers are more responsive to maintenance and aesthetic
concerns. Consequently, older detention basins are being
retrofit with native vegetation and shoreline stabilization
measures.
The most notable detention retrofitting project in the re-
gion is a demonstration project funded in part by the U.S.
EPA through Section 319 of the Clean Water Act. This
project involved an older dry-bottom basin in the Village of
Flossmoor, approximately 30 miles south of Chicago (Price
and Dreher, 1995). The basin had a failed outlet structure,
due to sediment clogging, and a paved low-flow channel
between its principal inlet and the outlet. Retrofitting in-
volved the excavation of stilling basins at the two basin
inlets, excavation of a permanent pool at the outlet, instal-
lation of a new multi-orifice outlet structure, and revegeta-
tion of the basin bottom and side slopes with native wet-
land and prairie vegetation. The retrofit basin provides
substantially improved pollutant removal and improved
hydraulic control of small storms, and requires substan-
tially less maintenance than the former basin. Local resi-
dents have indicated their satisfaction with the aesthetics
of the retrofit basin, as well.
130
-------
Converting, or "Daylighting," Storm Sewers
The term "daylighting" refers to the elimination of a storm
sewer or culvert and its replacement with open channel
flow. On the principle that open drainage systems provide
certain natural and aesthetic functions not provided by ar-
tificial, underground systems.
Benefits
There are several potential benefits of converting closed
pipes to open channels. These include improved water
quality and hydrologic functions as well as aesthetic ben-
efits. These benefits are optimized when the open chan-
nel is designed as a natural, unlined swale or stream.
Several water quality benefits are likely to result from
storm sewer daylighting. By running stormwater through
an open, vegetated channel, runoff pollutants can be fil-
tered and transformed by a combination of physical and
biological processes. These processes, similar to those
occurring in natural swale and stream systems, are con-
strained in closed pipes by inadequate light and the ab-
sence of natural substrates.
Improved hydrologic functions also are likely to result
from sewerdaylighting. For example, flow in an open swale
will have some opportunity for infiltration, thereby enhanc-
ing natural recharge and baseflow. Natural open channels
also can better dissipate flow velocities, potentially reduc-
ing downstream flooding and channel erosion.
Aesthetic benefits are most readily appreciated at the
point where a storm sewer or culvert discharges to a re-
ceiving stream or lake. Eliminating storm sewer bulkheads,
in particular, is likely to enhance the visual appeal of a
bank or shoreline, and enhance recreation in the waterbody
(Dreherand Price, 1997).
Local Examples
There are few reported examples of storm sewer
daylighting in northeastern Illinois, although there is increas-
ing interest in the concept among watershed managers.
One documented project is the conversion of several hun-
dred feet of large diameter storm sewer serving downtown
Barrington, northwest of Chicago. As part of a redevelop-
ment project, the Village removed the storm sewer and
replaced it with a meandering wetland swale. The objec-
tives of the project were to enhance the quality of the dis-
charge into nearby Flint Creek and to improve the appear-
ance of the property from an adjacent park and a planned
trail (Price, 1997).
Several area watershed groups are currently discussing
opportunities for daylighting. In particular, the Friends of
the Chicago River, as part of a comprehensive watershed
management project, have identified storm sewer
daylighting as a remedial best management practice
(BMP). It is notable that in some areas of the North Branch
Chicago River watershed, nearly all of the historical sur-
face drainage system has been replaced by storm sew-
ers. In this context, storm sewer retrofitting will not only
benefit the river but also may educate local residents to
the advantages of a natural drainage system looks like.
References
Dreher, D.W. and T. Price. Reducing the Impacts of Urban
Runoff: The Advantages of Alternative Site Design
Approaches. Northeastern Illinois Planning Commis-
sion, April 1997.
Dreher, D. and L. Heringa. Restoring and Managing Stream
Greenways: A Landowner's Handbook. Northeastern
Illinois Planning Commission, April 1998.
Northeastern Illinois Planning Commission. Natrual Land-
scaping Sourcebook. May 1997.
Northeastern Illinois Planning Commission. Stormwater
Detention Basin Retrofitting: Techniques to Improve
Stormwater Pollutant Removal and Runoff Rate Con-
trol, 1995.
Price, T. and D. Dreher. Flossmoor Stormwater Detention
Basin Retrofit Report: A Demonstration of Detention
Modifications to Improve Nonpoint Source Pollution
Control. Northeastern Illinois Planning Commission,
1995.
Price, T. Flint Creek Watershed Restoration Projects Re-
port with contributions from the Villages of Barrington
and Lake Zurich, Lake County Forest Preserve Dis-
trict, Citizens for Conservation, and Natural Areas Eco-
system Management. Northeastern Illinois Planning
Commission, September 1997.
131
-------
Impacts of On-site Sewage Systems and Illicit
Discharges on the Rouge River
Barry Johnson, P.E., M.S.
Camp Dresser & McKee
Detroit, Michigan
Dean Tuomari
Wayne County Department of Environment
Detroit, Michigan
Raj Sinha
Wayne County Department of Health
Detroit, Michigan
The focus of the Rouge River National Wet Weather
Demonstration Project (Rouge Project) is to clean up the
Rouge River in southeast Michigan. The Rouge River wa-
tershed includes portions of the City of Detroit and 47 com-
munities west and northwest of Detroit. Water quality sam-
pling and models show that eliminating combined sewer
overflows (CSOs) alone will not ensure that water quality
standards are met or that the river can be used for all pur-
poses the public desires. The information presented in this
paper shows that non-stormwater sources, such as on-
site sewage systems and illicit discharges, are major con-
tributors to the contamination of the river.
On-site Sewage Disposal Systems
On-site sewage disposal systems (OSDS) exist in ur-
ban areas of the Rouge River Watershed and are contrib-
uting to surface and groundwater pollution. In Wayne,
Oakland, and Washtenaw Counties, OSDS requirements
exist only for the installation of such systems; operation
and maintenance are the responsibility of the owners.
Rouge River OSDS failure rates documented in surveys
conducted in 1994, 1995, and 1997, varied between 17
and 55%. The 1995 and 1997 studies evaluated 528 resi-
dential OSDS for failures, which were identified by the fol-
lowing:
• observation of sewage discharging from the area of
the OSDS
• observation of liquid on the ground surface of the dis-
posal field
• identification of a pipe draining sewage from the dis-
posal field area
• heavy vegetation on or near the OSDS
• detection of dye in surface water downstream from the
septic tank after dye was placed in the septic tank.
Some surface waters in sewered communities have been
found to be unsafe for human contact due to high levels of
E. co//bacteria. Sources of E. co//here include CSOs, sani-
tary sewer overflows and leaks, illicit connections, wildlife
excrement, and failing OSDS. Illicit dumping of septic
wastes from recreational vehicles may also contribute to
occasionally high E. co//counts. Some surface waters that
drain areas which are not served by sanitary sewers have
also been found to be unsafe for human contact due to
high E. coli bacteria counts. These unsewered areas are
served by OSDS. Other potential sources of E. coli bacte-
ria here are illicit discharges through pipes that drain to
surface water, wildlife excrement, and agricultural opera-
tions.
In order to perform the OSDS surveys, it was necessary
to identify the locations of systems installed in the study
areas. While local health departments issue permits for
OSDS installations, three of the four health departments
in the Rouge River Watershed have not entered permit
information into computerized databases. Local commu-
nities did not have records of OSDS. Each of the surveys
required the development of a database of OSDS permit
information to help locate systems in the field. Census data
from 1990, which included information about the numbers
of OSDS by city block, were used to identify areas served
by OSDS. Information available from water utility billings
from each community was also included in the OSDS da-
tabases, along with the results from the field surveys. Re-
sults from compiling the databases were surprising to lo-
cal governments. These included the following:
• The 1990 census data showed that there are more
than 1,700 OSDS within Detroit city limits. City offi-
cials were surprised at these figures, since on-site
sewage disposal systems are illegal in Detroit.
• Further checking indicated that there are areas in the
City not served by sewers. Although a City policy ex-
132
-------
ists requiring payment for sewer connections, no fol-
low-up has occurred to verify that connections have
taken place.
Utility billings were screened for three communities in
the study area to identify households that were paying
for water but not for sewer service. Homeowners who
were billed for sewer service (e.g., were connected to
sewers) were not surveyed. Results through Septem-
ber 1997, revealed the following:
Total Number
of Homes
Community Contacted
Homes with Homes with
City Sewer Total Homes OSDS that
Connection with OSDS are Failing
A
B
C
152
239
53
18
22
15
134
217
38
28
45
6
• Of the 444 homes contacted, 55 reported that they
are connected to the city sewer but are not paying
sewer charges.
• Of the 389 homes surveyed with OSDS, 79 had failing
systems (20% failure rate).
• None of the communities were aware that homes were
connected to sewer systems but not being billed for
service. These communities are losing revenue by not
recovering costs of operation and maintenance of the
sewer system from these customers.
Wayne County Study
The Rouge Project Office funded a grant to Wayne
County to conduct visual surveys of OSDS for homes lo-
cated along a Rouge River tributary that drains into an
area being considered for canoeing. Because of high E.
coll bacteria counts, canoeing on the river has been dis-
couraged. Through October 31,1997, the County had con-
ducted surveys for 427 homes to identify signs of OSDS
problems. Of these, 90 systems have been described as
failing or potentially failing—a failure rate of 21%. Typical
descriptions from the field notes were as follows:
• Sewage backup in the home.
• Gray water discharging to the ground surface.
• Standing water on top of the gravel seepage field.
• Mushy area, associated with the back end of an ap-
parent seepage field.
• Illicit connection and undersized septic tank (100 gal-
lons) drained by a trench type (long single perforated
pipe) seepage field.
• Black sludge residue and toilet paper debris around
surface of the septic tank covering.
• Growth of cattails, wet marsh on the face of a down-
ward sloping hill.
Oakland County Study
Another study, which took place in the Oakland County
portion of the Rouge River Watershed, included dye test-
ing septic tanks and stream sampling for fecal coliform, E.
coll bacteria, and benthic macroinvertebrates. Study re-
sults are as follows:
• Of 49 surface water sampling sites, 43% had a daily
geometric mean for E coll bacteria of 1,000 or more
per 100 milliliter of sample.
• The macroinvertebrate study was done to determine
the water quality of streams in the survey area. A scale
was developed to rate macroinvertebrate and water
quality. The results in the study area ranged from 7,
which indicates poor water quality, to 20, which is con-
sidered good water quality.
• Dye testing conducted in 1994 showed that 53% of
the homes tested had discharges to the river.
• An optical brightener test to detect laundry waste was
conducted at the river sites where dye was collected.
These were all negative.
•Dye testing conducted in 1995 showed a 39% failure
rate for OSDS in the communities surveyed.
Future Direction
The future direction of this effort is to establish, in coop-
eration with local health departments, an on-site sewage
management program in each community. Communities
are also encouraged to address on-site sewage systems
in applications for general stormwater permits issued by
the State of Michigan under the National Pollutant Dis-
charge Elimination System (NPDES) Program. Septage
disposal problems are being addressed with septage haul-
ers and disposal facilities.
Costs
Grant expenditures for the Wayne County and Oakland
County surveys were $105,000 and $61,000, respectively.
This includes amounts spent by agencies to administer
the grants, conduct the investigations, and complete nec-
essary reports and other documentation. Additional costs
were realized by communities required to extend sewers
to problem areas and homeowners who were required to
correct failing OSDS or connect to available sewers.
Illicit Connections
From 1987 through 1996, Wayne County investigated
approximately 3,340 businesses and industries for illicit
connections to the storm sewer system. Approximately 9%
of the facilities inspected were found to have illicit connec-
tions. The elimination of these improper discharges has
diverted raw sewage and other pollutants from the river to
the wastewater treatment plant. Findings of the investiga-
tion are as follows:
133
-------
• An average of 2.6 improper connections were found
at businesses that had illicit connections.
• The majority of illicit connections in non-residential fa-
cilities were drains connected to storm sewers. These
included floor drains, trench drains, interior catch ba-
sins, oil separators, machine process waterdrains, and
sump pumps. The categories of illicit connections found
were floor drains (46%), sinks (20%), washing ma-
chines (15%), toilets (11%), and a variety of others
(8%).
• A method to prioritize the investigation was developed
based on the Standard Industrial Classification (SIC)
of businesses. The prioritization method was success-
ful for locating illicit connections. It was not helpful in
locating illicit discharges of E coli.
• The use of aerial, infrared, and thermal photography
to locate discharges that have a higher temperature
than that of the stream, or locations where algae might
be concentrated, is in the experimental phase. The
aerial infrared experiment also examines soil tempera-
tures, land surface moisture, and vegetative growth.
Assumptions are that (1) a failing OSDS will have in-
creased moisture in the surface soil, (2) the area will
be warmer, and (3) vegetation will grow faster than
the surrounding area. These differences should be vis-
ible in the digital data. Analysis of data collected has
been hampered by a lack of resources to conduct field
work needed to develop computer references.
•To date, there have been no definite correlations
among field tests for ammonia, anionic surfactants
(detergents), and E. coli.
•Field crews and members of the public have identified
a significant number of improper discharges to the river
through visual observations.
•Stable isotopes of oxygen and hydrogen have been
used to determine the presence of sanitary sewer water
in discharges.
• Visual observations and liquid flow testing indicate that
160 manholes and outfalls have suspicious discharges.
• Based on these findings, the estimated number of
potential illicit violations in the entire Rouge River
Watershed is 5,260.
It is estimated that 51 million gallons of liquid will be dis-
charged from illicit connections within the Rouge River
Watershed. Field work performed during dry weather (72
hours without precipitation) identified 160 manholes and
outfalls that had ammonia readings of 1.0 or greater, or
had visible conditions that were cause for further investi-
gation. All of these manholes were investigated for ammo-
nia, anionic surfactants, and E. coli. The bacteria results
showed that 16 locations had E co//bacteria counts greater
than 5,000 per 100 ml. These locations will have a de-
tailed investigation to locate the source. Many of the areas
with suspicious discharges are residential areas. Munici-
palities will be requested to participate in finding improper
connections.
Work performed in 1997-98 will focus on locating sources
of E co//that are impacting the Rouge River from Nankin
Dam to Merriman Road, a distance of approximately 1.5
miles. This is a prime area for recreational activity, which
would be significantly increased if the river was safer for
human contact. The work will begin at the 16 manholes/
outfalls with high E coli bacteria counts mentioned above.
Manholes located upstream of the sampling sites will first
be tested. Each highly suspect source will be dye tested
to confirm whether or not it is contributing to high E coli
counts. Sampling ofTonquish Creek and its tributaries that
drain into the proposed canoeing site will be conducted
during dry weather to determine if the same process will
be needed in this area.
An example of going upstream from a trouble spot to
locate the source of pollution is the investigation of a storm
sewer relief drain. An 11-foot storm sewer had been under
suspicion for several years. In 1997, samples were col-
lected from manholes located upstream from the discharge
point. Samples collected in June from one of the laterals
connected to the sewer had E coli counts of 8,160 and
9,600 per 100 ml. Additional samples were taken five days
later. Levels of E. co//from samples taken progressively
upstream in the storm sewer where the 9,600 per 100 ml
count was found were 12,560; 24,000; 160,000; and 9,600
per 100 ml. A lateral of this sewer had a result of 4,800 E.
coli per 100 ml. The manhole with the 160,000 per 100 ml
count was found to be the "hot spot."
The results of this sampling activity were shared with
city officials who decided to have the sewer televised.
However, the tapes did not show any suspicious connec-
tions. Plans were then made to begin dye testing at homes
located next to the storm sewer. Before beginning the pro-
cess of dye testing, another sample was taken at the trouble
spot to have current information. The results of that sample
indicated less than 8 E. coli bacteria per 100 ml. Following
discussion with the city, it was agreed that dye testing would
be postponed. It was also agreed that residents would be
informed of the sampling activities that had taken place on
their street. At this point, it was felt that the high E. coli
count from the initial sample may have been due to an
incident of someone dumping wastes directly into the
sewer. A letter was sent to residents asking them to let the
city or county know if they had knowledge of any practices
that could have resulted in the high E. coli counts. As a
result of these sampling activities, this investigation, and
community interest, sampling continues on a monthly ba-
sis on this street.
Future areas to be checked will be identified based on
citizen complaints, a review of manhole and outfall sam-
pling to determine contributing conveyances, and instream/
insewer sampling to localize the area. Using Rouge Project
CIS, maps have been prepared for tracking the sampling
of manholes and outfalls. These maps and the sampling
134
-------
data help municipalities identify and prioritize areas that
need to be further investigated.
Future Direction
The future direction of illicit connections/discharges is to
have each community in the Rouge River Watershed com-
mit to actively exploring illicit connections/discharges.
Grants and assistance from county agencies are available
to communities and agencies. As part of an application for
a General Stormwater Permit from the State of Michigan
under the NPDES program, a community is required to
develop an Illicit Discharge Elimination Plan. The Rouge
Project assists communities in preparing these applica-
tions. Elements of the Illicit Discharge Elimination Plan
recommended to be included are (1) a legal basis for the
program, (2) how problem areas will be identified, (3) how
the sources will be pinpointed, and (4) how to achieve cor-
rection, evaluation, and reporting.
Costs
The budget for the Illicit Detection Investigations Pro-
gram in 1996-97 was $735,000. The budget for the 1997-
98 program is $599,000. Besides field work, this budget
includes trials of different methods of investigation, test-
ing, and subcontracting for special studies. There is also a
significant cost for grant administration. Not included are
costs likely to be incurred by businesses to correct illicit
connections or the cost to communities of televising sew-
ers to locate illicit connections. The 1997-98 program pro-
vides nine full-time-equivalent employees. Of these, six
perform investigations and water sampling.
135
-------
Stormwater Management in an Environmentally-Sensitive
Urban Bushland in Sydney, Australia
Dr Stephen Lees
Executive Officer, Upper Parramatta River Catchment Trust
Sydney, New South Wales, Australia
Introduction
This paper outlines and discusses the challenges faced,
and the techniques employed, by the Upper Parramatta
River Catchment Trust in developing and executing an in-
tegrated stormwater management strategy in an environ-
mentally sensitive bushland reserve in the middle of the
Sydney metropolitan area, in New South Wales, Australia.
The key element of the strategy was a 30-metre high
concrete wall and associated structures, which form a large
flood detention basin in the bushland reserve. Completed
in mid 1996, the flood basin protects over 300 properties
in several residential and commercial areas from major
floods.
As part of the overall project, complementary measures
such as bushland regeneration, creek bank stabilisation,
waterquality monitoring and sediment and litter traps we re
implemented in the reserve to overcome serious existing
degradation of the bushland and protect it from further
degradation.
The adopted strategy was considered to be the best
possible compromise between the need to protect the
bushland environment and the need to protect homes, of-
fices and factories from flooding. This paper briefly out-
lines the history of the project and some of its more note-
worthy features.
Background
The Upper Parramatta River watershed or catchment
forms the headwaters of Sydney Harbour in the city of
Sydney, in the State of New South Wales, Australia. As
shown in Figure 1, the catchment is located in the centre
of the Sydney metropolitan area, between 20 and 30
kilometres west of the Sydney central business district.
The outlet of the upper catchment is at a weir separating
the freshwater and estuarine sections of the river. This is
located just downstream of the Parramatta central busi-
ness district, the main commercial centre for western
Sydney.
The area of the catchment is 110 square kilometres and
has a population of 230,000. Most of the catchment's
70,000 properties are single detached residences, although
there are extensive commercial and industrial areas and
an increasing number of multiple occupancy dwellings.
Much of the catchment is urbanised, although there are
significant areas of remnant vegetation in bushland re-
serves and urban forests, most located along the creeks.
Figure 2 shows that the two main tributaries of the
Parramatta River are Toongabbie Creek which drains the
west and south of the catchment and Darling Mills Creek
which drains the northeast.
The upper Parramatta River catchment includes portions
of the areas of four local authorities (called local councils
in Australia). Baulkham Hills, Blacktown and Holroyd cover
the upslope areas in the catchment's north, west and south
respectively, whilst Parramatta covers the catchment floor.
Although the catchment has experienced flooding since
the earliest days of European settlement from 1788, the
problem was compounded by rapid urban development of
the catchment in the 1960s and 1970s. At that time the
hydrologic impacts of urbanisation were not appreciated
and there was a lack of cooperation among the local coun-
cils. The growing flood threat only became apparent dur-
ing a series of storms in the late 1980s, which inundated
hundreds of properties many times. Detailed flood studies
showed that, in storms only marginally larger than those
experienced, substantial areas including much of the
Parramatta central business district would be flooded.
The historic Lennox Bridge overthe Parramatta River at
Parramatta increases the risk of serious flooding in the
Parramatta central business district. Constructed in 1837,
this sandstone arch bridge is the third oldest bridge in Aus-
tralia. Unfortunately, its arched waterway opening means
that, as the river level rises, less and less additional water-
way is available to pass flows.
Hydraulic studies showed that once the river level
reaches the top of the arch opening, as almost occurred in
the 1986 and 1988 floods, the river would break its banks
and quickly flood substantial areas of the central business
district. The flood risk, and the consequential development
136
-------
Wisemans Ferry^
Gosford
O
UPPER PARRAMATTA RIVER Castle Hill
CATCHMENT^.^Y/ .
Blacktown
7
(
Figure 1. Location of catchment in relation to Sydney
137
-------
••^•4—
V .—^.S^**-**-?- /Castle Hill I /\
.v^-^j \/'v J>K,
Kings Lesley ?o V f /*ULKMAM £ CHILLS ^/ y
N
Scale
0 1
Figure 2. Location of existing flood detention basins in catchment
restrictions imposed by the Parramatta Council, caused
the affected parts of the central business district to be-
come rundown.
A proposal in 1991 to reduce peak flood levels by de-
molishing Lennox Bridge was narrowly voted out by
Parramatta City Council. Soon after, the state government
put a 'permanent conservation order' on the bridge, effec-
tively ruling out that option.
Attention then turned to alternative solutions, particularly
in the Darling Mills Creek sub-catchment. Although this
sub-catchment occupies only one-third of the upper catch-
ment, it contributes half the storm flows because of its
higher rainfalls and steeper slopes. Moreover, as shown
in Figure 2 there were seven large flood detention basins
and several smaller basins in the Toongabbie Creek sub-
catchment; there were none along Darling Mills Creek.
For most of its length, Darling Mills Creek flows through
a heavily vegetated and steep-side sandstone valley up to
100 metres deep contained within a publicly owned
bushland reserve, Excelsior Reserve. Urban development
surrounds Excelsior Reserve on all sides. Despite this, the
dense vegetation and the deep valley, together with mas-
sive sandstone cliffs, create a sense of wilderness which
many local residents make use of forbushwalking and quiet
relaxation.
The extensive urban development surrounding Excel-
sior Reserve has also caused the creek's water quality to
deteriorate and the bushland either side of Darling Mills
Creek to become degraded. The reserve was infested with
weeds, there was severe creek bank erosion in places,
and considerable litter and nutrients were swept into the
reserve by stormwater. Despite voluntary bushland regen-
eration work by local residents, Baulkham Hills Shire Coun-
cil has been unable to allocate the funds needed to over-
come the degradation.
Investigations
Because the issues involved two local council areas
Baulkham Hills and Parramatta, investigations into a pos-
sible solution were sponsored and managed by the Upper
Parramatta River Catchment Trust, the catchment man-
agement authority for this area.
138
-------
The trust had been established in 1989 following repre-
sentations by local authorities and the public in the wake
of several major floods. It is a state agency funded by an
annual levy on all catchment properties. The trust's char-
ter is to coordinate flood mitigation, water quality and re-
lated catchment management activities in accordance with
the NSW Government's integrated catchment management
policy. It has a part-time board of twelve directors (nomi-
nated by the four local councils and relevant state govern-
ment agencies) and a permanent staff of seven.
The future of the trust is currently being considered in
the context of a planned catchment management body for
the entire Sydney Harbour catchment.
Initial studies by the trust showed that a large flood de-
tention basin in the bushland reserve on Darling Mills Creek
was the only viable way to protect from floods over 300
properties in the Parramatta central business district and
four other flood-liable areas along Darling Mills Creek and
the Parramatta River.
This immediately posed three significant challenges be-
cause it meant:
• building a large structure in an environmentally sensi-
tive bushland;
• storing floodwater temporarily in a heavily vegetated
reserve (until then detention basins had only been built
on grassed playing fields); and
• creating a large detention basin in the upstream local
council area (Baulkham Hills) to protect properties in
the downstream local council area (Parramatta) from
flooding.
Community Consultation — Support and
Opposition
From the outset, a community group formed to protect
the bushland reserve signalled its strong opposition to the
proposal for a large flood basin in the reserve. Despite its
strident opposition, the group participated in the workshops
and meetings held whilst the proposal was being formu-
lated and its impacts assessed. Other groups from flood-
liable areas further downstream were vocal in expressing
support for the flood basin.
A project steering committee was formed with represen-
tatives of the trust, both councils, government agencies
and groups supporting and opposed to the flood basin.
The committee met regularly during the course of the in-
vestigations. In addition, two community workshops were
conducted to obtain community feedback at critical stages.
Finally, a two-day Value Management Workshop to assess
the most favourable alternative was conducted by an in-
dependent facilitator. Progress reports were published in
the trust's quarterly newsletter delivered to all catchment
households.
Investigations and Design
Although not legally required, the trust decided to pre-
pare an environmental impact statement (EIS) to support
a formal application to Baulkham Hills Shire Council to
implement the stormwater strategy, including the large
detention basin. This ultimately involved some 40 sepa-
rate environmental, social, economic and technical stud-
ies over two years at a total cost approaching $500,000
(Australian dollars). The issues addressed ranged from Ab-
original archaeology, acoustics and air quality, to water
quality, weed control and zoology.
Of these issues, the most controversial was the likely
impact of the basin on vegetation in the impoundment area.
It was claimed that the raised flood levels inside the basin
would spread weeds to higher elevations in the bushland
reserve, whilst raised soil moisture levels would eventu-
ally kill off mature trees. This issue proved difficult to re-
solve because of the apparent absence of other detention
basins in bushland areas. In general, studies of other is-
sues found that impacts of the basin would be minor and
could be effectively mitigated.
A key part of any EIS is the assessment of all feasible
alternatives to the favoured option. Opponents of the large
basin sought to frustrate the EIS by repeatedly proposing
alternatives that, it was claimed, would avoid the need for
the basin. Each had to be examined carefully. In all, 20
alternatives or groups of alternatives were assessed, in-
cluding different large basins, groups of small basins, bridge
modifications, a flood tunnel by passing Parramatta, ac-
quisition of flood-liable properties and/or compensation.
The first stage of the evaluation confirmed that only a
large basin in Excelsior Reserve would protect all at-risk
communities at an affordable cost. The second stage evalu-
ation showed that the basin site near Loyalty Road, North
Rocks, was clearly superior on environmental, social and
financial criteria. This conclusion was confirmed at the
Value Management Workshop run by an independent fa-
cilitator.
Different types of basin walls (dams) were also exam-
ined. On technical, environmental and financial criteria, it
was found that a mass concrete wall constructed with roller-
compacted concrete (RCC) would be best. The concept
design was refined and detailed plans prepared. The de-
sign team included Ernest Schrader of the US, the world's
foremost expert in the roller-compacted concrete technique.
Mr. Schrader visited for a week during the design work,
and again during the construction.
The basin wall was to be 23 metres (at the spillway) to
30 metres (at the abutments) high, and 110 metres long at
its maximum height. It would comprise 23,000 cubic metres
of RCC, with pre-cast concrete panels on its external faces.
Its upstream face would be vertical. Its downstream face
would consist of a series of steps and have an overall slope
of 0.8 to 1. A central spillway, with walls either side, was
designed and model tested to safely pass even the largest
139
-------
possible flood. A 2.5 by 2.7-metre culvert allows passage
through the wall and contains a 1-metre square low-flow
channel.
Some key issues and concerns identified during prepa-
ration of the EIS are listed in Table 1, together with how
each was addressed.
The basin would have a maximum flood storage capac-
ity of 1.5 million cubic metres. Detailed hydrologic and hy-
draulic studies conducted by the trust showed that the flood
basin would reduce the peak flow in the critical 1 in 100
(1 %) annual exceedance probability (AEP) storm by 75%,
reducing the number of flood-liable properties further down-
stream from 313 to 75. However, all detained floodwater
would drain away within 3 to 6 hours of the rainfall easing;
and flows in the creek would be unaffected for more than
99% of the time.
Development Consent
In early September 1994 the trust submitted its devel-
opment application to Baulkham Hills Shire Council. This
was supported by an EIS comprising a 350-page main re-
Table 1. Darling Mills Creek Stormwater Management Strategy -Addressing Key Concerns
Issue or Concern How Addressed
Most flood mitigation benefits
in another local council area
(Parramatta).
Fear that temporary flooding of
bushland would kill or degrade
vegetation.
Visual intrusion of large man-
made structure into natural
setting.
Loss of bushland area due to
structure.
Raised water levels in basin will
spread weeds to higher levels
in valley.
Basin wall will block use of track
along bank of creek by hikers
and animals.
High-velocity water discharging
through culvert in basin wall (up
to 15 metres per second) will scour
downstream creek banks.
Construction truck movements
up and down narrow unsealed
track to site will cause unaccept-
able noise, dust and erosion.
Construction noise will disturb
residents living in nearby houses.
Need to construct basin wall
quickly to minimize risk of flooding
of construction works by rises in
creek.
Concern about safety of basin
wall in an extreme flood.
Significant areas of bushland
will be cleared to provide for
stockpiles, storage and batching.
Hikers may be trapped in basin
by quickly rising floodwaters.
Flood basin 'packaged' with various environmental measures to address serious degradation of the
Excelsior Reserve bushland and the creek.
Flood studies showed that all stored water will drain away within 3 to 6 hours of heavy rainfall easing.
Baseline surveys made of vegetation and creek channel against which future changes can be assessed.
Similar detention basin in Adelaide Hills of South Australia found to have caused no significant harm to
upstream bushland after 30 years.
Basin wall sited within a creek meander so that the wall is only visible within 50 metres upstream and
downstream.
Texture, width and colour of external panels designed to blend in with shadows from nearby tall trees.
Basin wall constructed using roller-compacted concrete to minimize its 'footprint' — only 2 hectares in 300-
hectare reserve.
All weeds removed from the basin area and up slope to prevent the spread of weeds. Ongoing bush
maintenance to control weed regrowth.
Culvert through the bottom of the wall allows normal creek flows to flow through, and enables hikers and
animals to pass from one side of the wall to the other.
Dissipater structure and stilling basin designed and model tested to control the high-velocity flow out of the
culvert under flood conditions and reduce its velocity before the floodwaters discharge into the downstream
creek.
A Rotec conveyor system was imported from the USA to deliver RCC and conventional concrete from a
temporary batching plant near the reserve edge to the basin wall. The conveyor zigzagged its way
between the trees, avoiding the need to remove any large trees.
Vehicle access to the construction site was through an adjoining industrial area. Construction was limited
to daylight hours five and half days per week. A 24-hour per day 'hot line' operated by a specialist
consultant received and dealt with all inquiries and complaints.
Precast concrete panels acted as formwork for placement of the RCC in layers and as the permanent
external facing of the basin wall.
Basin wall has a central spillway, with training walls on either side, capable of passing the probable
maximum flood. Steps on downstream face help dissipate energy of overtopping floodwaters. Design was
approved by state agency responsible for dam safety.
Areas able to be cleared strictly limited by contract with stiff penalties for non-compliance. RCC aggregate
blended off-site. Materials delivered to site only when required.
Studies showed rate of rise of floodwaters increased, but not unduly hazardous. Creek-bank walking track
upgraded. Bypass walking track constructed up and around basin wall. New footbridge built over creek.
Signs erected indicating egress routes.
140
-------
port and three volumes of specialist working papers, each
of about 300 pages.
Because of strong opposition to the project from the
small, but determined group, and the possibility that its
decision could be appealed in the courts, city council was
careful to allow everyone to have a say. The proposal was
publicly exhibited and comments on the proposal and the
EIS were invited. Letters of objection were received from
25 households and letters of support from 19. Council held
a public meeting to discuss the proposal. About 60 people
attended — equal numbers for and against the proposal.
Council also had a retired judge conduct a mediation con-
ference at which groups supporting and opposing the
project put their case; but no compromise could be found.
At about this time the trust was advised of a similar large
flood detention basin in the Adelaide Hills of South Austra-
lia, which temporarily inundated natural bush. An inspec-
tion by council members and senior staff showed healthy
mature trees growing in areas subject to regular inunda-
tion and no significant infestation by weeds.
Finally, after seven months of comment and delibera-
tions, the trust's application was approved unanimously
by the council, subject to over 100 conditions previously
agreed upon. Because government funding was required
to help finance the project, the EIS was also submitted to
the Commonwealth Government and, in due course, ap-
proved.
Construction
To avoid needless delay in commencing construction,
detailed design plans had been prepared whilst the EIS
was being finalised, before development consent was
granted.
Within a month of development consent, detailed de-
sign plans were completed and tenders for construction
called. Within three months, a $6 million contract to con-
struct the basins wall was awarded and a consultant ap-
pointed to supervise the construction. The other environ-
mental and structural measures, which formed part of the
overall Stormwater Management Strategy, were carried out
under separate contracts or by direct trust supervision of
contractors. Construction was subject to severe environ-
mental conditions reflecting the environmental sensitivity
of the site, all monitored closely by the relevant agencies.
The main difficulty experienced during the 12-month con-
struction period was frequent wet weather for four months.
This caused the exposed excavation to be flooded 20 times,
requiring exhaustive cleanup after each flood event.
Some of the more successful construction features were
the pre-blended aggregate, the RCC mix, the conveyor
system used to deliver RCC to the basin wall and the pre-
cast panels used both as formwork and the permanent
facing of the wall.
Upon completion of the basin wall in July 1996, all dis-
turbed areas were restored using previously salvaged
plants, mulch, topsoil and rocks.
Other measures carried out as part of the overall project
included:
• an alternative walking track around the basin wall site,
incorporating a new timber bridge over the creek, a
set of steps down a sandstone rock face and a formal
viewing area;
• detailed vegetation transect and creek cross section
surveys against which future changes can be as-
sessed;
• removal of all weeds from the 10-hectare basin im-
poundment area and regeneration with suitable na-
tive plants, then maintenance for at least five years;
• extensive creek bank stabilisation to allow four-wheel
drive vehicle access to the basin wall for maintenance;
• testing of water quality upstream and downstream of
the basin wall site before and during construction;
• construction of a CDS (continuous deflective separa-
tion) pollutant trap and sediment traps on gullies lead-
ing into the reserve; and
• survey, trial excavation and ongoing monitoring of sev-
eral rock shelters with Aboriginal archaeological po-
tential within the basin impoundment area.
Conclusions
The largest flood detention basin in New South Wales,
Australia, has been constructed in a degraded, environ-
mentally sensitive bushland reserve a few kilometres north
of the Parramatta central business district in western
Sydney. The project not only addressed flooding, but also
deteriorating water quality and bushland degradation. The
local council approved it after exhaustive studies showed
that a large basin was the only feasible way to cost-effec-
tively achieve the flood mitigation objectives and that any
adverse impacts on the bushland could be avoided or
minimised. The environmental assessment undertaken was
undoubtedly the most comprehensive ever carried out for
a stormwater project in New South Wales.
The project illustrates the changing nature of urban
stormwater control projects: the comprehensive investiga-
tion of all possible impacts, the detailed evaluation of al-
ternatives, the community involvement, its multi-objectives,
the strict environmental controls and the use of new con-
struction techniques to minimise environmental harm. The
increasing requirements for such projects mean that, in
future, their proponents will have to accept the consider-
able challenges involved, and allow the necessary time
frame and budget.
141
-------
Can a Steel Plant be Clean?
Nigel Ironside
Auckland Regional Council, Auckland, New Zealand
AlistairAtherton
Fletcher Challenge Steel Ltd, Auckland, New Zealand
Introduction
This paper discusses stormwater management, and in
particular stormwater quality control, from a heavy indus-
trial site located on the upper reaches of the Manukau
Harbour, in Auckland, New Zealand.
Founded in 1962, Pacific Steel Ltd., a business unit of
Fletcher Challenge Steel, is New Zealand's second larg-
est steel manufacturing plant, and its largest recycling op-
eration. The plant processes 200,000 tonnes of scrap steel
including some 60,000 car bodies annually to produce a
range of wire rod and reinforcing steel bar products. The
Pacific Steel site covers some 20 ha (45 acres) and in-
cludes both steel manufacturing and scrap metal recovery
operations. Prior to development, the site was under pas-
ture.
The site is located on the southern shores of the Mangere
Inlet, in the upper reaches of the Manukau Harbour (Fig-
ure 1). The outer Mangere Inlet is recognised as important
for marine vegetation, and as a high tide roost for thou-
sands of international migratory and New Zealand endemic
wading birds, including a number of threatened species.
Significant improvements in the quality of Pacific Steel's
stormwater discharge have been achieved overthe past 5
years following the construction of a stormwater pond/wet
Figure 1. Map location
142
-------
and treatment device, and the implementation of a com-
prehensive stormwater management plan for the site.
The stormwater treatment device is used as a demon-
stration site for stormwater treatment from a heavy indus-
trial area in the Auckland Region, and consequently has
been extensively monitored by both Pacific Steel Ltd. and
the Auckland Regional Council (ARC). This paper outlines
the results of this monitoring. The paper also highlights
how the site's stormwater management practices have
improved, particularly in response to the results of this
monitoring. Finally the paper discusses the practical ex-
periences gained in the operation of a stormwater treat-
ment device within a heavy industrial site.
New Zealand's Statutory Environmental
Framework
By way of background, New Zealand's environmental
statutory framework is set out in the Resource Manage-
ment Act 1991 (RMA).This is an omnibus piece of legisla-
tion having consolidated some 56 pieces of legislation re-
lating to the environment. The purpose of the RMA is to
"promote the sustainable management of natural and physi-
cal resources." The RMA also sets out the powers, duties,
and functions of the various authorities responsible for
implementation of the Act.
The ARC is the environmental protection agency for the
Auckland Region. With respect to stormwater manage-
ment, the ARC has responsibilities for minimising natural
hazards such as flooding, and for stormwater quality is-
sues.
The RMA is an effects-based piece of legislation, and
requires the effects, both positive and negative, of any pro-
posed activity to be identified prior to commencement of
that activity. Once identified, the RMA requires that any
adverse effects are, as far as possible, avoided, remedied,
or mitigated.
While focusing on the effects of a given activity, the RMA
also enables a "best practical option" (BPO) approach to
be taken to the discharge of contaminants to air, water, or
land, or an emission of noise. This BPO approach is de-
fined as the best method for preventing or minimising the
adverse effects on the environment having regard to,
amongst other things:
1. the nature of the discharge or emission and the sensi-
tivity of the receiving environment to adverse effects
2. the financial implications on people and society
3. the effect on the environment of that option when
compared to other options
4. the current state of technical knowledge and the like-
lihood that the option can be successfully applied
ARC Stormwater Management Program
The ARC initiated its stormwater quality control
programme in the late 1980s, in response to growing con-
cerns overthe impact contaminated urban runoffwas hav-
ing on both the freshwater and marine receiving environ-
ments.
Auckland is characterised by the fact that it sits astride
two major harbours. Its freshwater catchments are gener-
ally short and steep, with most discharging to low energy
estuarine and upper harbour areas. Any contaminants that
wash off the land, therefore, are rapidly deposited in these
low-energy marine environments, and accumulate with
time.
Given the variability of stormwater quality through time
and from different land uses, the ARC adopted a "best prac-
tical option" approach to its management and, in particu-
lar, the treatment of stormwater.
At an early stage in the development of the stormwater
quality control programme, the ARC prepared design guide-
lines for stormwater treatment devices (ARC 1992).
Through this design manual, guideline removal efficien-
cies for stormwater treatment of 75% for suspended sol-
ids are promulgated. The ARC also requires all new de-
velopment or redevelopment to address stormwater qual-
ity on a case by case basis (ARC 1995).
The ARC has established, in conjunction with a number
of interested parties, a range of representative stormwater
treatment demonstration sites, from which to monitor the
effectiveness of a range of treatment devices under
Auckland conditions. In addition to monitoring their effec-
tiveness, the ARC has used these devices to undertake
research to further characterise and quantify stormwater
related impacts in the Region. One such site has been the
Pacific Steel pond/wetland described in this paper.
Pacific Steel Ltd. - Stormwater Management
Historically, stormwater from the site was collected and
discharged, largely untreated, via three stormwater out-
lets (Figure 2).
Following detailed investigations in the late 1980s, a
stormwater treatment pond was commissioned in 1992,
designed to treat runoff from the entire site up to a 1-in-5-
year duration storm flow. The pond is located in the south-
western corner of the site, and discharges via a single
outfall to the "southern inlet," a small tidal arm of Mangere
Inlet (Figure 3). The retrofitting of the stormwater treat-
ment pond, described in more detail below, was accepted
at the time by the ARC as the best practical option for
stormwater quality control for the site.
In addition to the construction of the treatment pond,
Pacific Steel Ltd. has introduced a number of day-to-day
site management measures under the auspices of a
stormwater management plan. The plan is designed to
manage the site's stormwater system and minimise the
initial contamination at source. It establishes protocols and
sets frequencies fora range of site practices, such as regu-
lar storm drain inlet cleaning, street sweeping and dust
suppression, installation of waste handling facilities,
143
-------
Previous Pipe Network
Southern Outfall
Figure 2. Previous stormwater network.
New Pipe Network
Silt Pond and
Drainage from Yard Area
Figure 3. Current stormwater network.
wastewater audits, staff training, and emergency spill re-
sponse plans. The implementation of the stormwater treat-
ment and the management plan has dramatically increased
the awareness of workers on site to stormwater related
issues.
Sources of Contamination
The principal types of contamination of concern include:
sediment, oils and grease, and a variety of heavy metals.
Of the 20 ha site, some 45% is covered by paving, 20% by
roofs, and the remaining 35% by stone covered storage
yards.
Given the nature and activities on site, sources of po-
tential stormwater contamination are varied and include:
1. Runoff from roads and roofs
2. Steel scrap stockpiles
3. Steel making waste stockpiles
4. Car shredder waste
5. Steel manufacturing (dust and fumes)
6. Truck washes
Typical contaminant concentrations in the site's runoff
are shown in Table 1.
The variation in stormwater quality inflow to the pond,
as outlined in Table 1, stems largely from the site condi-
tions and the prevailing weather conditions. The 1990 re-
sults characterise stormwater runoff from the southwest-
ern part of the site, which did not include runoff from the
dirtier scrap recovery areas. Similarly the 1992 result
characterised stormwater from two-thirds of the site, again
without some of the scrap recovery area contribution. The
1994-1996 results represent the average stormwater qual-
ity as determined by monthly grab samples over that pe-
riod. This sampling would have encountered a range of
climatic conditions and is considered to represent the longer
term average inflow concentrations for the site. As with
the 1992 results, the 1997 results are the average inflow
concentrations from sampling of specific storm events.
The variability of contaminant inflows and the influence
of individual storm events is highlighted. A more detailed
analysis of the long term monthly data has suggested that
the inflow contaminant concentrations are reducing with
time (ARC unpublished data). This is principally attributed
to the improved site management practices carried out
under the Stormwater Management Plan. The performance
of the pond itself is discussed below.
Treatment Pond Design Characteristics
The treatment pond incorporates a permanent pond and
a constructed wetland, as well as an oil trap and an emer-
gency overflow (Figure 4). The pond was designed in ac-
cordance with standard stormwater management practices.
The pond is some 200m long and has an overall volume of
some 4,750 m3, with an additional live storage of approxi-
mately 4,200 m3. This volume is in excess of that required
to meet ARC'S guideline of 75% suspended solids removal
(i.e., approximately 4,450 m3). Total cost of the stormwater
pond and site upgrade was in the order of $NZ 1M in 1992
($US 0.6M).
The treatment device was initially commissioned in a
staged manner as the re-routing of the site's drainage sys-
tem took place. It was intended that only stormwater and
emergency cooling water overflows pass through the pond.
However, recent work related to PCB contamination in the
outflow of the pond has identified other waste streams,
such as truck wash effluent, which is also discharged to
the pond. The impact of these additional waste streams
on overall pond performance is currently being investigated.
Operational Experience With Treatment
Pond
Experience to date has shown the pond to be effective
at removing suspended sediments and other contaminants
in the site's runoff. The pond is trapping an average of 3
144
-------
Table 1. Typical Pacific Steel Ltd. Stormwater Contaminant Concen-
trations, as Measured at the Inlet to the Stormwater
Treatment Pond through time (g/m3).
Parameter
19901
19922
1994-19963 19974
Suspended solids
Total Oil and Grease
Copper (total)
Zinc (total)
Zinc (soluble)
19
11
0.018
0.18
0.09
101.1
-
0.14
1.6
0.2
77
35
0.14
2.94
0.37
210
-
0.48
7.4
0.23
1 Average Stormwater quality prior to treatment pond from "cleaner"
part of the site (Bioresearches, 1990)
2 Average inflow over six storm events to new pond prior to full
diversion of entire site flows (Leersnyder, 1993)
3 Average ARC monthly grab sampling 1994-1996 (ARC unpublished
data)
4 Average inflow concentration over four storm events in 1997 (NIWA,
1997)
Settling Pond
Most contaminants (silts &
metals) settle out in this
200m long, 15m wide and
1.5m deep pond. The
average time it takes
water to flow through the
pond and wetland system
is 9 days.
Access Berm
Inlet
Stormwater enters trap
where any oil and large
particles are removed.
Stilling Basin
Outlet Pipeline
Discharge to tidal inlet
that connects with
Harania Inlet and
Manukau Harbour.
Artificially
Created Wetland
Emergency Overflow
Figure 4. Pacific Steel Stormwater treatment pond
145
-------
tonnes per week of suspended solids which otherwise
would have been discharged to the wider receiving envi-
ronment. In addition, as much of the site is covered, the
site's storm drain inlets also trap significant quantities of
material which is removed on a regular basis. A more thor-
ough assessment of the pond's contaminant removal per-
formance is given below.
In the course of the 5 years of operation, a number of
operational issues and problems have become evident and
required addressing. Principal amongst these is the anaero-
bic conditions which can occur from time to time, turning
the pond completely black and foul smelling. It is unclear
where the high biological oxygen demand (BOD) water
suspected of causing this problem is sourced, although
recycled beer cans are suspected. An aeration device has
been installed at the inlet of the pond, to aerate the inflow,
particularly under low flow conditions in summer. This has
largely overcome the problem, although close monitoring
of the condition of the pond is required, as a "crash" can
occur relatively suddenly.
Water from immediately above the outflow is collected
and used on site for dust suppression during dry periods.
Further enhancements to improve sediment removal due
to the presence of elevated PCB concentrations, previ-
ously undetected and which have come to light due to tox-
icity screening of the effluent, have recently resulted in the
installation of two continuous cartridge filters at the outlet
of the pond. This is discussed in more detail below.
Stormwater Pond Monitoring
As indicated, the pond has been the subject of a range
of monitoring since its construction. The results of the spe-
cific monitoring investigations are outlined below.
Contaminant Removal Efficiency
Intensive performance monitoring of the treatment pond
was carried out initially, quantifying the mean concentra-
tion reduction and the event mean concentration for vari-
ous contaminants. This early monitoring estimated similar
removal efficiencies for each method (Leersnyder, 1993).
Therefore, long-term pond efficiency monitoring has utilised
mean concentration reductions on a monthly basis.
Table 2 compares the mean concentration reduction of
the pond during the initial two year commissioning period
(1992), the subsequent 1994-1996 period, and during re-
cent toxicity studies (1997) undertaken on the pond. The
1992 and 1997 studies sampled pond performance over
four individual storm events each.
It is worth noting the relatively high removal efficiencies
maintained since commissioning except for soluble Zn. This
is in spite of the progressive increase in contaminant loads
to the pond since the initial commissioning period. The
variability in pond performance is presumably related to
the influent variability found in Table 1. However, the long
term performance is considered to be acceptable from an
Table 2. Mean Concentration Reduction,'
Initial
Parameter
Suspended Solids
Copper (total)
Zinc (total)
Zinc (soluble)
Lead (total)
COD
Oil and Grease
Commisioning
Period
(1 992)
80
97
92
85
97
9
ND
Monthly
Sampling
(1994-1996)
73
79
98
78
87
31
9
Toxicity Testing
(1 997)
88
92
95
31
94
ND
ND
ND - Not Determined
engineering feasibility point of view. The toxicity of the in-
flow and outflows under these conditions and the impact
on the wider environment are discussed below.
Information relating to the relative contaminant removal
efficiencies of both the settlement pond and the constructed
wetland are illustrated in Table 3.
Wetland Sediment and Plant Tissue Quality
A range of monitoring investigations were conducted on
the wetland component of the treatment device some 3
years after commissioning. These investigations included
wetland conditions, the present concentrations of metals
and total petroleum hydrocarbons in the wetland sediment,
and the levels of metals in the dominant species of wet-
land plants (Bioresearches 1996). The results are dis-
cussed broadly below.
1. Wetland Condition
The wetland was commissioned with six species of plants
including:
• Juncus articulatus
• Baumea articulata
• Cotula coronopifolia
• Eleocharis sphacelata
• Schoenoplectus validus
• Baumea juncea
At the time of sampling (3-1/2 years after planting), open
water occupied approximately 50% of the total area of the
wetland compartment. The wetland has subsequently de-
veloped further and now covers almost the total area.
Of the six original plants, it appears the taller plants were
the most common, with the tall sedge B. articulata having
the greatest cover. The smaller sedge B.junceawas infre-
quent, while C. coronopifolia, which was expected to oc-
cur around the taller plants, was totally absent.
In general, the wetland was in a healthy condition and
the species diversity in the wetland had increased mark-
146
-------
Table 3. Percentage Contaminant Removal Efficiencies Showing the Relative Proportions Removed by Both the Settlement Pond and the
Constructed Wetland Components (1994-1996)
Reduction Inlet to Outlet
Proportion Removed by Pond
Proportion Removed by Wetland
Suspended
Solids
73%
93%
7%
Copper
79%
89%
11%
Zinc
(Total)
98%
82%
18%
Zinc
(Soluble)
78%
52%
48%
Lead
87%
87%
13%
COD
31%
94%
6%
edly with the introduction of opportunistic plants, since
commissioning. The wetland is also utilised by birds such
as pukeko (swamp hen), shags (cormorants), and ducks.
Maintenance was recommended, and subsequently un-
dertaken to remove small colonies of the invasive grass
species Glyceria maxima and raupo Typha orientalis for
fear of their rapid spread and potential to exclude other
rushes and sedges.
2. Sediment Grain Size
Generally the wetland substrate, which had accumulated
since commissioning, was dominated by sediments less
than .15 mm in diameter (i.e., very fine sands, silts, and
clays). For sediments at the inlet end, the proportion of
this finer material was 71.5% of the recently accumulated
sediments, while the outlet end sediments contained 79.4%
of the fine material. The greater difference was due to mark-
edly higher clay concentrations at the outlet end of the
wetland, reflecting the wetland's ability to remove some of
the smaller sized particles. Based on these findings, it was
calculated that the wetland was responsible for about 9%
of the overall sediment removal. This agrees well with the
estimates based on the long-term monthly sampling (Table
3).
3. Sediment Quality
The sediment quality information indicates that signifi-
cant quantities of contaminants are being retained in the
wetland, and that during the early life of the wetland, re-
moval at the inlet end would appear to be more rapid. Fur-
ther testing indicated that all the differences between wet-
land inlet and outlet sediment concentrations were statisti-
cally significant. The percentages of constituents at the
outlet end in comparison with those at the inlet end were:
cadmium 23.5%
copper 37.5%
lead 21.2%
zinc 21.9%
total petroleum hydrocarbons 24.5%
With further analysis of the data it became clear that
particular areas within the wetland contained higher levels
of contaminants than others. The mounded areas intro-
duced in the original design to "baffle" flows, appear to
have increased channelling and short circuiting through
the wetland. As a result of these findings, more attention is
being given to the placement of such structures in future
wetland designs.
Wetland sediment quality was compared to sediment
quality guidelines for the protection of freshwater aquatic
life (Persaud et al., 1992). The results indicate that 50% of
the readings were above the lowest effects level (LEL) but
below the severe effects level (SEL). While some 47.6%
of results were above SEL levels, with the greater
exceedance of SEL levels occurring at the inlet end
(Bioresearches 1996). At present, there is no indication of
abnormal plant growth or dieback which could be attrib-
uted to an increase in contaminant levels, and none would
be expected given the generally higher level of robustness
of these aquatic plants over other aquatic organisms and
the limited uptake of contaminants by the plants them-
selves, as described below.
Although it is not possible to relate sediment contami-
nant levels to precise effects on the dominant wetland
rushes and sedges, changes to the existing biological con-
dition are expected to occur in a progressive fashion
through time. A regular programme of monitoring (i.e. sedi-
ment levels, tissue levels, and plant vigour) is recom-
mended to quantify this change and provide information
on the optimum replacement intervals for the wetland.
4. Plant Tissue Quality
Plant tissue of the wetland plants were analysed for
heavy metal concentrations. Table 4 summarises the av-
erage levels for all species. The data indicate that the high-
est concentrations occurred in Juncus articulatus, and that
total metal loads are dominated by zinc concentrations,
which are an order of magnitude higherthan the other metal
levels measured.
Unfortunately, no baseline data on tissue concentrations
was collected prior to planting. However the plant tissue
concentrations were compared to "control" plants from the
broad geographical area of the original stock, and collected
from areas outside of major industrial or other
anthroprogenic influences. The assumption is that the "con-
trol" plants reflect broadly the metal concentrations of the
plants priorto planting. Given these assumptions, the plant
tissue concentrations outlined in Table 4 were found to
reflect the pattern of concentrations in the "control" plants
(Bioresearches 1996).
The highest average increase per species has occurred
in the tissues of Juncus articulatus and Baumea juncea
(13 times and 9 times, respectively) while the remainder
of the plant species have shown an increase of 4 to 5 times.
147
-------
Table 4. Average Metal Levels in Stormwater Wetland Plants
Plant
Species
J. articulatus
B. juncea
S. validus
E. sphacelata
B. articulata
Cadmium
0.22
<1 .033
<0.029
<0.015
<0.058
Chromium
4.03
1.53
0.73
0.28
0.43
Copper
38.3
2.7
4.6
4.2
2.3
Lead
31.0
3.6
0.71
0.85
1.32
Mercury
0.13
0.29
0.024
0.17
0.016
Nickel
4.9
1.1
1.1
1.1
1.0
Zinc
516.0
125.7
91.0
71.3
39.3
Average
O.071
1.4
10.4
7.5
0.13
1.8
168.7
With the inlet and outlet areas combined, the average
plant metal concentration is only 8% of that found in the
sediments; i.e., the accumulation of metals in the sediments
is not paralleled by a similar rate of uptake by the plants,
and the major mechanisms of metal removal in the wet-
land appear to be physico-chemical (flocculation, adsorp-
tion, settlement, filtration by plant stems) rather than bio-
logical.
Influent/Outflow Toxicity Testing
An extensive range of toxicity tests was conducted on
both the influent and outflow from the treatment pond (NIWA
1997). The study measured influent and outflow chemical
contaminant levels under storm conditions, and toxicity to
a variety of freshwater and marine species. In addition,
suspended particulate matter (SPM) in the outflow was
collected by means of a centrifuge and used to assess the
toxicity to sediment-dwelling marine species. Together,
these chemical and toxicological components allow a risk
assessment to be undertaken for potential ecological im-
pacts on the marine receiving water environment.
1. Water Column Toxicity
The results of the pond water toxicity showed relatively
high variability of the influent and outflow toxicity, although
there was an observed decline in toxicity across the pond,
with the Pacific Steel outflow generally classified as "slightly
toxic" to "moderately toxic." In addition, the marine spe-
cies appeared to be more sensitive than the freshwater,
with the greatest response exhibited by a marine diatom
and echinoderm.
Dilutions required to mitigate the outflow toxicity ranged
from 11 to 46-fold (based on algae and echinoderms, re-
spectively). These results suggest that of the test species
currently available the algae and echinoderms, are the
more sensitive species for monitoring stormwater impacts
on the marine environment of Auckland.
Comparisons between the measured water column tox-
icity and the expected toxicity based on chemical analysis
(totals) showed the measured toxicity to be much lower
than expected. Acute (short term) toxic effects would have
been expected for both the inlet and outflow on most oc-
casions, based on the guideline exceedances for a num-
ber of metals. Similarly, some toxicity was observed on
occasions when criteria were not exceeded, suggesting
that contaminants other than metals may be contributing
to the measured toxicity. Concentrations of ammonia con-
tributed a maximum of 42% to the observed outflow toxic-
ity, based on comparison with published chronic exposure
guidelines.
Total and soluble contaminant concentrations are com-
pared with acute and chronic freshwater and marine crite-
ria in Table 5. The acute and chronic metal criteria
exceedances over the four storm events monitored are
shown in Figures 5 and 6, respectively.
The low measured water column toxicity suggests that
the bioavailability of contaminants in both the total and
soluble phases is low. One theory is that the high dissolved
carbon values in the pond may have markedly reduced
the bioavailability of the metals and subsequent measured
toxicity on most occasions. Further work is programmed
to look at this.
2. Sediment Toxicity
The objective of the sediment toxicity tests was to as-
sess the toxicity of contaminants adsorbed onto the sus-
pended particulate material (SPM), which passes through
the pond, on benthic estuarine/marine organisms. Sedi-
ment samples were obtained from centrifuging the dis-
charge of the pond.
The first test scenario attempted to simulate a "worst
case" scenario appropriate to Auckland of 3 mm of annual
contaminated estuarine sediment deposition. The second
scenario involved the dilution of the SPM sediment with
clean estuarine mud to simulate a likely deposition sce-
nario, whereby discharged material is reworked with the
surface layer of the estuarine sediments. SPM dilutions of
5x, 15x, and SOxwere used.
The sediment toxicity under the "worst case" scenario
showed low acute survival for amphipods and juvenile shell-
fish. However, given the high particulate carbon and nitro-
gen levels in the SPM sediments, it is unclear whether the
observed toxicity was due to the contaminant levels or the
anoxic conditions which developed in the sediments caus-
ing them to turn black and smell strongly of hydrogen sul-
phide. Microscopic examination of the SPM sediments
showed high numbers of algal cells and protozoa present.
The acute survival for juvenile shellfish in the second
series of tests showed no adverse effects at all dilutions.
148
-------
N
ra
T3
LLJ
E
o
§ 5
LU
E
o
§ 5
^t
"c
to
>
LU
E
o
•E
LLJ
E
o
CO
•E
LLJ
E
o
00 t
00 O
8 8
O) CD
T- in
0 0
00 C\l
CO 00
T- 0
o o
™ 1
o <
0 <
<
O) r
0
o <
CO
•O O)
N 0)
D O
D 0
3> L,
- OM
J CO
D O
83
00
0
00
CO
o
00
8
o
0
in
CO
O)
0
0
O)
8
o
0
-
in
CO
o
8
o
0
CO
CO
!
o
8
o
0
CO
0
8
o
0
in
in
OM
00
0
o
,- CO
0 0
•si cd
N in
0 0
™
0
00
CO
0
o
0
in
00
in
0
o
V
R
0
CO
^
o
I °
°- CL
•E
LU
E
o
o ^ OM
ra
o
'c
o
^
o
T3
(5
•g-gg
to
o
'c
o
o '^ ro o o o
QJ O J- QJ QJ QJ
2 o>£ ^ ^ ^
ra
CD
o ra ra ra
,,, T3 T3 T3
¥ tu tu to
x x x
LU LU LU
149
-------
E
15>
10 1
1 -,
0.1 -,
S 0.01
c
-------
The amphipods, however, displayed threshold effects and
enabled extrapolation of an EC50 response level of 70%
SPM, and prediction of a threshold response (nominal
EC10 response) concentration of 10% SPM.
Contaminant levels in the SPM sediments exceeded
median effects sediment guideline levels (ER-Ms) for most
organic and inorganic contaminants in the pond (Table 6).
Interestingly, Pacific Steel SPM sediments were also rela-
tively enriched in some metals (Cd, Hg, Pb), chlordane,
and polyaromatic hydrocarbons (PAHs) when compared
with sediment samples taken from within the pond near
the outlet. This suggests that there may be higher levels of
some contaminants exported from the pond than would
be indicated by comparative size fractions in the sediments
near the outlet.
The relative toxicity of the various contaminants was
compared by dividing the measured SPM concentration
by the median effects guidelines (ER-M). The ratios indi-
cate that for Pacific Steel, PCBs followed by DDTs repre-
sent the greatest organic contaminant risk, with zinc and
lead the more potentially toxic metals. As a result of the
elevated PCB concentrations measured during the toxic-
ity investigations, considerable effort has been expended
by Pacific Steel to identify the sources of PCBs on site and
to prevent them from entering the stormwater system. In
addition to the control of PCB contamination throughout
the site, a pilot Mikroclean 1 micron nominal cartridge fil-
tration device has recently been installed on the outlet to
filter the discharge continuously. The performance of this
filtration device is expected to result in a reduction of aver-
age suspended solids and PCB concentrations in the dis-
charge of approximately 60%. These measures do not
cover the discharge from the emergency overflow, which
operates under larger, less frequent storm events.
The measured toxicity responses were in general agree-
ment with the expected effects given the contaminant lev-
els found in the sediments. Given this, the receiving water
sediment dilution required to prevent significant adverse
effects was estimated based on the chemical data and the
guideline exceedence, together with the toxicity results.
This suggests that a sediment dilution of up to 115x for
sediments leaving the Pacific Steel Pond would be required
to pevent significant adverse effects.
Biological Survey of the Receiving
Environment
Ecological surveys of the tidal inlet which acts as the
receiving environment for the stormwater treatment pond
discharge, were undertaken in 1991 and 1995. Samples
were collected for benthic invertebrate analysis and de-
scriptions made of the marine vegetation present, as well
as the use of the inlet by birds and fishes.
The southern inlet has a former Pacific Steel reclama-
tion on the northern side while the southern side contains
a recreational area, disused industrial land, and hospital
land at the southwestern corner. Midway along the recla-
mation boundary is an inlet which receives inflow from a
weired creek which drains a residential area. The inlet also
receives stormwater from a heavy industrial area
Prior to the construction of the pond, the southern inlet
acted as a natural stilling basin for both the Pacific Steel
site and the stormwater discharges from the heavy indus-
trial catchment above. The presence of saline conditions
are thought to have increased the sediment deposition,
and chemical reactions, probably with sea water, had re-
sulted in the formation of precipitates which also settled
out. (Bioresearches 1988). Sediment concentrations found
in the inlet were elevated in comparison with the Mangere
Inlet and the wider Manukau harbour. The inlet was
characterised in 1991 as "heavily polluted" (Roan 1991)
In comparison with the conditions found in 1991, by 1995
the volume of freshwater entering the inlet had greatly in-
creased. Contaminants had also declined due to Pacific
Steel treatment pond and the closure of much of the heavy
industry in the upper catchment. The total number of
macroinvertebrate taxa recorded in 1995 was 25, com-
pared to five in 1991. In 1991, the five species identified
were typical estuarine organisms; however, the majority of
those recorded in 1995 typically inhabit areas which are
predominantly freshwater habitats (Bioresearches 1996a).
By 1995, much of the bare mud had been colonised by
a variety of wetland/salt marsh plants, including Cotulaand
a variety of rushes. This recolonisation is reinstating a tran-
sitional zone more typical of inlets in other parts of the
Mangere Inlet. It is also notable that this process is occur-
ring despite the high levels of contamination in the under-
lying sediments (Bioresearches 1996a).
Following the 1995 survey it was concluded that "there
was no indication that the treated stormwater discharge
has had a detrimental effect on the Inlet's biological condi-
tion, but has probably led to a more rapid development of
a predominantly freshwater fauna upstream from the man-
grove zone" (Bioresearches 1996a).
Conclusion
The quality of the Pacific Steel stormwater discharge has
improved dramatically since the commissioning of the
stormwater treatment ponds in 1992. The performance of
the ponds themselves has matched and even exceeded
their predicted design efficiencies. The treatment system
has settled down and is operating as expected, although
day-to-day management of the entire stormwater system
is required to ensure peak performance. At present, there
appears to be no observable adverse effect on the wet-
land plants, and additional performance benefits have been
clearly demonstrated from the inclusion of a constructed
wetland in the overall stormwater treatment device.
The immediate receiving environment has also improved
considerably in the past five years, as a result of both the
improvement of stormwater discharges and land use
151
-------
Table 6. Comparison of Stormwater Sediment Contaminant Levels with Manukau Harbour Levels and Sediment Quality Guidelines (after NIWA
1997).
Stormwater comparison - total metals (mg/kgDW)
Site
Pacific -SPM
Pacific - pond
C
15.7
5.1
N
2.2
0.5
Zinc
7820
5635
Cadmium
16.7
8
Copper
653
1702
Mercury
1.54
0.14
Lead
1430
724
Cumulative ER-M
exceedance
Reference Sediment
Reglan
Harbour site Comparison0
Manukau - W
ER-Ld
ER-M
Scenario I (undiluted)
Pacific SPM/ER-M
Scenario II (diluted 5x)
Pacific SPM/ER-M
2.4
0.23
0.73
(26% mud)
89.8
86
150
410
19.0
3.8
0.11
1.2
9.6
1.7
0.34
21.8
19.4
34
270
2.4
0.48
0.093
0.15
0.71
2.2
0.44
22
6.1
46.7
218
6.6
1.3
32
6.4
Site
Pacific -SPMa
Pacific - pondb
C
15.7
5.1
N
2.2
0.5
Tot
PCB
10753
10500*
DDTs
44.2
<50
Chlordanes
55
<6
PAHs
5353
1900
Dieldrin
19.5
<40.1
Cumulative ER-M
exceedance
Reference Sediment
Reglan
Harbour site comparison0
Manukau - W
ER-Ld
ER-M
Scenario I (undiluted)
Pacific SPM/ER-M
Scenario II (diluted 5x)
2.4
0.23
0.73
(26% mud)
22.7
180
59.8
12
0.1
22
1.6
46
9.6
1.9
0.1
0.41
0.5
6
9.2
1.8
112
5377
1700
9600
0.6
0.12
0.1
0.41
0.02
8.1
2.4
0.48
82
16
a SPM = suspended particulate matter, sediment collected by centrifuge
b pond sediment data from the 1995 survey (Nipper etal., 1995)
0 Data from Holland etal., (1993)
d sediment guideline values from Long et al., 1995, Long & Morgan 1991 (for chlordane and dieldrin)
G. Mills, NIWA, personal communication
Key = bold = exceedance of ER-M values; italic = exceedance of ER-L values
changes which have occurred in the catchment above. The
nature of the receiving environment has changed signifi-
cantly to one dominated by freshwater inflows, primarily
as a result of the Pacific Steel treatment pond discharge.
The immediate receiving environment is not displaying any
significant adverse effects as a result of the treated
Stormwater discharges, although should be remembered
that the area had been classified as "heavily polluted" in
1992, and that the improvements made to date have to be
compared with that impacted status.
With regard to the long-term implications of this treat-
ment device, the results of the toxicity testing paint a some-
what more confusing picture. Measured water column tox-
icity of the outflow was found to be only slight to moderate,
even though the measured contaminant concentrations,
and in particular soluble zinc and copper, exceeded both
freshwater and marine acute US EPA criteria. The reason
for these findings is being investigated further but it is clear
that something is reducing the bioavailability of these con-
taminants. It is also uncertain if this effect on the
bioavailability of the contaminants is a long-term phenom-
ena, or whether the contaminant loads discharged from
the pond will become available in the future, and poten-
tially give rise to adverse effects in the wider receiving en-
vironment.
The measured acute toxicity of the particulate material
leaving the pond showed low survival rates for all test or-
ganisms in the experiments attempting to test the worst
152
-------
case scenario. However the anoxic conditions which de-
veloped during the experiment are likely to have influenced
the result. The acute toxicity shown by amphipods with
increasing pond SPM sediment concentration suggests that
the discharge from the pond may have a longer term im-
pact on the wider receiving environment as contaminant
concentration levels build up.
Interestingly, the finding of elevated PCBs in the dis-
charge has led to a range of measures being implemented
on site which should see the sediment and associated
contaminant levels being discharged decline by as much
as 60% in the future. Work is on-going to assess the effec-
tiveness of these measures in the longer term.
References
ARC 1992. Stormwater Treatment Devices Design Guide-
line Manual. Technical Publication No. 10 Auckland
Regional Council, Auckland
ARC 1995. Proposed Auckland Regional Policy Statement:
Incorporating decisions on Submissions Auckland
Regional Council, Auckland
Bioresearches 1988. Pacific Steel Reclamation Habitat As-
sessment 1988
Bioresearches 1990. Pacific Steel Ltd. Settlement Pond
Treatment Data 1990.
Bioresearches 1996. Pacific Steel Ltd. Stormwater Treat-
ment Wetland Sediment and Plant Tissue Quality
Bioresearches 1996a. Pacific Steel Ltd. Biological Survey
of Tidal Inlet
Holland, et al. 1993. Variability of Organic Sediments in
Inter-tidal Sandflat Sediments from the Manakau
Harbour, New Zealand Archives of Environmental Con-
tamination and Toxicology 25:456-463
Leersnyder 1993. The Performance of Wet Detention
Ponds forthe Removal of Urban Stormwater Contami-
nants in the Auckland (NZ) Region. Unpublished MSc
thesis, University of Auckland
Long, et al. 1991. The Potential for Biological Effects of
Sediment-sorbed Contaminants Tested in the National
Status and Trends Program, National Oceanic and
Atmospheric Administration, Seattle, WA. 175 + ap-
pendix
Long, et al. 1995. Incidence of Adverse Biological Effects
within Ranges of Chemical Concentrations in Marine
and Estuarine Sediments. Environmental Management
19:81-97
Mills, G. Personal Communication
Nipper, et al. 1995. Assessment of Water and Sediment
Toxicity Associated with Stormwater Treatment Facili-
ties. Report Prepared for Auckland Regional Council
Environment and Planning Division, ARC 242, 55 pp
NIWA 1997. Assessment of Water and Suspended Sedi-
ment Toxicity to Freshwater and Marine Species As-
sociated with Stormwater Treatment Facilities. NIWA
Client Report ARC 70219 July 1997
Roan 1991. Baseline Survey forthe Upper Southern Sec-
tor of the Harania Inlet, Manukau Harbour. forPacific
Steel Ltd 16pp
Persaud et al. 1992. "Guidelines forthe Protection and
Management of Sediment Quality in Ontario" 1992
Ontario Ministry for Environment 26 pp
153
-------
Real World Modelling
A Case Study of the Silver Lake Watershed Project
Randell K. Greer, RE.
Delaware Dept. of Natural Resources and Environmental Control
Dover, Delaware
Abstract
The past decade has brought an increasing realization
that our aquatic resources are better managed at the wa-
tershed level. This often requires the application of water-
shed modelling to provide managers with the information
they need fordecision making. Once, this required the hard-
ware and services of consultants specialized in the field,
along with the associated costs. However, advances in
technology have brought computers powerful enough to
run such models to the desktop of virtually anyone. Addi-
tionally, just as the private sector has gone through the
"downsizing era," government has also been pressured to
do more with less. These two factors have combined in
such a way that staff members from local governments
are now being asked to perform these studies. In a perfect
world, data would be readily available for such studies,
and the modelling would proceed just like the text book
examples. Unfortunately, this is rarely the case. Modelers
are often faced with data gaps and other problems which
may not even come to light until well into the modeling
process. This presentation will address these issues in the
context of a case study of a watershed management project
conducted in the Silver Lake watershed in central Dela-
ware. It is hoped that a review of the problems encoun-
tered and ultimate solutions to those problems will be help-
ful to other "part-time modellers" finding themselves in simi-
lar situations.
Background
The Dover Silver Lake (DSL) watershed is centrally lo-
cated in the state (Figure 1). The 168-acre lake is within
the city limits of Dover, the state capitol. However, much of
the 20,000-acre watershed is in Kent County, outside the
city limits. There are four major sub-basins or "branches"
which drain to the lake. The Maidstone Branch proved to
be a significant sub-basin in terms of the modelling effort,
as will be discussed later in this paper.
The lake and watershed are within the Coastal Plain
Region and display many of the problems associated with
water bodies in this region. Excess nutrients, particularly
nitrogen and phosphorus, coupled with typical low sum-
mer base flow conditions have led to blue-green algal
blooms from the time the lake was first formed as a mill
pond in the late 1700's. Bathymetric, sediment and in-lake
water quality surveys indicate that the lake is well into the
eutrophication process, occasionally reaching
hypereutrophic conditions in the summer months.
The lake's fish community is dominated by less desir-
able rough fish, such as gizzard shad and carp, which con-
stantly stir the bottom sediments while feeding. This mix-
ing of the water column is further exacerbated by the fact
that the lake is a popular boating area and is the only in-
land lake in the state which still permits water skiing.
These factors combine to prevent the lake from being
used to its fullest potential. Sport fisherman bemoan the
fact that there are not more game fish to be caught in the
lake. Even those that are caught have consumption re-
strictions due to toxins found in fish tissue samples. Al-
though the city maintains a small swimming and beach
area at a park along the lake, it is closed so many times
during the summer due to health concerns that the city is
considering building a public pool instead. Review of health
department records indicates these closings have in-
creased over recent years. Finally, landowners along the
lake have to endure the smell and unsightly scum associ-
ated with major blue-green algae blooms which inevitably
occur every summer.
Although the existing land use is largely a mix of agricul-
ture and woodland, planners have predicted the watershed
is slated for population growth of up to 52% and house-
hold increases of up to 73% by the year 2020. Since much
of this growth will occur outside the water and sewer ser-
vice area provided by the City of Dover, it is expected that
on-site waste treatment facilities in the porous, relatively
high water table soils will be a major source of nitrogen
loadings to the lake. Sedimentation from active construc-
tion sites, fertilizer applications to lawns, and pet wastes
are also expected to be significant potential sources of non-
point source (NPS) pollution as urbanization of the water-
shed increases. In the fringe areas where development
has occurred, bioassessments have already indicated evi-
154
-------
McKee Run
Silver Lake
Dover Creek
Legend
• Sub-Basins
Fork Branch
McKee Run
Dover Creek
Direct Inflow
0 1 Mile 2 Miles
Figure 1. Dover Silver Lake Watershed
dence of impacts to stream systems due to higher
stormwater runoff volume and longer duration of flow.
Based on these projections, impacts to the lake and
streams in the watershed are not likely to subside and may
increase.
Initiation of the Modelling Effort
The City of Dover first approached the Department of
Natural Resources and Environmental Control (Depart-
ment) with documented water quality problems in the lake.
However, it was recognized that Kent County would have
to be a major player involving any recommendations deal-
ing with land use. An advisory committee was formed, with
representatives from the department, city, county, environ-
mental groups, builders/developers and other stakehold-
ers. It was decided that a watershed protection/restoration
plan was needed to keep the lake from further degrada-
tion. An important component would be the development
of a watershed model, which would allow decision-makers
to assess impacts to the lake based on various policies
and recommendations put forward in the plan.
Although a grant was procured from the National Oce-
anic and Atmospheric Administration (NOAA), it was ap-
parent that there would not be enough funding to contract
the modelling to an outside consultant. At this point, the
advisory committee enlisted the help of the department to
assist in the modelling effort. Department staff felt it would
be feasible to develop a "planning level" model, but that a
"deterministic" model would require additional resources.
It was decided that a "planning level" model would suffice,
155
-------
and the staff were told to proceed. The remainder of this
paper will describe some of the goals and methodologies
of the DSL modelling effort. Although these have been
separated into data issues and model issues for purposes
of discussion, the reader should be aware that many of
these issues must be addressed concurrently during the
modelling effort.
Data Issues
Collection
In most cases, data collection will be the first priority in
model development. A thorough review of existing infor-
mation and data should be done. Sources of such infor-
mation used in the DSL Watershed Project included:
• Previous studies done in the watershed under the
Clean Lakes Program
• An analysis of the lake dam conducted by the Corps
of Engineers (USACOE)
• A flood study conducted by the Federal Emergency
Management Agency (FEMA)
• Water quality data collected in the watershed as part
of the department's responsibility under the Environ-
mental Protection Agency's (USEPA) 305b report
• Soils information from the Natural Resources Conser-
vation Service's (NRCS) soil survey for Kent County,
DE
• Land use data from the department's CIS database
• Zoning maps from the City of Dover and Kent County
• Population projections from the State of Delaware
• Topography based on US Geological Survey (USGS)
7.5-minute quadrangles
• Rainfall data from the USGS
• Flow data from the USGS
The last two items are especially important in any wa-
tershed modelling effort. As will be discussed, it is impor-
tant to realize that in most cases, it will be necessary to
have two years' worth of rainfall and flow data for the cali-
bration/verification process. The data must also be avail-
able in a format which is usable in the chosen model. For
example, use of a continuous-type hydrologic model will
require rainfall and flow data in a time series format, such
as 15-minute interval recordings. Daily totals would not be
adequate for use with such models.
Gaps
Data gaps are almost certain to occur in any major wa-
tershed modelling effort. Filling those gaps is, of course,
dependent on the nature of the missing data. Gaps in rain-
fall and/or flow data must be addressed quickly, since
modelling cannot proceed without such data. As mentioned
previously, it is desirable to have two years' worth of data
for the hydrologic model. Obviously, if data are not already
available and must be collected, the modelling itself must
be put on hold until the data become available. Therefore,
rainfall and flow data can be considered "mission critical"
and should be among the first data collected.
For the DSL Watershed Project, rain gages were placed
in two locations within the watershed at project inception
to check for spatial and temporal variation in rainfall. In
addition, the USGS maintains a time-series rain gage at
its main office in Dover. A statistical analysis was conducted
on the data from the three gages to ensure that there were
no significant differences in the rainfall data. Once this was
verified, missing data from one station could be compen-
sated with data from another station with confidence.
Although it was desired to have stream gages located at
all four of the major branches feeding the lake, funding
only allowed contracting with the USGS to re-establish a
gage on the Maidstone Branch; however, the USGS main-
tains a gage below the Silver Lake dam which measures
the flow from the entire watershed and has been in con-
tinuous service for more than 50 years.
Although land use data were available forthe watershed,
they were not a perfect match with some of the land uses
used to establish pollutant loadings. Specifically, the initial
land use map lumped all agricultural land together. Since
pollutant loadings vary considerably among agricultural
land uses, it was felt that further breakdown of this cat-
egory would be necessary. Fortunately, the DSL watershed
is immediately adjacent to a sub-watershed which drains
to the Chesapeake Bay. The Chesapeake Bay Program
(CBP) has done extensive data collection for modelling
purposes. Land use data were collected from a remote
satellite platform as part of the USEPA EMAP Program.
Further, the CBP data had gone through a field truthing
QA/QC process. Field investigation indicated that the ag-
ricultural land use in the CBP's Choptank Segment was
similartothatofthe DSL watershed. This made it possible
to use data from the Choptank Segment for extrapolation
purposes in defining the various agricultural land uses in
the DSL watershed.
Ambient water quality data were available from several
STORET stations located in the watershed; however, there
was very little water quality data collected from runoff dur-
ing actual storm events. Initially the project team intended
to implement a stormwater runoff sampling program to pro-
vide this data. However, this proved to be a very costly
venture in terms of equipment, lab costs, and staff time.
Furthermore, the yearthe data were to be collected turned
out to be one of the driest in many years. It was decided
that an alternative was needed. One of the areas studied
during the National Urban Runoff Program (NURP) was
the Metropolitan Washington, DC area. This study became
the basis for the DSL urban land use water quality data.
Once again, the Chesapeake Bay Program provided wa-
156
-------
ter quality data from agricultural land uses. Locally col-
lected stormwater data were used for verification purposes.
Analysis
One of the most crucial aspects of the modelling pro-
cess is data analysis. This is the opportunity forthe modeller
to get to know and understand the hydrologic processes
of the watershed. This starts with an analysis of the rain-
fall. The modeller should compare the data being used for
the model with historic data to make sure that there are no
anomalies. Data used for the calibration run should also
be compared to that used forthe verification run. Variation
in this data is to be expected. However, if the variation is
large, it should be kept in mind that it may be difficult to
verify the model with a high degree of confidence.
Arguably, the flow data are the most important and should
be analyzed accordingly. The US Geologic Survey does
an excellent job of summarizing data from its own gages
in its annual "Water Resources Data" reports for both sur-
face and groundwater. This provides a good "first cut" analy-
sis but should be supplemented whenever possible by
closer scrutiny of the data itself. In the case of the DSL
Project, USGS time-series flow data were imported into a
spreadsheet program and graphed so that individual storms
could be analyzed. As mentioned previously, there were
two gages in the watershed. One was below the spillway
of the Silver Lake dam on the St. Jones River, while the
other was on the Maidstone Branch, a majortributary drain-
ing to the lake.
Although is was desirable to model the watershed as a
whole, the available flow data would make this difficult.
The St. Jones gage measured the outflow from the lake
and therefore represented flow which had been routed
through a reservoir. Inflow data to the lake was limited to
only one tributary, the Maidstone Branch. Furthermore, the
St. Jones gage captured approximately 150 acres of highly
urbanized land below the dam spillway which resulted in
an incidental peak in the hydrograph priorto the main peak
from the reservoir. This incidental peak, along with the res-
ervoir effect made the St. Jones data less desirable from a
watershed modelling standpoint. However, visual obser-
vation indicated that the flow data from the Maidstone
Branch, though lower in magnitude, was clearly correlated
to that forthe St. Jones. Statistical analysis confirmed this
correlation; therefore, the Maidstone Branch was used as
a "surrogate" forthe DSL watershed as a whole. This elimi-
nated the problems associated with the St. Jones flow data
and simplified the modelling process. The relationship with
the highest correlation was the "Rv" value or runoff coeffi-
cient. This correlation was valuable in modelling the future
land uses, as discussed later in this paper.
Analysis of the flow data also revealed some other rel-
evant hydrologic relationships. There were considerable
seasonal differences in the flow for similar storm events.
For example, a storm on April 14, 1993 having 1.26" of
rainfall created a peak discharge ten (10) times greater
than a similar storm on July 14,1993 having 1.50" of rain-
fall. This was apparently due in large part to the effect that
vegetation had on interception and transpiration. It was
also no doubt related to greater infiltration losses and avail-
able surface storage. Whatever the reason, this has im-
portant implications for modelling. If one were to rely only
on event-type models to represent the hydrology of such a
watershed, the results may differ significantly from reality.
Another important hydrologic relationship is the ratio of
the surface runoff to the base flow. A manual separation of
one year's flow data from the Maidstone Branch revealed
that the base flow actually accounted for approximately
50% of the total flow. This also has important implications
forthe modelling. From a quantity standpoint, the model
must have the capability to differentiate the surface flow
and base flow in order for a true calibration to be accom-
plished. From a quality standpoint, the capability to ac-
count for the base flow is essential to estimate loadings
from soluble constituents such as nitrogen.
Uncertainty
One of the most difficult tasks to deal with in any model-
ling effort that attempts to predict future impacts is that of
uncertainty. Forthe DSL Project, there were two areas of
uncertainty which had to be addressed: future land use
and the hydrologic response due to this change in land
use. While zoning maps often provide some basis for de-
termining future land use, the unincorporated areas of the
watershed were zoned almost entirely as agricultural-resi-
dential with an average density of 1 unit/acre on the exist-
ing maps. To complicate matters, the City of Dover and
Kent County were both in the midst of updating their com-
prehensive plans, which would not be finalized until after
the watershed project was scheduled to be completed.
Although planners had predictions for population growth
through the year 2020, the actual land use mix associated
with that growth was anybody's guess. After wrestling with
this dilemma for some time, staff decided to develop five
alternative build-out scenarios of increasing density. The
mix for these build-out scenarios was based on land use
mixes observed in adjacent, more highly urbanized areas.
Alternative 3 was considered the most likely build-out sce-
nario forthe watershed and consisted of the following land
use mix:
• 40% undeveloped
• 25% low-density residential
• 20% medium-density residential
• 10% high-density residential
• 5% commercial/industrial/institutional
The various build-out scenarios, of course, provided the
basis for modelling future impacts to the lake, both in terms
of water quantity and water quality. Predicting changes in
hydrology is one of the strengths of event-type modeling.
However, for continuous-type modelling, some input pa-
rameters are used for calibration purposes that are not
157
-------
strictly based on physical characteristics of the watershed.
Calibrating to an unknown condition is tenuous at best
unless one has considerable experience with that model.
Fortunately, one of the flow relationships was of value in
overcoming this obstacle. As mentioned previously, the
strongest correlation found during analysis of the flow data
was the runoff coefficient or "Rv" value. This is merely the
ratio of the runoff (r) to the precipitation (p) and can be
expressed mathematically as follows:
Rv = r/p
(1)
This relationship was also examined during the NURP
study using runoff data collected from the various study
areas. Researchers further found that the runoff coefficient
was related to the percent imperviousness (I) in the water-
shed. A regression analysis resulted in the following equa-
tion (adjusted R2= 0.71):
Rv= 0.05+ 0.009(1)
(2)
Since this relationship ties the runoff with the percent
imperviousness, the DSL Project staff were able to use it
as a basis for calibrating the continuous hydrologic model
for the future build-out scenarios.
Variability
As anyone who has worked with water quality data is
aware, there is a tremendous degree of variability in such
data. While it is generally accepted that most constituents
found in stormwater runoff exhibit a lognormal distribution,
it still takes a relatively large data set to characterize local
conditions within a comfortable level of confidence. This
proved to be cost-prohibitive in the case of the DSL Project
and was a major consideration in the staff's decision to
develop a "planning level" model as opposed to a deter-
ministic model. Statistical methodologies and Monte Carlo
simulations are an alternative, but still require a high de-
gree of confidence in the distribution function. In the end,
a simple approach using low, median, and high values was
used forthe water quality constituents. This approach was
also used for analyzing the effects of Best Management
Practices (BMPs) in the watershed based on low, median
and high removal efficiencies. A matrix was then devel-
oped forthe various build-out scenarios, constituent load-
ings and BMP removal rates. From this, it was possible to
assess impacts to the lake based on worst case, best case
and most-probable case scenarios.
Model Issues
Decision
The decision as to whether to do watershed modelling
should not be taken lightly. Modelling is a labor-intensive
undertaking, requiring a considerable commitment of staff
resources. In most cases, the time required for such an
effort will be underestimated. Alternative methods should
be explored whenever possible. Results from other water-
shed studies in the area may be just as valid for planning
purposes. This could preclude having to do modelling al-
together. In any case, modelling should be viewed as a
"least preferred option" for assessing a watershed.
Selection
If the decision has been made to proceed with a model-
ling effort, the next step is to determine which model to
use. This is usually based on what aspect the watershed
study is focused on. However, in some cases the choice
of a particular model may be predicated on the available
data. If time series flow data are not available, for example,
some models may not be an option. The capabilities of the
modellers must also be considered. Models range in com-
plexity from simple regression formulae to statistical mod-
els to computer programs capable of doing very sophisti-
cated hydrologic simulations. There is, unfortunately, no
single model currently developed which can be thought of
as providing "one-stop shopping." Some models, for ex-
ample, are geared to event-type modelling for assessing
flooding and water quantity impacts. Others operate in
continuous mode for assessing water quality impacts.
Some model runoff conditions, others model receiving
water interactions. The models used for the DSL Project
were as follows:
• Hydrologic Models
TR-20 (NRCS) - Single-event mode only, synthetic
rainfall; used to estimate runoff associated with ma-
jorstorms(i.e., 10-YR, 100-YR, etc.) forFEMAflood-
plain studies.
PCSWMM (EPA) - Single or continuous mode, ac-
tual rainfall; calibrated against actual flow records
to simulate hydrologic processes for Maidstone
Branch sub-watershed.
• Hydraulic Model
HEC-RAS (COE) - Used to estimate water surface
profiles for FEMA floodplain studies.
• Water Quality Models
PCSWMM (EPA) - Used to estimate pollutant load-
ings for single or continuous rainfall events based
on Event Mean Concentrations (EMC).
Excel (Microsoft)- Spreadsheet program used to es-
timate pollutant loadings based on both EMCs and
Export Coefficients.
• Receiving Water Model
Vollenweider Model - Regression formula used to
estimate the steady state phosphorus concentration
in Silver Lake.
• Trophic State Model
Carlson Trophic State Index (TSI) - Method used to
determine the likelihood that the lake displays a par-
ticular trophic state (i.e., mesotrophic, eutrophic,
etc.).
158
-------
Calibration/Verification
For hydrologic simulation computer programs such as
EPA's Stormwater Management Model (SWMM), it is es-
sential to calibrate the model to local data. If available, a
full year's worth of rainfall and flow data in a time-series
format would be preferred. The purpose of the calibration
run is to make sure that the output from the simulation is
consistent with observed data. This is largely a matter of
adjusting the various input parameters until the results from
the simulation agree with the observed data within an ac-
ceptable degree of variance. The modeller must be care-
ful not to adjust the parameters outside the range of ac-
cepted values just to get a better calibration. The adage
"garbage in, garbage out" certainly applies in this case.
Once the model is calibrated, a verification run should
be made with an independent data set. The verification
process assures that the model has not been merely opti-
mized to a single data set, but can give consistently realis-
tic results as the input data varies. The calibration and
verification process has often been referred to as the "what
is" stage of the model.
Prediction
Once the model has been calibrated and verified, it can
be used for assessing "what if scenarios. This is the pri-
mary purpose of any major watershed modelling effort. In
the case of the DSL Project, future impacts to the tributar-
ies and lake under various land use changes could be as-
sessed, along with potential mitigation methods. Some of
the results were as follows:
• Peak discharges, volume of flow, duration of flow, and
out-of-bank incidences could all be expected to in-
crease underthe various build-out scenarios. The per-
centage increase depended on the percentage of im-
perviousness.
• Underthe most probable build-out scenario, flow depth
increased 9 inches for a storm which under existing
conditions was just at bank-full elevation.
• As a result of the hydrologic changes, stream channel
erosion would be expected to increase, thus increas-
ing total suspended solids (TSS) loadings to the lake.
• Loadings of total phosphorus (TP) to the lake remained
relatively flat with increased build-out.
• Loadings of total nitrogen (TN) to the lake increased
slightly with increased build-out.
• Even under a best-case scenario, it was unlikely that
the lake would drop below eutrophic conditions.
Validation
A step which is often overlooked in some watershed stud-
ies is validation of the model results. This is not to be con-
fused with verification, which is more closely associated
with the functioning of the model itself. The validation pro-
cess is an attempt to assure that the results from the model
can be accepted. A model could conceivably be calibrated
and verified, yet still yield unacceptable results. The water
quality estimates from the models developed for the DSL
Project were validated by comparing them to observed
concentrations from several STORET stations in the lake.
The model results indicated that the average total phos-
phorus (TP) concentration in the lake for the existing con-
dition would be 0.22 mg/l. This compared favorably to the
average observed TP concentration of 0.17 mg/l, and was
well within the range of observed values. Thus, staff felt
the results from the models could be accepted, particu-
larly for planning purposes.
Conclusions
As environmental managers become more aware of the
need to manage water and aquatic resources at the wa-
tershed level, supporting staff are increasingly being asked
to provide information for decision makers. This often in-
cludes the use of watershed modelling. Such was the case
for the Dover Silver Lake Watershed Project located in
central Delaware (DDNREC, 1992, 1995a, 1995b). Staff
were enlisted to model a 20,000-acre watershed for the
purpose of preparing a watershed protection/restoration
plan. A critique was done at the completion of the project
and it is hoped that information gleaned from that critique
will be helpful to others finding themselves in similar situa-
tions. Basically, it was found that the modelling tasks tended
to be grouped as either data issues or model issues. Data
issues included collection, gaps, analysis, uncertainty and
variability. Model issues were decision, selection, calibra-
tion/verification, prediction and validation.
References
Delaware Department of Natural Resources and Environ-
mental Control, Division of Soil and Water Conserva-
tion. (1992). "Watershed Protection Strategy for the
Dover/Silver Lake/St. Jones Watershed." Dover, DE.
Delaware Department of Natural Resources and Environ-
mental Control, Division of Soil and Water Conserva-
tion. (1995a). "Technical Background Report for Silver
Lake Watershed." Dover, DE.
Delaware Department of Natural Resources and Environ-
mental Control, Division of Soil and Water Conserva-
tion. (1995b). "Technical Background Report for Silver
Lake Watershed, Appendix A." Dover, DE.
159
-------
Water Quality Modeling to Support the Rouge River Restoration
Edward H. Kluitenberg, RE.
Applied Science, Inc.
Detroit, Michigan
Gary W. Mercer, RE.
Camp, Dresser and McKee
Detroit, Michigan
Vyto Kaunelis
Wayne County Department of Environment
Detroit, Michigan
Abstract
The Rouge River National Wet Weather Demonstration
Program (Rouge Project) has taken on the challenge of
implementing river restoration efforts in a highly urbanized
watershed. The 467-square mile Rouge River Watershed
is located in southeastern Michigan, and encompasses 48
communities, including the City of Detroit. A significant
number of stormwater and combined sewer overflow (CSO)
controls are being installed within the watershed to ad-
dress Rouge River pollution reduction objectives.
A suite of hydrologic, sewer system and riverine water
quality models have been used to address technical ques-
tions that have been asked in Rouge River Watershed plan-
ning. This paper presents application of four of the models
used by the Rouge Project: 1) TRTSTORM, 2) Watershed
Management Model (WMM), 3) Stormwater Management
Model (SWMM), and 4) Water Quality Analysis Simulation
Program (WASP). The TRTSTORM model predicts annual
overflow statistics for various CSO control facilities. A simple
pollutant loadings model, the WMM evaluates and com-
municates the relative impacts of various stormwater con-
trols. SWMM is aiding the development of subwatershed
management plans by predicting relative changes in wet
weather river response for alternative controls. Finally, the
WASP event model predicts the highly transient dissolved
oxygen drops caused by CSO discharges, thus the ben-
efits for various levels of CSO control.
The Rouge Project models have been and continue to
be an important decision-making aid for the project. In
addition, the modeling approach used by the Rouge
Project, as well as several specific modeling tools, are
transferrable to other urban watershed management
projects.
Introduction
The Rouge Project is using four modeling tools to sup-
port river restoration efforts in the highly urbanized Rouge
River Watershed. The U.S. Environmental Protection
Agency (USEPA) sponsored the Rouge Project, in 1992,
to demonstrate effective solutions to wet weather water
quality problems in urban areas. Under the leadership of
the Wayne County Department of Environment, CSO con-
trols and stormwater best management practices (BMPs)
are being implemented within a watershed approach which
stresses an inclusive process of all stakeholders.
This paper presents an overview of how water quality
models are being used to answertechnical questions which
arise in the Rouge Watershed planning process. Applica-
tion of four specific models is discussed including each
model's role and sample results which illustrate how the
model could be applied in other watersheds. Several les-
sons learned in the Rouge Project modeling effort are also
presented.
The Rouge Watershed
The Rouge Watershed encompasses 467 square miles
in Michigan's greater Detroit metropolitan area and is home
to 1.5 million residents. The Rouge River has been identi-
fied as one of the most polluted rivers in the Great Lakes
basin. The Lower, Middle, Upper and Main Rouge River
branches total 127 miles in length, and comprise one of
the state's most publicly accessible rivers.
Multiple pollution sources have led to the gradual deg-
radation of waterquality and habitat in portions of the Rouge
River and resulted in use impairments. The primary prob-
lems include CSOs, nonpoint stormwater runoff, illicit con-
nections, failing septic tanks, stream bank erosion and in-
creased flow variability. The combined effect of these pol-
160
-------
lutants has led to depressed DO levels, whole-body con-
tact prohibitions, damaged aquatic habitat, fish consump-
tion advisories and poor aesthetics.
One-third of the CSOs in the watershed are being con-
trolled via 11 demonstration CSO basins, several of which
became operational in 1997. Performance of the demon-
stration facilities will be used to determine the appropriate
level of control for the remaining CSOs. The watershed
has been divided into 11 subwatersheds where advisory
groups are forming to address all other pollution sources
in a holistic fashion. Numerous stormwater BMPs, recre-
ation and habitat projects have already begun and more
are planned.
Modeling Approach
The modeling effort consists of a three-tiered approach.
Tier 1 consists of several small area models used to simu-
late flows, pollutant loads and concentrations from spe-
cific pilot projects or localized areas of study such as wet-
lands, swales, wet detention ponds and individual CSO
basins. Tier 2 consists of a simple pollutant loading model
and a detailed sewer system model that both simulate pol-
lutant generation by subareaforthe entire watershed. Tier
3 is the river models which simulate instream flows and
concentrations in the four main river branches, based on
the inputs from the Tier 2 detailed sewer system models.
Following are four examples of these models in use.
CSO Facility Performance
While the 11 demonstration CSO facilities were in the
design stages, the TRTSTORM model was developed to
provide some early predictions as to how these basins
would perform (Kluitenberg et al., 1994). The model was
used to address the following questions:
• How will the proposed CSO facilities, designed to sev-
eral different sizing criteria, perform relative to pre-
sumptive criteria in the USEPA CSO Policy (USEPA,
1994)?
• What annual pollutant load reductions are expected
from the proposed facilities?
The TRTSTORM model is a simple hydrologic mass
balance model which tracks CSO facility filling, treatment,
overflow, dewatering and decanting based on long-term
hourly precipitation records. It is a modified version of the
U.S. Army Corps of Engineers "Storage, Treatment, Over-
flow, Runoff Model" (Hydrologic Engineering Center, 1976).
The model generates annual performance statistics for
flows to the treatment plant (via interceptor), treated and
untreated overflows.
The model was used to show that all CSO facilities de-
signed to the demonstration sizing criteria should meet the
85% capture and fouroverflow peryear presumptive crite-
ria in the USEPA CSO policy. The model used assumed
treatment efficiencies to determine expected annual load
reductions for a number of pollutants at each facility. Fig-
ure 1 shows a summary of the predicted annual reduction
in biochemical oxygen demand (BOD) entering the receiv-
ing water for: one site-limited facility; five basins sized to
provide 20-minute detention of a 1-year, 1-hourstorm (dem-
onstration size); and two basins sized to capture a 1-year,
1-hourstorm (Michigan Department of Environmental Qual-
ity (MDEQ) size). The results make it clear that for either
of the two sizing criterion, annual load reduction is strongly
governed by capture and is fairly insensitive to basin treat-
ment efficiencies.
Pollutant Loading Analysis
The Watershed Management Model (WMM) is being
used to estimate annual pollutant loading. In each
subwatershed, WMM is being used to address the follow-
ing questions and to communicate technical findings to
stakeholders in an easy-to-understand fashion.
100
90
80
70
60
50
40
30 -
20-
10-
0
D Treatment
D Capture
Site Limited
Demonstration Size MDEQ Standard
Figure 1. Annual percent reduction in BOD for various basin sizing criteria.
161
-------
• What are the relative contributions of different pollut-
ant sources in the subwatershed?
• What pollutant load reductions can be expected with
various stormwater BMPs and CSO controls?
• How will expected land use changes impact pollutant
loads at the bottom of the subwatershed?
The Rouge Project recently completed development of
WMM for Windows (Rouge Program Office, 1997) which
is being provided to each community for its own use in
subwatershed planning efforts. WMM calculates pollutants
loads for each source of flow (baseflow, stormwater run-
off, CSOs and point sources) in each watershed subarea
using annual flow volumes and event mean concentrations
(EMCs) assigned to that specific source. The model
projects annual pollutant loads by subarea. Various com-
binations of stormwater BMPs and CSO controls can be
selected in specific geographic areas to determine the
overall resulting pollutant reductions for a particular man-
agement plan.
WMM was used early in the project as a prioritization
tool to develop pie charts showing the major pollutant
sources in each subwatershed. It was recently used as an
analysis and communication tool in three detailed
subwatershed management studies. Figure 2 is a sample
of WMM results for BOD in the Middle 3 subwatershed,
where it was used to show the cumulative effect of two-
phase CSO control and two different stormwater manage-
ment plans.
Watershed Hydrology/River Hydraulics
The Rouge Project has developed a continuous, grow-
ing-season model of the entire watershed and the major
river branches using the USEPA Stormwater Management
Model (SWMM) (Huber et al., 1992). The model is used as
the hydraulic driver for the riverine water quality model. It
has also been used to assess river hydraulic impacts for
issues which arise in the subwatershed planning efforts.
Questions it has addressed include:
• How will expected land use changes impact instream
hydraulics (flow rates, volumes, depths and velocities)?
• How will proposed CSO control facilities impact
instream hydraulics?
• What combination of stormwater BMPs and CSO con-
trols will reduce instream peak flow rates to workable
levels for suitable fish habitat?
The SWMM RUNOFF block is used to model the hy-
drology of all storm sewered areas and areas with natural
drainage. An existing SWMM RUNOFF/TRANSPORT
model, the Greater Detroit Regional Sewer System Model
(Camp Dresser & McKee, Inc., June, 1994), is used to
model all CSOs entering the river. Inflow hydrographs from
both these models comprise all inputs to the one-dimen-
sional river model, which is simulated with the SWMM
TRANSPORT block. A continuous simulation with the full
model was calibrated to 6 months of 15-minute data col-
lected with a network of rain and stream flow gages. We
assume that these data coincide with a 40- or 50-year
analysis.
As part of the Upper 2 Subwatershed Management
Study, the model was used to evaluate several scenarios
including the cumulative impact of future land use projec-
tions, complete CSO control, placement of regional ex-
tended dry detention ponds throughout the subwatershed.
A fourth scenario involved placement of such ponds at only
a few select locations in the subwatershed instead of ev-
erywhere. The average increase in peak flow rates for a
range of typical storms is shown in Figure 3 for one sample
location. The results clearly show that the existing high
800-
700 -
I 600-
§ 500-
o
£ 400-
in
0 300-
CD
200 -
100 -
!
i
^^^^
D Stormwater •
• CSO
D Baseflow
Baseline Phase I CSO Phase II CSO SW Plan 1 SW Plan 2
Figure 2. Middle 3 subwatershed WMM model results - average annual BOD load.
162
-------
..Future Ph?.s.e.LLQS_O__ ...Dry Detention.. ..Select Dry...
Land'Use""' """Detention"
Modeling Scenario
Figure 3. Model river flow rate compared to existing conditions - Upper 2 subwatershed - Bell Branch at Beech Daly.
flow rates and velocities and the resultant bank erosion
problems will worsen, but that both types of regional de-
tention could be used to accommodate future land use
changes and reduce peak stream flows and velocities be-
low existing conditions.
Instream Water Quality
Building on the SWMM quantity model, a riverine water
quality model of the Lower, Middle, Upper and Main Rouge
River branches was developed using the Water Quality
Analysis Simulation Program (WASP) EUTRO model
(Ambrose et al., 1993). While the model was originally
developed and calibrated as a continuous model of eight
pollutants, it has evolved to its current, primary role as an
event model to simulate the CBOD-DO interaction which
results from CSOs, including the sudden transitory DO
drops which have been observed in the Rouge River. The
model is currently being used to address the following ques-
tions:
• Will various CSO control alternatives eliminate the tran-
sitory DO drops caused by high CBOD in CSO dis-
charges?
• What wet weather DO impairment will remain after all
CSO controls are in place?
• How much will dry weather DO improve after controls
eliminate most of the sediment oxygen demand (SOD)
contributed by CSOs?
The water quality model developed is shown schemati-
cally in Figure 4. Stormwater inputs are simulated with the
SWMM RUNOFF build-up/washoff algorithms. CSO inputs
are assigned concentrations based on the time from when
overflow begins, based on typical "pollutograph" shapes
from monitoring data. The Rouge Project also developed
a new model code linking the SWMM TRANSPORT river
hydraulic output to the WASP model. Portions of the model
have been calibrated to several heavily monitored wet
weather events.
The model was utilized to evaluate two alternative CSO
basin sizes in Oakland County on the main branch of the
Rouge. For one of the calibrated wet weather events, the
instream DO improvement was determined by modeling
the impact of complete CSO control with three CSO ba-
sins sized to the demonstration criterion. The simulation
was also repeated assuming the basins were enlarged to
the MDEQ standard sizing criterion. The simulated instream
DO shown in Figure 5 illustrates that the demonstration
size basins improve the DO sag enough that it no longer
falls below the 5 mg/l standard forth is event. It also shows
the marginal improvement which would have been
achieved if the MDEQ basin sizing criterion were used,
approximately doubling the size of each of the facilities.
The model of the entire main branch of the Rouge was
also used to simulate dry weather DO, which is primarily
driven by SOD and reaeration. For the first phase of CSO
control and also for complete control, model SOD was re-
duced to approach that of in-situ SOD measurements made
in river reaches which were not CSO impacted. The re-
sults in Figure 6 show that CSO controls will provide a
significant benefit to dry weather DO, but that some DO
impairment will remain in selected river reaches which are
somewhat impounded.
Instream performance monitoring began in 1997. The
monitoring is intended to show whether effluent from the
demonstration facilities will cause any remaining water
quality impairment. The water quality event models will be
used in the future as part of the analysis of the monitoring
results.
Model Findings
Many findings have arisen out of the Rouge Project,
several of which the models helped bring to light. Model
findings are given below.
• The impairments caused by wet weather pollution are
certainly not limited to wet weather periods. In the
Rouge this is especially true for the CSO contributions
to SOD and the resultant dry weather DO impairment.
163
-------
RUNOFF Stormwater
Runoff Model
Pollutants
RUNOFF/TRANSPORT
Combined Sewer System
Overflow Model
Flows \
' Flows
TRANSPORT River
Hydraulic Model
Hydraulics
WASP River
Water Quality Model
Pollutants
Figure 4. Rouge Tier 3 model schematic.
— - Baseline
MDEQ Size Basins
Demo Size Basins
13-Jun
14-Jun
15-Jun
16-Jun
17-Jun
Figure 5. Modeled DO for CSO control alternatives - Main Rouge at 8 Mile Rd.
— - Baseline
— Phase I CSO Control
Full CSO Control
10 20 30
Main Rouge River Miles Downstream of Adams Road
Figure 6. Dry weather model instream DO for June 13, 1994 - Main Rouge from Adams to Greenfield.
164
-------
• In a predominantly urban watershed, increased stream
flow due to urbanization damages aquatic habitat and
causes bank erosion, logjams, sedimentation and in-
creased instream solids concentrations.
• Some dry and wet weather DO impairment is expected
to remain in the Rouge Watershed after all CSO are
controlled, simply due to nonpoint stormwater runoff.
Stormwater and CSO must be addressed together in
a holistic approach.
• It is expected that the Rouge Watershed CSO basins
sized to demonstration criteria will be adequate to elimi-
nate any resultant water quality impairments.
• Rouge Watershed standard practices for on-site de-
tention of stormwater do little to mitigate the develop-
ment-induced flow increases for small storms, with
attendant increased velocities and streambank erosion.
Lessons Learned
Over the course of the Rouge Project modeling effort a
number of lessons have been learned about the model-
ing. Several of the key lessons are:
• If possible, model selection and development should
not be performed until the specific questions to be ad-
dressed by the model are well formulated.
• A simple loadings model such as WMM can be a good
technical resource, but it may be even more important
as a tool for communicating technical findings.
• Urban rivers dominated by stormwater runoff present
a unique modeling challenge as the difficulty of moni-
toring nonpoint sources means there are not well-de-
fined inputs for the model.
• Models should not be used to try to answer every ques-
tion. Many questions can still be answered via analy-
sis of monitoring data.
Conclusions
The Rouge project is successfully using a suite of four
water quality modeling tools to address technical ques-
tions raised in watershed management planning. The
Rouge Project models, modeling approach and findings
are a resource that is transferable to other urban water-
sheds.
References
Ambrose, R.B., T.A. Wool and J.L. Martin, The WaterQual-
ity Analysis Simulation Program WASPS, Part A: Model
Documentation, Environmental Research Laboratory,
Office of Research and Development, U.S. Environmen-
tal Protection Agency, Athens, GA, September 1993.
Camp Dresser & McKee, Inc, DWSD Greater Detroit Re-
gional System Report, Prepared for Detroit Water &
Sewerage Department, City of Detroit, Ml, June 1994.
Huber, W C. and R. E. Dickinson, Stormwater Management
Model, Version 4: User's Manual, EPA/600/3-88/001 a,
USEPA, Environmental Research Laboratory, Athens,
GA, October 1992.
Hydrologic Engineering Center, Storage, Treatment, Over-
flow, Runoff Model "STORM, "U.S. Army Corps of Engi-
neers, 1976.
Kluitenberg, E.H., and C. Cantrell, Percent Treated Analysis
of Demonstration Combined Sewer Overflow Control
Facilities, Technical Memorandum RPO-MOD-TM17,
Prepared for Wayne County Rouge Program Office,
Wayne County, Ml, October 1994.
Wayne County Rouge Program Office, User's Manual: Wa-
tershed Management Model Version 1.0, Technical
Memorandum RPO-NPS-TM27, Wayne County, Ml,
December 1997.
U. S. Environmental Protection Agency, Combined Sewer
Overflow Control Policy, Office of Water, EPA830-B-94-
001, April 1994.
165
-------
Overview of Urban Retrofit Opportunities in Florida
Michael Bateman, Eric H. Livingston, and John Cox
Stormwater Management Program
Florida Department of Environmental Protection
Tallahassee, Florida
Abstract
With the implementation of Florida's State Stormwater
Rule in February, 1982, stormwaterdischarges serving new
development or redevelopment were required to be treated
by the incorporation of site appropriate best management
practices (BMPs) into the project's Stormwater manage-
ment system. The implementation of this program has
greatly reduced the impact of Stormwater discharges on
aquatic resources, especially given Florida's rapid growth
which has seen the state's population grow from 9,746,224
in 1980 to an estimated 14,700,000 in 1997. However,
Stormwater discharges from development existing before
1982 continue to contribute to the degradation of Florida's
water resources. This paper will review the institutional
framework the state has implemented to address
Stormwater problems associated with existing land uses.
Its primary focus will be to summarize several different
types of urban Stormwater retrofitting projects that have
been undertaken to reduce pollution from older Stormwater
discharges. For each project, we will review the type and
design of BMP, site characteristics, cost, and pollutant re-
moval efficiency.
Introduction
Florida is blessed with a multitude of natural systems,
from the longleaf pine-wiregrass hills of the panhandle, to
the sinkhole and sand ridge lakes of the central ridge, to
the Everglades "River of Grass", to the coral reefs of the
Keys. Abundant surface water resources include over 20
major rivers and estuaries along with nearly 8,000 lakes.
Plentiful groundwater aquifers provide over 90% of the
state's residents with drinking water. Add the state's cli-
mate and it is easy to see why many consider the Sun-
shine State a favored vacation destination and why the
state has experienced phenomenal growth since the 1970s.
Today, Florida is the fourth most populous state and is still
growing rapidly, although not at the rate of 900 people per
day (300,000 per year) that occurred throughout the 1970s
and 1980s.
Florida's natural systems, especially its surface and
groundwater resources, are extremely vulnerable and eas-
ily damaged. This is partially the result of the state's sandy
porous soils, karst geology, and abundant rainfall. The
negative impacts of unplanned growth were seen as early
as the 1930s, when southeast Florida's coastal water sup-
ply was threatened by saltwater intrusion into the fragile
freshwater aquifer that supplied most of the potable water
for the rapidly expanding population. By the 1970s, it was
becoming all too clear that unplanned land use, develop-
ment, and water use decisions were altering the state in a
manner that, if left unchecked, could lead to profound, ir-
retrievable loss of the natural beauty that brought residents
and tourists to Florida. Extensive destruction of wetlands,
bulldozing beach and dune systems, continued saltwater
intrusion into freshwater aquifers, creation of impervious
surfaces and resulting increase in Stormwater, and the
extensive pollution of the state's rivers, lakes, and estuar-
ies were only some of the negative impacts of this rapid
growth.
Fortunately, Florida's citizens and elected officials be-
came educated about these problems and began devel-
oping programs to protect and manage the state's natural
resources. Florida began serious and comprehensive ef-
forts to manage its land and water resources and growth
coincident with the increasing strength of the environmen-
tal movement in the nation and the state during the early
1970s. Overthe next25 years, Florida's natural resources
management programs have evolved substantially. Col-
lectively, the individual laws and programs enacted during
this period can be considered "Florida's Watershed Man-
agement Program." In many cases, these laws have been
integrated either statutorily, with revisions to existing laws,
or through the adoption of regulations by various state,
regional, or local agencies. A primary focal point has been
the management of nonpoint sources of pollution, espe-
cially urban Stormwater, since stormwaterdischarges are
responsible for over half of the pollution load entering
Florida's rivers, lakes, and estuaries (Camp et al., 1995)
Florida's Stormwater Treatment Program
Research conducted in Florida during the late 1970s
characterized Stormwater pollutants, provided cost and
benefit information on many types of Stormwater treatment
166
-------
practices, and determined the importance of stormwater
discharges as a major source of pollution. As a result, in
1979, the Florida Environmental Regulation Commission
adopted the state's first stormwatertreatment requirements.
In 1982, the state's stormwater rule was fully adopted, re-
quiring all new development and redevelopment projects
to include site appropriate BMPs to treat stormwater. This
technology-based program establishes a performance
standard of removing at least 80% of the average annual
post-development loading of total suspended solids for
stormwater discharged to most waters. Stormwater dis-
charges to the state's most pristine waters, known as Out-
standing Florida Waters, are required to reduce pollutant
loading by 95%. As a result of the implementation of
Florida's stormwatertreatment program, the impact of the
state's rapid growth on its water bodies has been greatly
reduced.
With the successful implementation of Florida's
stormwater treatment, wetlands protection, and growth
management programs to address the adverse impacts of
new development, the focus of Florida's watershed man-
agement program has shifted to cleaning up "older
sources," such as existing urban or agricultural land uses,
and to integrating program components to eliminate dupli-
cation and improve efficiency and effectiveness. This has
led to greater emphasis on more holistic approaches to
address cumulative effects of land use activities within a
watershed and to a greater emphasis on regional struc-
tural controls and the purchase or restoration of environ-
mentally sensitive lands. Key institutional aspects of this
changing focus include:
1985 Chapter 403.0893, F.S., is created as the only
surviving section of a stormwater management bill that was
developed over a ten month process. The bill was an at-
tempt to put into law a cost-effective, timely process to
retrofit existing drainage systems to reduce the pollutant
loadings discharged to water bodies. The only section en-
acted creates explicit legislative authority for local govern-
ments to establish stormwater utilities or special stormwater
management benefit areas. Today, over 80 Florida local
governments have implemented a stormwater utility to pro-
vide their stormwater programs with a dedicated source of
funding.
1987 Chapter 373, F.S., is revised to add a new section,
the Surface Water Improvement and Management (SWIM)
Act, which establishes six state priority water bodies. It
directs the state's five regional water management districts
(WMDs) to prepare a priority water body list and develop
and adopt comprehensive watershed management plans
to preserve or restore these water bodies. The bill pro-
vides $15 million from general revenue sources and re-
quires a match from the WMDs. Unfortunately, a dedicated
funding source is not established making the program de-
pendent upon uncertain annual legislative appropriations.
1988 The State Nonpoint Source Assessment and Man-
agement Plan, prepared pursuant to Section 319 of the
Federal Clean Water Act, is submitted to EPA and ap-
proved. This qualifies the state for Section 319 NPS Imple-
mentation grants which are used for BMP demonstration
projects and to refine existing NPS management programs.
The delineation of the state's ecoregions, selection of riv-
erine ecoregion reference sites, and modification of EPA's
Rapid Bioassessment Protocols and metrics for use in
Florida is initiated. This will provide the state with better
tools to assess cumulative impacts of stormwater dis-
charges and the effectiveness of stormwater management
efforts.
1989 Chapters 373 and 403, F.S., are revised as part of
the 1989 Stormwater Bill. The legislation clarifies the
stormwater program's multiple goals and objectives; sets
forth the program's institutional framework which involves
a partnership among DER, the WMDs, and local govern-
ments; defines the responsibilities of each entity; addresses
the need for the treatment of agricultural runoff by amend-
ing Chapter 187, F.S., to add a policy in the Agriculture
Element to "eliminate the discharge of inadequately treated
agricultural wastewater and stormwater;" further promotes
the watershed approach used by the SWIM Program; des-
ignates State Water Policy, an existing but little used DER
rule, as the primary implementation guidance document
for stormwater and all water resources management pro-
grams; and creates the State Stormwater Demonstration
Grant Program with $2 million in funding as an incentive to
local governments to implement stormwater utilities.
1990 Chapter 17-40, FAC, State Water Policy under-
goes a total revision and reorganization so that it can be
used as guidance by all entities implementing water re-
source management programs and regulations. Section
17-40.420 (now 62-40.432) is created and includes the
goals, policies, and institutional framework for the state's
stormwater management program. Key elements are:
• DER is designated as the lead agency with responsi-
bility for setting program goals, providing overall pro-
gram guidance, overseeing implementation of the pro-
gram by the WMDs, and coordinating with EPA, espe-
cially with the advent of the new NPDES stormwater
permitting program.
• WMDs are the chief administrators of the stormwater
regulatory program (quantity and quality). They are
responsible for preparing SWIM Watershed Manage-
ment Plans, which include the establishment of
stormwater pollutant load reduction goals (PLRGs),
and for providing technical assistance to local govern-
ments, especially with respect to basin planning and
the development of stormwater master plans.
• Local governments are the front lines in the stormwater/
watershed management program since they determine
land use and provide stormwater and other infrastruc-
ture. They are encouraged, but not required, to set up
stormwater utilities to provide a dedicated funding
source for their stormwater program. Their stormwater
responsibilities include preparation of a stormwater
master plan to address needs imposed by existing land
167
-------
uses and by future growth; operation and maintenance
activities; capital improvements of infrastructure; and
public education. They are encouraged to set up an
operating permit system wherein stormwater systems
are inspected annually to ensure that needed mainte-
nance is performed.
Important stormwater program goals include:
•Preventing stormwater problems from new land use
changes and restoring degraded water bodies by re-
ducing the pollution contributions from older stormwater
systems.
•Retaining sediment on-site during construction.
•Trying to assure that the stormwater peak discharge
rate, volume, and pollutant loading are no greater af-
ter a site is developed than before.
•An 80% average annual load reduction for new
stormwater discharges to most water bodies.
•A 95% average annual load reduction for new
stormwater discharges to Outstanding Florida Waters.
•Reducing, on a watershed basis, the pollutant loading
from older stormwater systems as needed to protect,
maintain, or restore the beneficial uses of the receiv-
ing water body. The amount of needed pollutant load
reduction is known as a "Pollutant Load Reduction Goal
orPLRG."
With the inclusion of the PLRG concept in State Water
Policy, Florida has an institutional framework to begin fo-
cusing efforts on the reduction of environmental impacts
from existing land uses. While the focus of this paper is on
urban stormwater, many projects to reduce stormwater
pollution from agricultural activities have been undertaken,
including the construction of tens of thousands of acres of
wetlands to help restore the Everglades.
The rest of this paper will include short summaries of
successful urban stormwater retrofitting projects that have
been undertaken in Florida. These are representative of
the different types of structural BMPs that are being used
to reduce the impacts of urban stormwater discharges on
state's waters. However, it is also important to remember
that nonstructural pollution prevention programs are also
a crucial element of urban retrofitting. Educational efforts,
whether signage associated with a structural retrofit project
or statewide efforts such as "Florida Yards and Neighbor-
hoods," are essential in reducing "pointless personal pol-
lution" and in gaining the support of citizens and elected
officials for stormwater management programs.
168
-------
Urban Stormwater Retrofitting Project Fact Sheet
Lake Jackson Megginnis Arm Regional Stormwater System
Northwest Florida Water Management District
Watershed Area: 2200 acres
Watershed Land Use: 1191 acres - Low-medium density Residential 213 acres - Roads
102 acres - High density Residential 207 acres - Open
469 acres - Commercial
Project Overview: Studies in the mid-1970s of Lake Jackson in Leon County, Florida, determined that Stormwater from
the rapidly urbanizing Megginnis Arm watershed and from the construction of Interstate 10 were responsible for the
lake's water quality degradation. In 1983, the NWFWMD and the FDER cooperatively designed and constructed, using
EPA Clean Lakes grant and state funds, an experimental regional Stormwater treatment system. The system consists of
a 20 acre wet detention pond with a heavy sediment basin at the inflow, a 4.2 acre sand filter system, and a 5.7 acre,
three cell constructed wetland. The pond originally was sized for 150 acre-feet of storage, representing the runoff from a
2.5 inch storm in the watershed. Continued urbanization of the watershed resulted in greater volumes of Stormwater,
thereby reducing the system's effectiveness. Therefore, the system was enlarged in 1989-90 to increase the storage
volume by 31.7% thus providing 173.8 acre-feet of storage, or storage for 1.02 inches of runoff from the watershed. In
1992, the sand filter system was completely renovated, including new distribution pipes and sand filter media. Finally, in
1990-92, over 112,000 cubic yards of sediments which had accumulated in the bottom of Megginnis Arm were removed
and the littoral areas of the arm were replanted with native macrophytes and trees.
Project Cost: Original construction - $2,664,389 Filter renovation - $80,000
Pond expansion - $253,643 Dredging Megginnis Arm - $990,311
Educational signs - $19,565 Educational program - $40,000
Educational Component: Educational exhibits were installed at five public boat landings on Lake Jackson to increase
public awareness about Stormwater pollution, the regional Stormwater treatment system, and the dredging project. The
NWFWMD created the "Teacher's Guide to Stormwater Runoff in the Lake Jackson Watershed" and a video entitled "In
Search of Old Bigmouth" as resource materials for local teachers. These are used in conjunction with a field trip program
for local schools which provides students with "hands on" experience in water quality monitoring and the operation of the
regional Stormwater treatment system. More than 3,000 students have participated in this program.
Project Evaluation: About 6,000 cubic yards of materials were dredged from the heavy sediment basin after three years
of operation with additional material removed during the system's expansion. Monitoring data shows that in normal
operation, the system can reduce total volume by 30% and reduce loadings by over 90% for solids, 70% for total
nitrogen, 80% for total phosphorus, and 50% for orthophosphorus.
169
-------
Urban Stormwater Retrofitting Project Fact Sheet
Lake Ella Alum Injection System
City of Tallahassee Stormwater Utility
Watershed Area: 157 acres
Watershed Land Use: 13 acres - Residential
115 acres - Commercial/Residential
11 acres - Commercial
15 acres - Open
3 acres - Church
1 acres - Street
Project Overview: In 1985, a lake restoration project was initiated in Lake Ella, a shallow, 13.3 acre hypereutrophic
"lake" which receives Stormwater runoff from a 157 acre highly impervious watershed. Due to its highly developed and
urban watershed, and because of the low permeability of the watershed's clay soils, it was determined that traditional
Stormwater treatment BMPs could not be used. Instead, chemical treatment of runoff was evaluated using various
chemical coagulants including aluminum sulfate (alum), ferric salts, and polymers. Jar tests determined that alum con-
sistently provided the highest removal efficiencies and produced the most stable end product. Consequently, a prototype
alum injection system was designed where liquid alum was injected within storm sewers on a flow weighted basis.
Standard triplex metering pumps are used as the injection pumps, each individually regulated by sonic flow meters
attached to the storm sewer lines to be treated. Many of the smaller storm sewers were combined to reduce the points of
discharge into the lake from 17 to ten. Six of these ten inputs, representing 95% of the average flow, are equipped with
alum injectors. Alum is pumped from a 6000 gallon alum storage tank into individual one inch PVC underground carrier
lines to the point of injection. The alum mixes with Stormwater as it travels through the storm sewers, passes through a
fine mesh trash trap, and is discharged into Lake Ella. The restoration project also included the removal of 50,000 yds3
of accumulated sand, debris, and muck from the bottom of Lake Ella and the recontouring of the lake's bottom with a
gradual slope toward the outfall control structure.
Project Cost: The city's Stormwater utility paid $744,000 for the Lake Ella restoration project, with the alum system
costing $200,400. At a cost of $137/dry ton of liquid alum, annual chemical costs for alum injection are approximately
$10,000 per year.
Project Evaluation: Pre- and post-alum injection monitoring is summarized below:
Parameter Before After Parameter Before
After
PH
Total Nitrogen
BOD
Secchi Depth
7.41
1 876 ug/l
41 mg/l
0.5 m
6.43
41 7 ug/l
3.0 mg/l
2.2m
DO
Total Phosphorus
Chlorophyll-a
Florida TSI
3.5 mg/l
232 ug/l
1 80 mg/m3
98
7.4 mg/l
26 ug/l
5.1 mg/m3
47
Alum sludge accumulation rate: 0.33 cm/yr
Pollutants in sediments are much more tightly bound after alum injection system.
170
-------
Urban Stormwater Retrofitting Project Fact Sheet
Project Smart - Stormwater Reuse Demonstration
City of Winter Park and the University of Central Florida
Watershed Area: 8.13 acres
Watershed Land Use: 6.84 acres Impervious Residential/Commercial
1.29 acres Greenspace
84% impervious with 42% DCIA
Project Overview: Lake Mendsen is a small urban constructed pond which has been altered significantly over many
years and also receives untreated urban Stormwater runoff. The primary discharge from the pond occurs to two drainage
wells. The demonstration project was implemented to try to reduce the amount of untreated Stormwater which is dis-
charged to the pond and ultimately the Floridan Aquifer by detaining a portion of the first flush of Stormwater so that it can
be used for irrigation purposes or "reuse."
An area of the pond (approximately 0.7 acres) which receives Stormwater from two existing outfalls was isolated from the
main pond by the construction of a berm and weir system. The isolated area serves as a surface water reservoir for the
irrigation system. Accumulated sediments and invasive exotic vegetation also were removed from the area and the
bottom was recontoured. The resulting littoral zone was planted with five species of native aquatic macrophytes. Instru-
mentation was installed to monitor rainfall, irrigation pumping rates and volumes, and discharge volumes from the reser-
voir to the main body of the pond.
Project Cost: The entire project cost $143,000, although capital costs for the irrigation pump and system was only about
$4,600. Funding for the project was provided by the DER Pollution Recovery Trust Fund ($79,000) and by the city of
Winter Park and the University of Central Florida which provided $64,000 in money and in-kind services.
Project Evaluation: A mass balance was performed for the reuse pond over a study period of 358 days. The average
irrigation rate for the study period was approximately 1.07 inches per week over the 1.25 acre greenspace. The overall
mass balance demonstrated that 55% of the incoming runoff was reused and not discharged into Lake Mendsen. Based
on Florida rainfall statistics and Stormwater characterization data, this translates into an annual Stormwater pollutant
load reduction of over 80% for all pollutants. The project also resulted in a real economic benefit. Annualized cost
savings for irrigation were calculated to be approximately $3,300 per year, based on the reuse of Stormwater versus the
use of potable water from the city of Winter Park.
171
-------
Urban Stormwater Retrofitting Project Fact Sheet
Lake Greenwood Urban Wetland
City of Orlando Stormwater Utility
Watershed Area: 527 acres
Watershed Land Use: 275 acres - Residential 28 acres - Open
210 acres - Commercial/Industrial 14 acres - Water
Project Overview: The Greenwood Urban Wetland was built to alleviate flooding and to treat Stormwater runoff prior to
discharge to drainage wells which flow to the Floridan Aquifer. The system is designed to detain the runoff from 2.5
inches of rainfall. Approximately 300,000 cubic yards of material was removed to create the system which enlarged the
surface area of the "lake" from four to thirteen acres. Weirs were constructed to control water levels and establish three
ponds to maximize Stormwater detention. The average water depth is 5.1 feet, the storage volume is 66 acre feet, and
the hydraulic residence time is 22.7 days. The lakes have a 25 to 30-foot-wide littoral shelf which was planted with over
82,000 plants often species of native macrophytes. The lakes are connected by marsh flowways and the system also
includes a "riverine floodway" that allows large storms to bypass the lake system. The floodway is planted with seven
species of hardwood swamp trees. An upstream sediment/debris basin, pond aeration, and an irrigation system reusing
Stormwater are incorporated into the design to increase pollutant removal effectiveness. The reuse system allows the
City to irrigate the park and the adjacent city-owned cemetery with Stormwater instead of potable water, saving the city
$25,000 per year. In addition to providing flood protection and Stormwater treatment, the 26 acre Lake Greenwood Urban
Wetland park includes sidewalks, bridges, and green space passive recreation which is widely used by nearby residents.
Project Cost: $581,000 from the City of Orlando Stormwater Utility.
Project Evaluation: Preconstruction monitoring was conducted from May 19, 1987 through October 13,1988 to deter-
mine the trophic state of Lake Greenwood and to determine the potential loadings discharged to the lake's five drainage
wells. The preconstruction Trophic State Index averaged 64 and was highly variable ranging from 12. 5 to 80.8 with five
months above 70. After construction, TSI values averaged 57 but no months had values above 66 and variability was
less with a range of 36.2 to 66.3. Treatment effectiveness of the system is summarized below:
TN N02 N03 NH4 TP OP Cd Cu Pb Zn
SedTrap 4% -76% 4% -100% 11% 7% 26% 19% 10% 6%
Wetland 11% 8% -13% 10% 62% 77% 0% 59% 60% 69%
172
-------
Urban Stormwater Retrofitting Project Fact Sheet
Packed Bed Wetland Filter System
City of Orlando Stormwater Utility
Watershed Area: 121 acres
Watershed Land Use: 75 acres - Commercial 18 acres - Roads
14 acres - Stadium/parking 6 acres - Open space
8 acres - Industrial
Project Overview: Clear Lake is 360 acres in size and Stormwater loadings from its three square mile watershed have
led to serious water quality problems. An innovative Stormwater treatment system was needed for this basin to both
reduce pollutant load and function within a limited area where multiple demands are placed on the use of land. The
constructed experimental Stormwater treatment train consists of:
• A 3.3 acre off-line wet detention pond with a sediment trap at the inlet.
• Construction of diversion weirs to shunt the first flush to the wet detention pond while allowing the remaining
Stormwater to bypass the system.
• Construction of 10 packed beds consisting of five crushed concrete and five granite media beds, vegetated with four
differing combinations of wetland plants.
•Installation of two pumps to supply water to the packed beds from both the wet detention system during storms and
from Clear Lake during dry periods.
• Control valving to allow for varied water flow rates through the packed beds.
• Automated flow meters and composite samplers to allow storm event sampling.
Project Cost: $917,464 including monitoring costs with funding from DEP through the State Stormwater Demonstration
Grant Program and from the City's Stormwater utility.
Project Evaluation: Monitoring was performed on the effectiveness of the overall system, the performance of the
individual beds, and the best flow rate at which to operate the system (30, 60, or 120 gal/min). Analysis of the individual
beds showed consistent removal across all beds for cadmium, copper, lead, zinc, total nitrogen, TKN, nitrite, total phos-
phorus, TSS, VSS, and fecal coliform. Among the remaining parameters, chromium, ammonia, nitrate, orthophosphorus,
IDS, and TOC, pollutant removals within bed 6 were consistently low at all three flow rates. Conversely, bed 5 exhibited
consistently high removals for the same parameters. The high flow rate was determined to be the best operating rate for
the system. Overall pollutant load reduction is presented below:
Parameter % Removal Parameter % Removal Parameter % Removal
Cadmium 80 Total Nitrogen 63 Total phosphorus 82
Chromium 38 TKN 62 Orthophosphorus 14
Copper 21 Ammonia 6 TDS 8
Lead 73 Nitrate 75 TSS 81
Zinc 55 Nitrite -9 VSS 80
Fecal Coliform 78 TOC 38
173
-------
Urban Stormwater Retrofitting Project Fact Sheet
Bath Club Concourse Stormwater Rehabilitation Project
Town of North Redington Beach, Pinellas County, Florida
Watershed Area: 2.12 acres
Watershed Land Use: Pre-project-100% Impervious Roadway/Parking
Project Overview: The Bath Club Concourse is a combination roadway and parking lot connecting Bath Club Circle and
Gulf Boulevard. Before the project, the Bath Club Concourse was totally impervious consisting of asphaltic pavement.
Untreated runoff from the Concourse and its associated drainage area was directed by sheet flow into a single storm
sewer inlet and discharged offsite, and ultimately to Boca Ciega Bay.
The objectives of this project were: (1) to maximize the amount of Stormwater runoff that could be infiltrated on-site,
thereby reducing the annual volume that is discharged off-site without any treatment; and (2) to demonstrate innovative
alternative approaches to treating Stormwater in highly urbanized areas where land for traditional BMPs is scarce and
very expensive. Drainage was redirected toward two new pervious concrete parking areas located in the center of the
Concourse. These are separated by an unpaved landscaping island that also provides infiltration. To maximize infiltra-
tion of the pervious concrete parking areas, two 150-feet-long underdrainswere installed in the eastern half of the project
to facilitate the drainage of the subsurface soils immediately beneath the pervious concrete.
Project Cost: Total cost was $147,015 with construction costing $118,380 and landscaping costing $13,345. Funding
was provided by a Section 319 NPS grant from DER, the SWFWMD SWIM Program, and the Town of North Redington
Beach.
Project Evaluation: The project improvements resulted in a significant reduction of direct discharge of Stormwater
runoff from the site. Calculations accounting for average annual rainfall and runoff, as well as pore space volume and
subsurface water flow, indicate that the improvements caused a 33% reduction in total on-site runoff volume between the
pre- and post-project conditions. Further, the volume of surface runoff discharging directly to Boca Ciega Bay was
reduced by about 75%. Calculated overall removal efficiencies forthe project are based on the efficiency of the underdrain/
filter system to remove pollutants and are indicated as follows:
Parameter Lead Zinc TSS BOD TP OrthoP TN
% Removal 73 72 73 61 49 26 65
174
-------
Urban Stormwater Retrofitting Project Fact Sheet
Sunset Drive Outfall Stormwater Rehabilitation Project
City of South Pasadena, Pinellas County, Florida
Watershed Area: 49 acres
Watershed Land Use: 21.6 acres - Residential Multifamily
20.1 acres - Commercial
7.4 acres - Residential Single Family
Project Overview: The Sunset Drive drainage basin is nearly fully developed and consists of approximately 55% imper-
vious area. Historically, Stormwater was collected and discharged untreated to a local storm sewer which connects to a
City of St. Petersburg storm sewer main. This storm sewer main ultimately discharges to Boca Ciega Bay.
The objectives of this project were: (1) to reduce Stormwater pollutant loading to Boca Ciega Bay by incorporating an in-
line sediment sump/oil and grease skimmer in the Sunset Drive storm sewer system before its junction with the larger
storm sewer main; and (2) to demonstrate innovative alternative approaches to treating Stormwater in highly urbanized
areas where land for traditional BMPs is scarce and very expensive. The sump was designed, to the extent possible, to
meet the current rule requirements for this type of system. Due to physical limitations, the design provided for the storm
sewer flow to be diverted to the area of an existing greenspace for treatment, prior to being diverted back to the main flow
path of the storm sewer. The greenspace, which is adjacent to the bay, was modified into an open, linear wet-sump,
which included energy dissipaters and a skimmer baffle. The project also included an attractive boardwalk around and
over the facility as well as plantings of salt marsh vegetation in the sump's littoral zone.
Educational Component: The architecture and location of the boardwalk serves to attract pedestrian traffic to the
project. Being located immediately in front of City Hall provides an excellent high-profile example of how local govern-
ment can cooperatively implement measures to reduce Stormwater pollution. Several interpretive signs provide informa-
tion regarding nearshore aquatic plants and animals and the value of Stormwater treatment.
Project Cost: Total cost was $115,000 with construction costing $83,131. Funding was provided by a Section 319 NPS
Grant from DER, the SWFWMD SWIM Program, and the City of South Pasadena. A grant from the Tampa Bay National
Estuary Program paid for the educational signs.
Project Evaluation: The project provides an opportunity to trap and retain sediment and other suspended materials as
small as 0.1 mm in diameter. A corresponding reduction in other urban pollutants typically associated with suspended
solids such as heavy metals, bacteria, and oxygen demanding substances can also be expected. The sediment load
reduction to Boca Ciega Bay is estimated to be approximately 24.5 cubic yards per year.
175
-------
Urban Stormwater Retrofitting Project Fact Sheet
EMS Stormwater Enhancement Project
Pinellas County, Florida
Watershed Area: 9.24 acres
Watershed Land Use: 9.24 acres - Mixed Use (85% Impervious)
Project Overview: The original Stormwater facility was constructed in accordance with regulations in 1990 to provide
Stormwater treatment and peak attenuation for the county's new Emergency Medical Services (EMS) complex. The
facility discharges indirectly into Boca Ciega Bay. The pond was designed to capture Stormwater and treat, using a sand
filter encased in a concrete vault, the first half-inch of runoff from the entire site. The facility was constructed with 4:1 side
slopes, 2 foot average water depth, and a 0.4 foot treatment prism for capturing and filtering runoff. Priortothe enhance-
ments, a monoculture of primrose willow dominated the entire perimeter of the pond.
The primary objective of this project was to demonstrate how Stormwater ponds can be designed to enhance their
aesthetic and wildlife habitat values while at the same time meeting their intended water quality treatment and/or flood
control purposes. The secondary objective was to actually improve the treatment effectiveness of the existing pond by
expanding and planting the pond's littoral zone, increasing the treatment volume between the control elevation and
overflow weir, and increasing the permanent pool volume, thereby increasing the residence time in the pond.
Educational Component: Due to the adjacent location of the County's Cooperative Extension Service, the project is
readily available for touring by anyone visiting the Extension Service. Educational display boxes at various locations
along the mulched path surrounding the pond provide information regarding the importance and function of Stormwater
treatment facilities. Also, as part of the project, the Extension Service produced a 28 minute educational video entitled
"Stormwater Ponds: The new Urban Wetlands." While the video discusses the importance of treating Stormwater, it
focuses primarily on the potential value of Stormwater ponds for providing improved urban wildlife habitat. The video is
used to inform groups such as homeowner associations, condominium associations, and civic associations, about
Stormwater pollution and management.
Project Cost: Total cost was $78,500, with construction costing $63,244. Funding was provided by a Section 319 NPS
Grant from DER, the SWFWMD SWIM Program, and Pinellas County.
Project Evaluation: By more than doubling the permanent pool volume of the pond, the pond's residence time was
substantially increased. The pond's treatment volume also was increased by 13.4%, from 0.50 inches of runoff to 0.57
inches. The increased residence time allows for longer periods of physical settling as well as biological activity. The
reshaping and replanting of the littoral shelf resulted in increased nutrient uptake.
176
-------
Urban Stormwater Retrofitting Project Fact Sheet
Jungle Lake Water Quality and Habitat Enhancement
Southwest Florida Water Management District
Watershed Area: 1000 acres
Watershed Land Use:
Project Overview: Walter Fuller Park is a highly used recreational/athletic park located in the western part of the city of
St. Petersburg, approximately 2.5 miles east of Boca Ciega Bay. Jungle Lake was excavated about 75 years ago to
provide fill for the construction of local roads. The 11.2 acre kidney-shaped lake received untreated stormwater from five
inflows and discharges to the bay via a single outflow. During most storms, runoff bypassed Jungle Lake and was
discharged directly to the bay. To improve the quality of water in the lake and that which is discharged to the bay, a BMP
treatment train was constructed. The system includes:
• A diversion weir so that most stormwater is routed into the lake for treatment instead of directly into the bay.
• Modification of the inflow ditches to create shallow sloughs vegetated with native aquatic macrophytes.
• Expansion of the lake to create littoral zones vegetated with macrophytes.
• Two partially submerged berms which produce a longer flow path, increase residence time, provide natural habitat,
and replace park uplands resulting from the lake perimeter modifications.
• Sediment sumps at the northeastern and southeastern inflows.
• An oil and grease skimmer on the outfall structure.
• Over 15,000 herbaceous plants consisting of 11 species, 170 trees, and 700 shrubs.
Project Cost: $328,000, which included $59,000 from the City of St. Petersburg and $269,000 from the SWFWMD
SWIM Program. About 51,000 cubic yards of fill were needed for the project, of which over half would need to be
imported at a cost of $3.80 to $4.75 a cubic yard. Since the area northeast of the lake was three feet higher than the
surrounding roads and fields, this area was excavated instead. Within this area, two soccer fields were designed and
constructed to provide the community with additional recreational facilities and to promote park usage. Even after the
sodding of the soccer fields and the installation of an irrigation system, over $35,000 was saved.
Educational Component: During the conceptual planning and design phases, the City and SWIM staff met with mem-
bers of the Jungle Lake Civic Association to obtain their input and to extend their ownership of the park to include the
stormwater improvements. The members assisted in the selection of plants and received a grant to supplement the
wetland and forest plantings. They also assisted in planting the vegetation and are participating in the educational,
maintenance, and monitoring aspects of the project. The site has on display eight educational displays that inform the
general public and students about stormwater issues and management. A teacher's manual was produced that can be
used in the classroom or to accompany the signs during school field trips.
177
-------
Urban Stormwater Retrofitting Project Fact Sheet
Oleander Avenue Stormwater Exfiltration Trench System
City of Daytona Beach, Florida
Watershed Area: 49 acres
Watershed Land Use: Single Family Residential -
23% Directly Connected Impervious Area (DCIA)
Project Overview: The Oleander Avenue watershed historically discharged untreated runoff to storm sewers that ulti-
mately discharged to the Halifax River. The area was also subject to periodic local flooding due to the inadequate
capacity of the conveyances. The primary objective of this project is to demonstrate the cost-effectiveness of using
exfiltration systems as a method of retrofitting Stormwater problem areas for future use within the city's beachside
community.
To alleviate the flooding problem and to reduce pollutant loading to the river, a perforated pipe exfiltration trench treat-
ment system was constructed. Site constraints limited the treatment volume to 0.75 inches over the DCIA which trans-
lates into a storage volume of 30,700 cubic feet. The 294 feet of exfiltration system is designed to accept the runoff from
a 5 year, 24 hour storm representing flows of from 1.5 to 17.5 cfs from the drainage area subbasins. Actual pipe sizes
varied from 19" x 30" to 29" x 45" to meet the design storm flow conditions. The rock filled trench measures 16 feet in
width and 2 feet in depth.
Project Cost: Total cost was $513,700 with construction costing $375,617. This represents a cost of approximately
$10,200 per acre. Funding was provided by DEP from a State Stormwater Demonstration Grant and from the City of
Daytona Beach.
Project Evaluation: The exfiltration trench appears to be functioning very well as water quality monitoring efforts have
failed to find any discharge from the system. Since exfiltration systems provide 100% treatment for all water which is
retained and exfiltated, this system will reduce the Stormwater pollutant loadings discharged to surface waters by at least
80%, since the trenches will eliminate the discharge from over 80% of the storms that occur. The project allowed the city
to identify the design and construction constraints associated with this type of treatment system as well as installation
costs for these systems. This knowledge will be used as the city retrofits other basins.
178
-------
Urban Stormwater Retrofitting Project Fact Sheet
Indian River Lagoon Baffle Boxes
Brevard County Surface Water Management
Project Overview: The Indian River Lagoon National Estuary Program identified stormwater discharges as the major
factor in the decline in the lagoon's health. In particular, reductions in the stormwater loadings of total suspended solids,
nutrients, and freshwater are needed to restore the lagoon. The county developed an innovative BMP, the baffle box,
which can be installed within existing rights-of-way as a way of retrofitting stormwater discharges where land is unavail-
able for traditional BMPs. Baffle boxes are large sediment traps that require regular maintenance. Sediment accumula-
tion rates vary depending on site characteristics such as drainage area, land use, soil type, slope, mowing frequency,
and base flow. The boxes accumulate from 500 to 50,000 pounds per month, and requires monthly cleaning in the wet
season and cleaning every two to three months in the dry season. By the end of 1997, the county had installed 31 baffle
boxes, with others under construction. As part of the implementation of the Indialantic area stormwater master plan, 11
baffle boxes currently are being installed and monitored. Three different designs are being evaluated to determine their
effectiveness including: (1) a two-chamber box for small pipes and drainage areas; (2) a three-chamber box for larger
pipes; and (3) two boxes in series, where one box currently exists and collects large amounts of sediment.
Project Drainage Area and Cost: The average cost of installing a baffle box is around $22,000 and the average clean
out cost is $450 (by private contractor). Funding from a Section 319 NPS Grant from DEP and from the County's
stormwater utility are paying for the Indialantic projects. These all serve mainly residential land uses. The construction
costs and watershed drainage area are summarized below:
Project
Drainage area
Cost
Project
Drainage area
Cost
Alamanda
Rivershore
Indialantic I
Monaco
Pinetree
1 .8 acres
7.2 acres
25 acres
54 acres
1 34 acres
$14,376
$9,463
$13,580
$32,835
$33,925
Franklin (2)
Riverside
Sunset Part
Puesta Del
Cedar Lane
36 acres
161 acres
24 acres
2.2 acres
0.9 acres
$33,362
$24,944
$23,422
$25,181
$25,027
Project Evaluation: The monitoring program for the 11 new baffle boxes will not begin until the spring of 1998. However,
previous assessments of the effectiveness of baffle boxes on 22 existing systems is shown below:
The county has also installed a continuous deflective separation unit, a new BMP from CDS Technologies of Australia.
This unit cost $55,000 to install and treats the runoff from a 40 acre watershed. This unit captures 100% offloatablesand
has been cleaned out twice resulting in the removal of 8,013 pounds of sediment.
179
-------
Urban Stormwater Retrofitting Project Fact Sheet
Oil and Grease Removal BMP Demonstration
City of Oakland Park, Florida
Watershed Area: 5 acres
Watershed Land Use: Mixed commercial and industrial (95% Impervious)
Project Overview: The City of Oakland Park received one of the state's Stormwater Demonstration Grants to develop
and monitor a prototype BMP for in-line removal of oil and grease from Stormwater using oil absorbent material. The
Northeast 40th Court site was chosen because inspection of the storm sewer system revealed substantial amounts of oil
and grease. These were attributed to the large number of automobile repair shops, paint shops, plating shops, and
similar businesses in the drainage area. The project consisted of characterizing the concentrations of oil and grease in
the Stormwater, a review of the material safety data sheets of three different oil sorbent materials, a laboratory bench
scale study of one of the oil sorbent materials, construction of the BMP system, and effectiveness monitoring. The final
BMP system included diversion box with a weir to direct runoff into the treatment system. As Stormwater enters the
treatment unit, flow is directed against an aluminum baffle imparting a slight rolling motion which causes floatables and
trash to be trapped against the baffle wall for easy removal. Upon entering the treatment chamber, velocity slows greatly,
allowing grit, sludge, and oil particulate matter to settle to the sloping bottom. The Stormwater is then redirected upward
through two cross-layers of the absorbent media, which are secured by being sandwiched between two aluminum
grates, where free oil and grease are removed via absorption into the material. The absorbent media chosen was
custom made by NewPig Corporation of Tipton, Pennsylvania. The product, called the Spaghetti Pillow, consists of
shredded strips of polypropylene packaged in tough, UV resistant mesh skin in the shape of a rectangular bag or pillow.
The two layers of media are placed perpendicular to each other to avoid short circuiting.
Project Cost: Total cost of the project was $260,870. This included $71,490 for the construction of the treatment system
and $189,380 for sampling equipment, consultant, and laboratory fees.
Project Evaluation: Inflow and outflow sampling of the system was conducted for ten storms between July 1994, and
April 1995. Storm event oil and grease concentrations ranged from 0 to 261 mg/l, with mean pollutant concentrations
ranging from 1.41 to 85.58 mg/l. Oil and grease mass removal efficiencies ranged from 71 % to 95%, while flows ranged
from Oto 1.75 cfs.The absorbtion efficiency of the filter media bags were measured twice. The amount of oil and grease
absorbed ranged from 1.7 pounds to 62.5 pounds, which represents an absorbtion efficiency of 110% to 470%.
180
-------
Urban Stormwater Retrofitting Project Fact Sheet
BMP Treatment Train in the Florida Keys
City of Key Colony Beach, Florida
Watershed Area: 268 acres
Watershed Land Use:
Project Overview: Recognizing the importance of reducing stormwater pollution in protecting its sensitive natural re-
sources, the City included in its comprehensive plan policies requiring the retrofitting of its existing drainage system.
With technical assistance from the DEP and the SFWMD, the City's consultant developed a stormwater master plan in
1993. The plan included plugging 28 existing stormwater outfalls and constructing a retention basin and swales with
raised inlets and exfiltration trenches which overflow into injection wells. Implementation of the master plan began in
1995, and is scheduled for completion by the year 2000. Phase 1 has been completed and Phase 2 will be completed by
the fall of 1998. The stormwater master plan calls for the construction of 82,146 linear feet of swales, 9 modified raised
swale inlets, about 60,000 linear feet of exfiltration trench, 35 inlet baffle systems to direct the first flush into the exfiltration
trenches, and 22 injection wells.
Project Cost: The total cost of the original stormwater retrofitting master plan was estimated to be $1.2 million. However,
the city's residents and elected officials decided that they did not want water standing in the swales, resulting in the
addition of the exfiltration trench system. To date, using funds from the city, the DEP, and a Section 319 Grant, the city
has implemented two phases of the master plan as shown below:
Basin Acres Swale (If) Sod (sF) Exf. Trench (If) Injection Wells Cost
4-1
5-2
2-2
2-5
2-8
2-11
3-1
5-1
8-2
0.66
3.03
3.50
3.00
2.13
1.76
3.69
26.47
0.02
827
521
1200
1200
934
566
878
4800
1306
29,257
21,304
14,000
14,000
11,000
11,000
37,000
371
445
269
1200
1200
934
878
3100
371
1
0
1
1
1
1
1
3
1
$72,200
$47,083
$148,112
$147,790
$129,600
$27,854
$113,175
$439,773
$72,174
Project Evaluation: Actual stormwater monitoring will not begin until the completion of Phase 2. By plugging the direct
stormwater discharges to surface waters and providing storage and treatment for the first 1.5 inches of runoff, the
stormwater volume and pollutant loadings will be substantially reduced. Modeling indicates that these will be reduced by
up to 75% from pre-project conditions.
181
-------
References
Camp, Dresser, and McKee. 1995. Final Report on the
City of Oakland Park Stormwater Demonstration Grant
Project. Submitted to Florida Department of Environmen-
tal Protection, Stormwater/NPS Management Section,
Tallahassee, FL.
City of Orlando Stormwater Utility Bureau. 1995. Final
Report on the Packed Bed Wetland StormwaterTreatment
System. Submitted to Florida Department of Environmen-
tal Regulation, Stormwater/NPS Management Section, Tal-
lahassee, FL.
England, G. 1997. Maintenance of Stormwater Retrofit
Projects. Pages 169-177, Proceedings of the Fifth
Biennial Stormwater Research Conference.
SWFWMD, Brooksville, Florida.
Greiner Engineering. 1993. City of Key Colony Beach
Stormwater Master Plan (Stormwater Retrofit Project).
Harper, H. 1990. Final Report on the Long Term Perfor-
mance of the Alum Stormwater Treatment System at
Lake Ella, Florida. Submitted to Florida Department of
Environmental Regulation, Stormwater/NPS Manage-
ment Section, Tallahassee, FL.
LaRock, P. 1988. Final Report on the Evaluation of the
Lake Jackson StormwaterTreatment Facility. Submit-
ted to Florida Department of Environmental Regula-
tion, Stormwater/NPS Management Section, Tallahas-
see, FL.
Macrina, J. J. and D. M. Vickstorm. 1985. Jungle Lake
Water Quality and Habitat Enhancement. Proceedings
of the Fourth Biennial Stormwater Research Confer-
ence. SWFWMD, Brooksville, FL.
McCann, K. and L. Olson, Orlando Stormwater Utility. 1994.
Final report on Greenwood Urban Wetland Treat-
ment Effectiveness. Submitted to Florida Department
of Environmental Protection, Stormwater/NPS Man-
agement Section, Tallahassee, FL.
Northwest Florida Water Management District. 1984. Fi-
nal Construction Report - Lake Jackson Clean Lakes
Restoration Project. Submitted to Florida Department
of Environmental Regulation, Bureau of Operations,
Tallahassee, FL.
Northwest Florida Water Management District. 1990. Fi-
nal Report on the Expansion of the Lake Jackson
Stormwater Treatment Facility. Submitted to Florida
Department of Environmental Regulation, Stormwater/
NPS Management Section, Tallahassee, FL.
Northwest Florida Water Management District. 1992. Lake
Jackson Regional Stormwater Retrofit Plan. Water
Resources Special Report 92-1. Havana, FL.
182
-------
Evaluating the Cost Effectiveness of Retrofitting an
Urban Flood Control Detention Basin for Stormwater Treatment
Peter Mangarella
Woodward-Clyde
Oakland, California
David Drury
Santa Clara Valley Water District
San Jose, California
Chee Chow Lee
Environmental Technology Institute,
Nanyang Technical University, Singapore
Richard Mattison
Kinnetic Laboratories Inc.
Santa Cruz, California
This paper describes the retrofitting of a flood control
basin in Sunnyvale, California and subsequent monitoring
to evaluate the pollutant removal effectiveness of the ret-
rofitted basin. The authors wish to thank the Santa Clara
Valley Urban Runoff Pollution Prevention Program, and
especially the City of Sunnyvale for their support and co-
operation in the conduct of this study.
Background
The northern portion of Santa Clara Valley experienced
significant subsidence during a period of excessive ground-
water pumping. In order to protect that area from flooding,
a system of levees and pump stations were built. Accord-
ing to a survey conducted in 1990, there are 17 munici-
pally owned and operated pump stations in Santa Clara
Valley (Woodward Clyde Consulting, 1990). These pump
stations generally consist of pumps, storage units such as
a sump or a detention basin, and inlet and outlet works.
Sumps and detention basins are designed to reduce the
capacity of the pumps that would otherwise be needed to
pass peak flood flows. These pump stations have gener-
ally been operated as single-purpose flood control facili-
ties. The pump operating schedules are designed such
that the pumps go on as soon as water begins to fill the
basin, with the goal of emptying the basin as soon as pos-
sible after the event. These facilities were examined for
their potential to provide water quality treatment in addi-
tion to flood control. One retrofitting option to achieve wa-
ter quality benefits would be to change the pump operat-
ing schedule in order to increase detention time and to
provide for a seasonal wet pond. Based on a preliminary
evaluation of the feasibility of retrofitting detention basins,
the Santa Clara Valley Urban Runoff Pollution Prevention
Program decided to conduct a pilot study to retrofit a facil-
ity, and conduct testing to measure water quality benefits
and costs (Woodward Clyde Consulting, 1994).
Site Description
In this pilot study, structural and operational retrofitting
was conducted on Sunnyvale Pump Station No. 2 located
just north of the junction of Route 237 and Calabazas
Creek. The pump station consists of four primary pumps
rated at 39 cfs capacity and one auxiliary electric pump (9
cfs). The detention basin area is 4.4 acres with a 30-acre-
ft capacity; it receives water from a 463-acre watershed
that consists of industrial park (30%), commercial (10%),
and residential (60%) land uses (Figure 1).
Retrofitting Actions
The basin, originally constructed as an in-line dry deten-
tion basin with pumped outflow, was retrofitted to operate
as an in-line extended detention basin with a seasonal wet
pool and pumped outflow. The retrofitting required one
operational and three structural changes. The detention
basin has an open channel and a submerged 36-inch pipe
(which was below the open channel) that connected the
inlet and outlet. In orderto minimize short-circuiting, a single
barrier of rock was placed in the channel and a riser was
placed over the entrance to the 36-inch pipe. A gabion weir
was installed at the outlet to provide better distribution of
flow from the basin into the outlet.
183
-------
I
-N-
I
—>• Direction of Flow
Elevations Relative to
Mean Sea Level
50
100
Water Quality Sampling Location
Sediment Sampling Location
Feet
1" = 70feet
Project No.
87201 15M
Santa Clara Valley Nonpoint
Source Pollution
Control Program
Woodward-Clyde Consultants
Figure 1. Sunnyvale Pump Station No. 2 showing sampling locations.
Operational changes consisted of modifying the pump
schedule to create a 2-foot permanent pool at the outlet,
provide a temporary pool over the depth range of 2 to 2.4
feet, and provide slow release for water depths above 2.4
feet. The 2.4-foot maximum for the temporary pool was
estimated based on a flooding analysis to ensure that
modifications to the pumping schedule did not cause a sig-
nificant increase in the 100-year flood levels in the deten-
tion basin.
Monitoring
Flow data and water quality samples were obtained us-
ing automated samplers located in the pump station and
in the pipe entering the basin. Flow-weighted composite
samples were obtained from eight storm events over three
wet seasons (October through April) between March 1991
and April 1993. Characteristics of the monitored storms
are given in Table 1.
Water sampling was conducted for total suspended sol-
ids (TSS), selected heavy metals (total and dissolved cad-
mium, chromium, copper, lead, nickel, and zinc), and oil/
grease. Table 2 shows how the metals concentrations at
the inlet to the basin compare with data collected from other
urban monitoring stations that sampled relatively single-
land use catchments. Data from the detention basin inlet
were most similar to residential/commercial data. Two
rounds of sediment samples were also collected at three
locations in the basin (near inlet, midway between inlet
and outlet, and near outlet) to characterize sediment
grainsizse and chemistry in the basin.
No data were obtained prior to retrofitting the facility to
estimate "before and after" performance. However, based
on the design and operation of the facility, pre-retrofit treat-
ment performance was predicted to be quite low (Wood-
ward Clyde Consulting, 1990).
Monitoring Results and Pollutant Removal
Effectiveness
Because of difficulties in obtaining consistent and reli-
able flow measurements, pollutant removal effectiveness
was estimated based on the difference between influent
and effluent concentrations rather than loads (Table 3).
184
-------
Table 1. Storm Sampling Event Statistics for the Detention Basin
Storms
Rain (San Jose NWS gauge #7821)
Sampling
Event
SE17
SE20
SE21b1
SE23
SE24
SE25
SE27
SE28
Date
3/20/91
2/1 2/92
3/22/92
12/6/92
12/10/92
1/6/93
2/1 7/93
3/23/93
Duration
(hours)
11
17
19
16
6
29
60
14
Volume
(in)
0.4
1.3
0.6
1.1
0.9
0.5
2.2
1.1
Average
Intensity
(in/hour)
0.04
0.08
0.13
0.07
0.15
0.02
0.04
0.08
Peak
Intensity
in/hour)
0.10
0.20
0.20
0.20
0.30
0.10
0.20
0.20
Antecedent
Dry Period
(days)
2
3
1.3
39
4
5
6
7
1 Consists of two small storms.
Note: The median event volume for the period from 1948 to 1989 was 0.49 inches.
Table 2. Comparison of Median Total Metal Concentrations in Stormwater at the Detention Basin Inlet to other Santa Clara County Monitoring
Stations
Concentrations (|ig/L)
Parameter
Cadmium
Chromium
Copper
Lead
Nickel
Zinc
Detention Basin
Inlet Station
n = 8
1.1
11.5
24.0
37.5
20.5
180.0
Residential -
Commercial
Land Use Stations
(L1 , L3, L4, L5, L6)
n = 21
1.0
16.0
33.0
45.0
30.0
240.0
Industrial
Land Use Station
(L2)
n = 25
3.9
24.0
50.5
90.5
46.0
1150.0
Open
Land Use Station
(L7)
n = 4
0.3
10.5
11.0
2.0
5.0
5.0
Average removal of TSS was 50%. Removal of total chro-
mium averaged about 30%, removal of total copper and
zinc was about 40%, and removal of total lead and total
nickel averaged about 50%. The data indicated that the
basin did not remove dissolved metals or hydrocarbons.
Estimates for removal of total cadmium were not made
because the concentrations in the influent and effluent were
very low (at or near the laboratory detection limit).
Sediment samples were also collected at various loca-
tions within the basin (see Figure 1) were tested for se-
lected heavy metals (Table 4). Sediment concentrations
were higher near the inlet, but none of the samples con-
tained metals at concentrations exceeding hazardous
waste criteria. Estimated sediment accumulation rates were
low, as expected for a fully urbanized area, with cleanout
frequencies estimated at between once every 10 or 20
years.
Cost-Effectiveness Evaluation
The amortized cost over 20 years for retrofitting, opera-
tions and maintenance (including sediment disposal) was
estimated at $8,200 per year. Based on this cost and as-
suming flow rates were typical, the cost effectiveness of
removing the metals was estimated. For example we esti-
mated that, 1.1 Ib. of copper could be removed per$1,000
spent on retrofitting. This compares well with an estimate
of 1.5 Ib. of copper for $1,000 spent on street sweeping.
However, the potential for significantly reducing heavy
metals loads to the Bay due to retrofitting the existing flood
control facilities is minimal because only a small portion of
the watershed is served by detention facilities. Even if the
other existing flood control facilities were retrofitted, and
achieved improved removals comparable to those mea-
sured in the pilot study, the net reduction in copper would
only amount to approximately 100 pounds per year, which
is less than 1% of the estimated mean annual copper load
to the Bay (14,000 Ib.).
Conclusions
• Flood control basins, especially those with pumped
outlets, may be good candidates for retrofitting for water
quality control without increasing flood control risk.
• Metal removals measured in the retrofitted basin
ranged between about 30-50% depending on the metal
and about 50% for TSS. The basin did not appear to
be effective in removing oil/grease.
• Concentrations of metals in the sediments tended to
be higher near the inlet, but were well below hazard-
ous waste criteria.
• Amortized costs for retrofitting the basin were about
$8,200/year, and based on the pollutant removal per-
185
-------
Table 3. Summary of Inlet and Outlet Concentrations for Selected Pollutants at the Detention Basin
SE17
B1 - Inlet
B2 - Outlet
Reduction
Cadmium (|ig/L)
Total Dissolved
0.4 <0.2
0.2 <0.2
-
Chromium (|ig/L)
Total Dissolved
3.6 1.8
2.7 1.1
25%
Copper (|ig/L)
Total Dissolved
8.7 5.4
6.8 4.7
22%
Lead (|ig/L)
Total Dissolved
6.4 2.2
3.4 1
47%
Nickel (|ig/L)
Total Dissolved
1.7 <2
1.7 <2
0%*
Zinc (|ig/L)
Total Dissolved
46 28
26 19
43%
TSS
(mg/L)
12
7.3
39%
TH
(mg/L)
97
120
-
TO&G
(mg/L)
1.5
1.4
7%
SE20
B1 - Inlet 6.6 1.3 12 1 24 3 45 1 16 1 180 19 90 110 0.2
B2-Outlet 4.8 2.5 61 9 3 10 1 4 1 73 22 24 63 <0.2
Reduction - - 50% 0 63% - 78% - 75% - 59% - 73%
SE21b
B1-Inlet 1.1 0.2 18 1 24 2 53 <1 25 <1 180 5 140 - -
B2-Outlet 1.5 <0.2 14 1 16 2 35 <1 19 <1 120 7 93
Reduction -* - 22% - 33% - 34% - 24% - 33% - 34%
SE23
B1-Inlet 1 0.2 11 <1 27 5 30 1 13 3.9 190 41 74 100 0.7
B2-Outlet 0.6 <0.2 8.3 1.4 12 4.7 12 <1 5.8 2.2 82 45 31 90 0.5
Reduction - - 25% - 56% - 60% - 55% - 57% - 58%
SE24
B1 - Inlet 1.6 <0.2 21 1.1 40 2.1 76 <1 42 9.6 270 22 180 140 0.6
B2-Outlet 1.3 0.2 15 8.6 24 5 40 1.4 29 15 160 31 96 140 3.5
Reduction - - 29% - 40% - 47% - 31% - 41% - 47%
SE27
B1 - Inlet 1 0.5 6.3 1.4 14 5.4 13 <1 83 63 70 35 30 110 1.6
B2-Outlet 0.6 0.4 4.9 1.7 8.9 4.5 6.6 <1 25 20 47 26 15 220 1.3
Reduction - - 22% - 36% - 49% - 70% - 33% - 50%
Average
Reduction - - 29% - 42% - 53% - 51% - 44% - 50%
186
-------
c
QJ
C
g
c
'S
Q
>
o
CL
o
o
^~ r^ CD
o) oo r^
000
T- •* 00
to to to
O O 00
888
o CD r^
O) oo r^
•* co •*
in •* •*
goo
C\l C\l T-
co r^ in
OM T- CO
oo in o
co in T-^
CM T- TJ-
CO CO CO
i*^ to to
C\l T- OM
CO ^ -^'
in in csi
o o o
ro ro ro
'o 'o 'o
t t t
tn tn tn
III
in in in
to ~
^ S ^
T- CO C\l
Q Q Q
(D in co
000
CD o r^
in r^ in
O C\l f
CD •»- CN
OM
888
o o o
TJ-" CD" o"
CO CO CO
o r^ co
T- CO f
000
O C\l CO
•^ o o
in co
^" CD O)
c\i d d
CO CO CO
LO CD CD
i i i
, , ,
, , ,
ro ro ro
'o 'o 'o
t t t
tn tn tn
CO CO CO
O) O) O)
CN CN CN
r-i-r-
to ~
'S -S =
T- CO C\l
Q Q Q
0
O
O
C\l
t
o
8
,
o
o
in
OM
8
in
C\l
o
o
c
g
'ra
'c
c
o
0
E
"o
f—
OJ
1
1
"ro
is
1
ro
c
ra
<" c\l
2 (D
O =
0 i-
-1—*
187
-------
formance, yielded cost-effectiveness values that were for Pollution Control Potential. Prepared for Santa Clara
somewhat comparable to street sweeping. Valley Nonpoint Source Pollution Control Program.
July.
• Because of the limited number of basins and the small
portion of the watershed served by those basins, ba- Woodward-Clyde Consultants. 1994. Sunnyvale Detention
sin retrofitting would reduce watershed loads of met- Basin Demonstration Project. Prepared for Santa Clara
als by only about 1%. Valley Nonpoint Source Pollution Control Program.
November.
References
Woodward-Clyde Consultants. 1990. Evaluation of Exist-
ing Stormwater Pump Stations in Santa Clara County
188
-------
Retrofitting to Protect Drinking Water Reservoirs from the
Impacts of Urban Runoff
James D. Benson and Melissa Beristain
New York City Department of Environmental Protection
Valhalla, New York
To meet federal Surface Water Treatment Rule (SWTR)
requirements and Filtration Avoidance mandates, the New
York City Department of Environmental Protection has
developed a proactive program to manage and protect the
Kensico Reservoir and its watershed. The prime compo-
nents of the program are aggressive stormwater and wa-
terfowl management, sewer system inspection and repair,
an in-reservoir turbidity curtain, reservoir dredging, and
hazardous spill containment. Protecting water quality in
the Kensico Reservoir is imperative because it is the final
impoundment for 90% of the city's unfiltered water supply
before it enters the distribution system. This paper focuses
on the stormwater management element, which targets
pathogens and turbidity, the key pollutants regulated by
the SWTR that are conveyed to the reservoir by stormwater.
The first phase of the project, watershed assessment, site
selection, and stormwater management facility screening
and design, is complete. Lessons learned and recommen-
dations for planning similar efforts are summarized in this
paper. Construction of the stormwater facilities is sched-
uled to begin in the spring of 1998 and to be completed in
two phases, over a five-year period. Baseline stormwater
water quality data will be collected until the facilities are
constructed. Data will be collected from select stormwater
facilities once they are operational in order to assess the
effectiveness of the program.
Introduction
New York City has placed great emphasis on protecting
and improving the quality of its drinking water supply
through watershed protection and management programs.
This paper describes one such program developed and
implemented by the New York City Department of Envi-
ronmental Protection's (NYCDEP) Bureau of Water Sup-
ply, Quality and Protection.
The city's drinking water supply system is one of the
largest in the world, supplying 1.45 billion gallons of water
each day to 9 million city and upstate residents. The entire
watershed covers 1,969 square miles and comprises 19
reservoirs and three controlled lakes (lakes in which the
city has water ownership), and numerous wetlands water-
courses and intermittent streams. Land use, topography,
hydrology and political climates vary dramatically within
and among the system's three watersheds: the Delaware,
the Catskill and the Croton (Figure 1). One reservoir, the
Kensico, is integral to managing the unfiltered Catskill and
Delaware systems because it serves as the final impound-
ment for Catskill and Delaware water before it enters the
distribution system. On average, approximately 1.3 billion
gallons flow through the Kensico Reservoir each day. This
accounts for 90% of the system's daily demand. For this
reason, it is important to control the quality of stormwater
entering the reservoir from developed land.
The U.S. Environmental Protection Agency recognizes
the importance of the Kensico Reservoir and has required
the NYCDEP to implement a watershed management and
protection plan that targets fecal coliform bacteria and tur-
bidity. Plan elements include aggressive stormwater and
waterfowl management programs, sewer system inspec-
tion and repair, an in-reservoir turbidity curtain, and haz-
ardous spill containment. This paper focuses solely on the
stormwater management element which includes hazard-
ous spill containment. Brief descriptions of the other com-
ponents follow.
Sewer Inspection and Repair
The sewer system within the watershed, including type
and size of pipe and manhole locations, was mapped. Of
the 95,000 feet of sewer line in the watershed, 55,000 feet
were installed before 1970 and are more prone to defects.
The older sections of sewer line were inspected for poten-
tial sources of exfiltration, and cross or illicit connections.
No illicit connections were discovered; 39 segments and
three manholes were found to be in need of repair. The
town of Mount Pleasant and Westchester County are com-
pleting the repairs under intermunicipal agreements.
Waterfowl Management
The waterfowl management program is designed to elimi-
nate or reduce the numbers of geese and gulls roosting
and defecating in or near the surface water through haz-
ing, using noisemakers, motorboats, hovercraft, and bird
distress tapes; shoreline meadow management and physi-
cal barriers; and Canada geese egg depredation. The pro-
gram, implemented August 1 through March 31, also in-
189
-------
Rensselaer
County
Schohane
County
Chenango
County
Columbia
County
Out chess
County
Catskill/Delaware Watershed
Croton Watershed
Rivers and Reservoirs
Catskill Aqueduct
Croton Aqueduct
Delaware Aqueduct
Shandaken Tunnel
Distribution System
1 Suffork
Nassau !, CountV
ens County
10
10
20
30 Miles
DSP
Figure 1. New York City's water supply system.
eludes research into new methods of bird control and on-
going assessments of program effectiveness. Although
labor intensive, the waterfowl management program is a
permanent program because it eliminates the greatest
source of fecal coliform bacteria.
Turbidity Curtain
The turbidity curtain (like a silt fence 750 feet long with
floats on top and weights on the bottom) installed at the
mouth of Malcolm Brook in the southwest section of the
Kensico Reservoir successfully directs turbidity and fecal
coliform bacteria conveyed to the reservoir away from the
Catskill Upper Effluent Chamber. Maintaining high-quality
water in the effluent chamber is critical as water is con-
veyed directly to the distribution system from the cham-
ber. Water entering the chamber is constantly monitored
to determine compliance with the SWTR. Since the cur-
tain has been so effective, it is a permanent program.
Reservoir Dredging
The channels leading to the reservoir's two effluent cham-
bers, and the sediment deltas at the mouths of Malcolm
190
-------
and Young Brooks will be dredged in 1998. Dredging will
eliminate the potential for accumulated sediments to be
resuspended during storms, potentially causing a contra-
vention of turbidity and fecal coliform bacteria water qual-
ity standards.
The Kensico Reservoir Stormwater Management Project
is designed to reduce fecal coliform bacteria and turbidity
delivered to the reservoir by controlling and treating
stormwater. As turbidity is a direct measure of suspended
solids in the water column, the stormwater component tar-
gets sources of suspended solids. The first phases of the
project, assessment of the watershed, site selection, and
the screening and design of stormwater control and treat-
ment facilities, were completed in December 1997. Fund-
ing is in place to construct the stormwater facilities over a
5-year period, beginning in March 1998. The NYCDEP has
committed to maintaining, monitoring, and evaluating the
performance of the facilities.
Phase I: Watershed Assessment
The Kensico Reservoir watershed occupies approxi-
mately 13 square miles and includes four suburban towns
in Westchester County, New York, plus a small portion in
Fairfield County, Connecticut (Figure 2). To assess
stormwater pollutant loading in the Kensico watershed, the
reservoir basin's physical characteristics, including land
use, soils, topography, vegetation and reservoir tributar-
ies, were inventoried and digitally mapped. The watershed's
topography is hilly and rolling, and overtwo-thirds of it con-
tains slopes greater than 8%. Almost one-third of the land
area is used as passive open space, and approximately
one-fifth of the land area is developed with low-density
residential uses (Figure 2). The remaining land area is pri-
marily active open space, farmland and commercial/busi-
ness. As water quality is, in part, a function of the amount
of impervious surfaces in the watershed, of greatest con-
cern is developed land directly adjacent to the effluent
chambers that convey drinking water to the consumers.
Phase II: Stormwater Remediation Needs
Assessment and Management Plan
Development
A preliminary assessment of stormwater remediation
needs in the Kensico watershed was conducted by evalu-
ating tributary water quality data, land use/impervious sur-
faces, SWMM model predictions of runoff quantity and qual-
ity, and field observations of existing erosion. That evalua-
tion concluded that 73 of the watershed's 148 sub-basins
have a relatively high potential to contribute fecal coliform
bacteria and suspended solids to the reservoir. Using the
criteria listed below, reservoir tributaries in 19 of the 73
sub-basins were prioritized for stormwater remediation.
Preliminary Stormwater Remediation Evaluation Criteria
• proximity to reservoir effluent chambers
• known or potential sources of pollutants
• quality and quantity of stormwater runoff
• presence of wetlands
• topography
• property ownership
• observed erosion
Based upon these criteria, conceptual designs were pre-
pared for 88 stormwater management facilities and ero-
sion controls within 19 sub-basins. The conceptual designs
were the basis for an environmental evaluation and im-
pact statement.
The conceptual stormwater management plan was then
refined by applying the selection criteria (bulleted below)
in combination with the results of detailed field investiga-
tions, maintenance requirements and site constraints. A
total of 57 stormwater management facilities were sited to
reduce erosion, manage peak stormwater flows, allow for
settling of sediments and coliform die-off, and ultimately
reduce pollutant loads delivered to the reservoir (Figure
3). During the process of developing preliminary facility
designs, property owners required that five facilities be re-
designed, and denied permission to construct the facilities
at three sites. Ultimately, 44 engineered designs were fi-
nalized. Facility types included 10 extended detention ba-
sins, 14 segments of stream channel stabilization, 13 sta-
bilized outlets, one area of parking lot stabilization, and
one sand filter system. Road stabilization and drainage
improvements were incorporated into stilling and deten-
tion basins and sand filter designs. Hazardous spill con-
tainment is being addressed in coordination with a major
road improvement project that will significantly alter drain-
age along the Interstate 684 and Route 120 corridors which
abut the reservoir. The conceptual plan includes four ex-
tended detention basins that will serve as spill contain-
ment facilities, and containment booms to be deployed at
the 22 storm drain outlets along I-684 in the event of a
spill.
Site Selection and Conceptual Facilities Evaluation Cri-
teria
• Do the site and the facility meet the intent of reducing
pollutant loads?
• Does the facility minimize impact to environmental re-
sources and achieve measurable water quality ben-
efits?
• Does the existing condition warrant engineered im-
provements?
• Are there property ownership/permission constraints
which make implementation impractical or impossible?
• Have any watershed/land use conditions or assump-
tions changed since issuance of the Final Environmen-
tal Impact Statement which affect the appropriateness
of the facility and/or the site?
• Are there likely to be permit issues which will compro-
mise the viability of the practice?
191
-------
Town Boundaries
Kensico Subbasin Boundaries
Kensico Reservoir
Streams
„• v Intermittent Streams
/\jf Perennial Streams
Land Use Classifications
Commercial/Retail
Public Assembly
Interior V\foter Body
Manufacturing/Warehouse
Mixed Use
Office
Active Open Space
Passive Open Space
Low-density Residential
Medium-density Residential
Transportation Corridor
Unclassified
Mount
Pleasant
Malcolm Brook
Subbasin
Catskill Upper
Effluent ._
Chamber
Shaft 18
Effluent
Chamber
2000 0 2000 4000 6000 Feet
Connecticut
Figure 2. Land use in the Kensico Reservoir Watershed.
192
-------
Stormwater Facility Types:
A Extended Detention Basin
-------
• Are the maintenance and/or operation requirements
of the practice, so burdensome as to effectively make
the practice inappropriate?
Having met the final screening criteria, each facility was
designed to minimize on-site and downstream impacts
without sacrificing water quality benefits. For example, the
designs incorporated existing topography, avoided wetland
encroachment, and incorporated emergent wetland fea-
tures necessary for long-term maintenance and features
to discourage waterfowl attraction. The attention given to
minimizing disturbances and subsequent on- and off-site
impacts was a crucial component of enlisting the support
of the community, regulatory agencies and private prop-
erty owners.
Phase III: Implementing the Stormwater
Management Plan in a Developed Area
Generating Community Support
The final plan proposes to construct 44 stormwater man-
agement facilities and erosion controls on public and pri-
vate property. Immediately after proposing the 88 concep-
tual facilities, NYCDEP identified the property owners, and
launched an outreach campaign to explain and generate
support for the project. The ultimate goal of the campaign
was to secure legal permission to gain access to design,
construct, and maintain the facilities on private property.
Securing permission to construct 18 facilities on private
land holdings from 32 land owners has been a challenging
aspect of the project. Alternate sites in the same sub-ba-
sin were pursued where access to private property was
denied during preliminary design development.
Obtaining Regulatory Approvals
An expert advisory panel was enlisted to review con-
ceptual plans and facility designs for the highest priority
sub-basin, Malcolm Brook, which discharges in the imme-
diate proximity of the Catskill Upper Effluent Chamber.
Panel members were volunteers from academia and gov-
ernment agencies that are actively involved in planning
and implementing stormwater management projects. The
panel's comments helped shape the plans and designs,
and were used to support certain aspects of the plan when
applying for regulatory approvals.
Municipal support for the project and regulatory approval
to construct the facilities were also needed. Initially, this
involved a series of explanatory meetings with the town
supervisors, engineers and planners. Once support forthe
conceptual project was obtained, the standard applications
for local permits and approvals were submitted. A similar
process of "pre-application" meetings was followed with
federal and state permitting agencies. The pre-application
meetings set the stage for the relationship between
NYCDEP, the municipalities and regulatory agencies, and
allowed the agencies to comment on the designs before
they were finalized and permit applications were submit-
ted. The goal of the pre-application process was to mini-
mize the need for design revisions and to avoid delays
during the regulatory approval process.
Modeling Water Quality Benefits
Waterquality modeling predictions can provide valuable
supporting information when developing stormwater man-
agement plans, if sufficient data are available. The U.S.
Environmental Protection Agency's Stormwater Manage-
ment Model (SWMM) was used to simulate runoff volumes
and turbidity and fecal coliform bacteria loading in select
tributary sub-basins of the Kensico Reservoir. The model
predicted pollutant loads under existing conditions and fu-
ture build-out conditions in the year 2010, with and without
the project. With the projected increase in impervious sur-
faces, results from modeling estimated that future
stormwater loads of turbidity and fecal coliform bacteria
inputs to the reservoir will increase by 16% and 21%, re-
spectively. Model predictions also estimated that construc-
tion and operation of stormwater facilities will reduce fu-
ture inputs of turbidity and fecal coliform bacteria by 23%
and 15%, respectively, when compared to future loads
without the stormwater controls. The overall effect of the
program on reservoir water quality will be less than the
benefit associated with the targeted tributaries. Thus, for
example, the model predicts much higher reductions in
turbidity at the discharge of Malcolm Brook, than for the
reservoir as a whole, 95% and 9.9%, respectively. Model
predictions of anticipated water quality benefits in individual
basins are listed in Table 1. In addition to the predicted
reductions, the extended detention basins will attenuate
peak rates of stormwater discharge and reduce peak con-
centrations of pollutants delivered to the reservoir. The
results predict that the plan will have substantial water
quality benefits.
Constructing, Operating, Maintaining and
Monitoring Facilities
NYCDEP recognizes the need for an aggressive main-
tenance program to ensure that the stormwater facilities
function as originally designed. Prior to construction, in-
spection and maintenance plans and contracts to carry
out the plans will be in place. Further, water quality moni-
toring stations have been incorporated into facility designs
and studies have been designed that will evaluate the per-
formance of the stormwater controls.
The stormwater plan will be implemented sub-basin by
sub-basin, with construction scheduled to begin in the
spring of 1998. The construction schedule was prioritized
using criteria that included severity of erosion, water qual-
ity benefits, proximity to the effluent chambers, and per-
mitting and property owner constraints.
Conclusions, Recommendations and
Challenges
The Kensico Reservoir watershed stormwater manage-
ment plan will improve water quality in the reservoir by
controlling and treating stormwater runoff in select tribu-
taries. An aggressive public outreach campaign, design-
ing the facilities to minimize site and resource disturbances,
and providing for proper maintenance of stormwater con-
trols and monitoring effectiveness, were high priorities for
194
-------
Table 1. SWMM Model Predictions of Tributary Loads of Turbidity and Fecal Coliform Bacteria
Future Load without Plan
Future Load with the Plan
Tributary
Sub-basin
Malcolm Brook
N1
N2
N3
N4
N5
N12
Bear Gutter Creek 5
Bear Gutter Creek 8
Whippoorwill
E11
Turbidity
5% increase
104% increase
9% increase
1 4% increase
103% increase
30% increase
107% increase
60% increase
76% increase
11% increase
0% increase
Fecal Coliform
Bacteria
6% increase
95% increase
1 1 % increase
10% increase
1 06% increase
23% increase
not modeled
59% increase
73% increase
not modeled
0% change
Turbidity
95% reduction
91% reduction
81 % reduction
63% reduction
90% reduction
84% reduction
68% reduction
77% reduction
95% reduction
11% increase
96% reduction
Fecal Coliform
Bacteria
72% reduction
60% reduction
41% reduction
38% reduction
52% reduction
54% reduction
not applicable
59% reduction
64% reduction
not applicable
70% reduction
NYCDEP. The stormwater management plan will be used
as a template for similar efforts in other urban reservoir
watersheds, and in NYCDEP's overall stormwater man-
agement, mitigation, and cost-sharing programs. Program
recommendations are as follows:
• An aggressive outreach campaign is needed to se-
cure support for the project, get permission to include
privately-owned land in the retrofit program, and ob-
tain regulatory permits and approvals. The campaign
should begin during conceptual plan development and
continue through facility construction and operation.
• The pre-application review process can streamline the
permitting and approval process.
• Water quality modeling results can support the selec-
tion and prioritization of sites and facility types.
• The inspection and maintenance requirements should
be defined and commitments to carry out the require-
ments should be obtained prior to construction.
• The advisory panel formed to review conceptual plans
and facility designs should be fully informed of water-
shed conditions, jurisdictional constraints and agency
capabilities.
• Design and construction contract bid documents and
payment processes should be clearly defined such that
all parties itemize work units in the same manner.
• A contractor should be selected that is experienced in
watershed assessment, as well as application of the
remediation programs likely to be warranted in the area.
• Facility designs should maximize water quality ben-
efits and minimize disturbances to natural resources.
Water quality monitoring capabilities should be included
in facility designs.
Acknowledgments
RoyF. Weston, Inc. of New York (Valhalla, NY) was con-
tracted to develop the Kensico Water Quality Control Pro-
gram and conceptual stormwater management plan.
Hazen and Sawyer, PC. (Manhattan, NY) was contracted
to reevaluate the conceptual stormwater management plan,
prepare engineering designs and construction cost esti-
mates for stormwater management and erosion control in
the Kensico Reservoir Watershed, and prescribe inspec-
tion and maintenance requirements.
195
-------
Empirical Modeling Approaches for Establishing
Nutrient Loading Goals for Tampa Bay
Anthony Janicki and David Wade
Post, Buckley, Schuh & Jernigan, Inc.
St. Petersburg, Florida
The Tampa Bay National Estuary Program (TBNEP) has
guided and supported the development of an empirically
based approach to setting external nitrogen load targets
for Tampa Bay, FL. Working closely with the scientists and
resource managers of the TBNEP, the authors used the
available data to elucidate relationships among loadings,
waterquality, and the subsurface light environments. These
relationships were then applied to the development of de-
fensible pollution load management targets for a Compre-
hensive Conservation and Management Plan (CCMP) for
Tampa Bay.
Background
Tampa Bay is the largest estuary in the state of Florida.
It extends approximately 35 miles into the west central
coast of Florida (Figure 1), and is 5 to 10 miles wide along
the majority of its length. Surface water flow from the 2,276-
square-mile watershed is provided by the Hillsborough,
Palm, Alafia, Little Manatee, and Manatee Rivers and over
40 minor tributaries. The mainstem of the bay is greatly
affected by the exchange of seawaterand nutrients to and
from the Gulf of Mexico.
The biological systems of the estuary are characterized
by both submerged and emergent vegetated habitats. The
emergent vegetated habitats are dominated by a mosaic
of mangrove forests and saltmarshes. Seagrass meadows
are the dominant submerged vegetation of the estuary, and
comprise Thalassia, Syringodium, Halodule, and Ruppia.
Due to development of its watershed, Tampa Bay expe-
rienced increases in pollutant loadings, declines in water
quality, and loss of seagrass acreage between 1950 and
the early 1980s. Long-term observations suggested that
along with direct physical destruction for development
(Janicki etal., 1994), pollutant loadings to Tampa Bay, and
the associated decline in water quality, have contributed
to a reduction in the extent of naturally occurring seagrass
meadows (Lewis et al., 1985; Avery, 1991; Lewis et al.,
1991). A major part of the poor water quality impacts on
seagrasses is thought to be the attenuation ofdownwelling
sunlight by excess concentrations of phytoplankton and
suspended solids in the water column. Hence, the seagrass
plants do not receive enough sunlight to remain healthy,
and they perish. Recent applied research concerning the
role of water quality in the loss of seagrass meadows was
discussed in Morris and Tomasko (1992), Stevenson et al.
(1993), and Batuik et al. (1992). Figures 2 and 3 present
the extent of the loss of seagrass acreage between 1950
and 1990.
Due to the cooperative efforts of local and regional re-
source managers, the Tampa Bay Agency on Bay Man-
agement (ABM, 1989), private interests, and concerned
citizens, pollutant loadings have been reduced and water
quality has been improving in the bay since 1984. With the
implementation of a number of different management ac-
tions in the early 1980s, including the implementation of
advanced wastewater treatment technologies, chlorophyll
levels have been declining from 1985 to present (Figure
4). Research on the recovery has suggested that a lag
may exist in the recovery of seagrass meadows relative to
the decline in seagrasses, and that seagrasses may con-
tinue to recover at a relatively slower rate (Johansson,
1991).
In 1991, the TBNEP was initiated as a cooperative pro-
gram to continue this process of reducing pollutant loads,
improving waterquality, and restoring lost habitat. To pro-
vide an objective focus for pollutant load management tar-
gets, the participants in the program selected the acreage
of seagrass meadows in the bay as a living resource bench-
mark by which progress could be measured. Recent ob-
servations by the Southwest Florida Water Management
District have indicated that seagrass acreages have in-
creased from 1990 to 1995, following the period of improved
waterquality. Thus, the important management questions
for the TBNEP became:
1) How many acres of seagrass should be restored to
return the bay to a restored state?
2) At what level should pollutant loads be managed to
reach the seagrass acreage defined by question 1?
The TBNEP developed an ad hoc political consensus
regarding the first question by establishing quantitative
196
-------
Study Area Location
27°45'
27°30'
27°30'
82°45'
82=30'
Projection UTM
Datum NAD 27
Map Prepared by Costa! Erwironmental/PBS&J
Map Publication No. S9820601
Figure 1. Tampa Bay National Estuary Program study area location.
seagrass restoration acreage targets. The target was set
by mapping and enumerating the reduction in seagrass
acreage from 1950 to 1990, and subtracting from this acre-
age the amount of seagrass that was permanently lost to
physical impacts (e.g., channel dredging, island creation,
borrow pits) and not likely to be restored (Janicki et al.,
1994). The work for this paper was completed to answer
the second question regarding what level of pollutant loads
would be consistent with meeting this acreage target.
Objectives
The specific objective of this work was 1) to use the avail-
able data to document the relationships among loadings,
water quality, and the subsurface light environments, and
2) to apply these empirical relationships to estimate pollu-
tion load management targets that would result in the main-
tenance of suitable light levels to restore the historical
seagrass acreage within the bay.
A paradigm was developed to illustrate the management
of nitrogen loads to effect changes in the acreage of
seagrass meadows (Figure 5). Using this paradigm, ex-
ternal nitrogen loads to the bay result in increased nitro-
gen concentrations in the bay. The increased nitrogen con-
centrations lead to increased chlorophyll concentrations.
The increased chlorophyll concentrations result in de-
creased depths to which surface light can penetrate in
sufficient levels to maintain seagrass meadows, and the
197
-------
Projection: UTM
Datum: NAD 27
Seagrass Present
Seagrass Absent, Deep
Seagrass Absent, Shallow
Map Prepared by Costal Environmental/PBS&J
Map Publication No. S9820801
Figure 2. 1950 seagrass distribution in Tampa Bay.
198
-------
Projection UTM
Datum: NAD 27
Seagrass Present
Seagrass Absent, Deep
Seagrass Absent, Shallow
Map Prepared by Coastal Environmental/PBS&J
Map Publication No. S9820901
Figure 3. 1990 seagrass distribution in Tampa Bay.
199
-------
Q.
E
o o o
in ^t co
Bn) E-[[Ai)dQjO[ip
(fl
±
200
-------
Nitrogen Export
To Gulf of Mexico
Average
Nitrogen Concentrations
STEP1
Internal Sources and Sinks
of Nitrogen
Phytoplankton Export
To Gulf of Mexico
Average
Chlorophyll-a Concentrations
Average
Turbidity and Color
STEP 2
Figure 5. Empirical target setting approach paradigm.
Average % of Surface Light
at Target Depth
decreased light levels result in reductions in the extent of
seagrass meadows starting with the deeper meadows.
Thus, by using the available data to elucidate the links
of this paradigm, the TBNEP was able to establish nitro-
gen load targets that would provide suitable light condi-
tions to a desired number of acres of bay bottom that could
potentially support seagrass meadows.
Methods
The first step in meeting this objective was to use the
available data to quantify each of the links in the para-
digm. Monthly external nitrogen load data were calculated
and compiled by the TBNEP (Zarbock et al., 1994). Monthly
water quality data, including nitrogen concentrations, chlo-
rophyll concentrations, and light penetration data were used
from 1986 to 1990, from data recorded by the Environ-
mental Protection Commission of Hillsborough County.
Seagrass data from the Southwest Florida Water Manage-
ment District, restoration target areas from the TBNEP, and
bathymetric data from the National Oceanographic and
Atmospheric Administration (NOAA) were combined to
quantify the restoration areas.
The final step in the approach was to apply the quantita-
tive relationships of the paradigm to compute the nitrogen
loading targets which were consistent with meeting the
seagrass acreages targets.
Results
The available monthly data were sufficient to develop
the links of the previously described paradigm.
Paradigm Link of Nitrogen Loads to
Nitrogen Concentrations
A significant and useful relationship between monthly
external nitrogen loads and monthly nitrogen concentra-
tions in the bay (i.e., the first link in the paradigm) was not
observed in the available data. Thus, a modification to the
original paradigm was made to link the external nitrogen
loads directly to chlorophyll concentrations. The relation-
ships between nutrient loads and water quality have been
the subject of classical limnological research (Vollenweider
1968,1975,1976) and more recent research for estuaries
(Riley 1972, Boynton et al., 1982).
Paradigm Link of Nitrogen Loads to
Chlorophyll Concentrations
Statistically significant relationships between monthly
external nitrogen loads and monthly chlorophyll concen-
trations were observed in the available data, and were
enumerated using data from 1986 to 1994. Monthly spe-
cific intercepts were used in the regression models to pre-
clude problems with potential seasonal autocorrelation in
the data. Figure 6 presents an example of these relation-
ships. The overall R-square value for these relationships
was 0.69.
201
-------
50-
40-
30-
O
S
,o
20-
10-
1000
2000 3000
Total Nitrogen Load (Meteric Tons/Year)
4000
4-
2 3^
o>
^ A jM^1 A A Y A J
A 4A|^A**|? 4^ *^jli HA"A
1 -. AJ
I
10
I
20
30
40
Chlorophyll-a (,ug/l)
Figure 6. Resulting empirical relationships for links of the paradigm: a) the relationship of external nitrogen load to chlorophyll concentration
in Hillsborough Bay, and b) the relationship of chlorophyll concentration to light attenuation in Hillsborough Bay.
The bay was divided into four segments for all steps of
the paradigm. For each segment, the quantity of external
nitrogen delivered to it was calculated as the input from
the watershed and airshed plus the amount contributed
through circulation from the other bay segments. The ni-
trogen loads from the watershed and airshed were calcu-
lated in a separate TBNEP study (Zarbock et al., 1994).
The nitrogen contributed from circulation from the other
bay segments was calculated by solving a series of steady-
state dilution equations, using salt as a conservative tracer
substance, and the observed salinity data for each seg-
ment and the Gulf of Mexico with the freshwater inflow
data for each segment. An empirically fit parameter was
used to correct these salt-derived equations for the
nonconservative nature of nitrogen. This was done by com-
paring total nitrogen concentrations predicted by applying
the conservative substance-derived dilution equations to
actual nitrogen loads and comparing the predicted nitro-
gen concentrations with observed nitrogen concentrations.
These relationships were statistically significant (overall
R-squareofO.68), and indicated that for the period of 1986
to 1994, an average annual amount of approximately 8 to
9% more nitrogen was input to the upper segments of the
bay than could be explained by loadings alone. A possible
202
-------
explanation for this observation is that nitrogen from the
fine sediments of these segments was being liberated to
the waters above. The nitrogen in the sediments likely ac-
cumulated during the pre-1985 period, during which exter-
nal loads were higher.
Paradigm Link of Chlorophyll
Concentrations to Light Attenuation
Statistically significant relationships were observed be-
tween observed monthly chlorophyll concentrations and
light attenuation. As described below, the depth to which
20.5% of surface light is estimated to penetrate was se-
lected as the light attenuation measurement for these re-
lationships. As forthe previous regression models, monthly
specific intercepts were used to preclude potential sea-
sonal autocorrelation in the observed data. Figure 6b pre-
sents an example of the relationship between these data.
The overall R-square value for these relationships was
0.67. The TBNEP recognized that turbidity was also an
important determinant of light penetration in the bay. How-
ever, chlorophyll concentrations are currently of greater
concern to bay managers, and a very useful and statisti-
cally significant model could be developed using chloro-
phyll alone as an explanatory variable.
Paradigm Link of Light Attenuation to
Seagrass Restoration
Because a large database of light attenuation data paired
with seagrass condition data was not available, this final
link of the paradigm was derived from observations from
independent research conducted by the participating
TBNEP scientists. ATBNEP study conducted forthe South-
west Florida Water Management District by Mote Marine
Laboratory (Dixon and Leverone, 1995) collected photo-
synthetically active radiation (PAR) measurements at the
deep edges of seagrass beds in the lower portion of the
bay for one year. After subtracting an average measured
bottom reflectance of 2%, the TBNEP scientists reached a
consensus that an annual average 20.5% of surface light
was needed for seagrass survival in Tampa Bay.
Application of the Results to Nitrogen Load
Targets
The empirically derived links of the paradigm were ap-
plied as a single tool to answer the second question de-
fined above (i.e., at what level should nitrogen loads be
managed to meet the seagrass acreage restoration tar-
gets set by the TBNEP).
A series of candidate 15-year nitrogen load schedules
were applied to the integrated models and the resulting
chlorophyll values were estimated. These chlorophyll val-
ues were then used to estimate the depth to which 20.5%
of surface light would penetrate. The NOAA depth data
and seagrass restoration areas were then applied to a
detailed computer mapping system to calculate the acre-
age of seagrass restoration areas which would be illumi-
nated at 20.5% of surface light under the estimated chlo-
rophyll levels. As described previously, seagrass restora-
tion areas were defined as those areas of the bay bottom
which had seagrass meadows in 1950, did not have
seagrass meadows in 1990, and were not physically al-
tered so as to preclude the restoration of seagrass given a
proper light regime.
References
Agency on Bay Management. 1989. Chlorophyll—a Target
Concentration Proposed for Tampa Bay. Prepared by
Task Force on Resource-based Water Quality Assess-
ment, Tampa Bay Regional Planning Council Agency
on Bay Management. St. Petersburg, FL. 10 pp.
Avery, WM. 1991. Status of Naturally Occurring and Intro-
duced Halodule wrightii'm Hillsborough Bay. In: BASIS
II: Tampa Bay Area Scientific Information Symposium 2.
Eds. S.F. Treat and PA. Clark. Tampa, FL. pp. 177-178.
Batuik, R.A., R.J. Orth, K.A. Moore, WC. Dennison, J.C.
Stevenson, L.W Staver, V. Carter, N.B. Rybicki, R.E.
Hickman, S. Kollar, S. Beiber, and P. Heasly. 1992.
Chesapeake Bay Submerged Aquatic Vegetation Habi-
tat Requirements and Restoration Targets: A Techni-
cal Synthesis. 186 pp.
Boynton, W.R., WM. Kemp, and C.W Keefe. 1982. A com-
parative analysis of nutrients and other factors influ-
encing estuarine phytoplankton production. In: Estua-
rine Comparisons. Ed. V.S. Kennedy. Academic Press.
pp. 69-90.
Dixon, L.K. and J.R. Leverone. 1995. Light requirements
of Thalassia testudinum in Tampa Bay, Florida. Sub-
mitted to Surface Water Imrpovement and Manage-
ment Program, Southwest Florida Water Management
District by Mote Marine Laboratory, Sarasota, FL. Mote
Marine Laboratory Technical Report Number 425.
Janicki, J., D. Wade, and D. Robison. 1994. Habitat pro-
tection and restoration targets for Tampa Bay. Prepared
for Tampa Bay National Estuary Program. TBNEP
Technical Report No. 07-93.
Johansson, J.O.R. 1991. Long-term trends of nitrogen load-
ing, waterquality, and biological indicators in Hillsborough
Bay, Florida. In: BASIS II: Tampa Bay Area Scientific
Information Symposium II. Eds. S.F. Treat and PA. Clark.
Tampa, FL. pp. 157-176.
Lewis, R.R. Ill, M.J. Durako, M.D. Moffler, and R.C. Phillips.
1985. Seagrass Meadows of Tampa Bay - a Review. In:
BASIS. Proceedings Tampa Bay Area Scientific Infor-
mation Symposium. Eds. S.F. Treat, J.L. Simon, R.R.
Lewis III, and R.L. Whitman Jr. Tampa, FL. pp. 210-246.
Lewis, R.R. Ill, K.D. Haddad, and J.O.R. Johansson. 1991.
Recent areal expansion of seagrass meadows in Tampa
Bay, Florida: Real bay improvement ordrought-induced?
In: BASIS II: Tampa Bay Area Scientific Information Sym-
posium II. Eds. S.F. Treat and P.A. Clark. Tampa, FL.
pp. 189-192.
203
-------
Morris, L.J. and D.A. Tomasko. 1992. Proceedings and
conclusions of workshops on submerged aquatic veg-
etation initiative and photosynthetically active radia-
tion. Melbourne, FL. 244 pp.
Riley, G.A. 1972. Patterns of production in marine
ecosytems. In: Ecosystem System and Structure. Ed.
J.A. Wiens. University of Oregon Press.
Stevenson, J.C., L.W. Staver, and K.W. Staver. 1993. Water
quality associated with survival of submerged acquatic
vegetaion along an estuarine gradient. Estuaries
16:346-361.
Vollenweider, R.A. 1968. Scientific Fundamentals of the
Eutrophication of Lakes and Flowing Waters, with
Particluar Reference to Phosphorus and Nitrogen as
Factors in Eutrophication. OECD Technical Report
DAS/CS1/68.27. 158pp.
Vollenweider, R.A. 1975. Input-Output Models. Schweiz
Z. Hydrologie. 37:53-84.
Vollenweider, R.A. 1976. Advances in defining critical load-
ing levels for phosphorous in lake eutrophication. Mem.
1st. Ital. Idrobiol. 33:53-83.
Zarbock, H, A. Janicki, D. Wade, D. Heimbuch, and H.
Wilson. 1994. Estimates of total nitrogen, total phos-
phorous, and total suspended solids loading to Tampa
Bay, Florida. Prepared for Tampa Bay National Estu-
ary Program. Technical Report No. 04-94.
204
-------
Alum Treatment of Stormwater Runoff -
An Innovative BMP for Urban Runoff Problems
Harvey H. Harper, Ph.D., RE. and Jeffrey L. Herr, RE.
Environmental Research & Design, Inc.
Orlando, Florida
Eric H. Livingston
Florida Department of Environmental Protection
Tallahassee, Florida
Alum treatment of stormwater runoff was begun in 1986
as part of a lake restoration project at Lake Ella in Talla-
hassee, Florida in 1986. Our system provides treatment of
stormwater runoff entering the lake by injecting liquid alum
into major stormsewer lines on a flow-weighted basis dur-
ing rain events. When added to runoff, alum forms non-
toxic precipitates of AI(OH)3 and AIPO4 which combine with
phosphorus, suspended solids and heavy metals, caus-
ing them to be rapidly removed from the treated water.
The alum stormwater treatment system resulted in imme-
diate and substantial improvements to water quality in Lake
Ella which led to implementation of similar systems on other
urban lakes. There are currently 23 alum stormwater treat-
ment systems either operational or under construction in
Florida, and one experimental system in Seattle, Wash-
ington.
Alum treatment of stormwater runoff has consistently
achieved a 90% reduction in total phosphorus, 50-70%
reduction in total nitrogen, 50-90% reduction in heavy
metals, and >99% reduction in fecal coliform. Ultimate
water quality improvements in the receiving water body
have been related to the percentage of total inputs treated
by the system. Heavy metal and phosphorus associations
with alum floe have been shown to be extremely stable
over a wide range of pH and redox conditions.
In general, alum treatment of runoff is substantially less
expensive than traditional treatment methods and often
requires no additional land purchase. Recent designs have
incorporated automatic floe collection and removal systems
with disposal to drying beds or sanitary sewer.
Introduction
The addition of alum to water results in the production of
chemical precipitates which remove pollutants by two pri-
mary mechanisms. Removal of suspended solids, algae,
phosphorus, heavy metals and bacteria occurs primarily
by enmeshment and adsorption onto aluminum hydroxide
precipitate according to the following net reaction:
Al+3
6H2O
AI(OH)3(S) +3H3O+
Removal of additional dissolved phosphorus occurs as a
result of direct formation of AIPO4 by:
Al+3 + HnP04n-3 -» AIP04(s)
Hn+
The aluminum hydroxide precipitate, AI(OH)3, is a ge-
latinous floe which attracts and adsorbs colloidal particles
onto the growing floe, thus clarifying the water. Phospho-
rus removal or entrapment can occur by several mecha-
nisms, depending on the solution pH. Inorganic phospho-
rus is also effectively removed by adsorption to the AI(OH)3
floe. Removal of particulate phosphorus is most effective
in the pH range of 6-8 where maximum floe occurs (Cooke
and Kennedy, 1981). At higher pH values, OH~ begins to
compete with phosphate ions for aluminum ions, and alu-
minum hydroxide-phosphate complexes begin to form. At
lower pH values and higher inorganic phosphorus concen-
trations, the formation of aluminum phosphate (AIPO4) is
favored.
In 1985, a lake restoration project was initiated at Lake
Ella, a shallow 13.3-acre hypereutrophic lake in Tallahas-
see, Florida, which receives untreated stormwater runoff
from approximately 163 acres of highly impervious urban
watershed area. Initially, conventional stormwatertreatment
technologies, including retention basins, exfiltration
trenches and filter systems, were considered for reducing
available stormwater loadings to Lake Ella in an effort to
improve water quality within the lake. Since there was no
available land surrounding Lake Ella that could be used
for construction of traditional stormwater management fa-
cilities, and the purchasing of homes and businesses to
acquire land for construction of these facilities was cost-
prohibitive, alternate stormwatertreatment methods were
considered.
205
-------
Chemical treatment of stormwater runoff was evaluated
using various chemical coagulants, including aluminum
sulfate, ferric salts, and polymers. Aluminum sulfate (alum)
consistently provided the highest removal efficiencies and
produced the most stable end product. In view of success-
ful jartest results on runoff samples collected from the Lake
Ella watershed, the design of a prototype alum injection
stormwater system was completed. Construction of the
Lake Ella alum stormwater treatment system was com-
pleted in January 1987, resulting in a significant improve-
ment in water quality.
The alum precipitate formed during coagulation of
stormwater can be allowed to settle in receiving
waterbodies or collected in small settling basins. Alum pre-
cipitates are exceptionally stable in sediments and will not
redissolve due to changes in redox potential or pH under
conditions normally found in surface waterbodies. Over
time, the freshly precipitated floe ages into even more stable
complexes, eventually forming gibsite. The solubility of dis-
solved aluminum in the treated water is regulated entirely
by chemical equilibrium. As long as the pH of the treated
water is maintained within the range of 5.5-7.5, dissolved
aluminum concentrations will be minimal. In many in-
stances, the concentration of dissolved aluminum in the
treated water will be less than the concentration in the raw
untreated water due to adjustment of pH into the range of
minimum solubility.
Since the Lake Ella system, alum stormwater treatment
systems have been constructed in Florida for Lake Dot,
Lake Rowena and Lake Lucerne in Orlando; Lake Osceola,
Lake Virginia North and Lake Mizell in Winter Park; Lake
Cannon in Polk County; Channel 2 Drainage Canal in
Pinellas Park; Celebration Town Lake in Celebration; Lake
Holden in Orange County; Lake Tuskawilla in Ocala; and
a set of five separate systems has been installed at Lake
Maggiore in St. Petersburg. An experimental treatment
facility has also been constructed in the Lake Sammamish
watershed in Seattle, Washington. In addition to these
projects which are currently operational, additional projects
are currently under design in Winter Park, Orlando, Largo,
Tampa and Clearwater. The first project to treat stormwater
discharged to a brackish water became operational in Janu-
ary 1998 in the City of St. Petersburg, Florida.
Alum treatment of stormwater runoff has now been used
as a viable stormwater treatment alternative in urban ar-
eas for over 10 years. Over that time, a large amount of
information has been collected relative to optimum sys-
tem configuration, water chemistry, sediment accumula-
tion and stability, construction and operation costs, com-
parisons with other stormwater management techniques,
and floe collection and disposal. A summary of current
knowledge in these areas is given in the following sec-
tions.
System Configuration
Once alum has been chosen as an option in a stormwater
retrofit project, extensive laboratory testing must be per-
formed to verify feasibility and to establish design param-
eters. The feasibility of alum treatment for a particular
stormwater stream is typically evaluated in a series of labo-
ratory jar tests conducted on representative runoff samples
collected from the project watershed area. This extensive
laboratory testing is necessary to determine design, main-
tenance and operational parameters such as the optimum
coagulant dose required to achieve the desired water qual-
ity goals, chemical pumping rates and pump sizes, the need
for additional chemicals to buffer the pH of receiving wa-
ter, post-treatment water quality characteristics, floe for-
mation and settling characteristics, floe accumulation, an-
nual chemical costs and storage requirements, ecological
effects, and maintenance procedures. In addition to deter-
mining the optimum coagulant dose, jar tests can also be
used to determine floe strength and stability, required mix-
ing intensity and duration, and design criteria for dedicated
floe settling basins.
In a typical alum stormwater treatment system, alum is
added to the stormwater flow on a flow-proportional basis
so that the same dose of alum is added to a gallon of
stormwater flow regardless of the discharge rate. A vari-
able speed chemical metering pump is typically used as
the injection pump. If a buffering agent, such as NaOH, is
required to maintain desired pH levels, a separate meter-
ing system and storage tank will be necessary. The opera-
tion of each injection pump is regulated by a flow meter
device attached to each incoming stormwater line to be
treated. Data from each stormwater flow meter is trans-
formed into a 4-20 mA electronic signal which instructs
each metering pump to inject alum according to the mea-
sured flow through each individual line. Mixing of the alum
and stormwater occurs as a result of turbulence in the
stormsewer line. If sufficient turbulence is not available
within the stormsewer line, artificial turbulence can be gen-
erated using aeration or physical stormsewer modifications.
Mechanical components for the alum stormwater treat-
ment system, including chemical metering pumps,
stormsewer flow meters and electronic controls, are typi-
cally housed in a central facility which can be constructed
as an above-ground or below-ground structure. A 6,000
gallon tank is typically used for bulk alum storage. Alum
feed lines and electrical conduits are run from the central
facility to each point of flow measurement and alum addi-
tion. Alum injection points can be located as far as 3000 ft
from the central pumping facility. Early designs for alum
stormwater treatment systems utilized individual chemical
metering pumps and stormsewer flow meters for each point
of alum addition. However, in an effort to reduce overall
system costs and complexity, current alum stormwater
treatment systems often feed alum to multiple points us-
ing a single chemical metering pump and control valves.
Water Chemistry
In general, construction and operation of alum
stormwater treatment systems has resulted in significant
improvements in waterqualityfortreated waterbodies. The
degree of observed improvement in water quality is directly
206
-------
related to the percentage of annual hydraulic inputs treated
by the alum stormwater treatment system. A comparison
of pre- and post-modification water quality for three typical
alum stormwater treatment systems is given in Table 1,
including Lake Ella and Lake Dot (which provide treatment
for approximately 95-96% of the annual hydraulic inputs
entering these lake systems), and Lake Osceola (which
provides treatment for only 9% of the annual hydraulic in-
puts entering the lake system).
Operation of the alum stormwater treatment systems
resulted in a decline in pH within each of the three
waterbodies, ranging from a reduction of approximately 1
unit in Lake Ella to 0.6 units in Lake Osceola. A pH reduc-
tion of only 0.1 unit was observed for the Lake Dot treat-
ment system which injects alum in combination with so-
dium hydroxide to control pH levels within the lake. In ad-
dition, significant improvements in dissolved oxygen were
observed in both Lake Ella and Lake Dot. Alum treatment
of stormwater runoff resulted in a 78% reduction in total
nitrogen concentrations in Lake Ella, a 55% reduction in
Lake Dot and a 4% reduction in Lake Osceola where only
a small portion of the annual hydraulic inputs are treated.
The majority of the total nitrogen removal observed is a
result of reducing concentrations of dissolved organic ni-
trogen and particulate nitrogen, since alum is generally
ineffective in reducing concentrations of inorganic nitro-
gen species, such as ammonia or NOX. Alum stormwater
treatment resulted in a substantial reduction in measured
concentrations of orthophosphorus and total phosphorus
in each of the three lake systems, with total removals of
89%, 93% and 30% for Lake Ella, Lake Dot and Lake
Osceola, respectively. Alum stormwater treatment also
reduced in-lake concentrations of BOD in each of the three
lake systems, with a reduction of 93% in Lake Ella, 84% in
Lake Dot and 22% in Lake Osceola.
Alum stormwater treatment has been extremely effec-
tive in reducing concentrations of chlorophyll-a in receiv-
ing waterbodies, with a reduction of 97% in Lake Ella, 89%
in Lake Dot and 13% in Lake Osceola. Reductions in mea-
sured concentrations of chlorophyll-a occur as a result of
enmeshment and precipitation of algal particles within the
water column of the lake by alum floe as well as phospho-
rus limitation created by low levels of available phospho-
rus in the water column. Substantial increases in Secchi
disk depth were observed in Lake Ella and Lake Dot, and
to a lesser extent in Lake Osceola, with improvements of
340% in Lake Ella, 212% in Lake Dot and 9% in Lake
Osceola. Based upon the Florida TSI Index (Brezonik,
1984), Lake Ella and Lake Dot have been converted from
hypereutrophic to oligotrophic status, with a conversion
from eutrophicto mesotrophicthe case in Lake Osceola.
A graphic history of total phosphorus concentrations in
Lake Lucerne, which was retrofitted with an alum
stormwater treatment system in June 1993 that provides
treatment for approximately 82% of the annual runoff in-
puts into the lake, is given in Figure 1. Priorto construction
of the alum stormwater treatment system, total phospho-
rus concentrations in Lake Lucerne fluctuated widely, with
a mean concentration of approximately 100 u,g/l. Follow-
ing start-up of the alum treatment system, total phospho-
rus concentrations began to decline steadily, reaching equi-
librium concentrations of 20-40 u,g/l. A slight increase in
total phosphorus concentrations was observed during the
last half of 1995 when the system was off-line due to light-
ning damage. When system operation resumed in June
Table 1. Comparison of Pre- and Post-Modification Water Quality Characteristics for Typical Alum Stormwater Treatment Systems
PARAMETER
UNITS
LAKE ELLA
BEFORE AFTER
(1974-85) (1/88-5/90)
LAKE DOT
BEFORE AFTER
(1986-88) (3/89-8/91)
LAKE OSCEOLA
BEFORE AFTER
(6/91-6/92) (2/93-12/96)
# of Samples
Lake Area
Watershed Area
Percent of Annual
Hydraulic Inputs
Treated
15
11
15
12
46
PH
Diss. O2 (1 m)
Total N
Total P
BOD
Chlorophyll-a
Secchi Disk Depth
Diss. Al
Florida TSI Value
s.u.
mg/l
|ig/l
H9/I
mg/l
mg/m3
m
|ig/l
7.41
3.5
1876
232
41
180
0.5
-
98
(Hyper-
eutrophic)
6.43
7.4
417
26
3.0
5.1
>2.2
44
47
(Oligotrophic)
7.27
6.6
1545
351
16.8
55.8
<0.8
-
86
(Hyper-
eutrophic)
7.17
8.8
696
24
2.7
6.3
2.5
65
42
(Oligotrophic)
8.22
8.8
892
37
4.4
24.8
1.1
18
61
(Eutrophic)
7.63
8.8
856
26
3.4
21.7
1.2
51
56
(Meso-
trophic)
13.3ac
57 ac
95
5.9 ac
305 ac
96
55.4 ac
153ac
9
207
-------
Lake Lucerne
Total Phosphorus
180
1990
1991
1992
1993
1994
1995
1996
Date
Figure 1. Trends in total phosphorus concentrations in Lake Lucerne before and after alum treatment of stormwater runoff.
1996, total phosphorus concentrations returned to equilib-
rium values of approximately 20 u,g/l.
In general, measured concentrations of heavy metals
have been extremely low in value in all waterbodies retro-
fitted with alum stormwater treatment systems, with no vio-
lations of heavy metal standards. In addition, measured
levels of dissolved aluminum have also remained low in
each lake system. Mean dissolved aluminum concentra-
tions for Lake Ella, Lake Dot and Lake Osceola have aver-
aged 44 u,g/l, 65 u,g/l and 51 u,g/l, respectively. Although
there is no standard for dissolved aluminum in the State of
Florida, the U.S. EPA has recommended a long-term av-
erage of 87 u,g/l for protection of all species present in the
U.S. The solubility of dissolved aluminum is regulated al-
most exclusively by pH. As long as the pH of the treated
water can be maintained in the range of 6.0-7.5 during the
treatment process, dissolved aluminum concentrations will
remain at minimal levels.
Floe Accumulation
Laboratory investigations have been conducted on
stormwater runoff collected from a wide range of land uses
typical of urban areas to quantify the amount of alum floe
generated as a result of alum treatment of stormwater runoff
at various treatment doses. After initial formation, alum floe
appears to consolidate rapidly fora period of approximately
6-8 days, reaching approximately 20% of the initial floe
volume. Additional consolidation appears to occur over a
settling period of approximately 30 days, after which col-
lected sludge volumes appear to approach maximum con-
solidation (Harper, 1990).
Estimates of maximum anticipated sludge production,
based upon literally hundreds of laboratory tests involving
coagulation of stormwater runoff with alum at various
doses, and based upon a consolidation period of approxi-
mately 30 days, are given in Table 2. At alum doses typi-
cally used fortreatment of stormwater runoff, ranging from
5-10 mg/l as Al, sludge production is equivalent to approxi-
mately 0.16-0.28% of the treated runoff flow. Sludge pro-
duction values listed in Table 2 reflect the combined mass
generated by alum floe as well as solids originating from
the stormwater sample.
Field investigations have also been performed in lake
systems receiving alum treated stormwater runoff to docu-
ment the accumulation rate of alum floe within the sedi-
ments by visual inspection of sediment core samples col-
lected in clear acrylic tubes at selected monitoring sites in
each lake. A comparison of observed and predicted floe
accumulation rates in lake systems receiving stormwater
treatment is given in Table 3. Each of the listed lakes has
been receiving alum treatment for five years or more. The
primary predicted settling area for floe accumulation was
determined by evaluating lake bottom topography and
stormsewer inflow characteristics. Predicted floe accumu-
Table 2. Anticipated Production of Alum Sludge from Alum Treatment
of Stormwater at Various Doses
Sludge Production1
Alum Dose
(mg/l asAI)
5
7.5
10
As Percent Of
Treated Flow
0.16
0.20
0.28
Per 106
Gallons Treated
214ft3
268ft3
374ft3
1 Based on a minimum settling time of 30 days
208
-------
Table 3. Comparison of Observed and Predicted Floe Accumulation
Rates in Lake Systems with Alum Stormwater Treatment
Lake
Lake Ella
Lake Lucerne
Lake Osceola
Predicted Predicted
Settling Area Accumulation Rate
50% of lake bottom
areas 10 ft or deeper
50% of lake bottom
1 cm/yr
3.3 cm/yr
0.5 cm/yr
Observed
Accumulation
Rate
0.33 cm/yr
none
none
lation rates are based upon the anticipated floe production
rates summarized in Table 2.
Annual floe production in Lake Ella was predicted to be
approximately 1 cm/yr over 50% of the lake bottom. How-
ever, floe accumulation evaluations performed in 1990 in-
dicate an observed accumulation rate of approximately 0.33
cm/yr, approximately one-third of the predicted accumula-
tion rate. The reduced observed accumulation rate is
thought to be a result of additional floe consolidation over
time and incorporation of the alum floe into the existing
sediments. The observed post-treatment floe accumula-
tion rate in Lake Ella is similar to the pre-treatment sedi-
ment accumulation rate in Lake Ella resulting from the ex-
tremely high algal production prior to the lake restoration
efforts in 1985. Sediment accumulation in Lake Lucerne
was anticipated to occur in areas 10 ft or deeper, with a
predicted accumulation of 3.3 cm/yr. However, no sedi-
ment accumulation was observed at any of the 10 fixed
monitoring locations within the lake which have contrib-
uted data on an annual basis since start-up of the alum
treatment system. A similar conclusion has been reached
in Lake Osceola which has no visible floe accumulation
after approximately five years of alum stormwater treat-
ment. Both Lake Lucerne and Lake Osceola appear to in-
corporate alum floe into the existing sediments with no vis-
ible surface floe layer.
Construction and O&M Costs
A summary of construction and annual operation and
maintenance (O&M) costs for existing alum stormwater
treatment facilities, with treated watershed areas ranging
from 64 ac to 1450 ac, is given in Table 4. Construction
costs for alum stormwater treatment systems have ranged
from $75,000 to $400,000, depending upon the number of
outfalls to be retrofitted. In general, the capital cost of con-
structing alum stormwater treatment systems is indepen-
dent of the watershed size since the capital cost for con-
structing a treatment system for a 100-acre watershed is
identical to the cost of constructing a system to treat 1000
acres at the same location, although annual O&M costs
would differ. The average capital cost for existing alum
stormwater treatment facilities is $245,998.
Estimated O&M costs are also provided in Table 4 and
include chemical, power, manpowerfor routine inspections,
and equipment renewal and replacement costs. Opera-
tion and maintenance costs for existing alum stormwater
treatment systems range from $5,500 to $26,298 per year.
Construction costs and annual O&M costs are also included
on a per acre treated basis for comparison with other
stormwater treatment alternatives.
Comparison with Other Stormwater
Treatment Alternatives
In general, removal efficiencies obtained with alum
stormwater treatment are similar to removal efficiencies
obtained using a dry retention or wet detention stormwater
management facility. A comparison of treatment efficien-
cies for common stormwater management systems is given
in Table 5 (Harper, 1995). Estimated removal efficiencies
for alum treatment exceed removal efficiencies achieved
in dry retention fortotal phosphorus and TSS, but fall short
of dry retention fortotal nitrogen and BOD. Dry retention
Table 4. Summary of Construction and O&M Costs for Existing Alum Stormwater Treatment Facilities
Project
Lake Ella
Lake Dot
Lake Lucerne
Lake Osceola
Lake Cannon
Channe 2
Lake Virginia North
Celebration
Lake Holden
Lake Tuskawilla
Lake Rowena
Lake Mizell
Lake Maggiore (5)
Webster Avenue
Lake Virginia South
Merritt Ridge
AVERAGES
Area
Treated
(ac)
158
305
272
153
490
84
64
158
183
311
538
74
1450
91
437
195
310
Construction
Cost/System
($)
200,400
250,000
400,000
300,000
135,000,
180,000
242,000
300,000
292,000
242,000
75,000
300,000
400,000
130,000
288,000
201,575
$ 245,998
Estimated
Annual
O&M Cost
($)
_
-
16,000
6500
13,100,
-
-
25,000
-
19,627
-
15,389
21,450
12,397
-
26,298
$17,307
Construction
Cost Per
Area Treated
($/ac)
1268
823
1472
1959
276
2144
3769
1898
1598
111
139
4049
1379
1423
659
1033
$1542
Annual O&M
Cost Per
Area Treated
($/ac)
_
-
59
43
27
-
-
158
-
63
-
208
74
136
-
135
$100
209
-------
Table 5. Estimated Removal Efficiencies for Common Stormwater
Management Systems
Type of System
Estimated Removal Efficiencies (%)
Total N Total P TSS BOD
Dry Retention 80 80 80 80
(0.50-in runoff)
Wet Detention 20-30 60-70 85 50-60
Wet Detention with 0 60 >90 90
Filtration
Dry Detention 10-20 20-40 60-80 03-50
Dry Detention with 0-20 0-20 60-90 0-55
Filtration
Alum Treatment
50-70
>90
>95
60
may be the more effective common stormwater manage-
ment technique in use today, if other things are equal. But
certainly, removal efficiencies achieved with alum treatment
exceed removal efficiencies obtained using wet detention,
wet detention with filtration, dry detention, or dry detention
with filtration. Alum is the best choice where space for re-
tention cannot be built.
Alum treatment of stormwater runoff compares favor-
ably with other stormwater treatment alternatives with re-
spect to both initial capital construction costs and annual
O&M costs. A comparison of certain costs for alum
stormwater treatment and equivalent retention facilities is
given in Table 6. Initial capital construction costs and an-
nual O&M costs for three existing alum stormwater treat-
ment facilities are compared with the estimated cost for
construction of an alternate retention facility for treatment
of the first 0.5 in of runoff. Each of the alternate retention
facilities would require purchase of land in heavily urban-
ized areas which if available, would be expensive. The cost
listed forthe alternate retention facilities include land costs
only and not actual construction costs. Estimated annual
Table 6. Comparison of Certain Costs for Alum Stormwater Treatment
and Equivalent Retention Facilities
Alum Treatment
System
Equivalent Retention Facility
Location
Annual Land Annual
Area Capital O&M Area Land O&M
Treated Costs Costs1 Required2 Costs Costs3
(ac) ($) ($) (ac) ($) ($)
Lake
Lake
Lake
Osceola
Lucerne
Cannon
88
210
490
235,000
420,000
135,000
6500
16,000
13,100
3.0
7.3
17.0
1 ,500,000"
3700,000"
850,000=
9000
21
51
,900
,000
Includes chemical costs, weekly inspection, and $1000 for supplies
and parts
Based on equivalent treatment of 1 inch of runoff and a 3-ft deep
pond
Based on $3000/acre for O&M (Ref; FOOT)
Based on a land cost of $500,000/acre
Based on a land cost of $50,000/acre
O&M costs for retention pond maintenance, such as rou-
tine mowing, weed control and trash removal, is higher
than the estimated O&M costs forthe alum treatment sys-
tems which include chemicals, weekly inspections, and
parts and supplies.
Floe Collection and Disposal
Although virtually all existing alum stormwater treatment
systems allow for floe settling in receiving waterbodies,
and although only beneficial aspects of alum floe accumu-
lation have been observed, current alum treatment sys-
tem designs feature collection and disposal of floe. Where
possible, sump areas have been constructed to provide a
basin for collection and accumulation of alum floe. The
accumulated floe can then be pumped out of the sump
area, using either manual or automatic techniques, on a
periodic basis. Several current treatment systems provide
for automatic floe disposal into the sanitary sewer system
at a slow controlled rate. Since alum floe is virtually inert
and has a consistency similar to that of water, acceptance
of alum floe on a periodic basis poses no operational prob-
lem for wastewater treatment facilities. A schematic of a
settling pond designed forthe Lake Virginia system is in-
cluded in Figure 2.
A recent design for collection of floe discharging from a
submerged pipe in a lake system is also illustrated in Fig-
ure 2. The floe containment area consists of a fabric mesh
sized to allow water flow while trapping floe particles. The
floe is then collected in the sump area in the bottom of the
containment area and pumped on a periodic basis to the
sanitary sewer system or adjacent drying bed. Drying char-
acteristics for alum sludge are similarto a wastewater treat-
ment plant sludge. A drying time of approximately 30 days
is sufficient to dewater and dry the sludge, with a corre-
sponding volume reduction of 80-90%. Dried alum sludge
has chemical characteristics suitable for general land ap-
plication or in agricultural sites, as outlined in Chapter 62
of the Florida Administration Code (Florida Department of
Environmental Protection, 1996).
Conclusions
Alum treatment of stormwater runoff has emerged as a
viable and cost-effective alternative for providing
stormwater retrofit in urban areas. Based upon 10 years
of experience with alum stormwater treatment, the follow-
ing conclusions have been reached:
• In lake systems where a large percentage of the an-
nual runoff inputs are retrofitted with an alum treat-
ment system, alum treatment has consistently
achieved a 90% reduction in total phosphorus, 50-70%
reduction in total nitrogen, 50-90% reduction in heavy
metals, and >99% reduction in fecal conforms. Ulti-
mate water quality improvements in the receiving
waterbodies are highly correlated with the percentage
of total inputs treated by the system.
• The observed accumulation rate of alum floe in the
sediments of receiving waterbodies are to be substan-
210
-------
Inflow
Concrete
sump pad
Setting pond
FLOC pump station
J. 4" PVC FLOC disposal main
to sanitary sewer system
4" fabric formed
concrete sump area
Alum FLOC barrier
4" Aluminum
support post
Lake
Osceola
Shoreline / FLOC
Barrier connection
A
Shoreline
Figure 2. Typical schematics of floe collection and disposal systems.
tially lower than the predicted accumulation rate due
to additional floe consolidation overtime and incorpo-
ration of alum floe into the existing sediment.
Construction costs for alum stormwater treatment sys-
tems are largely independent of the watershed area
to be treated and depend primarily upon the number
of outfalls to be retrofitted.
In general, removal efficiencies obtained with alum
stormwater treatment are roughly similar to removals
obtained using a dry retention, and better than alter-
native stormwater management facilities.
Alum treatment of stormwater runoff is often substan-
tially less expensive than other stormwater treatment
alternatives with respect to both initial capital costs and
annual O&M costs.
Several innovative designs have been developed for
collection of alum floe in sump areas and containment
areas, with floe disposal made to sanitary sewer or
adjacent drying beds.
References
Brezonik, P.L. 1984. "Trophic State Indices: Rationale for
Multivariate Approaches." In Lake and Reservoir Man-
agement, pp. 427-430, EPA-440/5-84/001.
Cooke, G.D., and Kennedy, R.H. 1981. Precipitation and
Inactivation of Phosphorus as a Lake Restoration Tech-
nique. EPA-600/3-81/012.
Florida Department of Environmental Protection. 1996.
"Florida Administrative Code, Chapter 62" Tallahassee,
FL.
Harper, H.H. December 1990. "Long-Term Performance
Evaluation of the Alum Stormwater Treatment System
at Lake Ella, Florida." Final Report submitted to the
Florida Department of Environmental Regulation, Project
WM339.
Harper, H.H. October 1995. "Pollutant Removal Efficiencies
for Typical Stormwater Management Systems in Florida."
In Proceedings of the 4th Biennial Stormwater Research
Conference (Sponsored by the Southwest Florida Wa-
ter Management District) pp. 6-17, Clearwater, FL.
211
-------
An Eight-Step Approach to Implementing
Stormwater Retrofitting
Richard A. Claytor, Jr. RE.
Center for Watershed Protection
Ellicott City, Maryland
What are retrofits and why are they important? In the
quest for watershed protection and restoration, watershed
professionals are constantly seeking new tools for control-
ling stormwater runoff and associated adverse impacts.
Stormwater retrofits are among the most promising of these
tools. Retrofits are structural stormwater management
measures for urban watersheds designed to help lessen
accelerated channel erosion, reduce pollutant loads, pro-
mote conditions for improved aquatic habitat, and correct
past mistakes. Simply put, these best management prac-
tices (BMPs) are inserted in an urban landscape where
limited stormwater controls existed.
Retrofits come in many shapes and sizes from large re-
gional retention ponds that provide a variety of controls to
small on-site facilities providing only water quality treat-
ment for smaller storms. At least some kind of practice
can be installed in almost any situation. But fiscal restraints,
pollutant removal capability, and watershed capture area
must all be carefully considered in any retrofit selection
criteria.
Restoration versus Retrofitting
Stormwater retrofits should be applied along with other
available watershed restoration strategies for reducing
pollutants, restoring habitat and stabilizing stream morphol-
ogy as part of a holistic watershed restoration program.
While some professionals rightfully assert that true water-
shed restoration is not feasible, the term is applied here
as simply a concerted strategy to install a functional native
biological community in a stream, lake or river. Some of
the many watershed restoration strategies include:
• Stabilizing stream channel morphology
• Improving aquatic habitat within urban streams
• Replacing or enhancing riparian cover along urban
streams
• Promoting pollution prevention source controls within
the watershed
• Recolonizing streams with native fish communities
Many, if not most, of these components should be
planned in conjunction with an urban retrofit program. With-
out establishing a stable, predictable hydrologic water re-
gime that regulates the volume, duration, frequency, and
rate of flow, many of these other strategies may be disap-
pointing failures. To successfully restore a stream's over-
all aquatic health, stormwater retrofitting is an essential
element.
Table 1 presents a step-by-step approach to stormwater
retrofitting developed by our Center for Watershed Pro-
tection staff overthe past several years. An eight-step pro-
cess is briefly discussed and several case studies from
the author's experience emphasize particular points. At the
conclusion of the eight-step process, two additional case
studies are presented in more detail to illustrate some of
the many real world challenges of implementing retrofit
projects.
Step 1. Watershed Retrofit Site Inventory
The first step in getting retrofits "in the ground" is the
process of locating and identifying where it is feasible and
appropriate to put them. This involves a process of identi-
fying as many potential sites as rapidly as possible. The
best retrofit sites fit easily into the existing landscape, are
located at or near major drainage or stormwater control
facilities, and are easily accessible. For example, almost
every urban area has some type of existing pond or other
feature that might be adaptable for retrofitting. In many
newer neighborhoods, dry stormwater detention facilities
are present for flood control. In older neighborhoods there
are often aesthetic ponds, or other water features which
can make suitable retrofits. Table 2 lists some of the most
likely spots for locating facilities, some common applica-
tions, and applicable case studies.
Usually the first step is completed in the office using
available topographic mapping (a 5-ft contour interval is
quite satisfactory), low altitude aerial photographs (where
available), storm drain master plans, and land use maps
(zoning or tax maps are best). Scouting for potential sites
should follow the guidance discussed above in Table 2.
Two important tasks need to be undertaken before ventur-
212
-------
Table 1. Basic Elements of a Stormwater Retrofitting Implementation Strategy
Step Elements
Purpose
Preliminary Watershed Retrofit Site Inventory
Field Assessment of Potential Retrofit Sites
Prioritizing of Sites for Implementation
Public Involvement Process
Retrofit Design
Permitting
Construction Administration and Inspections
Maintenance Plan
First cut at identifying potential retrofit sites
To verify that sites are feasible and appropriate
To set up a priority for implementing future sites
To solicit comments and input from the public and
adjacent residents on potential sites
To prepare construction drawings for specific facilities
To obtain the necessary approvals and permits for
specific facilities
To ensure that facilities are constructed properly in
accordance with the design plans
To ensure that facilities are adequately maintained
Table 2. Locations for Stormwater Retrofits
Location
Type of Retrofit
Case Study
Existing Stormwater detention
facilities
Immediately upstream of existing
road culverts
Immediately below or adjacent to
existing storm drain outfalls
Directly within urban drainage and
flood control channels
Highway rights-of-way and
cloverleafs
Within large open spaces, such
as golf courses and parks
Within or adjacent to large parking
lots
Usually retrofitted as a wet pond or Stormwater
wetland capable of multiple storm frequency
manaagement
Often a wet pond, wetland, or extended
detention facility capable of multiple storm
frequency management
Usually water quality-only practices such as
sand filters, vegetative filters or other small
storm treatment facilities
Usually small-scale weirs or other flow
attenuation devices to facilitate settling of
solids within open channels
Can be a variety of practices, but usually
ponds or wetlands
Can be a variety of practices, but usually
ponds or wetlands capable of multiple storm
frequency management
Usually water quality-only facilities such as
sand filters or other organic media filters
(e.g., bioretention)
Wheaton Branch, Sligo Creek, Wheaton, MD - multi-
cell wet pond with extended detention
Epsilon Pond, Redland, MD - dry extended detention
facility
Long Quarter Branch, Towson, MD - gravel based
wetland filter
Indian Creek, College Park, MD - instream concrete
weir flow attenuation! device
Bear Gutter Creek, Route 22, Armonk, NY - combina-
tion wet pond and Stormwater wetland
MeisnerAve Retrofit, Staten Island, New York City-
micro-pool extended detention facility
Kettering Subdivision, Prince Georges Co., MD -
Bioretention practices
ing into the field. First, the drainage area to each potential
retrofit site should be delineated and the potential surface
area of the facility measured. The drainage area can be
used to compute a capture ratio. Capture ratio is the per-
centage of the overall watershed that is being managed
by all retrofit projects. The surface area is used to com-
pute a preliminary storage volume of the facility. A shortcut
storage volume consists of multiplying two-thirds of the
facility surface area times an estimated depth (2/3 SA d).
These two values can be used as a quick screening tool.
In general, an effective retrofitting strategy must capture
at least 50% of the watershed and a minimum target stor-
age volume forwatershed retrofit is approximately 1/2 inch
per impervious acre within the watershed.
Step 2. Field Verification of Candidate Sites
Candidate retrofit sites from Step 1 are investigated in
the field to verify that they are feasible. This field investi-
gation involves a careful assessment of site specific infor-
mation such as presence of sensitive environmental fea-
tures, location of existing utilities, type of adjacent land
uses, condition of receiving waters, construction and main-
tenance access opportunities, and most importantly,
whether or not a desirable retrofit will actually work in the
213
-------
specified location. Usually a conceptual sketch is prepared
and photographs are taken.
One study that incorporated the principles of this pro-
cess was the Longwell Branch Stream Restoration Study
conducted in Westminster, Maryland in 1994. The Longwell
Branch watershed drains approximately two square miles
of moderate-to-high density commercial and residential
land. The area was built prior to stormwater management
requirements, and consequently the stream system was
suffering the typical urban impacts along much of its reach.
This investigation utilized a retrofit inventory form which
provided field investigators with specific information such
as topography, property lines and ownership, storm drain
outfall locations, drainage area, mapped utility locations,
and other important site design features. This data helps
field investigators decide if sites are to be retained or elimi-
nated from further consideration. Longwell Branch retained
approximately two-thirds of the identified sites. Those elimi-
nated involved conflicts with utilities, wetlands or adjacent
properties.
Step 3. Prioritize Sites for Implementation
Once sites have been located and determined to be fea-
sible and practical, the next step is to set up a plan for
future implementation. Even the best stormwater retrofit-
ting programs have limited capital budgets for individual
project design and construction. Therefore, it is prudent to
have an implementation strategy based on a prescribed
set of objectives. For example, in some watersheds, imple-
mentation may be based on a strategy of reducing pollut-
ant loads to receiving waters where the priority of retrofit-
ting might be to go after the "dirtiest" land uses first. But if
the strategy is oriented more toward restoring stream chan-
nel morphology, priority retrofits are targeted to capture
the largest drainage areas and provide the most storage.
Whatever the restoration focus, it is useful to provide a
scoring system that can be used to rank each retrofit site
based on uniform criteria. Atypical scoring system might
include scores for the following items:
• Pollutant removal capability (storage provided and type
of BMP)
• Stream channel protection capability (ability to control
modest flow events)
• Cost of facility (design, construction and maintenance)
• Ability to implement the project (land ownership, con-
struction access, permits)
• Potential for public benefit (education, location within
a priority watershed, visible amenity, supports other
pubic involvement initiatives)
• Percent of watershed capture
Step 4. Public Involvement Process
A successful project must involve the immediate neigh-
bors who will be affected by the changed conditions. Nearly
all retrofits require significant modifications to the existing
environment. Adry detention pond, for example, isforsome
a very desirable area in the community. It is a place to
walk the dog and only rarely is there any water in the facil-
ity. A wet pond or stormwater wetland retrofit, on the other
hand, may have large expanses of water and may have
highly variable water fluctuations. Adjacent owners some-
times resist these changes. In order to gain citizen accep-
tance of retrofits, affected persons must be involved in the
process from the start and throughout the planning, de-
sign and implementation process. Citizens who are in-
formed about the need for, and benefits of, retrofitting are
more likely to accept projects.
Still, some citizens and citizen organizations will never
support a particular project. This is why it is mandatory
that there be an overall planning process which identifies
projects early and allows citizen input before costly field
surveys and engineering are performed. Projects that can-
not satisfy citizen concerns may need to be dropped from
further consideration.
This step of a good retrofit program must utilize a good
public relations plan. Slide shows, or field trips to existing
projects, can be powerful persuasions to skeptical citizens.
Before any site goes forward to final design and permit-
ting, it should be presented at least once to the public.
Step 5. Retrofit Design
The design process is for some, including this author,
the most rewarding part of the process. Here, the concept
is converted from a dream to a construction drawing. De-
sign of retrofit projects incorporates the same elements as
any other BMP project including: adequate hydrologic and
hydraulic modeling, detailed topographic mapping, prop-
erty line establishment, site grading, structural design,
geotechnical investigations, erosion and sediment control
design, construction phasing and staging to name a few.
But BMP design for developing residential areas usually
follows a prescribed design criteria (e.g., control of the 2-
year storm or sizing for a specified water quality volume),
whereas retrofit designers must work backwards from a
set of existing site constraints to arrive at an acceptable
stormwater control.
Sometimes this process yields facilities that are too small
or ineffective, and therefore not practical for further con-
sideration. One such project in Gaithersburg, MD, was re-
cently proposed as a major stormwater wetland upstream
from an existing road culvert to control a 1,000 acre water-
shed. The problem was that only 1/20th of an inch of total
storage (.05") was obtainable. Clearly this facility would
have been a maintenance nightmare and likely would have
done little to remove pollutants or control downstream chan-
nel erosion. The City of Gaithersburg correctly decided not
to pursue the project even though it had already retained
a consultant and spent significant time and money on pre-
liminary design.
The key to successful retrofit design is the ability to maxi-
mize pollutant removal and channel erosion protection
214
-------
while limiting the impacts to adjacent infrastructure, resi-
dents or other properties. Designers must consider issues
like avoiding relocations of existing utilities, minimizing
impacts on existing wetland and forest, maintaining exist-
ing floodplain elevations, complying with dam safety and
dam hazard classification criteria, avoiding bad mainte-
nance situations, and providing adequate construction and
maintenance access to the site.
Step 6. Permitting
Permitting issues for retrofit projects often involve im-
pacts on wetlands, forests and floodplain alterations. Many
of these impacts are either unavoidable or necessary to
achieve reasonable storage targets. Permitting agencies
are primarily focused on ensuring that the impacts have
been minimized to the extent practicable and that the ben-
efits of the proposed project are clearly recognizable.
One recent project in New York City's Staten Island
Bluebelt illustrates this point. A larger detention facility is
being proposed for the Richmond Creek subwatershed to
control a 400-acre headwater drainage area. The facility
was initially conceived to provide a wet pond with wetland
elements and extended detention of runoff from the 1 -year
storm. The facility is proposed within the Bluebelt park sys-
tem where impacts to trees and wetlands were a major
concern to park personnel and regulatory agencies.
Several alternative designs were presented that mini-
mized wetland and forest impact while maximizing stor-
age volume to provide downstream channel erosion pro-
tection. The real balancing act was to achieve enough stor-
age to provide meaningful downstream channel protec-
tion and at the same time minimizing upstream impacts to
a mature forest and wetland. The final acceptable solution
consisted of a micro-pool wet pond with extended deten-
tion for the 1-inch rainfall event and a total disturbance
limit of about a half acre.
Step 7. Construction Administration and
Inspections
Like any major design project, proper construction in-
spection and administration is integral to a successful fa-
cility. This is especially important for retrofit projects. Ret-
rofitting often involves construction of unique or unusual
elements, such as flow splitters, underground sand filters,
or stream diversions. These practices are unfamiliar to
many contractors. Most publicly funded projects are
awarded to the low bidder, who may be qualified to do the
work but has never constructed projects of this nature be-
fore. Therefore, it is almost a necessity to retain the origi-
nal retrofit designer or other qualified professional to an-
swer contractor questions, approve shop drawings, con-
duct regular inspections, hold regular progress meetings,
conduct construction testing, and maintain construction
records.
Step 8. Maintenance Plan
Always the last element to be discussed, and often the
least practiced component of a stormwater management
program, maintenance is extremely important in retrofit
situations. The reasons are simple. Most retrofits are un-
dersized when compared to their new development coun-
terparts, and space is at a premium in urban areas where
many maintenance provisions such as access roads, stock-
piling or staging areas are either absent or woefully under-
sized.
Designers (see Step 5) must balance maintenance ac-
cess and storage volumes (forforebays, catch basins, and
debris trapping areas) with waterquality, flood control, and
the other constraints discussed above. But in Step 8, the
maintenance must be accomplished as designed.
Retrofit Case Studies
1. Example of Retrofitting an Existing Stormwater De-
tention Facility Wheaton Branch, Montgomery County,
MD
The Wheaton Branch facility, located near Wheaton, MD,
is arguably one of the best known modifications of a former
dry detention facility, retrofitted to provide waterquality and
channel protection controls. The facility, constructed in
1990, drains an 800-acre watershed that is over 50% im-
pervious. A unique design feature was the three-cell wet
pond (constructed around an existing sanitary sewer trunk
main) to provide water quality controls. Extended deten-
tion controls for the 1.5-inch rainfall event were incorpo-
rated for channel protection. The three-cell pond has a
complex flow path for both baseflows and small stormflows
to facilitate maximum settling of solids. Controls for larger
storms (i.e., two -100 year rainfall events) were balanced
against upstream backwater constraints and dam safety
considerations. Figure 1 illustrates the key operational and
design elements of the project.
The first cell of the facility, orforebay provided almost a
tenth of an inch of storage per impervious acre within the
watershed (this is too small for most retrofits). A 25-foot
wide access ramp with a level 30' by 30' pad was provided
for future dredging. During the design phase, it was esti-
mated that dredging of the forebay would be necessary
every five years or so. The first cleanout of the forebay
occurred in July 1997, a little over seven years after comple-
tion of the project.
The Wheaton Branch retrofit facility was part of the larger
Sligo Creek watershed restoration project. Downstream
habitat improvement and native fish restocking projects
accompanied the retrofit and have proved very successful
over their seven-year initial period. John Galli, (Galli and
Schueler, 1992)and his colleague Jim Commins (1992)
have published several reports and articles on the suc-
cess of the stream restoration efforts in Wheaton Branch.
Some important design lessons are also illustrated by
the Wheaton project. The existing hydraulic characteris-
tics of the facility were first analyzed to assess the control
originally provided. The original facility provided partial
control of the 2-, 10-, and 100-year storm and safely passed
215
-------
Pond
Embankment
Existing
Townhouse
Development
Flow path
during small
storm events
(< 0.3" runoff)
Figure 1. Wheaton Branch Retrofit key operation and design elements - Montgomery County, MD
216
-------
the probable maximum flood (PMF) through a massive
emergency spillway. The retrofit required a balancing act
to maximize water quality control, while maintaining enough
control for larger storms to avoid impacting downstream
houses or the 100-year floodplain.
Routing storms through the three-cell pond was ex-
tremely difficult due to the very low head conditions and
the unusual backwater created by downstream ponds. The
original pond bottom was excavated for much of the per-
manent pool storage (for pond and wetland components),
the emergency spillway was modified to maintain passage
of the PMF and the outlet control structure was completely
overhauled.
All of these measures added up to quite an expensive
project. The total cost for the facility, including engineer-
ing, construction, and construction inspection was approxi-
mately $800,000. Although this was certainly a healthy sum,
it equates to approximately $640,000 per square mile
($1,000/acre) of drainage area. This is a third less than
the typically quoted figure of approximately $1 million per
square mile of drainage for average retrofitting (Karouna,
1989).
2. Example of a Retrofit in a Highway Right-of- Way Bear
Gutter Creek, Westchester County, NY
The Bear Gutter Creek Retrofit is one of many projects
recently designed to protect the Kensico Reservoir (one of
the principal components of New York City's drinking wa-
ter system) from impacts of stormwater runoff. The Bear
Gutter watershed is approximately one square mile in area
and drains mixed land uses, having approximately 30%
impervious area, directly into the Kensico Reservoir. Note
that this is an unfiltered drinking water system serving mil-
lions of New Yorkers. The retrofit is located immediately
below a state road culvert and within the NY Route 22
right-of-way.
Interesting design features include a flow diversion weir
at the downstream end of an existing large diameter road
culvert which diverts baseflow and stormflow for up to the
1.5-inch rainfall event into a primary settling area. Storms
larger than the 1.5-inch rainfall are diverted to a stabilized
downstream channel below the facility. The primary set-
tling chamber is sized for about 1/3-inch per impervious
acre and has both a wet component and storm storage
above the wet pool. An existing 1 -1 /2-acre emergent wet-
land, adjacent to the facility receives runoff as a polishing
treatment below the primary settling chamber. See Figure
2 for an illustration of the facility and representation of de-
sign features.
The design criteria for the Bear Gutter Creek project (as
well as all of the Kensico project) was to provide a facility
' Inflow
Primary Settling
Chamber
NY State
Route 22
/
:#£
V" M
_ J_ ^'^
o1"//^ _ Low F|ow Retease str
• Inflow
•1,400 Feet
to Bear Gutter
Kensico Reservoir
12"Dia.
Coconut
Fiber Rolls
with Stone
Outlet Windows •
Existing
Emergent/Shrub
Scrub - Non-Tidal
Wetland (Essentially
Undisturbed)
Legend:
Flowpath for Events < 1.5' Rainfall
Wetland Limits
RAG Jan '98
Figure 2. Bear Gutter Creek Retrofit - Illustration and representation of design features
217
-------
with the minimum storage volume necessary to maximize
particulate settling, and provide long detention times to
allow for fecal coliform dieoff. An original design concept
called for siting the facility in the middle of the 1-1/2 acre
wetland. Unfortunately very little space was available within
the road right-of-way or anywhere else outside of the ex-
isting wetland. The solution was to use a flow diversion
structure coupled with a concrete weir and baffle to maxi-
mize the length of the flow path within the primary settling
chamber and then utilize the wetland as a "polishing" treat-
ment. Coconut rolls were specified within the wetland to
encourage additional detention for controls of larger storms.
Summary — Is Retrofitting Really That
Complicated?
The answer to this question might seem elusive. Retro-
fitting can be a daunting task, and usually not an inexpen-
sive one. The key to a successful local program is to fol-
low a systematic and straightforward process toward imple-
mentation. The eight-step process presented above is cer-
tainly not the only way to get projects built. Some jurisdic-
tions identify and construct pilot projects first and then ex-
pand a program from there. Others spend much more time
on planning and public involvement. Whatever the focus,
retrofitting is still more of an art than a science, and plan-
ners and designers who take an approach geared toward
innovation will go a long way toward successfully plan-
ning, designing, and building stormwater retrofit projects.
References
Commins, J. and Stribling, J. 1992. Wheaton Branch Ret-
rofit Project: 1990-91 Biomonitoring Project. ICPRB
Report No. 92-1. Interstate Commission on the
Potomac River Basin, Rockville, MD
Galli, J. and Schueler, T. 1992. Wheaton Branch Stream
Restoration Project. In: Watershed Restoration
Sourcebook. Metropolitan Washington Council of Gov-
ernments, Washington, D.C.
Hazen and Sawyer, Inc. and Center for Watershed Pro-
tection, 1997. Final Site Plan for BMP 67, Sub-Basin
BGC-5. Kensico Watershed Stormwater Best Manage-
ment Facilities. Westchester, NY
Karouna, N. 1989. Cost Analysis of Urban Retrofits. Un-
published Report. Metropolitan Washington Council of
Governments. Washington, D.C.
Watershed Protection Techniques, Vol. 1, No. 4. 1995.
Center for Watershed Protection. Ellicott City, MD
218
-------
Identifying Wetland Restoration Opportunities in the Rouge River Watershed
Donald L. Tilton
Tilton & Associates, Inc.
204 E. Washington St.
Ann Arbor, Michigan
Abstract
This report discusses factors to be considered in identi-
fying wetland restoration opportunities in an urban river
watershed. The discussion is based on a recent study of
wetland restoration opportunities in the Rouge River Wa-
tershed in southeast Michigan, funded by the United States
Environmental Protection Agency through the Rouge River
National Wet Weather Demonstration Project. Wetland
ecosystem restoration or creation in urban settings pre-
sents certain unique challenges compared to similar
projects in rural or undeveloped areas. Identifying appro-
priate sites for wetland restoration in urban settings re-
quires special consideration of the unique characteristics
of urban environments. Environmental challenges fre-
quently encountered in urban settings include contaminated
sources of water, contaminated soils, severe hydrologic
conditions, unsuitable adjacent land uses, and severely
disrupted plant and animal communities in existing wet-
lands. Social and economic issues that influence urban
wetland restoration projects include opportunities for rec-
reational uses of restored wetlands and environmental
education in wetland areas. When wetland restoration
projects account for these factors, remarkable wetland
ecosystems can be restored in urban areas.
Introduction
Wetland restoration has become a widely used part of
many river restoration efforts. Wetland ecosystems pro-
vide many functions to a watershed, such as fish and wild-
life habitat, water quality improvement, flood water stor-
age, and passive recreation. In an urban environment,
these wetland functions have usually been severely im-
pacted over the years so there are benefits to be gained
by restoring wetland habitat as part of any river restoration
effort.
While restoration of wetland habitat has been an active
part of resource management for many years, most wet-
land restoration efforts have been focused in rural areas
as part of wildlife management programs. Organizations
such as the US Fish and Wildlife Service, state wildlife
departments, and private conservation organizations such
as Ducks Unlimited have restored thousands of wetland
acres over the years. The techniques for wetland restora-
tion in agricultural environments have been developed over
many years and are based on many experimental ap-
proaches. The state of the art is such that wetlands can be
restored to provide habitat for specific plant and animal
communities. Several guidebooks are available for the
habitat requirements of certain wildlife species.
Since the 1970s, research has been conducted on the
design and construction of wetlands that provide functions
other than fish and wildlife habitat. The majority of this re-
search has focused on the use of wetlands forwastewater
treatment, but there have also been some notable efforts
to design wetlands for the specific purpose of abating
nonpoint source pollution. The wastewater treatment field
has evolved to the point where design manuals have been
published detailing the design and construction of several
different types of treatment wetlands. Supplementing these
manuals is a series of guidebooks describing design guide-
lines for stormwater treatment wetlands (Schueler, 1992).
While the information and techniques developed for
wetland restoration in wildlife management and nonpoint
source pollution treatment is valuable, the application of
this information to urban wetland restoration projects needs
to be critically evaluated. The urban ecosystem is limited
by physical, chemical, and biological characteristics that
influence the types of wetland ecosystems that can be prac-
tically restored. The urban ecosystem is sufficiently differ-
ent from undisturbed ecosystems that design manuals need
to be modified to reflect the urban environment. The fol-
lowing information is meant to highlight some of the key
aspects of the urban ecosystem that should be consid-
ered when undertaking wetland restoration projects in ur-
ban settings.
Physical and Chemical Factors
Frequency and duration of flooding. The hydrologic char-
acter of urban rivers and streams frequently consists of
widely varying fluctuations in flow rate and water level and
frequent flooding. While wetland habitats are also frequently
flooded, the frequency and duration of flooding in an ur-
219
-------
ban setting is usually greater than in undisturbed rivers
and needs to be considered. The flooding frequency of
certain urban rivers may be severe enough to preclude
the establishment of forested wetlands. Forested wetlands
in northern areas of the United States require a period of
relatively low water levels in order to survive.
The Rouge River is a river ecosystem that is character-
ized by a diversity of hydrologic conditions with a diversity
of frequency and duration of flooding. Some reaches of
the river experience flooding out of the river channel an
average of eight times a year predominantly in the early
spring. Water level fluctuations accompanied by these flood
events typically represent a rise in river water level of six
feet. Wetland restoration sites that were identified in the
Lower Rouge River were all within the floodplain, but at
locations where the water level fluctuations were typically
less than three feet. The decision to locate wetland resto-
ration sites in areas with moderate flood elevations was
made in an effort to reduce the adverse impact of severe
frequent flooding on waterfowl nesting success. All of the
wetland restoration designs accounted for the frequency
and duration of flooding and were designed to allow flood
waters to inundate restored wetlands, thus imitating the
relationship between riparian wetlands and flooding con-
ditions in the river.
Water level fluctuations. In locating wetland restoration
sites, one should consider the magnitude of water level
fluctuations. Wetland habitat that is flooded by three to four
feet of water may be detrimental by flooding wildlife popu-
lations, especially if the flood occurs during the waterfowl
nesting season. Water level fluctuations in wetlands have
been recognized by wildlife biologists as important aspects
of wetland management, and they frequently manage wet-
lands by draining the basin to encourage plant diversity.
The wetland restoration sites in the Rouge River were ex-
pected to receive stormwater runoff during periods of pre-
cipitation, but the character of the watershed draining into
the wetlands was such that the water level in the wetland
would not recede between storm events. While permanent
open water wetlands are important to certain species of
wildlife, the wetland inventory of Wayne County indicated
that open water wetlands were relatively common in the
vicinity of the wetland restoration site. The design of the
wetland restoration was, therefore, intended to produce a
wetland that would have moderate water level fluctuations.
This was accomplished by designing inlet diversion struc-
tures that would divert only a portion of each storm event
into the wetland, thus allowing for minor water level fluc-
tuations typical of undisturbed wetlands along the river.
Simplified hydrologic models of the water balance for the
wetland were used to determine the volume of water to be
diverted into the wetland with each storm event.
Flow rate. Wetland restoration projects should minimize
the flow rate of water flowing into the wetland because of
the damage caused by high velocity to habitat and fish
and wildlife populations. Receiving streams in urban set-
tings are frequently subjected to waterflowing at high veloc-
ity, which causes significant adverse impacts. Wetlands
may be significantly impacted in a similar manner unless
measures are taken to mitigate the potential impacts. One
effective technique that has been used in the Rouge River
is to provide a regulated inlet structure that restricts the
amount of water that flows into the wetland. Bypass struc-
tures are used to protect the wetland from severe velocity
and the accompanying damage to fish and wildlife popula-
tions.
Water quality. Wetland restoration efforts can be severely
constrained in urban efforts unless the water quality of ur-
ban runoff is considered. While wetland habitats have been
recognized as potential nutrient sinks, wetlands can also
be damaged by severe nutrient loading. Several ap-
proaches were used to manage the potential adverse im-
pact of excessive nutrient loading. First, the wetland res-
toration sites in the Rouge River are protected by sedi-
mentation basins prior to the wetland that remove exces-
sive sediment and nutrient loading. The basins were de-
signed with a restricted outlet that allows for floating mate-
rial and sediment to be retained in the sediment basin.
The effect of such sediment basins is to limit sediment load
to the wetland and to provide a location where sediment
can be removed during routine maintenance and opera-
tion. The second approach was to design wetland restora-
tion projects that planned for plant communities adapted
to high nutrient loading. In practice this meant that there
was a predominance of emergent wetlands and a mini-
mum of open water and submergent wetland types.
Bioaccumulative contaminants. Several recent reports
have shown that levels of bioaccumulative contaminants,
such as heavy metals, are higher in fish and wildlife in
stormwater ponds compared to populations in wetlands
not exposed to urban stormwater. The Canadian Wildlife
Service and Environment Canada (Wren, et al.,1997) re-
ported that persistent chemicals bioaccumulate in sedi-
ments, water, and wildlife and, in some locations, the chemi-
cal concentrations exceeded the Ontario Sediment Qual-
ity guidelines for low effect level for aquatic animals. The
report indicated that definitive estimates of exposure and
effects are required to clearly understand the risk to wild-
life populations. Similarly, in a study of fish species in
stormwater treatment ponds in Florida, (Campbell 1994)
reported that red ear sunfish had significantly higher lev-
els of cadmium, nickel, copper, lead, and zinc compared
to fish from control ponds. Bluegill and largemouth bass
collected from stormwater ponds also had significantly
higher levels of heavy metals compared to control ponds.
Wetland restoration projects in the Rouge River ac-
counted for the potential accumulation of contaminants by
incorporating several protective measures. Wetland sites
with heavy industrial uses in the watershed were avoided.
When wetland sites with runoff from commercial areas were
selected, sediment basins were designed to trap contami-
nants bound to sediment particles priortothe runoff enter-
ing the wetland. Restored wetland habitats were also de-
signed to minimize the potential for incidental ingestion of
220
-------
contaminated sediment by fish and wildlife species by lim-
iting the area of open water available for fish and water-
fowl feeding.
Biological Factors
Target wetland communities. Wetland restoration efforts
should be based on the type of wetland habitats that con-
tribute to the restoration of a diverse river ecosystem. In
certain river ecosystems, emergent wetland habitats de-
signed to abate nonpoint pollution sources are important,
while in other regions it may be more appropriate to re-
store riparian forests that support wildlife species and sta-
bilize river banks. The type of wetland to be restored should
be determined prior to searching for potential sites and
should be based on an open review of wetland resources
and natural resource restoration goals for the river eco-
system.
Maps of the wetlands of the Rouge River indicated that
approximately 70% of the wetlands in the watershed have
been lost. The analysis also showed that the most com-
mon type of remaining wetland is forested wetland habitat
concentrated along the river corridor. A review of vegeta-
tion maps of the vegetation prior to European settlement
indicated that while forested wetlands were present, there
were also extensive areas of emergent wetlands in the
watershed. Emergent wetlands are recognized as being
effective in abating nonpoint sources of pollution and are
important areas for fish and wildlife populations. Based on
the historical vegetation data, and the goal to manage
nonpoint source pollution of the river, a decision was made
to restore emergent wetland habitats to the greatest ex-
tent possible.
Existing habitat. Whenever possible wetland restoration
sites should be connected or linked to existing wetland
habitat. An important benefit of this approach is that plant
and wildlife populations in the existing wetlands have an
opportunity to migrate into the new wetland habitat. The
existing wetlands maps for the Rouge River watershed
were used to identify existing habitat along the river corri-
dor. Wetland maps were then compared to maps of hydric
soils to indicate the potential forwetland restoration. Each
existing wetland was visited and assessed for wildlife use
and condition of existing habitat. The result of this analy-
sis of existing wetland habitat and quality was that wet-
land sites were restored adjacent to forested wetlands that
had viable populations of amphibians, reptiles, and water-
fowl.
Nuisance plant, fish, and wildlife species. Wetland res-
toration projects are frequently subjected to significant
damage by nuisance plant, fish, and wildlife species. Carp
can destroy plantings of submerged and emergent plant
species and cause severe turbidity due to feeding activity.
The resultant turbidity can suppress development of sub-
merged aquatic plant communities. Grazing of wetland
plants by geese can be severe and has in some cases
resulted in complete removal of plants in newly planted
wetlands. Nuisance plant species include invasive wetland
plant species, such as purple loosestrife and reed canary
grass, which can dominate a wetland, preventing the de-
velopment of a diverse plant community. On the Rouge
River, carp have been controlled at wetland restoration sites
by limiting the connection of the wetland to the river. When
it was necessary to connect wetlands to the river, a chain
link fence was installed across the connection to the river
to exclude adult carp. However, carp were introduced into
several restoration sites during a flood and manual meth-
ods of control have been necessary. Canadian geese dam-
age has been managed by stringing rope with flags across
the wetland at 50-foot intervals until the wetland plant com-
munity was established. The ropes seem to disturb geese
from landing on the wetland surface thus preventing graz-
ing in the wetland. After the plant community was stabi-
lized, the ropes were removed.
Ecological traps. Ecological traps are areas of attractive
habitat that represents an unsafe condition for plant or
animal populations (Gates and Geysel, 1978). Species
mortality in ecological traps frequently exceeds reproduc-
tion due to the hazards represented in the habitat. A
stormwater treatment basin with steep side slopes and
permanent open water is an example of an ecological trap.
Waterfowl nests located within the storage elevation of the
retention basin can be destroyed during precipitation events
and result in adverse impacts to reproduction of the spe-
cies. Wetland restoration in urban rivers needs to be
planned in such a way that ecological traps are not cre-
ated.
Ecological traps were prevented in the Rouge River wet-
land restoration areas by linking the new wetlands to ex-
isting habitat that offered an opportunity for a sustainable
population of plant and animal species. Buffer areas of
natural vegetation were created around all new wetlands
to provide upland habitat forwetland species that require
upland habitat for part of their life cycles. Wetland restora-
tion sites surrounded by urban or industrial property were
avoided. Finally, all restored wetland areas were linked in
one manner or another with wildlife corridors to allow mi-
gration of wildlife species.
Socioeconomic Factors
Environmental education. Wetland restoration projects
represent an opportunity to provide outdoor classrooms to
local schools and recreation departments. Natural habitat
in urban settings are rare and if a restored wetland area is
planned near a school, the opportunity to explore coop-
erative arrangements with the school may be worthwhile.
The wetland restoration sites in the Rouge River have been
used by the science classes studying aquatic ecology and
soil chemistry, as well as art and creative writing classes.
The wetland is located within one mile of the high school
and is being used by additional classes and programs each
year.
Passive recreation. The emphasis of wetland restora-
tion is the renewal of wetland functions, such as fish and
wildlife habitat, waterquality protection, or flood water stor-
age. When the opportunity arises, however, significant
public benefits can be realized if passive recreation is in-
221
-------
tegrated into the wetland restoration project. In the Rouge
River project, we found that wetlands that had a trail or
public access site were used by residents in the neighbor-
hood. During monitoring visits, the residents would share
with us their experiences in the wetland regarding wildlife
observation or water level fluctuations. Signage at future
wetland restoration projects will assist residents in under-
standing the restoration project and the benefits gained
from restoring wetlands.
References
Campbell, K.R. 1994. Concentration of Heavy Metals As-
sociated with Urban Runoff in Fish Living in Stormwater
Treatment Ponds. Archives of Environmental Contami-
nation and Toxicology 27, 352-356.
Gates, J.E. and L.W. Geysel. 1978. Avian vent dispersion
and fledgling success in field forest ecotones. Ecol-
ogy 59, 871-883.
Schueler, T.R. 1992. Design of stormwater wetland sys-
tems: guidelines for creating diverse and effective
stormwater wetlands in the mid-Atlantic region. Met-
ropolitan Washington Council of Governments. 134 p.
Wren, C.D., C.A. Bishop, D.L. Stewart and G.C. Barrett.
1997. Wildlife and contaminants in constructed wet-
lands and stormwater ponds: current state of knowl-
edge and protocols for monitoring contaminant levels
and effects in wildlife. Technical Report Series Num-
ber 269. Canadian Wildlife Service.108 p.
222
-------
Taking Root: Sowing and Harvesting the Seeds of
Public Involvement and Education
Josephine Powell and Noel Mullett
Wayne County Department of Environment
Detroit, Michigan
Zachare Ball
Environmental Technology and Consulting, Inc.
The Rouge River, a tributary to the Detroit River, in south-
east Michigan, has been well documented as a significant
source of pollution to the Great Lakes System. The Rouge
River Watershed spans approximately 467 square miles
involving 48 communities, three counties, and over 1.5
million residents. The eastern portion of the watershed
consists of much of the old industrial areas of Detroit and
Dearborn. The western and northern portions consist of
newer suburban communities and areas under heavy de-
velopment pressure.
This paper discusses the products, programs, and part-
nerships used by the Public Involvement team of the Rouge
River National Wet Weather Demonstration Project (Rouge
Project) to: 1) first increase watershed awareness in Rouge
River watershed residents; 2) educate them about the pol-
lution sources to the Rouge River; and then 3) involve them
in restoration of the Rouge River by showing them that
small changes in their daily activities can help restore the
river.
Even before the inception of the Rouge Project fouryears
ago, it was clear that a comprehensive public involvement
and education program was necessary to support Rouge
River restoration activities. Asurvey of watershed residents
in 1994 determined that, while few people viewed the
Rouge River as a viable resource because of its pollution,
the majority broadly supported efforts to improve its qual-
ity. The survey determined that a grassroots approach
coupled with a top down strategy was needed. A public
outreach strategy, based on the survey, used the philoso-
phy that communication with Rouge River watershed stake-
holders must be continual, consistent, truthful and always
two-way.1
The strategy identified seven stakeholder groups that
must be educated: the general public, local government
officials, industry and business, environmental and com-
1For information about Rouge Project products or stategy, contact Josephine A.
Powell, Wayne County (Michigan) Department of Environment, Detroit, Ml 48226
Telephone: 313-224-3620.
munity groups, the technical community, the media, and
schools.
So, for instance, as local officials were being educated
about the Rouge River's problems and possible solutions,
the Public Involvement team was developing easy-to-read
fact sheets, brochures and posters for the general public.
At the same time, pilot pollution prevention programs in
three watershed neighborhoods and two business areas
were developed to fashion a watershed-wide program and
fine-tune appropriate messages for the general public and
businesses.
It became clear that several themes had to be used to
educate watershed residents. First of all, stakeholders had
to learn that they lived in a watershed containing 48 com-
munities and parts of three counties. This is a diverse wa-
tershed where land uses range from undeveloped land and
farms on its western edge to heavily urbanized industrial
sections on its eastern boundary. Household incomes are
just as diverse: in Bloomfield Hills, in Oakland County, the
average household income is $150,001 while Highland
Park, in Wayne County, has an average household income
of $9,805. Furthermore, stakeholders needed to under-
stand that the pristine, rural tributary in northern Oakland
County was part of the same watershed as the brownish,
urban river flowing past the massive Ford Rouge Plant
complex in Dearborn.
In addition, since many of those surveyed thought the
pollution in the Rouge River was from industrial sources, a
public education campaign had to explain storm water and
non-point source pollution, and their impacts on the Rouge
River. Finally, the campaign had to explain to all stake-
holders their personal role in Rouge River restoration ef-
forts.
Products
A number of products were created over the past two
years to accomplish our goals. They include:
Activity BooAvThis 12-page booklet was designed for
elementary school students. It contains word searches,
223
-------
crossword puzzles and connect-the-dot pictures that en-
tertain while delivering a message about non-point source
pollution, watershed awareness and personal responsibil-
ity. Over 30,000 books have been delivered to area schools
through requests from teachers in six targeted areas in
the watershed, to schools that participate in the Friends of
the Rouge Education Project and to teachers who attend
education conferences. In addition, 10,000 Rouge River
posters, featuring wildlife that live in or nearthe river, were
distributed to area schools. The books and posters are
also used as a handout at watershed fairs and other com-
munity events.
Placemats: Restaurants in the watershed were ap-
proached about using a Rouge Project placemat that fea-
tured a drawing by a fourth-grade student who won a
Friends of the Rouge poster contest. The back of the
placemat listed Rouge Friendly pollution prevention tips.
Close to 50,000 placemats were distributed to watershed
restaurants and were enormously popular, according to a
survey of restaurant owners. A second placemat, recently
developed, features watershed education, and promotes
recreational activities in the Rouge River. Over 50,000
placemats have been used by area restaurants with an-
other 50,000 on order. Currently, the Rouge Project is ne-
gotiating with a popular fast-food restaurant to feature a
Rouge River placemat in its restaurants during Earth Day
week in April, 1998.
Theater Advertising: Three public service announce-
ments were developed for a Rouge River awareness cam-
paign that ran on theater movie screens prior to the fea-
tured film, for five months. The first one promoted Rouge
Rescue, an annual river clean-up event sponsored by the
well-respected grassroots organization, Friends of the
Rouge; the second featured a Great Blue Heron as an
example of wildlife that lives along the Rouge River; and
the third featured Rouge Friendly household tips.
Newspaper Insert: A local community newspaper chain
agreed to re-visit the Rouge River for a 10-year update to
a section it produced in 1986 entitled " Our River: We dis-
covered it; We settled along its banks; We built homes,
farms and factories; And slowly, steadily we began to kill
it." The update was a 12-page award-winning insert that
reached 160,000 households in 12 communities in the
Rouge River Watershed. Entitled "Changing Currents,"the
section documented the positive changes occurring in the
Rouge River Watershed; discussed pollution prevention
and Rouge-friendly behavior; and included a watershed
map that was later reproduced as a poster.
Door Hanger: To augment storm drain stencilled mes-
sages organized by Friends of the Rouge, a door hanger
in the form of a fish was developed as a leave-behind in-
formation piece for neighborhood residents. Text on the
door hanger explains storm drain stenciling activities and
lists Rouge Friendly home care activities. To date, over
10,000 door hangers have been distributed.
In addition, an ongoing media campaign has been in-
strumental in educating the public that they live in the Rouge
River Watershed; that everyone is part of the pollution prob-
lems in the Rouge and can be part of the solution; and that
with everyone's help, the Rouge River can be restored.
Programs
Pollution prevention programs were pilot-tested in three
watershed neighborhoods and two business areas to fine-
tune public information materials and program elements.
They are:
The Rouge Friendly Neighborhood Program: This pollu-
tion prevention program was piloted in watershed neigh-
borhoods in a number of areas of the watershed to pro-
mote education, river stewardship and storm drain stencil-
ing activities. One pilot area, Brightmoor, is a Detroit neigh-
borhood where the Main Rouge flows through a large park
that borders the area. This densely populated, urban neigh-
borhood is bisected by a business strip of predominately
auto service businesses and has been plagued by illegal
dumping and disinvestment. The local neighborhood or-
ganization is very active in the annual Rouge Rescue event
sponsored by Friends of the Rouge.
Since the inception of the Rouge Friendly Neighborhood
Program, the residents organized a business association
that holds monthly meetings attended by city officials from
the police department, the department of public works and
the local neighborhood city hall and officials from the county
environmental department. Environmental issues are a
standing agenda item at the meetings and several busi-
nesses have qualified to be Rouge Friendly Businesses,
which will be discussed below. The residents, in conjunc-
tion with Friends of the Rouge, hold a spring and fall Rouge
Rescues, and in conjunction with the city parks depart-
ment, received a Rouge Project community grant to re-
move several log jams from the river, stabilize river banks,
plant wildflowers and native grasses and restore nature
trails in the park.
In another pilot neighborhood in a more upscale area,
residents have adopted a wetlands that adjoins their sub-
division, participated in storm drain stenciling activities and
promoted Rouge Friendly lawn and garden tips.
The Rouge Friendly Business Program: The Rouge
Friendly Business Program is a companion pollution pre-
vention program to the neighborhood program and was
modeled after a similar program in Bellevue, Washington.
It is aimed at educating small-to-mid-sized businesses to
recognize that they can positively impact the Rouge River
by making small changes to daily business practices. Since
auto-related businesses are very common in the Rouge
River Watershed, an automotive services roundtable, made
up of representatives of automotive service associations
and the local chamber of commerce and a businessman,
was convened. The group met for nearly a year to review
draft materials, make suggestions about program promo-
tion and help mold the program before it was implemented.
In addition, the industry representatives promoted the pro-
gram in their publications and helped recruit businesses
224
-------
to participate in the program. Currently, there are 20 Rouge
Friendly businesses. In addition, a pilot partnership has
been formed with the state Environmental Assistance Di-
vision and RETAP program to aim a similar effort at the
metal finishers industry. Currently in the planning stage is
the creation of similar roundtables for food services and
the construction services.
Based on what was learned from the neighborhood and
business pilots, the Rouge Project public involvement team
has fine-tuned its watershed awareness messages and is
now promoting them via slogans such as:
• Use your head, you live in a watershed
• Storm drains aren't garbage cans
• When it comes to pollution, every home is waterfront
property
• Everyone is part of the problem and needs to be part
of the solution
• Simple changes can make big differences
These messages were incorporated into new tools, like
magnets and brochures, that were developed in the past
several months. A display, created to stress these mes-
sages, includes a watershed map and pictures of Rouge
Friendly activities and tips. In the past year, the display
has been exhibited at 41 watershed events where public
involvement staff made contacts with approximately 27,000
people.
Partnerships
In order to give our messages momentum, a Rouge
Public Involvement Team was established that has formed
numerous partnerships with many stakeholder groups.
Those partners include:
• Watershed townships and cities that have begun to
use our "Storm Drains" display at their events.
• A resource recovery authority that has coupled its back-
yard composting/yard waste reduction messages with
Rouge River Watershed water quality information and
pollution prevention techniques.
• A newspaper chain that writes about Rouge restora-
tion and awareness activities almost weekly. The news-
paper chain also features a Rouge River guest col-
umn written monthly by watershed stakeholders.
• Neighborhood organizations that distribute Rouge ma-
terials and spearhead stewardship activities for their
section of the Rouge.
• Business associations that helped develop Rouge
Friendly business materials and then spread the word
about Rouge Friendly activities to their membership.
Business owners were also instrumental in restora-
tion activities by making Rouge Friendly activities part
of their daily business practices.
• Friends of the Rouge, a grassroots organization, that
has spread a stewardship message through several
programs, many in conjunction with the Rouge Project.
The Friends of the Rouge Education Project teaches
water quality testing in 100 watershed schools and
oversees a student sampling day of the Rouge River
every May. Students are encouraged to create displays
and videos for a Student Congress later in the month.
The organization also sponsors an annual Rouge Res-
cue clean-up in several watershed communities and
has begun to organize a second clean-up in the fall.
• The Rouge River Remedial Action Plan Advisory Coun-
cil, composed of representatives from local govern-
ment, business, the general public and non-profit or-
ganizations, that helped create a Rouge River Recre-
ational Guide map to be distributed in the spring of
1998. In addition, Rouge Project public involvement
staff participate in various subcommittees of the RRAC
and facilitate public education efforts of these commit-
tees.
• The League of Women Voters which awarded a wet-
lands education fellowship to a member who is a resi-
dent of a Rouge Friendly neighborhood. She has spo-
ken several times locally about the benefits of wet-
lands and made presentations on behalf of the Rouge
Project. League of Women Voters members have also
helped distribute Rouge River educational materials.
• The Greening of Detroit which co-sponsored an Arbor
Day event with the Rouge Project for 200 elementary
school children incorporating a tree planting event with
a demonstration of Rouge Friendly tips.
• Oakland Community College, where an environmen-
tal studies teacher has made the Rouge River a regu-
lar part of her lesson plans. She also has given her
students extra credit if they perform volunteer activi-
ties relating to Rouge River awareness.
Although quantifying our success has been difficult, we
do have a baseline survey that showed that people are
committed to doing what they can do to improve the Rouge
River and a future survey will evaluate if the awareness of
watershed residents has increased since the inception the
public involvement activities. In addition, efforts are un-
derway to better record and report information such as
areas stenciled, stream miles adopted, the number of
Rouge Rescue sites, the number of participating Rouge
Education Project schools, and other volunteer efforts and
opportunities as a way of measuring success and docu-
menting compliance with the public education requirements
of Michigan's voluntary storm water general permit.
Meanwhile, we are heartened by the following:
• The popularity of the Rouge River Activity Book; it has
been a big hit at local teachers' conferences as well as
in watershed schools.
• The number of personal contacts we've made at area
community events.
225
-------
• The popularity of the two Rouge River placemats with
restaurant owners and their customers.
• The interest of local elected officials and municipal em-
ployees who ask for their own copy of our display, and
their willingness to distribute Rouge materials.
• The use of our ideas and products by other watershed
groups nationwide.
• The increased use of the Rouge River for recreational
purposes as people learn of its potential.
Research Findings, Recommendations and
Lessons Learned
Listed below are various recommendations that the
Rouge Project public involvement team would encourage
others to consider when developing public education and
involvement materials or programs. These recommenda-
tions are based on the research findings or advice received
from others, including a Watershed Management Peer
Review.
• To effectively inform and involve watershed residents,
four central themes should be kept in mind: (1) estab-
lishing and maintaining a two-way flow of information
with key stakeholder groups; (2) establishing or ex-
panding educational mechanisms aimed at children;
(3) building partnerships and utilizing communication
networks and resources of existing organizations; and
(4) devising ways to use the various public media to
best advantage.
• Utilize third parties. As much as possible, try to have
information come from sources such as non-govern-
mental organizations or environmental groups, the
state environmental regulatory agency, EPA, and rec-
ognized experts from the academic/scientific commu-
nity.
• Peer-to-Peer communication/education is also vital.
People respond positively to suggested, behavioral
changes if they are being promoted by individuals
whom they know and trust or regard as their peers.
Balance public information materials with public in-
volvement programs. Merely providing information is
not sufficient. Programs and activities must be devised
and implemented that actually involve citizens and
other stakeholder groups in restoration efforts.
Recognize and appreciate that each target audience,
or stakeholder group, will need to be taken through a
"communications continuum" of Awareness, Under-
standing, Involvement and Action.
Scale is important. To the extent practicable, efforts
should be directed to the local level. Implementing edu-
cation and involvement programs at the right scale and
bringing together stakeholders who share a common
resource is critical to the success of public education
programs. The slogan "Think Globally, Act Locally"
surely applies.
Avoid artificial barriers. Involvement opportunities (par-
ticularly in the early stages) should not be limited to
only a few activities. To the extent practicable, indi-
viduals should be able to choose among a menu of
activities or involvement opportunities; and within any
given involvement program, flexibility should be offered
to allow adaptation to local conditions and personal
preferences.
Lead/educate by example. Educational efforts should
not be limited to the general public or community stake-
holder groups. Local government agencies must rec-
ognize the impact that their own municipal operations
have on pollution prevention. Internal staff awareness
and education programs should be developed in the
areas of equipment and vehicle maintenance, golf
course and public lands maintenance, facilities main-
tenance, and land use planning, to name a few.
Keep it fun and celebrate small successes. Remem-
ber, change doesn't happen overnight: it happens in
bits and pieces.
226
-------
The Tollgate Wetland's Educational Experience
Patrick E. Lindemann
Ingham County Drain Commissioner
Mason, Michigan
The Tollgate Drainage District Sewer Separation Project
(TDDSP) involves a watershed of 234 acres, 550 residen-
tial homes, ten commercial properties, over 500 apartment
units, and four governmental agencies. The Project in-
volved the separation of a combined sewer system, and
creation of a wetland detention basin. In addition to its
stormwater detention uses, the wetland serves as a wild-
life refuge, learning center, and, a local point of public out-
reach to bring the community together.
Introduction
Community organizers have long proclaimed the ben-
efit of involving stakeholders in all projects in order to en-
sure their success. Residents of the Tollgate Drainage Dis-
trict/watershed were incorporated early in the 'planning pro-
cess. Methods by which they were incorporated varied.
Guiding principles for educating residents and encourag-
ing their involvement were that 1) an investment in atti-
tude and behavior change is a good investment, and 2)
retrofitting water resource protection into urban environ-
ments will be more successful when accompanied by edu-
cation and public outreach.
This project included a concept development and plan-
ning phase which suggested 1) the implementation of an
education and public outreach program during project con-
struction, and 2) the establishment of a long-term assess-
ment and maintenance phase. During the planning pro-
cess, public outreach was accomplished mainly through
mailings and public meetings These meetings minimized
negative feelings among the stakeholders.
Conflicts existed between the City of Lansing (Michigan)
and Lansing Township due to the City's mandate that the
Township separate the stormwater and sanitary water, as
part of the City's sewer separation project. The traditional
method for disposing of stormwater would be to take it to
the nearest river. This cost was estimated and determined
to be excessive, and likely to put the Township in financial
difficulty. But once the City struck an agreement with the
Township to take their sanitary sewer water, the Township
had to develop some outlet for their stormwater.
TDDSSP employed an unusual method of dealing with
the stormwater. Rather than piping it to the river, the water
was retained in the neighborhood and a wetland complex
was designed to treat the water on site. This method cost
approximately one-third of the traditional method of dis-
posing of stormwater, a savings of $15 million or more.
However, this solution would work only if the residents,
or "owners" of the project, and primary stakeholders,
changed their social behavior to reduce pollution loading,
and act as owners of the wetland complex. In orderto bring
about changes in behaviors, the residents were given in-
formation about how their actions affected the outcomes
of the stormwater management system and its costs.
For purposes of the remainder of this paper, the out-
reach/education approach is divided into two phases. The
first is "during construction," or short-term, and the second
is "after construction," or long-term. Figure 1 shows the
relationship between education/public outreach and con-
struction, and accomodates its impact on the project dur-
ing the process, and on long-term assessment and main-
tenance.
Goals of Education and Public Outreach
The public outreach and education goals "during con-
struction" were two-fold. The first was to facilitate commu-
Concept Development and Planning
Education
and
Public Outreach
Project
Construction
Long Term Assessment and Maintenance
Figure 1. Diagram of Tollgate Drain Project.
227
-------
nity participation and ownership of the multi-objective
project. The second was to change attitudes about the
project. To achieve these goals, it was necessary to study
the demographics of the resident population, and to learn
about their attitudes toward portions of the project.
Three surveys were designed. The first survey indicated
long-term residency, conservative beliefs and distrust of
public officials. Residents also expressed opposition to
change, and did not consider themselves to be environ-
mentalists. It was also learned that their knowledge of en-
vironmental issues was limited. For example, when asked
if they knew what non-point source pollution is, only 17%
of respondents said yes (Figure 2). The survey also indi-
cated that over 65% of respondents fertilized their lawns
with a regular program, and half of those were applied by
professional lawn care companies. Our baseline data, col-
lected via water quality tests performed prior to construc-
tion, showed excessive amounts of fertilizer in the
stormwater.
We concluded that public outreach had to be conducted,
for both short and long-term outcomes, to bring about atti-
tude change and a stronger sense of environmental stew-
ardship. A number of approaches were employed to make
this happen.
Education and Public Outreach During
Construction
Public meetings were organized during both the plan-
ning and construction processes. These meetings served
two main purposes. They assisted in informing the resi-
dents about the oncoming project, and they gave stake-
holders an opportunity to vent their anxieties.
A neighborhood network was established to bring the
community together. The Tollgate District was divided into
11 different sections, and block captains were chosen.
Block captains formed phone trees which were utilized to
disseminate updates on construction activity, organize
I 83%
D17%
Figure 2. Tollgate survey response to the question, "Do you know
what non-point source pollution is?"
neighborhood meetings and block parties. These meet-
ings and block parties were used as forums to educate
residents.
An on-site office was opened in an apartment in the
neighborhood. It was staffed with field inspectors; a hot-
line was established, and information was distributed from
it. Two recent college graduates were hired to work there.
They "leafleted" the neighborhood each week, with updates
on street closures, paving schedules, and other current
project information. They became a positive project pres-
ence and were available to residents to assist in aspects
of daily living that were impaired by the construction, and
answered questions as they arose.
Tours were offered to residents, at their convenience,
and explanations were offered. Twenty-four hour responses
were available for all problems during the construction
phase.
The cost of this during construction outreach was a frac-
tion of the total project. This up-front investment in out-
reach achieved our goals in a cost-effective manner, as
Figure 3 explains.
Besides the pre-construction survey already mentioned,
surveys were also done "during construction" and "after
construction". The first survey was conducted door-to-door
and of the possible 554 single-family homeowner respon-
dents, 549 completed the survey of demographics, behav-
iors and attitudes. The second and third surveys also con-
centrated on attitudes and behaviors. Attitudes toward the
project changed from 35% positive before construction
began to 81% positive after construction was completed.
The public outreach will continue, long term, and will
contain several component parts. First, further surveying
will be conducted. Surveys will determine the success of
public outreach in changing social behaviors and attitudes.
Second, water quality testing will occur on a regular
schedule over the next seven years. Should public out-
reach be successful, there should be a reduction in pollu-
tion loading. Stormwater quality should improve, as atti-
tudes and behaviors change. Data recovered from water
monitoring will become part of a feedback loop designed
to provide information and education about successes and
failures.
Third, staff members of the drain commission have been
trained in all aspects of this project, and are prepared to
respond to stakeholders in ways that encourage contin-
ued partnership in the project. Maintenance needs of the
wetlands and the storm/sewer system should decline as
behaviors change. Should behaviors toward pollution load-
ing not change, maintenance costs will ascend, which we
do not want to happen.
Fourth, education will be ongoing and will not be limited
to this watershed. It will include a variety of participants
and programs. Partnerships have been formed with local
environmental groups such as Urban Options, which has
228
-------
• Staffing
• Field Office
• Supplies
• Hotline
Total
$27,000
$2400
$1000
$300
$30,700
Less than 1/3 of 1 % of the Total Project Budget
Figure 3. Education and public outreach investment (during
construction).
agreed to 1) send representatives to neighborhood block
parties and teach composting techniques, 2) work with lo-
cal landscapers to demonstrate environmentally friendly
landscaping, and 3) provide direct technical assistance to
residents. Local grade schools and other educational in-
stitutions have been contacted and teams have been es-
tablished to develop curriculum items on urban non-point
source pollution issues such as wetland and stormwater
management, pollution reduction, and ecosystem diver-
sity. The Tollgate wetlands will become a living classroom
for field trips for students of all ages.
A reptile and amphibian roundup will take place this
spring. Specimens of varying reptiles and amphibians will
be caught, tagged, and released in the Tollgate wetlands
by second and fifth grade classes from a nearby grade
school. This exercise is but one example of programs de-
signed to involve parents and children in understanding
the complexities and needs of urban wetlands, and the
creatures that live therein.
Other local environmental groups are participating by
building birdhouses and educational kiosks along the path-
way adjacent to the wetlands, and providing information to
assist in educating wetland users from the broader com-
munity. Service clubs will be invited for box lunch tours.
This broader outreach will have an impact on the attitudes
toward the overall sewer separation project in the City of
Lansing.
Conclusion
Planning for this project included a public outreach and
education plan. The savings to the overall project's initial
and long-term costs were substantial. We developed a well-
informed public who are willing to take ownership and who
understand the need to change behaviors. Retrofitting es-
tablished urban environments will be more successful when
accompanied by education and public outreach. The ulti-
mate success or failure of most projects depends on the
willingness of the residents to change their long-standing
practices pertaining to wetlands and runoff.
229
-------
Minneapolis Chain of Lakes Phosphorus Reduction Strategy
Jeffrey Lee
Minneapolis Park and Recreation Board
Minneapolis, Minnesota
Abstract
The Minneapolis Chain of Lakes is a group of five heavily
used urban lakes located approximately 1.5 miles south-
west of downtown Minneapolis, MN. The 8,000-acre wa-
tershed was developed over the last 100 years and is now
completely urbanized. Watershed development and human
activities have lead to water quality degradation over the
last 40 years. In 1991, a diagnostic study of urban runoff
and lake water quality lead to the development of an 8-
year, multi-million dollar watershed management effort. A
key component of the management strategy was quantifi-
cation of phosphorus load reduction goals. The watershed
management plan is based on reductions of inflow phos-
phorus loads to the lakes to attain long-term water quality
goals. The phosphorus load reduction strategies were de-
veloped for each of the lakes based on lake water quality
modeling and watershed load analysis. The load reduc-
tion goals established for each of the lakes are (percent of
total watershed load): Brownie -10%; Cedar- 40%; Lake
of the Isles - 20%; Calhoun - 30%; and Harriet - 20%. The
two main structural best management practices (BMPs)
being implemented are wetland/pond systems and grit
chambers. The structural BMPs, in conjunction with in-
creased street sweeping and public education, are being
implemented to reduce phosphorus loading from this fully
developed watershed.
The Minneapolis Chain of Lakes are located 3 miles
southwest of downtown Minneapolis, MN. The Chain's five
lakes are the central natural resource feature of the Min-
neapolis Chain of Lakes Regional Park. The regional park
receives over 2.25 million visitors per year, making it the
second most heavily used regional park in the Twin Cities
Metropolitan Area. The chain receives urban runoff from a
fully developed 8,000-acre watershed that includes por-
tions of Minneapolis and the adjoining suburban commu-
nities of St. Louis Park and Edina. Major land use catego-
ries presently include residential development (51%), in-
dustrial/commercial (19%), and open space (14%). The
lakes in the Chain (Brownie, Cedar, Isles, Calhoun and
Harriet) are interconnected with navigable channels, cul-
verts or pumping systems. The lakes and the watershed
drain from north to south, flowing from the upper chain
(Brownie to Calhoun), where water is pumped to Harriet,
which discharges to Minnehaha Creek, a tributary to the
Mississippi River. Demand for this study and watershed
management effort arose from public concerns over in-
creased recreational use of the lakes and attendant deterio-
ration in water quality.
Diagnostic Study
The first step in designing the watershed diagnostic pro-
gram was an assessment of previous watershed runoff
studies. Sixteen subwatersheds were selected for
stormwater runoff monitoring in 1991. The selected
subwatersheds accounted for 86.8% of the runoff from the
watershed. Seventy-three percent of the residential area
and 21% of the commercial area in the watershed were
included in the monitored areas. Two subwatersheds, the
Bass and Twin Lake drainage areas, drain through wet-
land complexes prior to discharge to the lakes. These
subwatersheds were combined with other sample sites for
continuous runoff monitoring.
Twin Cities
Metropolitan Area
Figure 1. Chain of Lakes watershed location map (after Osgood,
1998)
230
-------
Minneapolis Chain of Lakes watershed concentration
ranges for nutrients were similar to those of the Nation-
wide Urban Runoff Program (NURP) (Oberts, 1983). Ex-
port rates developed from this study were typically below
those found in other studies (Wilson and Brezonik, 1998).
The stormwater runoff characteristics of importance are
listed below:
• The range of pollutant concentrations were consistent
with concentrations found in the NURP studies.
• The mass of phosphorus and nitrogen input into the
Chain of Lakes from stormwater in 1991 was high
enough to cause water quality degradation.
• Most of the measured pollutants exhibit some type of
seasonal trend, with most concentrations highest in
the spring of the year following snow melt.
• Most of the measured pollutants do not show any ma-
jor trends among watersheds.
• The heavy metal concentrations in stormwater, while
at times quite high, rarely exceed state water quality
standards. Relatively low heavy metal concentrations
reflect the primarily residential nature of the Chain of
Lakes watershed (Wilson, 1993).
Data collected during a 1991 diagnostic study indicated
that the lakes exhibit physical and chemical processes in-
dicative of moderately eutrophic lakes, including
hypolimnetic oxygen depletion, anoxicsediment phospho-
rus release, and late-summer decline in water transpar-
ency. Based on a variety of trophic indicators, Jensen and
Brezonik (1998) concluded that Lake Calhoun, Lake of the
Isles, and Cedar Lake are eutrophic and that Lake Harriet
is mesotrophic. The water quality studies also examined
historical water quality data and the results of
paleolimnological studies; this data indicated that water
quality in the lakes has significantly declined since the early
1900s, but has been relatively constant since the early
1970s. The lakes seem to have attained a degree of equi-
libria with watershed nutrient inputs (Brugam and Speziale,
1983; Lee, 1993).
Vollenweider (1976) examined lake eutrophication with
regard to areal based watershed loadings of phosphorus
and the impact upon the trophic status of a lake with re-
gard to lake mean depth, surface area (and, as such, vol-
ume) and flushing rates. Loading levels which exceed
Vollenweider's permissible level would lead to increases
in lake productivity and accelerated eutrophication. Based
on the 1991 loading levels, Lake Harriet's loadings are only
slightly above the permissible level. Loadings to Lake of
the Isles exceed the permissible level by a factor of 9.3,
Cedar Lake exceeds permissible levels by a factor of 5.1,
and Lake Calhoun exceeds the permissible level by a fac-
tor of 3.8. Lake of the Isles had the highest average phos-
phorus concentration over the 1991 - 1996 growing sea-
son, followed by Lake Calhoun, Cedar Lake, and Lake
Harriet (Jontz and Lee, 1998).
Watershed Management Plan Development
Phosphorus and, to a lesser extent, nitrogen, are impor-
tant in the eutrophication process in lakes. Excessive in-
puts of phosphorus to lakes will lead to an increase in fer-
tility and subsequent algal blooms. Findings of the 1991
Storm Water Runoff study reinforced these concerns and
were useful in designing a watershed management plan.
Barr Engineering Company (1992) and Lee (1993) used
computer models to calculate the pollutant concentrations
associated with particular land uses within the Chain of
Lakes watershed. Those concentrations became the ba-
sis for determining the estimated total annual mass of pol-
lutants (loadings) contributed to the lakes.
Loading Rates (kilogram/hectare of land/year)
Total Phosphorus Total Nitrogen
Open/Green Space Land Use
Residential Land Use
Comm/Mixed Use
0.13
0.54
0.41
1.59
2.52
2.61
Storm event data was reduced using FLUX (Walker
1986), an interactive computer program that calculates
pollutant load and flow weighted mean concentrations
(FWMC). FLUX was used with the continuous flow records
and parameter concentrations to develop a FWMC and
loading (in kg/yr) for sites where both flow and sample
analysis data were available. Annual event mean concen-
tration and literature values were also used to refine FWMC
and load for the monitored and unmonitored watersheds
(Oberts, 1983; Oberts, 1990; Bannerman, etal., 1983).
In 1991, stormwater contributed 50% of the Lake Harriet
phosphorus budget, and the pumped discharge from Lake
Calhoun provided another 21%. Stormwater loadings ac-
counted for over 80% of the phosphorus input to the upper
Chain of Lakes. The total mass of phosphorus and nitro-
gen input to the Chain of Lakes from stormwater in 1991
was much lower than would be expected based on pub-
lished nutrient yields from other urban watersheds
(Mulcahy, 1989).
Table 1. Observed and Modeled Lake Water Quality Conditions for
1991 -1995 and Watershed Phosphorus Loads for 1991 from
BATHTUB
Five-year Mean Annual Conditions
Lake Cedar
Isles
Calhoun
Load Source (kg phosphorus/yr)
Precipitation
External load
Internal load
Other1
Total (kg/yr)
In-Lake Water Quality2
Observed Total P (ng/L)
38.5
220.3
76
9.6
344.4
49
23.4
167.8
34.3
66.5
292
59
95.5
536.1
465
112.3
1208.9
40
1 Advective/diffuse inflows
2 Observed based upon 1991 -1995 average lake conditions and 1991
monitored loadings
continued
231
-------
Table 1. Continued
Predicted Water Quality (BATHTUB model outputs)
Lake Cedar Isles
Calhoun
Load Source (kg phosphorus/yr)
Precipitation
External load
Internal load
Other1
Total (kg/yr)
In-Lake Water Quality3
Predicted Total P (|ig/L)
38.5
220.3
76
9.6
344.4
33.8
23.4
167.8
34.3
66.5
292
51.1
95.5
536.1
465
112.3
1208.9
33.2
3 Predicted water quality based upon BATHTUB model outputs.
Observed TP concentrations greater than predicted concentrations
indicates internal phosphorus loading.
Water budget and nutrient concentrations from the storm
water sampling were input to the BATHTUB lake model to
determine lake interactions. BATHTUB is a lake model de-
signed to facilitate the application of eutrophication mod-
els to lakes and reservoirs (Walker 1986). BATHTUB does
nutrient and water balance calculations, and predicts wa-
ter quality conditions using empirical relationships devel-
oped by Walker (1985). The model also provides a break-
down, by subwatershed, of nutrient loads to each lake.
The BATHTUB model provides diagnostic variables that
summarize the physical, chemical, and biological interac-
tions occurring in the lake. These assessments were used
to evaluate the management options for each of the lakes.
A critical part of the management plan was the develop-
ment of lake water quality goals and watershed loading
reduction strategies by a Citizens Advisory Committee (Lee,
1995; Derby and Lee, 1998). The long-term goals recom-
mended by the committee are based on an analysis of
historical, existing and predicted lake conditions. The goals
represent water quality conditions that are close to pre-
dicted historical levels and are attainable by reducing wa-
tershed phosphorous loads (Heiskary and Walker, 1988;
Wilson and Walker, 1989). The short-term goals represent
immediate lake improvements, attainable through water-
shed management and inlake manipulations, while the
long-term goals require the lakes to equilibrate with re-
duced watershed loads.
The long-term (5-10 year) water quality goals, historic
water quality, and 1992 average conditions for each lake
are:
Lake
Predicted
Historical
Total Phosphorous*
1992 Summer
Mean
Total Phosphorous
Long-term
Mean TP
Water Quality
Goal
Brownie**
Cedar
Lake of the Isles
Calhoun
Harriet
24
21
28
18
19
34
48
74
54
44
35
25
40
25
20
Predicted historical phosphorus derived from Vighi and Chiaudani
(1985) which uses lake water chemistry, specifically alkalinity, to
predict unimpacted water quality.
Brownie Lake conditions based upon 1991 data.
In the Chain of Lakes, as in many lake systems, phos-
phorus is the limiting nutrient. Reducing watershed load-
ings of phosphorus and interrupting internal phosphorus
recycling will reduce blue-green algae numbers and de-
crease the frequency of nuisance algal blooms. Thus, the
watershed management plan is based on reducing phos-
phorus loads. Phosphorus load reduction strategies were
developed for each of the lakes, based on lake water quality
modeling and watershed load analysis. The load reduc-
tion goals (as percent of total watershed load) established
for each of the lakes are as follows:
Lake
Phosphorus Load Reduction
Brownie
Cedar
Lake of the Isles
Calhoun
Harriet
10%
40%
20%
30%
20%
Implementation of Best Management
Practices
Priority management areas were chosen based on the
results of the 1991 Storm Water Monitoring Program and
subsequent modeling efforts. Due to the primary concern
for lake eutrophication problems, phosphorus loads were
used to determine priority watersheds. The priority water-
shed management areas were ranked, based upon an-
nual phosphorus load.
The best management practice selection process in-
cluded a series of analyses that factored in a range of
watershed and BMP characteristics. The pollutant of con-
cern (phosphorus) dictated the main emphasis of the se-
lection process. In many cases, the watershed character-
istics of each subwatershed area, as well as site availabil-
ity and characteristics of each site, limited the size and
type of BMP. For larger pond/wetland systems, the BMP
space requirements were such that large areas of open
space were needed; in all cases, these were parklands.
Consideration for adjoining neighborhoods and environ-
ments/parks required that all BMPs be designed to aes-
thetically blend in and become a neighborhood asset rather
than a "mosquito infested swamp," as many opponents
feared. The cost of BMP installation, and amortized costs
on the basis of pollutant removal efficiency over the life of
the project, were also factored into each BMP decision.
Finally, the maintenance effort for each BMPoverthe long
term dictated which agency would be responsible for long-
term management.
As a result of the BMP selection process, the following
watershed best management practices were selected as
part of the watershed management plan:
Non-structural BMPs
• Water Quality Education
• Catch Basin Stenciling
• Lawn Care Education
232
-------
• Street Sweeping/Cleaning
• Goose Population Management
• Erosion Control Ordinance
Structural BMPs
• Grit Chamber Construction
• Stormwater Wetland/Detention Basin Construction
The Minneapolis Chain of Lakes Clean Water Partner-
ship Implementation Project began in 1994. Implementa-
tion of watershed management practices was through the
cooperative efforts of the Minneapolis Park & Recreation
Board, the cities of Minneapolis and St. Louis Park, the
Minnehaha Creek Watershed District, and the Minnesota
Pollution control Agency. As first steps, regenerative air
street sweepers were purchased and a watershed-wide
education program was implemented. The Minneapolis City
Council enacted and implemented a construction erosion
control ordinance for all construction activities in 1996. Con-
struction of structural BMPs in the Chain of Lakes water-
shed began in 1995. Total projected project costs for the
seven year program are detailed in Table 2.
In-lake management techniques were also implemented
as part of the comprehensive management effort. Prior to
finalizing the management plan, the in-lake management
alternatives were subjected to the analysis suggested by
Cooke, et al., (1993). As part of the comprehensive man-
agement plan, whole-lake alum treatments have been com-
pleted (Cedar and Isles) or are scheduled (Calhoun) for
the lakes. The plan also includes annual aquatic macro-
phyte harvesting.
Structural Best Management Practices
Priority subwatersheds were targeted for BMP installa-
tion. The structural BMPs, in conjunction with increased
street sweeping and public education, are being imple-
mented to reduce phosphorus loading. Grit chambers and
constructed wetland/detention basins will be responsible
Table 2. Minneapolis Chain of Lakes - Clean Water Partnership
Implementation Project Costs for 1994 - 2001 (after Panzer,
1998)
Expenditures
Cost
Education $374,500
Watershed Management Practices
Grit chambers 757,580
Street cleaning 926,500
Stormwater ponds & wetlands 4,594,500
Other 7,500 $6,286,080
In-Lake Improvements
Alum 237,950
Erosion protection 284,890 $522,840
Monitoring Programs
Lakes 465,000
Storm water 433,740
Beach 37,490 $36,230
Total Expenditures $8,119,650
forthe largest reductions in watershed phosphorus loads
to the lakes.
Grit Chambers installed within the existing storm sew-
ers settle sediment from storm water. Grit chambers have
been or will be constructed in the storm water drainage
systems of priority subwatersheds that do not have suffi-
cient open space for detention basins or wetlands. The
efficiency of sedimentation (and corresponding pollutant)
removal is directly related to the size of the structure —
the larger the structure the greater the pollutant removal.
Stormwater runoff was modeled using the P-8 model to
predict the pollutant removal by grit chambers for flows
expected during a typical wet year, dry year, and normal
year (HDR Engineering, 1992). The results of this model-
ing were:
Total Suspended Solids
Total Phosphorus
Lead
COD
3% to 58% removal
1%to26% removal
2% to 56% removal
1%to35% removal
The range in pollutant removal reflects the range of sizes
of grit chambers modeled (10 different sizes). The highest
removals were achieved for a small subwatershed which
was 100% impervious. The systems are designed for a
10-year storm (2.29 inches/hour) with a target reduction of
33 - 42% for total suspended solids (Asgain, 1994).
The greatest pollutant loading occurs during late winter
snow melt and early spring rainfall. Therefore, it is recom-
mended that these structures be cleaned two times per
year: (1) in the fall or winter, in order to have maximum
capacity for settling solids in the snowmelt and (2) in the
spring, in orderto clean solids accumulated from the snow-
melt and to maximize capacity for the remainder of the
year.
The major BMP expenditures forthe Chain of Lakes wa-
tershed management effort have been constructed wet-
lands/detention basins. The priority subwatershed analy-
sis results showed that most of the phosphorus entering
Cedar Lake came from the subwatershed draining through
Twin Lakes. Twin Lakes (a shallow wetland) receives run-
off from about 1,600 acres of residential and commercial
land, discharging over half of Cedar Lake's water and over
60% of the external phosphorus load. The area treated by
this project represents about 85% of Cedar Lake's con-
tributing subwatershed.
A 1994 feasibility study evaluated several BMP alterna-
tives and selected a combination of measures to address
non-point source pollution:
1. Construction of a wet detention pond/wetland in Twin
Lakes Park to treat runoff from approximately 1,600
acres before entering Twin Lakes.
2. Excavation of the Twin Lakes basin to improve the
lake's aesthetic quality and increase phosphorus re-
moval effectiveness (before discharge to Cedar
Lake).
233
-------
3. Construction of a wet detention pond and shallow
wetland on property owned by Minneapolis Park &
Recreation Board, immediately upstream from Cedar
Lake. This would further treat the discharge from Twin
Lakes and the direct runoff diverted from Cedar Lake
(Panzer, 1998).
Modeling results for the Cedar Meadows/Twin Lakes
watershed detention project indicated that the system
would remove 43% of the Cedar Lake watershed phos-
phorus load. Work on the Cedar Lake watershed project
was completed in the spring of 1996.
A similar wet detention pond system is proposed for Lake
Calhoun in 1998. The project is currently in the final de-
sign phase. The Lake Calhoun detention system will treat
runoff from an 897-acre subwatershed draining from the
southwest. This subwatershed currently contributes 37%
of the total phosphorus load to Lake Calhoun. Modeling
for the southwest Lake Calhoun subwatershed detention
system indicates that the system would remove 48% of
the subwatershed phosphorus load and 13% of total Lake
Calhoun watershed phosphorus load.
Conclusions
The use of a wide array of tools made possible the de-
velopment of realistic and attainable water quality goals
for the Chain of Lakes. The realities of urban limnology,
that is, the magnitude of stresses placed on the aquatic
resources, require that new, alternative BMP strategies be
developed. In fully developed watersheds such as the Min-
neapolis Chain of Lakes, retrofit opportunities have ne-
cessitated the conversion of open space (usually parkland)
to detention systems. This conversion requires that aes-
thetic and neighborhood considerations play a major role
in the design process.
The practice of urban limnology requires the adaptation
of ecological concepts and modeling tools to describe and
predict stresses placed on urban lakes. Monitoring and
modeling efforts priorto implementing the watershed man-
agement plan informed goal setting and allowed for pre-
diction of water quality improvements due to management
practices. It is widely understood that BMPs alone will not
prevent lake water quality degradation due to watershed
urbanization. In the case of the Chain of Lakes, historical
data showed that most of the stresses occurred during wa-
tershed development. The lakes have reached a new equi-
libria with current watershed loadings, such that reductions
in external and internal phosphorus loads can be expected
to cause measurable improvements in water quality. Thus,
while BMPs cannot be expected to prevent declines, imple-
mentation of watershed practices after lakes have reached
this new equilibria can be expected to lead to improve-
ments when implemented in conjunction with in-lake man-
agement.
The benefit derived from attaining these goals is im-
proved lake trophic status, which will lead to a reduction in
algae blooms and aesthetically improved lake conditions.
Attainment of the goals will shift Lakes Calhoun and Ce-
dar from their present eutrophic conditions to slightly me-
sotrophic states, while keeping Lake Harriet mesotrophic.
Lake Calhoun and Cedar Lake present the best opportu-
nities for water quality improvement through implementa-
tion of watershed management practices. Monitoring re-
sults suggest that shifts in water quality are already occuring
(see Figure 2).
Lake monitoring results for 1996-97 show that Cedar
Lake and Lake of the Isles water quality is now approach-
ing the lakes' goals (see Table 3). The 1993-97 period in-
cluded years of higherthan average precipitation and some
of the water quality improvements noted may be due to
increased lake flushing. Monitoring over the next 3 years
will further indicate the stability of water quality changes
due to watershed management.
100
80
60
40
20
0
Lake Calhoun
1991 1993 1995 1997
1992 1994 1996
Year
200
150
100
50
0
Lake of the Isles
150
100
50
Cedar Lake
1991 1993 1995 1997
1992 1994 1996
Year
1991 1993 1995 1997
1992 1994 1996
Year
Figure 2. Growing season total phosphorus data for 1991 -1997 for the Minneapolis Chain of Lakes (values in |ig/L; mean, 25th and 75th
percentile box plot).
234
-------
Table 3. Post-implementation Watershed Load Projections, Predicted
In-lake Total Phosphorus Concentrations, and Mean Growing
Season Total Phosphorus Concentrations (1996-1997)
Lake
Cedar
Isles Calhoun
Load Source (kg phosphorus/yr)
Precipitation
External load
Internal load<5»
Other
Total (kg/yr)
Loading reduction (kg phosphorus/yr)
38.5
154.4
38
9.2
240.1
104.3
23.4
155.2
17.15
54.4
250.15
41.85
95.5
366.7
232.5
107.3
802
406.9
In-lake Water Quality
Predicted Total P (|ig/L)<6>
Lake goals TotalP (|ig/L)
G.S. Mean Total P (ng/L)
(96-97)
30
25
27.5
42
40
37
22
25
27.5
(5)
(6)
Internal load reduced by 50% (representing average
reduction over life of treatment).
Predicted values derived using Dillon and Rigler (1974).
References
Asgain, M. (1994). Methodology and assumptions used in
the design of grit chambers. City of Minneapolis Sewer
Planning and Design. Minneapolis, MN.
Bannerman, R., K. Baun, and M. Bohn (1983). Evaluation
of Urban Nonpoint Source Pollution Management in
Milwaukee County, Wisconsin. Wisconsin Department
of Natural Resources, Madison, Wl.
Barr Engineering Company 1992. Minneapolis Chain of
Lakes Clean Water Partnership Project - Stormwater
Monitoring Study. Prepared for the Minneapolis Park
and Recreation Board.
Brugam, R.B. and B.J. Speziale (1983). Human disturbance
and the paleolimnological record of change in the zoop-
lankton community of Lake Harriet, Minnesota. Ecol-
ogy 64(3):578-5Q-\.
Cooke, G.D., E.G. Welch, S.A. Peterson, and PR. Newroth
(1993). Restoration and Management of Lakes and Reser-
voirs. Second edition. Lewis Publishers: Boston, MA.
Derby, E. and J. Lee (1998). Citizen participation in water
quality management. In: E. Derby, J. Lee, and D. Pilger
(Eds.) Minneapolis Lakes and Parks — Special Ses-
sion Proceedings. North American Lake Management
Society International Symposium. Minneapolis, MN.
Dillon, PJ. and F.H. Rigler (1974). The phosphorus-chlo-
rophyll relationship in lakes. Limnology and Oceanog-
raphy 79:767-773.
HDR Engineering, Inc. (1992). Storm Water NPDES Part
II Monitoring and Modeling Results - City of Minne-
apolis Final Report. Prepared for City of Minneapolis
Public Works Department. Minneapolis, MN.
Heiskary, S.A. and W.W. Walker, Jr. (1988). Developing
phosphorus criteria for Minnesota lakes. Lake and
Reservoir Management 4(%):1 -10.
Jensen, K. and P. Brezonik (1998). Water quality trends
in the Minneapolis Chain of Lakes. In: E. Derby, J.
Lee, and D. Pilger (Eds.) Minneapolis Lakes and Parks
— Special Session Proceedings. North American Lake
Management Society International Symposium. Min-
neapolis, MN.
Jontz, J. and J. Lee (1998). An overview of the monitoring
program and water quality of Minneapolis lakes 1992
-1996. In: E. Derby, J. Lee, and D. Pilger (Eds.) Min-
neapolis Lakes and Parks — Special Session Proceed-
ings. North American Lake Management Society In-
ternational Symposium. Minneapolis, MN.
Lee, J.T. (1993). Minneapolis Chain of Lakes: Phase 1 —
Diagnostic Report (final). Minneapolis Park and Rec-
reation Board Water Resources Studies #93-1. Min-
neapolis Park and Recreation Board, Minneapolis, MN.
Lee, J.T. (1995). Public Awareness. In: R. Kirshner, K.
Soulliere, and M. Murphy (Eds.) Empowering Water-
shed Stakeholders: A Train Workshop Proceedings.
Northeast Illinois Planning Commission, Chicago, IL.
Mulcahy, J.P. (1989). Phosphorus export in the Twin Cities
Metropolitan Area. Prepared for the Minnesota Pollu-
tion Control Agency. Metropolitan Council of the Twin
Cities Area, St. Paul, MN.
Oberts, G. (1983). Surface Water Management: Evaluation
of Nationwide Urban Runoff Program. Metropolitan
Council Publication No. 10-83-127, December 1983.
Oberts, G. (1990). Design considerations for management
of urban runoff in wintry conditions. Proceedings of
International Conference on Urban Hydrology under
Wintry Conditions. Narvik, Norway.
Osgood, D. (1998). Lakes of the Twin Cities Metropolitan
Area. In: E. Derby, J. Lee, and D. Pilger (Eds.) Minne-
apolis Lakes and Parks — Special Session Proceed-
ings. North American Lake Management Society In-
ternational Symposium. Minneapolis, MN.
Panzer, M. (1998). The Minneapolis Chain of Lakes resto-
ration project watershed best management practices.
In: E. Derby, J. Lee, and D. Pilger (Eds.) Minneapolis
Lakes and Parks — Special Session Proceedings.
North American Lake Management Society Interna-
tional Symposium. Minneapolis, MN.
Vighi, M. and G. Chiaudani (1985). A simple method to
estimate lake phosphorus concentrations resulting
from natural background loadings. Water Research
79(8):987-991.
Vollenweider, R.A. (1976). Advances in defining critical
loading levels for phosphorus in lake eutrophication.
Mem. 1st. Ital. Idrolbiol. 33:53-83.
Walker, W.W. (1985). Empirical Methods for predicting
Eutrophication in Impoundments; Report 3, Phase II:
235
-------
Model Refinements. Prepared by William W. Walker,
Jr., Environmental Engineer, Concord MA, for the US
Army Engineers Waterways Experiment Station,
Vicksburg, MS.
Walker, W.W (1986). Empirical Methods for predicting
Eutrophication in Impoundments; Report 4, Phase III:
Applications Manual. Prepared by William W. Walker,
Jr., Environmental Engineer, Concord MA, forthe U.S.
Army Engineers Waterways Experiment Station,
Vicksburg, MS.
Wilson, C.B. and W.W. Walker, Jr. (1989). Development
of lake assessment methods based upon the aquatic
ecoregion concept. Lake and Reservoir Management
5(2) :11-22.
Wilson, G.J. (1993). Analysis of Urban Stormwater Qual-
ity from the Minneapolis Chain of Lakes Watershed.
M.S. Thesis. University of Minnesota, Minneapolis,
MN.
Wilson, G. and P. Brezonik (1998). Analysis of urban
stormwater quality from the Minneapolis Chain of
Lakes watershed. In: E. Derby, J. Lee, and D. Pilger
(Eds.) Minneapolis Lakes and Parks — Special Ses-
sion Proceedings. North American Lake Management
Society International Symposium. Minneapolis, MN.
236
-------
Restoration in the Sunshine: Retrofitting the Watersheds of
Two Urban Lakes in Florida
Craig W. Dye, Keith V. Kolasa, and K. Lizanne Garcia
Southwest Florida Water Management District
Brooksville, Florida
Because of its karst geology and generally flat terrain,
Florida is dotted with numerous shallow lakes. The shal-
lowness of the lakes, combined with generally warm tem-
peratures and long periods of sunlight, make these lakes
particularly susceptible to anthropogenically accelerated
eutrophication, resulting in poor habitat and water quality.
In response to citizen complaints regarding degraded wa-
ter quality of two lakes in Pinellas County, the Southwest
Florida Water Management District and two local govern-
ments conducted diagnostic/feasibility studies to develop
restoration strategies and projects. Although the technical
aspects of retrofitting the watersheds and restoring the
lakes were daunting, the inter-governmental coordination,
regulatory, and social hurdles were far more difficult to
overcome and have caused significant delays to restora-
tion activities. This paper will review the watershed retrofit
and restoration activities for Lakes Maggiore and Semi-
nole, and highlight several unique technical solutions pro-
posed forthe projects. The many pitfalls encountered while
conducting these two projects, and strategies for avoiding
them, will be discussed.
Introduction
Florida's landscape is dotted with over 7,000 generally
shallow lakes. Many of these lakes are found within urban
areas or areas undergoing rapid urbanization, particularly
in the Central Florida corridor which includes the metro-
politan areas of Tampa/St. Petersburg, Lakeland/Winter
Haven, and Orlando. Because of Florida's subtropical lo-
cation and climate, the growing season for algae and other
aquatic plants is typically year round. Adding pollutant and
nutrient-laden stormwater runoff to these shallow urban
lakes creates the potential for accelerated eutrophication
and accompanying problems of algal blooms, fish kills,
cattail and hydrilla infestations, and the general degrada-
tion of the ecological and aesthetic qualities of these sys-
tems. Although Florida is a leader in stormwater regula-
tion and treatment, state stormwater regulations have been
in place only since 1984, and most of the urban develop-
ment that is responsible for degrading these urban lakes
occurred long before.
Techniques for effectively treating stormwater (e.g., wet
detention) generally are well established, but depend on a
minimum amount of open land. The difficulties of retrofit-
ting stormwater systems in urban areas are thus largely a
matter of having undeveloped land available upon which
to construct treatment systems. Obviously, undeveloped
land is infrequently available, and if it is, it is typically quite
expensive. Although the technical difficulties of retrofitting
urban watersheds can be daunting, the political and regu-
latory hurdles can be even more challenging. Overlapping
municipal and agency jurisdictions and responsibilities fre-
quently result in "turf wars" that can slow restoration projects
and ultimately increase the cost of implementation. Ad-
ministrative requirements within local governments and
agencies to develop agreements for funding and project
contracts can also significantly delay and increase overall
retrofit costs.
In this paper, case studies of two urban lake restoration
projects currently underway in Pinellas County, FL, will be
presented. These two projects demonstrate the great op-
portunities, as well as the prodigious challenges encoun-
tered while attempting to retrofit highly developed urban
watersheds and improve the ecological and aesthetic quali-
ties of important urban resources.
Lake and Watershed Descriptions
Lakes Maggiore and Seminole are located in Pinellas
County, FL (Figure 1). Lake Maggiore is a 386-acre lake
located just south of downtown St. Petersburg, while 684-
acre Lake Seminole is near the cities of Largo and Semi-
nole in central Pinellas County (Figure 1). Land use in the
watersheds of both lakes (Table 1) is largely residential,
light industrial, and commercial, although there are public
parks located on the shores of both lakes. Additionally, there
is a large upland habitat preserve (Boyd Hill Nature Park)
on the western shore of Lake Maggiore. Of the developed
portions of the watersheds of both lakes, only a small frac-
tion (approximately 11%) receives stormwater treatment
based on 1984 state regulations.
Both lakes support boating, fishing, and water skiing,
although the degraded conditions have curtailed these ac-
tivities, particularly in Lake Maggiore. Lake Maggiore was
the site of a major annual hydroplane race until a driver
237
-------
Pinellas Co.
N
S
Figure 1. Lake Maggiore and Lake Seminole.
Table 1. Land Use Data for Lake Maggiore and Lake Seminole Watersheds, Pinellas County, Florida.
Lake Maggiore1
Acres Percentage
Lake Seminole2
Acres
Percentage
Commercial/lndustrial/Public
Pond
Undeveloped/Open/Recreational
Low Density
Medium Density
High Density
Overall Total
94
64
1448
671
690
87
2297
4.1%
2.8%
63.2%
29.3%
30.1%
3.8%
100%
510
N/A
2554
1766
521
267
3489
15%
N/A
73%
50%
15%
8%
100%
1 Source - Harper, H. and J. Herr, 1994.
2 Source - Southwest Florida Water Management District and Pinellas County Department of Environmental Management, 1992.
was thrown from his boat during a crash, and subsequently
developed a serious waterborne infection which resulted
in hospitalization. Since then, conditions on the lake have
been deemed by the City of St. Petersburg to be too dan-
gerous for water contact sports.
Diagnostic/feasibility studies, conducted for both lakes
in the early 1990's, identified the source and magnitude of
nutrients and pollutants entering the lakes. Not surpris-
ingly, both lakes have high levels of nitrogen, phosphorus,
and chlorophyll (Table 2) and are highly productive. Florida
Trophic State Indices (TSI) were calculated for each lake
and annual averages ranged between 75 and 85 for both,
indicating eutrophicto hypereutrophic conditions. Compari-
sons of water quality with other Florida lakes (Table 3) in-
dicate that both lakes are among the most degraded in the
state. Biological communities of both lakes were dominated
by undesirable species offish, algae, macrophytes (espe-
cially cattails), and invertebrates.
Large monocultures of cattails in both lakes have con-
tributed to general habitat degradation and significant addi-
238
-------
Table 2. \Nater Quality Data for Lake Maggiore and Lake Seminole, Pinnellas County, FL.
Lake Maggiore1
1992-1997
Lake Seminole2
1991-1997
Parameter
Median
Range
Median
Range
Total Phosphorus (mg/L)
Total Nitrogen (mg/L)
Chlorophyll a (ng/L)
Secchi Depth (m)
Florida Trophic State Index
0.059
2.42
87.8
0.28
82.7
0.005-0.394
0.437-9.04
1 4.8-620
0.11-1.1
55.4-104
0.10
1.90
62.6
0.38
80.8
0.02-1.01
0.02-7.53
5.34-144
0.15-0.9
36.4-111.4
1 Source - Southwest Florida Water Management District, unpublished data.
2 Source - Pinellas County Department of Environmental Management, unpublished data.
Table 3. Comparison of Water Quality Data for Lake Maggiore and Lake Seminole (Pinellas County, FL) to Water Quality Data for Florida Lakes
Reported by FDEP (Friedemann and Hand, 1989). The Florida Percentile Shows the Ranking for Each Parameter among the Percentile
Distribution Listed by FDEP for Median Values for their Florida Lake Database for 1970 to 1987.
Lake Maggiore1
Median Value
1992-1997
Percentile of FL Median Values
Total Phosphorus (mg/L)
Total Nitrogen (mg/L)
Chlorophyll a (ng/L)
Secchi Depth (m)
Florida Trophic State Index
0.059
2.42
87.8
0.28
82.7
45 (864 lakes)
86 (781 lakes)
94 (550 lakes)
3 (782 lakes)
97 (756 lakes)
Lake Seminole2
Median Value
1992-1997
Percentile of FL Median Values
Total Phosphorus (mg/L)
Total Nitrogen (mg/L)
Chlorophyll a (ng/L)
Secchi Depth (m)
Florida Trophic State Index
0.10
1.90
62.6
0.38
80.8
67 (864 lakes)
80 (781 lakes)
88 (550 lakes)
5 (782 lakes)
96 (756 lakes)
1 Source - Southwest Florida Water Management District, unpublished data.
2 Source - Pinellas County Department of Environmental Management, unpublished data.
tion of organic materials to the sediments. Additionally, the
profusion of tall stands of cattails have generated numer-
ous citizen complaints regarding obstructed views of the
lakes (particularly Lake Maggiore) and the role cattails play
as havens for mosquitoes, rodents, and undesirable rep-
tiles (alligators and snakes).
Recommendations from both diagnostic studies centered
on treating stormwater runoff, coupled with specific in-lake
restoration projects, to substantially improve water quality
in the two lakes. For Lake Maggiore, this required whole
lake dredging and retrofitting the watershed to apprecia-
bly reduce nutrient sources to the lake. For Lake Semi-
nole, it is expected that some dredging, along with retrofit-
ting, will be required.
Restoration Approach - Lake Maggiore
To initiate the Lake Maggiore restoration project, the city
approached the district to fund a diagnostic/feasibility study.
The district agreed to fund a $160,000 study with in-kind
contributions from the city. The city's Engineering Depart-
ment was designated to manage the entire lake restora-
tion project. Their first task was to establish an advisory
committee charged with providing input to, and oversee-
ing the progress of, the diagnostic study. The committee,
which included local citizens, representatives from envi-
ronmental organizations, and government agencies re-
viewed the results of the diagnostic study and selected a
TSI of 60 as the target water quality goal for the lake.
To achieve the target TSI for Lake Maggiore, it was de-
termined that retrofitting as much of the watershed as pos-
sible was required. This was obviously a technically chal-
lenging and expensive proposal because of the lack of
available undeveloped land for construction of standard
wet detention systems. Although there is substantial un-
developed city-owned land surrounding the lake, includ-
ing a public park and the Boyd Hill Nature Preserve, a strict
city ordinance prohibits stormwater treatment facilities on
these lands. Thus, a creative and less land-intensive solu-
tion was required, and the ultimate treatment option cho-
sen was alum injection.
Alum (aluminum sulfate) has long been used in drinking
water treatment and wastewater treatment to remove
239
-------
particulate matter and phosphorus (P). In Florida, the use
of alum to remove P from stormwater runoff entering lakes
has been effective as a restoration tool for numerous ur-
ban lakes, particularly in the Orlando area. Treatment typi-
cally consists of injecting a calibrated dose of alum solu-
tion into a stormwater stream to precipitate the dissolved
and particulate P and thereby reduce the in-lake P con-
centration; and ultimately reduce algal populations within
the lake. Alum injection systems are compact, and this was
viewed as a desirable feature for installation at Lake
Maggiore. All the pumps, control panels, piping, and alum
storage tank could be placed in a very small area, approxi-
mately the size of a one- or two-car garage. Additionally, a
number of stormwater outfalls to the lake could be treated
from a single system through a manifold. Finally, the cost
of constructing alum injection systems was considerably
less than condemning and purchasing property for con-
struction of standard wet detention systems.
Five alum injection sites were constructed to treat over
63% of the runoff from the watershed and reduce the
amount of P entering the lake by 80%. Construction was
completed in the fall of 1997, and the injection devices
should be in operation by February of 1998. Stormwater
from the remaining untreated portion of the watershed
(eastern side of the lake) will be directed through a wet
detention system to be constructed as part of the proposed
dredging project.
Whole lake dredging was proposed as another major
restoration project for Lake Maggiore. The lake bottom is
covered with over 2.3 million cubic yards of organic de-
posits and fine sediments which are a source of nutrients
and create significant oxygen demand during the warmer
months. Thus, to reach the TSI goal set forthe lake it would
be necessary to remove the bulk of these sediments in
addition to treating the stormwater. The land requirements
needed for this project were particularly difficult to meet
since it was not possible to use the city-owned land around
the lake for processing or disposal of the dredge spoil.
The cost of transporting the spoil to a landfill or other offsite
property was prohibitive, so a unique solution was pro-
posed to dewaterthe sediments in an effort to reduce spoil
disposal costs.
The proposed solution for dewatering the sediments,
developed for the phosphate industry to settle out clay
particles, involved introducing organic polymers into the
dredge spoil, followed by screening the material to sepa-
rate out the aqueous fraction. Untreated dredged material
from Lake Maggiore is expected to have a solids content
of 4-6%, while the polymerized material is expected to have
a solids content of 10-15% immediately after treatment and
30-40% after rapid drying. This treated spoil material could
be easily transported within seven days of dredging
(weather permitting).
Despite reducing the amount of spoil to be transported,
it still would not be possible to cost- effectively transport all
the sediments to a landfill. Therefore, an option was in-
vestigated that involved pumping a portion of the dredge
spoil to several old borrow pits located within the city. Be-
cause of opposition from the public and the State Depart-
ments of Transportation and Environmental Protection, this
option was abandoned. Subsequently, another option was
developed that proposed that a portion of the lake could
be filled with the dewatered spoil to create a public park,
and a wet detention system could be built to treat the
stormwater runoff from the eastern watershed. Although
the proposal included filling in 34 acres of the 386 acre
lake to accomplish this task, it was believed that the water
quality and habitat benefits were worth transforming that
amount of lake bottom. However, regulatory and permit-
ting agencies did not completely agree with our cost/ben-
efit analyses and at this time the permit is still being re-
viewed by the US Army Corps of Engineers. When, and if,
the project is approved, it should take 18 months to two
years to complete.
Several smaller projects were proposed and have been
implemented for Lake Maggiore, including the construc-
tion of a new operable outfall structure from the lake to
allow future drawdowns and sediment consolidation, the
removal of large cattail stands on the eastern shoreline
and revegetation of the site with desirable aquatic plants,
and the purchase of an aquatic plant harvester to remove
cattails and other undesirable aquatic plants on an ongo-
ing basis.
Restoration Approach - Lake Seminole
As with Lake Maggiore, a local government, Pinellas
County, sought funding from the District for a diagnostic/
feasibility study of Lake Seminole. The District expended
over $400,000 on a detailed study to document the sources
and magnitude of nutrients and pollutants entering the lake.
Pinellas County put its Department of Environmental Man-
agement (DEM) in charge of the project. Like the city, DEM
formed an advisory committee comprising similar interest
groups and agencies. Once the study was completed, the
committee recommended a water quality target TSI of 60,
the same as for Lake Maggiore.
One of the conclusions of the diagnostic study, that the
eutrophic condition of the lake was the result of untreated
stormwater runoff entering the lake, was similar to that of
the Lake Maggiore study. In contrast to Lake Maggiore,
the Lake Seminole study identified only small pockets of
organic sediment within the lake basin and, thus, no large
dredging project was proposed. Retrofitting the Lake Semi-
nole watershed posed the same difficulties encountered
at Lake Maggiore, since there was only a small parcel of
open land potentially available for standard wet detention
treatment systems. The parcel, which the county purchased
for $1.9 million, comprised five usable upland acres on the
shore of the lake and was the last open land available
around the lake.
The county has not yet settled on one type of retrofit
system forthe watershed. Rather, a consultant is prepar-
ing a watershed management plan that is expected to pro-
240
-------
pose a mix of retrofit solutions. The original diagnostic study
identified several sub-basins within the watershed that
contribute significant amounts of nutrients to the lake, and
within which several remedial actions could be imple-
mented immediately. On the five-acre parcel, a wet deten-
tion system was proposed that would incorporate a state-
of-the-art system employing a residence time of nearly 14
days. At another site, an existing retention pond was rede-
signed to improve its treatment capabilities by reducing
the slope, meandering the banks of the pond, and planting
native wetland vegetation. An operable structure has been
installed at the pond outfall to allow for drawdowns and
cattail removal. Finally, untreated runoff from a 15-acre
residential area will be routed through a refurbished wet-
land on the property of a local junior college. Like the five
acre parcel, this system will provide a 14-day residence
time prior to discharge. Once the watershed management
plan is complete, additional retrofit projects will be proposed
and implemented.
A unique retrofit project currently being considered for
the restoration of Lake Seminole incorporates two state-of
the-art technologies. This project proposes to link a vortex
trash and sediment collector (CDS Technology) with an
alum injection system. The vortex system would remove
large sediment particles and trash while the alum injection
system would inactivate the soluble phosphorus in the run-
off. Both systems require only small amounts of land for
installation and may fit into existing stormwater collection
systems. Such integrated systems may be especially well
suited to the Lake Seminole watershed, since many of the
streets on the western (heavily developed portion) shore
of the lake collect stormwater and discharge directly to the
lake through pipes at the end of each street. A total of 75
such discharges have been identified along the western
shoreline of Lake Seminole.
Like Lake Maggiore, Lake Seminole has expansive
growths of cattails that have degraded the littoral habitat
and caused citizen complaints about obstructed lake views.
The county experimented with several cattail removal/
revegetation projects and, as with the city, was provided
with an aquatic plant harvester purchased by the district.
These projects met with some difficulty. Particularly the
revegetation portion, where grass carp, introduced earlier
to eradicate hydrilla, ate the newly planted vegetation. Ul-
timately plastic fences were constructed to bar the carp
from these sites until the new vegetation became properly
established; however, cattail removal has proven to be an
expensive and problem-plagued undertaking.
Lake Seminole discharges over a fixed crest weir to Boca
Ciega Bay during high water periods. A recommendation
from the diagnostic study was to replace the weir with an
operable structure so the lake could be periodically drawn
down for sediment consolidation purposes and cattail re-
moval. This structure is currently in the design stage and
is expected to allow a maximum three feet fluctuation in
lake levels.
Success is a Four Letter Word
Although funding such large scale lake restoration
projects is often a difficult undertaking, all three entities
involved in both projects (the district, the City of St. Pe-
tersburg, and Pinellas County) initially were enthusiastic
about providing adequate funding. The district provided
$5 million for each lake while the city and county were
responsible for an additional $5 million each for their spe-
cific project. Ten million dollars for each project seems
generous, but as the recommendations of the diagnostic
studies reached the design stage it was obvious that timely
implementation of the retrofit and in-lake restoration
projects was going to be difficult to fully achieve. And it is
well known that "time is money."
For the restoration of Lake Maggiore, the city had to
win the confidence of a very active environmental group
associated with the Boyd Hill Nature Preserve. Previous
adversarial interactions with city government had caused
this group to be suspicious of any "environmental improve-
ment" project sponsored by the city. Although the city had
disbanded the advisory committee, a number of meetings
were held with this group and homeowner committees to
allay concerns and, ultimately, these groups endorsed the
overall project; however, they vehemently opposed any
use of public lands (park land) for either retrofit or dredg-
ing. This opposition had serious financial implications for
both aspects of the restoration in that (1) land would not
be available to construct small ponds at the alum injec-
tion sites for collecting the precipitate prior to disposal at
a landfill and (2) there would be no land available around
the lake or any other nearby public land on which to place
the dredge spoil. Both these difficulties required time and
project redesign to solve.
The alum pond problem was resolved when the city
Engineering Department approached the owners of a
nearby private golf club and sought their approval to use
the course's water features as alum collection ponds. In
exchange for improvements to several drainage features
on the golf course, club officials agreed to the city's re-
quest, and the alum injection project was able to move
forward.
In contrast, the dredging was (and still is) far more chal-
lenging. In addition to the lack of available land, the dredg-
ing project ran into serious regulatory difficulties. Although
it would seem the environmental benefits and public in-
terest being served would help propel the project quickly
through the permitting process, it became obvious that
this was wishful thinking. In fact, because this project was
not a private venture but was government sponsored, it
seemed as though the restoration team was compelled to
jump additional regulatory hurdles to ensure compliance
with every last letter of the law. Dredging the lake pre-
sented a number of regulatory issues that were capable
of being resolved relatively quickly with the state agency
responsible for issuing the permit; however, because there
was no spoil disposal site within the watershed, the project
was redesigned to place the spoil on 34 acres of lake bot-
tom adjacent to the eastern shore. This aspect of the
project was and continues to be a permitting challenge,
since placing spoil on sovereign submerged land in Florida
is severely restricted and requires the approval of the Gov-
241
-------
ernor and Cabinet. Additionally, Lake Maggiore and all of
the natural waterbodies of the county are considered Out-
standing Florida Waters. This special protective state des-
ignation limits any developmental activities within the
waterbody, especially dredging and filling.
To make this design acceptable to the permitting agen-
cies, additional stormwater treatment was incorporated
within the 34-acre spoil site, and the city designed a park
with extensive boardwalks and environmental exhibits.
However, the features designed into the site were of less
concern to the regulatory authorities than the fill itself. Ul-
timately the site was redesigned to allow only eight upland
acres for the park with the remaining 26 acres used for
constructed wetlands and littoral zones. The state permit
was obtained after over one and a half years of frustrating,
contentious meetings and failed compromises, including
three trips to the state capital to win approval of the Gover-
nor and Cabinet. Soon after the state permit was received
and bids were being prepared for a dredging contractor,
the Corps of Engineers (COE) asserted their displeasure
with the design, which has further delayed permitting (in
fact, the COE has still not issued a permit or ruling). The
economic feasibility of conducting this project is now in
question, since the various delays and project redesigns
(including transporting the bulk of the spoil ten miles for
disposal) have escalated the costs beyond the initial project
budget.
For Lake Seminole, the regulatory issues to date have
not been as onerous; however, the administrative difficul-
ties of implementing projects have been as frustrating as
the permitting problems for Lake Maggiore. Approval of
funding agreements and project contracts between Pinellas
County and the district was often a difficult and time-con-
suming ("time is money," again) effort because of the many
steps involved in the approval processes for both entities.
In one egregious case, a full year was spent ("time is
money") developing a contract with language acceptable
to lawyers and contract managers for the county and Dis-
trict. This issue was finally resolved when the District's
project manager requested a meeting of all lawyers and
contracts personnel from both entities. This illustrates the
need to account for the slow pace of administrative pro-
cesses in any restoration time-line.
Although the diagnostic study for Lake Seminole and
recommendations for restoration activities were completed
in 1992, project implementation has been slow. In addition
to the contract approval process, the county has taken a
very cautious approach to selecting projects for retrofitting
the watershed and restoring the lake. This is the result of a
proposed retrofit plan for another watershed in the county
that generated extremely negative outcries from the citi-
zens involved. Since then, county staff have developed a
very deliberate process for evaluating projects forthe Semi-
nole watershed and have provided numerous opportuni-
ties for public input. Although several retrofit projects were
selected for immediate design as a result of the 1992 di-
agnostic study, the county insisted on having a watershed
management plan developed before proceeding with ad-
ditional retrofit projects. In the preparation of this plan, the
consultant is duplicating some of the work conducted in
the original diagnostic study, and this plan will not be avail-
able until the summer of 1998.
Changes in project managers at both the district and
county and changes in project implementation responsi-
bilities at the county have resulted in miscommunication
and project coordination problems that have delayed the
overall restoration effort for Lake Seminole. In 1996, the
county shifted responsibility for conducting the project from
the DEM to the Engineering Department and simulta-
neously changed project managers. At about the same
time, the district also changed project managers. Overall,
these changes have been very positive, but have led to
delays in several projects while the personnel involved
became familiar with the projects and each other.
Unlike the city, the county had maintained an advisory
committee of some kind since the inception of the restora-
tion project in 1991. This committee originally consisted of
staff from various government agencies, as well as con-
cerned citizens. Initially, there was a great deal of enthusi-
asm from citizen members who had felt that once the di-
agnostic study was complete implementation of restora-
tion projects would begin and their lake would soon be
restored. Their enthusiasm was manifested in the estab-
lishment of a lake protection group that sponsored lake
cleanup days, stormdrain painting efforts, and lobbying of
the agencies involved to expedite restoration projects. Once
the actual project timelines became clear, they became
frustrated with the slow pace of implementation, their en-
thusiasm rapidly waned, and their attendance at the com-
mittee meetings ended. Now, the committee is composed
almost entirely of agency staff and consultants and is sin-
gularly lacking in citizen participation. As with Lake
Maggiore this lack of citizen involvement may hinder the
implementation of future projects.
Lessons Learned (The Hard Way!)
The first step of any lake restoration project is the comple-
tion of a thorough diagnostic/feasibility study to determine
the source and magnitude of pollutants influencing water
quality and to develop remedial strategies. This may seem
obvious, but many citizens and municipal officials see this
work as a waste of money since they claim to know the
cause of problems in their lake. The expenditure of mil-
lions of dollars on projects designed on anecdotal evidence
would be negligent on the part of any responsible lake
manager and should be vigorously resisted.
Once retrofit or in-lake restoration projects and their costs
have been identified, all possible funding options should
be sought. If the financing of a particular project can be
split between several entities, the project is often easier to
sell to those responsible for approving the funds since le-
veraging limited funding, especially in the form of a "pub-
lic/private" partnership, is attractive to most local govern-
ments. Again, this strategy would seem obvious, but other
242
-------
state, private, and federal funding sources are often over-
looked.
Solid intergovernmental cooperation and coordination
are essential to the timely and effective implementation of
any restoration project; however, these elements of effec-
tive communication are often lost in the "turf' battles that
can occur between and inside large agencies. If permits
are needed to implement a retrofit or in-lake project, en-
sure that all appropriate agencies and authorities (local,
regional, state, and federal) are included and participate
from the beginning. Costly delays to the Maggiore and
Seminole projects have resulted not only from classic bu-
reaucratic inertia, but also from not providing sufficient in-
formation to the permitting agencies responsible for ap-
proving the projects. More frequent meetings with all the
regulatory personnel involved may have helped identify
permitting problems, and thus lessened the time spent
seeking permits and developing alternative solutions for
unpermittable activities. However, the narrow interpreta-
tion by the regulating agencies of the rules governing the
projects and the negligible credit received for the "net en-
vironmental benefit" of the overall restoration remains a
source of extreme frustration forthe project managers and
citizens who will benefit from the projects.
For both Lakes Maggiore and Seminole, the most sig-
nificant lesson learned was the need to thoroughly edu-
cate the public, the regulatory/permitting staff, and the poli-
ticians as to the goals, objectives, and time-lines of the
overall projects. Garnering public support for lake restora-
tion is absolutely crucial in all aspects of the project, from
funding to implementation. Maintaining the public's inter-
est in a typically lengthy restoration process requires that
small, quickly implementable projects be proposed and
completed to demonstrate some immediate benefit to the
citizens. And, as noted earlier, adequately educating poli-
ticians and regulators will be the only way these projects
can be appropriately funded and permitted in a timely
manner.
Acknowledgments
The authors gratefully acknowledge the following indi-
viduals fortheir assistance in preparing this paper: Charles
Courtney, King Engineering Associates, Inc.; Marty Kelly,
Douglas Leeper, and Patricia Twardosky, Southwest Florida
Water Management District; Michael Link, City of St. Pe-
tersburg; Sandra McDonald, Nancy Page, and David
Talhouk, Pinellas County; Rick Powers, Bromwell Carrier,
Inc.; Jim Cockerill and Art Hill, CDS Technologies, Inc.
References
Friedemann, M. and J. Hand. 1989. Typical Water Quality
Values for Florida's Lakes, Streams, and Estuaries.
Standards and Monitoring Section, Bureau of Surface
Water Management, Florida Department of Environ-
mental Regulation, Tallahassee, FL. 33 p.
Harper, H. and J. Herr. 1994. Lake Maggiore Restoration
Stormwater Pre-treatment Ponds/Alum Treatment Sys-
tems. Environmental Research and Design, Inc., Or-
lando, FL. 375 p.
Southwest Florida Water Management District and Pinellas
County Department of Environmental Management.
1992. Lake Seminole Diagnostic/Feasibility Study. S.W
Florida Water Management District, Brooksville, FL.
310 p.
243
-------
Retrofit Study for the Lower Neshaminy Creek Watershed
George Townsend and Mary Beth Corrigan
Tetra Tech, Inc.
Fairfax, Virginia
Terri Bentley
Bucks County Planning Commission
Doylestown, Pennsylvania
David Athey, P.E.
Tetra Tech, Inc.
Christiana, Delaware
The Lower Neshaminy Creek watershed is a densely
developed area in the Bucks County, Pennsylvania, coastal
zone. Historically, stormwater best management practices
(BMPs) consisted of stormwater management basins de-
signed only to control flooding. Water quality was a sec-
ondary concern. Retrofit options to meet the water quality
control requirements of section 6217 of the Coastal Zone
Act Reauthorization Amendments (CZARA) have been
developed. Fifteen basins having different design features
and draining different land uses were selected as examples.
Retrofit recommendations focus on low-cost design and
maintenance options that landowners or local governments
can implement. Graphics were developed that demonstrate
the problems and illustrate the retrofit plans for each ba-
sin. The Bucks County Planning Commission plans to fund
the modifications to one basin as a demonstration project.
The hydrologicand water quality benefits associated with
a range of additional BMPs that are appropriate to Bucks
County were also evaluated. A matrix summarizing the
water quality benefits of these BMPs was developed. Imple-
mentation costs and maintenance requirements were ana-
lyzed. An easily understandable "how-to" document tar-
geted at local government officials and staff, developers,
and people who maintain stormwater management facili-
ties was developed to present the information on retrofit
options and costs, maintenance requirements, and BMP
selection.
Introduction
In July 1995 the Bucks County Planning Commission
(BCPC) entered into a grant agreement with the Pennsyl-
vania Department of Environmental Protection (DEP) and
the National Oceanic and Atmospheric Administration
(NOAA) to complete a study involving modifying options
for existing stormwater management basins in the Lower
Neshaminy Creek watershed. The study focused on
stormwater basins and their effect on stream water qual-
ity. By identifying problems and possible solutions forthese
basins, the BCPC hoped to develop guidance that could
be used by local municipalities to repair malfunctioning
basins or to learn to design better ones in the future. Guid-
ance was also developed on the selection of appropriate
BMPs for different development situations and the cost/
benefit of selected BMPs.
Study Area
The predominant land cover in the Lower Neshaminy
study area is single-family residential housing (2 to 4 dwell-
ing units per acre). Other significant amounts of land cover
in the area include multifamily residential buildings (>4 units
per acre) and commercial and manufacturing structures.
Minor land cover in the area includes transportation, utili-
ties, community service, military, recreation, agriculture,
mining, and vacant land. Projected future land uses in-
clude predominantly high-density residential, as well as
commercial, industrial, and resource protection.
Retrofit Options for Existing Stormwater
Basins
Some residential and commercial developments in the
Lower Neshaminy Creek watershed have been built with
one or several runoff controls on site. Most of these facili-
ties were designed to control flooding and other hydro-
logic impacts from development, with little consideration
forwaterquality control. These basins could be redesigned
and retrofitted to address both the water quality and water
quantity concerns in the watershed.
Stormwater Management Basins Used in
the Study
Seventeen stormwater management basins in the Lower
Neshaminy Creek watershed were assessed for potential
244
-------
retrofitting/redesign options (Figure 1). The survey of ba-
sin conditions was conducted during several site visits to
the watershed during February through July of 1995. The
immediate surrounding areas were also observed for prob-
lems that might have been caused by the basins, such as
flooding or eroded conditions. General information regard-
ing the physical appearance and condition of the basins,
and problems including erosion, improper maintenance,
and malfunctioning basin components, were noted.
Basin Summaries
After the site visits and review of the available site plans,
summaries were compiled to provide a general descrip-
tion of the conditions at each basin. Table 1, gives examples
of basin descriptive information, and the problems and
suggestions for remediation. Figure 2 presents the condi-
tions at an example basin and the types of retrofit options
developed. Table 2 is a summary of the problems observed
in the basins and potential solutions.
bOWER
SOUTHAMPTON
STUDY BASIN LOCATIONS
Note: All locations are approximate.
Source: Bucks County Planning Commission, 1994
Outlet to the
Delaware River
Figure 1. Map of study area, Lower Neshaminy Creek, stormwater management and water quality study.
245
-------
Table 1. Example Basin Summaries
Basin
Description
Problems
Suggestions
Basin 70 - Neshaminy
Square Shopping
Center
Basin 72- Highland
Avenue
Basin 76 - Old Lincoln
Highway
Basin 77 - Bensalem
Township Industrial
Park I
Basin 78 - Bensalem
Industrial Park II
Three-stage outlet structure
Basin appears to pick up runoff
from all of parking lot and entrance
road as well as rooftops (down-
spouts discharge to surface)
Perforated riser; designed for
water quality attenuation
Outlets to street system
Development very new
Concrete overflow spillway
Cigar-shaped (long and narrow)
with steep longitudinal slope
Good stand of grass
Approximately 8-inch diameter
outlet pipe
Purpose unknown; majority of
parking lot bypasses basin and
there is no evidence of roof
downspouts
Large outlet pipe
Outlet to enclosed system;
riprap overflow spillway
Steep side slopes
Unconventional shape - long
and linear with bulb at top
Two-stage outlet
Parking areas and rooftops
drain to the basin
Significant erosion at far inlet manhole
and catch basin at parking lot
Not well maintained: a lot of trash and
discard items
Fairly steep side slopes with dead
brush at several locations, very
marshy on bottom
No water quality component
Algae present in both the basin and
the discharge channel
Poorly maintained basin with poor
growth of grass
Needs to be regraded since invert of
outlet is higher than portions of basin
(causing ponding to occur)
Very steep slopes (slopes should be
less for water quality purposes.)
Erosion around riprap at inlet
No water quality component
Minor amount of marshy bottom
Inlet and outlet adjacent to each other
(short-circuit)
No water quality component
• Partially clogged inlets with some
ponding (both typical)
• No water quality component
• Many tall weeds at bottom, which
inhibit drainage
• Some erosion on side slopes
Perform maintenance (e.g., fix
erosion, clean debris)
Add orifice plate*
Excavate the basin and make it
into a pond
Add forebays at each inlet
Add an upstream basin
Regrade side slopes where
possible and add vegetation
Install oil and grit separators at
existing manholes
Install water quality or sand filter
inlets*
Regrade and reseed the slope
Plant additional vegetation
Modify slopes; perhaps add
terracing
Add orifice plate*
Add forebay at inlet
Intercept parking lot that bypasses
the basin
Add vegetation
Add water quality inlets*
Add series of riprap or timber
checkdams
Add orifice plate*
Regrade or add forebays to
minimize short-circuiting
Add vegetation
Add water quality inlets*
Perform routine maintenance
Add forebays to minimize short
circuiting at inlet 2
Regrade where possible
Add water quality inlets and/or
orifice plate*
'Flood protection from large storms may be lessened when orifice plates and/or water quality inlets and outlets are installed.
It is important to note that the observations and the rec-
ommended renovations to these facilities or sites were of-
fered for voluntary adoption by municipalities and facility
owners. The retrofit procedures described in the report were
not intended to be mandatory. The municipal officials and
facility owners were strongly encouraged to implement any
and all of the measures identified for the upgrades through
mutual agreement or cooperative effort. The result of this
implementation might be to reduce some types of basin
failures (e.g., outlet failure, side slope failure, scouring,
standing water in dry facilities), increase overall water qual-
ity associated with stormwater runoff in urban areas, and
reduce the need for long-term rehabilitation or repair of
these facilities.
Costs and Benefits Associated with Urban
Runoff Controls
Costs of nonpoint source controls vary from site to site
and area to area. Because of the variety of options avail-
able for controlling urban runoff, it is difficult to pinpoint
exact costs for runoff control. An important factorthat needs
to be considered in determining such costs is the cost of
maintaining facilities. Volume 2 of the Neshaminy Creek
Nonpoint Pollution and Wetlands Study (Bucks County
Planning Commission, 1994) presents information on the
relative cost and benefit of various nonpoint source con-
trol practices. EPA's Guidance Specifying Management
Measures for Sources of Nonpoint Pollution in Coastal
Waters (1993) also presents detailed information on rela-
246
-------
Algae
Dead Brush
Trash and
Debris
Marshy
Bottom
Eroded Manhole
and Inlet
T k i
Loading
Area
Parking
I
Shopping
Center
Basin 70
Neshaminy Square
Shopping Center
Before Conditions
Add Orifice
Plate to 3
Stage Outlet
Investigate Upstream Sources or Aerate
to Eliminate the Presence of Algae
Flatten Side Slopes
and Add Vegetation
Remove Dead Brush
on Sides and Reseed
Regrade to
Eliminate
Marshy Bottom
Clean Trash and
Debris
and Reseed Eroded
Areas
Repair Eroded
Manhole and
Inlet
Add Forebays at Inlets
. Loading Area —.
Add Water
Quality Inlets
Shopping
Center
Basin 70
Neshaminy Square
Shopping Center
After Conditions
Figure 2. Before and after retrofit conditions at an example basin.
247
-------
Table 2. Summary of General Problems and Solutions
Problem
Possible Solution(s)
Slide slope erosion
Erosion around outlet structure
Receiving stream erosion
Lack of water quality components*
Trash
Ponding
Sedimentation
• Regrade slopes
• Add vegetation
• Install energy dissipators and/or level spreaders
• Install energy dissipators/level spreaders
• Install riprap or similar stabilizing structure
• Orifice plates
• Perforated riser
• Sediment forebay
• Water quality inlets
• Vegetation
• Pollution prevention
• Periodic maintenance
• Trash screens
• Regrade basins
• Reset outlets where appropriate
• Plant wetlands species
• Sediment forebays
'Flood protection from large storms may be lessened when water quality controls are installed.
tive costs for urban runoff controls. Volume II of the
Neshaminy Creek Watershed Stormwater Management
Plan includes information on the maintenance costs asso-
ciated with various stormwater management practices.
General Costs and Benefits Associated with
Urban BMPs
The cost of design and installation is just a portion of the
overall cost of implementing structural BMPs. The addi-
tional costs that need to be considered include routine
maintenance, inspections, modifications if the system is
networking properly, and retrofitting, if necessary. It is dif-
ficult to anticipate and quantify the exact costs associated
with controlling urban runoff.
Cost should not be the only factor in choosing a BMP or
series of BMPs. The types of pollutants removed, the fea-
sibility of the BMP in light of site constraints, the amount of
maintenance needed, and aesthetics and safety are also
considerations. Table 3 provides a general overview of the
benefits and disadvantages of typical BMPs, pollutant re-
moval rates, and general construction and annual costs
associated with the BMPs. Table 4 summarizes the feasi-
bility and comparative costs for each of the BMPs. These
tables are meant as guides only; actual costs may vary.
General Findings
Field assessments were performed on 17 example
stormwater management basins in Bucks County, Penn-
sylvania. Recommendations for modifications or upgrades
were applied to the example basins in many forms, includ-
ing outlet modifications, grading, inlet modificafions, veg-
etation changes, fencing, and general maintenance pro-
cedures. Each basin was documented in "before" and "af-
ter" conditions, and changes were discussed in detail. Costs
were considered in an attempt to evaluate how "achiev-
able" these modifications could be, assuming that the lower
the relative cost, the more likely that the modification could
be done. Due to the desire to include a water quality com-
ponent, the structural modifications were evaluated from
the standpoint of how to improve water quality functions.
Modifications ranging from costly ($89,700) to reasonable
($10,000) were developed for each individual basin. Infor-
mation was also presented on the relative costs and feasi-
bility of installing a variety of other BMPs.
248
-------
Table 3. BMP Benefits and Costs
Beneficial
with some
Limitations
BMP Beneficial for: for:
Extended • Flood control -Water quality
Detention - • Erosion control
Dry Pond
Wet Pond • Flood Control -Water
• Erosion control
Vegetated -Water quality -Flood
Filter Strip • Erosion control
Advantages
• Provides peak flow
control
• Provides good particulate
removal
• Can serve larger develop-
ments
• Usually does not release
warm or anoxic water
downstream
• Provides excellent
protection from down-
stream erosion
• Can create wetland and
meadow habitat when
landscaped properly
• Provides peak flow control
• Cost-effective for larger,
more intensely developed
sites
• Enhances aesthetics and
can provide recreational
benefits
• Helps to prevent scour
and resuspensions of
sediments
• Provides good nutrient
removal
• Low maintenance
requirements
• Can be used as part of
the runoff conveyance
system to provide runoff
pretreatment
• Can reduce particulate
pollutant levels in areas
where runoff velocity is
low to moderate
• Provide urban wildlife
habitat
• Economical
Disadvantages
• Removal rates for
soluble pollutants are
low
• Not economical for
drainage areas less
than 10 acres
• If not adequately main-
tained, can become an
eyesore and health
hazard
• Improper design can
lead to significant
reduction in efficiency
• Extremely large storms
tend to "blow through"
the system, reducing
pollutant removal
• May not be economical
for drainage areas
less than 10 acres
• If not properly main-
tained can become an
eyesore and safety
and health hazard
• Requires considerable
space
• Not suitable for hydro-
logic soil groups A and
B in the NRCS classifi-
cation unless the pond is
lined, or inappropriate
soils are replaced with
more appropriate soils
• Possibility of release
of warm and anoxic
water which may
impact downstream
aquatic life
• Can concenterate
water, which signi-
ficantly reduces
effectiveness
•Ability to remove
soluble pollutants is
highly variable
• Limited feasibility in
highly urbanized areas
where runoff velocities
are high and flow is
concentrated
• Requires periodic
repair, regrading, and
sediment removal to
prevent channelization
Pollutants
Removed
(Average %
Efficiency)
• TSS (45%)
• Phosphorus (25%)
• Nitrogen (30%)
• COD (20%)
• Lead (50%)
•Zinc (20%)
• TSS (60%)
• Phosphorus (45%)
• Nitrogen (35%)
• COD (40%)
• Lead (75%)
•Zinc (60%)
• TSS (65%)
• Phosphorus (40%)
• Nitrogen (40%)
• COD (40%)
• Lead (45%)
•Zinc (60%)
General Cost
Construction
$0.50/ft3
Annual
$0.30/ft3
Construction
(pond < 1
million ft3)
$0.50/ft3
(pond > 1
million ft3
$0.25/ft3
Annual
$0.008 -
0.07/ft3
quality
Construction
(existing
vegetation)
$0/acre
(seeding)
$400/acre
(seed and
mulch)
$1 50/acre
(sod)
$1 1 ,300/acre
Annual - Natural
succession
(existing
vegetation)
$1 00/acre
(seed)
$1 25/acre
(seed and mulch)
$200/acre
(sod)
$700/acre
(continued)
249
-------
Table 3. Continued
Beneficial
with some
Limitations
BMP Beneficial for: for:
Grassed • Erosion control • V\teter quality
Swale • Flood control
Constructed • Flood control • Water quality
V\fetlands • Erosion control
Sand Filter -Water quality • Flood control
• Erosion control
Advantages
• Requires minimal land
area
• Can be used as part of the
runoff conveyance system
to provide pretreatment
• Can provide sufficient runoff
control to replace curb and
gutter in large-lot single-family
residential developments and
on highway medians
• Economical
• Can serve large develop-
ments; most cost-effective
for larger, more intensely
developed sites
• Provide peak flow control
• Enhance aesthetics and
provides recreational
benefits
• Prevents shoreline erosion
• Helps prevent scour and
resuspension of solids
• High pollutant removal
potential
• Provides high removal
efficiencies of particulates
• Requires minimal land area
• Provides flexibility to
retrofit existing small
drainage areas
• High removal of nutrients
Disadvantages
• Low pollutant removal
rates
• Leaching from culverts
and fertilized lawns may
actually increase the
presence of trace metals
and nutrients
• Not economical for
drainage areas less than
1 0 acres
• Potential eyesore and
health and safety
hazard if not properly
maintained
• Requires large land area
• Possible thermal and
anoxic discharge, which
could impact downstream
aquatic life
• May contribute to nutrient
loadins during vegetation
die-down periods
• Not feasible for drainage
areas greater than 5
acres
• Feasible only in areas
that are stabilized and
highly impervious
• Not effective as water
quality control for
intense storms
Pollutants
Removed
(Average %
Efficiency)
• TSS (60%)
• Phosphorus (20%)
•Nitrogen (10%)
• COD (25%)
• Lead (70%)
• Zinc (60%)
• TSS (65%)
• Phosphorus (25%)
• Nitrogen (20%)
•COD (50%)
• Lead (65%)
•Zinc (35%)
• TSS (80%)
• Phosphorus (60%A)
• Nitrogen (35%)
• COD (55%)
• Lead (80%)
•Zinc (65%)
• Oil and grease (75%)
General Cost
Annual - No
natural
succession
(existing
vegegation)
$800/acre
(seed)
$825/acre
(seed and mulch)
$900/acre
(sod)
$1400/acre
control
Construction
(seed)
$6.50/lin ft
(sod)
$20/lin ft
Annual
(seed)
$1/linft
(sod)
$2/lin ft
Construction
$50,000-
$100,000/acre
(This is based
on actual con-
struction costs
for a develop-
ment in northern
Delaware)
Construction
$5/ft3
Annual
$0.10-0.80/ft3
Sources include: MWCOG, 1992; Terrene Institute, 1994; USEPA, 1993, 1996.
250
-------
Table 4. Relative Costs and Feasibility
BMP
Relative Cost
Feasibility Factors
Extended Detention - Dry Pond
Wet Pond
Vegetated Filter Strip
Lowest cost alternative in its size range
Moderate to high compared to alternatives;
however, maintenance requirements tend to
be less than with dry ponds
Low comparative cost
Grassed Swale
Constructed Wetlands
Low compared to curb and gutter
Marginally higher than wet ponds
Sand Filters
Comparatively high construction costs;
requires regular maintenance
Good if used in conjunction with pretreatment
(e.g., sediment forebay, grassed swale)
Requires dedication of land that could otherwise
be used for buiding
Viable option if downstream flooding is a concern
Provides aesthetic benefits (which could be trans-
lated into economic benefits for developers) if
creatively designed and properly maintained
Requires dedication of land that could otherwise
be used for building
Viable option if downstream flooding is a concern
For use in areas where land can be dedicated for
stormwater runoff control
Better for new development than for retrofit in
already developed areas
Can be incorporated into the landscape of a
development, adding aesthetic value
Requires some maintenance (mowing, cleaning
trash)
More aesthetically pleasing than curb and gutter
Provides habitat
Can be used as a selling point for developments
Requires some maintenance (more until wetlands
become established)
Should be used in conjunction with other BMPs
(e.g., sediment forebay, swales, etc.) to
maximize wetland potential
Disposal of "dirty" sand may be a waste disposal
issue in industrial areas because of contents of
sand (hydrocarbons, heavy metals, etc.)
Good for use in areas where land is not available
for ponds (e.g., retrofit areas)
High TSS removal rate is a definite benefit
Sources include: MWCOG, 1992; Terrene Institute, 1994; US EPA, 1993, 1996.
References
Bucks County Planning Commission. 1994. Neshaminy Creek
Nonpoint Pollution and Wetlands Study. Volumes 1 and 2.
Doylestown, PA: Bucks County Planning Commission.
Metropolitan Washington Council of Governments (MWCOG).
1992. A Current Assessment of Urban Best Management
Practices: Techniques for Reducing Nonpoint Source Pol-
lution in the Coastal Zone. Washington, DC. Prepared for
the United States Environmental Protection Agency.
Terrene Institute. 1994. Urbanization and Water Quality. A
Guide to Protecting the Urban Environment. Washing-
ton, DC: Terrene Institute, in cooperation with the
United States Environmental Protection Agency.
United States Environmental Protection Agency
(USEPA). 1993. Office of Water. Guidance Specify-
ing Management Measures for Sources of Nonpoint
Pollution in Coastal Waters. Washington, DC: United
States Environmental Protection Agency.
United States Environmental Protection Agency
(USEPA). 1996. Office of Water. Economic Benefits
of Runoff Controls. Washington, DC: United States
Environmental Protection Agency.
251
-------
The Stormwater Management StormFilter
TM
James H. Lenhart, RE. and Bryan 0. Wigginton
Stormwater Management™
Portland, Oregon
Introduction
Since 1991, the StormFilter™ (formerly the CSF®
Stormwater Treatment System) has been treating
Stormwater runoff from small single retail sites to large ur-
ban parking lots, residential streets, urban roadways and
freeways. The StormFilter™ is a filtration Best Manage-
ment Practice (BMP) used for removal of pollutants from
Stormwater. The flow-through system is housed in concrete
vaults utilizing rechargeable filter cartridges which can hold
a variety of filter media. An assortment of patented and
commercial filter media is available to effectively remove
high levels of Stormwater pollutants. The appropriate me-
dia are selected based on the pollutants expected at the
site. The StormFilter™ offers the flexibility of changing to
different media if actual pollutant loadings/concentrations
differ from expectations.
System configurations include the Precast StormFilter™,
the Cast-ln-Place StormFilter™ and the Linear Storm Fil-
ter™. The precast and linear models utilize pre-engineered
precast concrete vaults for ease of design and installation.
The precast units can come with traffic-bearing lids for
placement in parking lots where they take up virtually no
land space. The cast-in-place filters are customized units
for larger flows and may be either covered or uncovered
underground units.
The StormFilter™ is designed to be especially effective
for the treatment of high pollutant concentration flows, and
particularly those storms early in the rainy season. In gen-
eral, the StormFilter™'s efficiency is highest when pollut-
ant concentrations are highest (Lenhart, 1998).
How It Works
The filters work by percolating Stormwater through the
cartridges containing filter media. The media trap particu-
lates and adsorbs materials such as dissolved metals and
hydrocarbons. Surface scum, floating oil and grease are
also removed, After passing through the filter media,
Stormwater flows into a pipe manifold cast into the floor of
the vault to an open channel drainage way.
The typical precast StormFilter™ configuration shown
in Figure 1 consists of an inlet bay with flow spreader, car-
tridge bay containing the flow cartridges, an overflow and
outlet bay (above outlet).The inlet bay serves as a grit
chamber and provides for flow transition into the cartridge
bay. The flow spreader provides for the trapping of
floatables, oils, and surface scum prior to entry into the
cartridge bay. Stormwater enters the cartridge bay through
the flow spreader, cascades over an energy dissipater and
begins to pond. The StormFilter™ is also designed with
an inline overflow which operates when the inflow rate is
greater than the flow capacity of the cartridges.
Filter Cartridge Hydraulics
Each cartridge plugged into the underdrain manifold is
about 19" tall and 19" in diameter and designed to treat a
flow of 15 gpm with 2.3 feet to total system head.
Once Stormwater begins to pond in the vault cartridge
chamber, it percolates through the media and begins to fill
the center drainage tube. As the center drainage tube fills,
the air is purged out the one way air relief valve located in
the filter top. Once the tube is filled with filtered water, there
is enough buoyant force on the float to lift it from its seat.
The filtered water is then allowed to flow out of the car-
tridge into the drainage manifold. As the column of water
moves down, the air valve snaps shut and primes a hang-
ing column of water. The hanging column of water then
draws the Stormwater horizontally through the media and
into the inner drainage tube with about 18 inches of suc-
tion head.
The filter cartridge (Figure 2) will continue to draw water
throughout the storm duration. At some point during the
storm, the outflow rate from the filter will exceed the inflow
rate and the water surface elevation in the vault will begin
to drop. This will continue until the water surface reaches
the lower lip of the hood. At that point the suction head will
violently draw air into the hood causing high energy turbu-
lence between the inner surface of the hood and the outer
surface of the filter media.
This high energy turbulence scrubs much of the accu-
mulated sediments from the surface of the filter causing 1)
a high concentration suspension which then settles, or 2)
a direct sloughing of sediment to the vault floor. Sediments
252
-------
Flow Spreader
\
Traffic Bearing Lid
Radical Flow
Cartridge
Underdrain Manifold
Outlet
Figure 1. Precast StormFilter™ diagram.
don't get into the underdrain area because all water arriv-
ing there has been filtered. This scrubbing action will par-
tially restore the permeability of the filter thus enhancing
its performance by increasing the filter life and decreasing
filter maintenance frequency and costs.
The Drainage Manifold
The drainage manifold is a pre-manufactured pipe sys-
tem that is shipped to either a concrete precaster or cast-
in-place concrete contractor. The manifold consists of a
series of 3"x2" tees on equidistant spacing with control
valves at the outlet. The control valves automatically regu-
late flow to 15 gpm per cartridge. The manifold is secured
in place according to plan and cast into the floor of the
facility (Figure 3). The removable cartridges are then
"plugged" into the manifold tees.
Basic System Design
The StormFilter™ is sized to treat the peak flow of a
design storm. Peak flows are typically determined by cal-
culations based on the contributing watershed hydrology
and using a design storm magnitude. The design storm is
usually based on the requirements set by the local regula-
tory agency. The size of a StormFilter™ is determined by
the number of filter cartridges required to treat the peak
stormwaterflow.
Each cartridge is designed to treat a peak flow of 15
gpm (or 30 cartridges/cfs). For example: a peak design
stormwater flow rate of 150 gpm would require 10 car-
tridges. The StormFilter™ typically requires 2.3 feet of head
differential between the invert of the inlet and the invert of
the outlet.
Depending on individual site characteristics, some fil-
ters are equipped with high and/or low flow bypasses. High
flow bypasses are installed when the calculated peak storm
event generates a flow which overcomes the overflow ca-
pacity or design capacity of the filter. Base flow bypasses
are sometimes installed to prevent continuous inflows
caused by ground water seepage.
Available Filter Media
CSF® Leaf Compost Media
In autumn deciduous trees begin to drop their leaves in
preparation forthe winter. In metropolitan areas the leaves
accumulate, clog storm drains, and may cause local flood-
ing. To prevent this, many cities have leaf collection pro-
grams. Leaves are either swept up or brought to drop-off
points and then transported to landfills, or municipal or com-
mercial composting facilities.
Using a feed stock of pure deciduous leaves (i.e., no
mixed yard debris such as prunings and grass), Stormwater
253
-------
Air Vent
Screen Cap
Air Relief Valve
Granular Media
Float
Outer Screen
Pleated Fabric Insert
Hood
Porous Center
Drainage Tube
Outlet Manifold
Figure 2. Filter cartridge.
Figure 3. Casting the pipe manifold into the floor of this cast-in-place vault.
254
-------
Management composts leaves collected by the City of
Portland, OR, over a period of eight months into a mature
stable compost. Stormwater Management then processes
the finished compost into a granular media which re-
sembles soil and has no odors. Once complete, the me-
dium appears to have the physical and chemical charac-
teristics desirable for the filtration of stormwater.
Pollutant Removal Mechanisms
There are three primary pollutant removal mechanisms
performed by the CSF® filter media: mechanical filtration
to remove sediments, chemical processes to remove
soluble metals, and adsorption properties to remove oils
and greases.
Sediments
The media are contained in a series of filter cartridges
which have a 7" thick layer of the media through which
stormwater passes. Sediments are filtered out on both the
surface of the filter and the surfaces of granules through-
out the media matrix. As sediments are removed from
stormwater runoff and accumulate on the surface of the
filter, the permeability will decline thus requiring facility
maintenance. Sediment removal will vary with particle size
distribution, but removal has been as high as 95%.
Soluble Heavy Metals
The media also acts as a chemical filter to remove dis-
solved ionic pollutants such as soluble heavy metals, in-
cluding lead, copper, and zinc. The mechanism of cation
exchange is provided by humic acids, which are produced
by the aerobic biological activity which occurs during
composting. Soluble heavy metal removal rates vary from
65% to 95%.
Oils & Greases
Removal of oils and greases (O&G) and other organic
compounds is facilitated by the high organic carbon con-
tent of the media. The organic carbon is oleophilic and
adsorbs free oil and grease. The system performs best
when O&G loadings are less than 20 mg/l. Measured re-
moval rates are as high as 80% (W&H Pacific, 1992).
Pleated Fabric Inserts
The pleated fabric insert is used primarily for sediment
control (Figure 4).The insert fits inside the standard car-
tridge leaving an annular space between the inside of the
insert and the drainage tube, which can be used for the
addition of granular media to remove selected soluble pol-
lutants.
The reusable insert is constructed of a durable fabric
which is easily cleaned with low-pressure water. Each in-
sert has a total of 75 square feet of surface area. Two fab-
ric pore sizes of 70 microns and 36 microns are available.
Sediment loading performance data for the two fabrics,
using granular media only, are presented in Graph 1.
The CSF® media showed a 50% decrease in flow rate
after the accumulation of 12 pounds of dry sediment per
cartridge. The 70 urn fabric filter insert prolongs the per-
meability of the system by three times over media alone
and the 36 um fabric increase the life by a factor of two. As
shown by the 50% flow rate decrease line, it would require
23+ pounds of sediment per cartridge to slow the flow rate
of the 36 urn fabric to 7.5 gpm and 31+ pounds per car-
tridge for the 70 um fabric.
Graph 2 shows the particle size distribution of the sedi-
ments used in these tests. TSS removal by the fabric fil-
ters is 100% of particles over 36 um, and some portion of
particles under 36 um (The sediments were taken from
existing stormwater quality facilities, dried and screened
through a #45 mesh).
Also note on Graph 1, the abrupt jumps in flow rate.
These jumps represent the flow recovery generated by the
self cleaning mechanism. Once the jump occurred, the
sediments that settled to the bottom were intentionally re-
suspended during testing to ensure that 100% of the sedi-
ments used were attached to the filter surface.
Table 1 presents estimated volumes of water treated as
well as the approximate dry sediment poundage that a typi-
cal 8' x 14' StormFilter™ containing 18 cartridges could
treat and capture respectively.
Values in Table 1 assume TSS = 75 mg/l, sediment com-
posed of 50% fine sand and smaller particles, and an 8' x
14' StormFilter™ containing 18 cartridges. The values also
assume the removal of 100% of suspended solids with no
sediment accumulation at the bottom of the vault.)
Figure 4. Pleated fabric inserts.
255
-------
Flow Rate vs. Ibs. of Sediment Accumulated in Filter (CSF/70um/36um)
0 5 10 15 20 25 30 35 40
Ibs. of Sediment Accumulated/Cartridge
Flow Rate Decrease by
~
Flow Rate Decrease by
50%
CSF Filter Media
70 um Filter Fabric
36 um Filter Fabric
Graph 1. Sediment accumulation reduces flow rates.
100
90
80
70
60
50
40
30
20
10
*S-
10
100
1000
Particle Size
Graph 2. 1997 185th particle size distribution.
Table 1. Predicted Sediment Removal Rates
CSF® Media
36 micron fabric
70 micron fabric
6.7% Decrease (decrease from 115 pounds sediment
15 to 14gpm)
180,000 gal. water
200 pounds sediment
320,000 gal. water
310 pounds sediment
500,000 gal. water
50% Decrease (decrease from 260 pounds sediment
15 to 7.5 gpm)
409,000 gal. water
425 pounds sediment
680,000 gal. water
560 pounds sediment
892,000 gal. water
256
-------
If the designer knows the maximum allowable decline in
system efficiency and the system is designed for the 15
gpm/cartridge, the graph and table can be used to estab-
lish the total sediment loading that occurs between main-
tenance cycles. (More water has passed through filters
when their permeability has decayed to 50% than when it
has decreased by only 6.7%.)
Perlite and Zeolite Media
Perlite is a "puffed" volcanic ash. This lightweight mate-
rial is commonly used as a filter for water filtration. Though
not chemically active, Perlite is very effective for removal
of fine particles due to its micro pores and blocky struc-
ture. Perlite can be used as a stand alone medium or in
conjunction with the pleated fabric insert. Perlite can also
be mixed with other media such as Granulated Activated
Charcoal (GAG), commercial cation exchange media, etc.
Perlite would inexpensively act on TSS while GAG, etc.
would act on organics.
Zeolites are naturally occurring minerals that exhibit cat-
ionic exchange properties. Some zeolites have the capac-
ity for anion adsorption. Blended perlite/zeolite media pro-
duced by Stormwater Management is suggested for wa-
tersheds where soluble phosphorus is of concern.
To demonstrate the removal efficiency of phosphate by
the zeolite, 2-inch column studies have been performed
with a varying matrix of zeolite mixed with a highly perme-
able perlite. The tables below present the data obtained
from several column studies and show the percent removal
of ortho-phosphate promoted by the zeolite. (Each table is
followed by a brief description of the sample number and
sample time.) The influent source was irrigation return water
from a commercial container nursery.
The samples labeled 10G-1 and 30G-1 represent the
first effluent samples taken from a 10-gram and 30-gram
column test respectively. These samples were taken after
approximately 100 ml of stormwater had passed through
each 2-ing diameter column. The samples labeled 10G-2
and 30G-2 represent samples taken after approximately
1000 ml of stormwater had been passed.
Table 2. Ortho-Phosphate Removal
Sample
Ortho-P
(mg/l)
Influent
0.33
Sample
10G-1
0.18
Sample
10G-2
0.25
Sample
30G-1
0.15
Sample
30G-2
0.18
% removal
45.4%
24.2%
54.5%
45.4%
Tables 3, 4, and 5 represent preliminary testing of phos-
phate (at three concentrations) removal.
Table 3. Ortho-Phosphate Removal
Sample Influent Sample 40.3
Sample 40.5
Ortho-P (mg/l)
% removal
0.37
0.20
49.5%
0.32
13.5%
The sample labeled 40.3 represents a composite sample
of the first 4 gallons of effluent passed through 40 grams
of zeolite with the 40.5 sample being the 5th gallon passed.
Table 4. Ortho-Pohsphate Removal
Sample Influent Sample 50.1 Sample 50.3 Sample 50.5
Ortho-P (mg/l)
% removal
0.38
0.15
60.5%
0.34
10.5%
0.36
5.3%
The samples labeled 50.1, 50.3, and 50.5 represent
samples taken of the 1st gallon, the 3rd gallon, and 5th
gallon of effluent respectively.
Table 5. Ortho-Phosphate Removal
Sample Influent
Sample 60.5
Ortho-P (mg/l)
% removal
0.37
0.19
48.6%
The effluent sample labeled 60.5 represents a combina-
tion of the entire 5 gallons of stormWater passed through
the column.
Though further studies are underway, these data indi-
cate a sorbtive capacity of 60 mg ortho-phosphate/kg of
zeolite media. We expect to achieve double that removal
rate soon. Depending on the mass concentration of zeo-
lite in the media, one is able to estimate the mass removal
of ortho phosphate between filter maintenance cycles.
Facility Maintenance
The primary purpose of the StormFilter™ is to filter out
and prevent pollutants from entering down gradient water
bodies. Like any effective filtration system, these pollut-
ants must be periodically removed to restore the
StormFilter™ to its full efficiency and effectiveness. Main-
tenance requirements and frequency are dependent on
the pollutant loading characteristics of each site. To assist
with maintenance issues, detailed Operation & Mainte-
nance Guidelines are available.
Maintenance services can be purchased completely, or
in part. Stormwater Management also provides tracking of
all installed systems, notifies system's owner of mainte-
nance needs and notifies the regulatory agency that the
system has been maintained.
Maintenance is usually performed late in the dry season
to rejuvenate the filter media and prepare the system for
the next rainy season. Maintenance activities can also be
performed mid-season in the event of a spill or excessive
sediment loading due to site erosion.The most common
maintenance schedule is once a year, but sites can and
do vary.
Maintenance involves replacing the cartridges with a
series of newly recharged cartridges. The old cartridges
are then cleaned and recharged with new media for re-
use. Systems with excess sediment accumulation on the
vault floor can be vactored out.
257
-------
Media residuals can be re-composted to reduce accu- References
mulated hydrocarbons and then used in landscaping, ero- . . . . „_. , _. .... . . _. ,
sion control applications or daily cover for landfills This Lenhart< J- P*[™{ Rem°^l: ls rt an Accurate Performance
sustainable process helps minimize total maintenance Measure- Waterworid Magazine, June 1998.
costs- W&H Pacific. Final: Compost Stormwater Treatment System,
March 1992.
258
-------
Bioretention: An Efficient, Cost Effective
Stormwater Management Practice
Larry S. Coffman
Prince Georges County, Department of Environmental Resources
Largo, Maryland
Derek A.Winogradoff
Planning Section, Programs and Planning
Largo, Maryland
Abstract
In 1993 the Department of the Environmental Resources,
Prince Georges County, Maryland, introduced the "Design
Manual for Use of Bioretention in Stormwater Manage-
ment". Bioretention facilities have become commonly re-
ferred to as rain gardens. They were designed as an alter-
native, cost effective Stormwater best management prac-
tice (BMP) which allows multifunctional use of green space
and landscaped areas for storage and treatment of runoff.
Bioretention BMPs are simply very shallow landscaped
depressions where runoff is infiltrated or filtered through a
soil/plant complex for treatment. Bioretention has been
successfully used as an alternative, cost-effective BMP for
commercial, industrial and residential applications address-
ing both landscape and Stormwater management objec-
tives. More recently, as a retrofit practice, the BMP has
been incorporated into green space, streetscapes, median
strips and parking lot islands. This paper summarizes re-
cently completed monitoring findings and the lessons
learned over the last five years with the use of bioretention
for urban Stormwater management.
General Bioretention Design Features
Bioretention was originally modeled afterthe hydrologic/
physical characteristics of an upland terrestrial forest or
meadow community (as opposed to a wetland) dominated
by facultative trees, with understory shrubs and herbaceous
upland plant materials. The BMP is strategically placed to
intercept drainage and is therefore, subject to repeated
hydraulic loading. Because of this, the designer must be
sure that the BMP will be well drained to maintain aerobic
conditions. Proper drainage can be achieved by infiltration
(where soils allow), underdrains, or both. Key factors in
the design and construction of bioretention facilities are
careful selection of plant material that can tolerate extreme
hydrologic regimes; good drainage to prevent anaerobic
conditions; safe conveyance of overflows; careful selec-
tion and control of backfill soils; and, careful inlet/ outlet
controls to prevent erosion.
Bioretention BMPs should be designed as "off-line" con-
trol devices where excess runoff and high-velocity flows
bypass the BMP minimizing erosion and flushing of land-
scape materials and debris. Stored runoff (ponded water
over the bioretention area) should exfiltrate over a period
of less than a day into the underlying soils and in some
cases, into an underdrain that discharges to an appropri-
ate conveyance system. Where soil infiltration rates are
lower than 1 inch/hour, or in order to extend the life of the
facility, underdrains should be used.
Bioretention areas typically consist of the following com-
ponents: shallow ponding area (6" or less), mulch layer (2-
3"), sandy planting soil (2-3'), a variety of plant materials
and where appropriate underdrains. The design of the BMP
can vary greatly to accommodate site constraints, ground
water recharge, soils, habitat/ecological objectives, water-
shed hydrology and aesthetics. The freedom in design
variation creates opportunities for site integration of natu-
ral features with man-made infrastructure. Specific con-
figuration and location is determined after site constraints
such as location of utilities, groundwater level, steep slopes,
underlying soils, existing vegetation and drainage are con-
sidered.
The drainage area for one facility should generally be
between 0.25 and 1 acre. Multiple bioretention facilities
are needed for larger drainage areas. The storage volume
of the facility will be determined by the desired level of
control (e.g. first half-inch of runoff) and dewatering capa-
bilities of the design. The BMP works best when there are
many facilities with small drainage areas. Large facilities
with large drainage areas tend to allow soils to remain satu-
rated creating anaerobic conditions, stressing the plants
and potentially reducing the pollutant removal effective-
ness for many pollutants.
Planting soil should be sandy loam, loamy sand or loam
texture and have clay content of 10% to 15% or less. The
pH of the soil should be between 5.5 and 6.5 to maximize
259
-------
plant growth and microbial activity forthe up-take and trans-
formation of pollutants. The planting soil should contain 3
to 5% organic content and other soil augments typically
needed to support the types of plants used.
Native plant species are recommended as they are gen-
erally more suited to regional climate, soils, hydrology, and
disease. The designer should assess aesthetics, site lay-
out, habitat objectives and maintenance requirements when
selecting plant species. After placing the trees and shrubs,
the ground cover and/or mulch should be established.
Ground cover such as grasses, legumes or flowers can be
used. Two or three inches of commercially available fine
shredded hardwood mulch or shredded hardwood chips
should be applied to provide protection from erosion, as
well as to enhance evapotransporation in the facility.
Monitoring
In an effort to refine bioretention design and determine
the pollutant removal efficiency of the BMP, a two-year
study was conducted by the University of Maryland, in con-
junction with Prince George's County, Department of En-
vironmental Resources and the National Science Foun-
dation. The study, known as the "Optimization of
Bioretention" and included both laboratory and field test-
ing.
The laboratory experiments were conducted in two small
(3 ft x 2.5 ft x 2.5 ft deep) and one large (10 ft x 5 ft x 3.5 ft
deep) bioretention "boxes". These boxes were filled with a
sandy loam soil topped with a 1 -inch layer of commercially
available shredded bark mulch and planted with creeping
junipers. Perforated pipe sampling ports were placed at
three depths (upper, middle and lower) within the larger
box to examine pollutant removal as a function of depth
and time of exposure. Two sampling depths were used for
the smaller box. Effluent samples were analyzed in the
University of Maryland's Environmental Engineering Labo-
ratory.
A synthetic runoff recipe was used for all testing to allow
for greater consistency and correlation of field and labora-
tory results. The runoff recipe was based on the average
pollutant loading taken from the county's 3-year wet
weather data collected as part of the NPDES stormwater
monitoring program. The hydraulic loading or rate of appli-
cation to the bioretention facilities was based on a typical
rainfall event in Prince Georges County (0.1 in/hr with a 6-
hr duration).
Average removals in the laboratory bioretention systems
are shown in Table 1. Sample contamination resulted in
no nitrate data forthe small boxes. In both cases, removal
of the heavy metals (copper, lead, and zinc) was excel-
lent, ranging from 93-99%. Effluent levels of phosphorus
showed 70-80% removal. Average TKN and ammonia re-
movals were 60-80% in the lower effluent ports, but sig-
nificant ranges were noted for each. The system removed
only a small amount of the nitrate (23%) and nitrate con-
centrations above the influent levels were noted from the
upper ports.
For the small bioretention boxes, flow from the bottom
began approximately 45 minutes after the runoff applica-
tion. It took about 2 hours for flow to occur from the bottom
ports in the large box. In both boxes, the head built up to
3-5.5 inches.
Field Experiment
Field monitoring was performed on an existing
bioretention facility. This facility was constructed with an
underdrain at a depth of about 2.5 ft., which was used for
the sampling port. The same runoff pollutant "recipe" was
applied in the field. The runoff application continued for
nearly six hours with samples taken every 25-30 minutes.
Results from the field experiment mirrored the labora-
tory results. For copper, lead, and zinc, nearly total removal
was achieved by the bioretention facility with removals
exceeding 95%. All metal concentrations in the effluent
were less than or near instrument detection limits (2 mg/L
for copper and lead, 25 mg/L for zinc). The total phospho-
rus removal was about 60%. TKN removal was about 50%.
The ammonia removal was somewhat variable but the
majority of the samples monitored indicate about 70% re-
moval. The removal for nitrate was only about 15% with
wide variation.
Summary of Monitoring Observations and
Findings
Heavy metal removal showed excellent agreement be-
tween the field and laboratory experiments. Removals were
Table 1. Summary of Percent Removal Rates for Laboratory Bioretention Systems
Large Box
U
M
L
Small Box
U
L
Cu
(Mfl/L)
90
93
93
91
97
Pb
(Mfl/L)
93
99
99
95
98
Zn
(H9/L)
87
98
99
93
97
P
(mg/L)
0
73
81
16
70
TKN
(mg/L)
37
60
68
55
76
NH4+
(mg/L)
54
86
79
(-7)
60
N03-
(mg/L)
(-97)
(-194)
23
-
-
TN
(mg/L)
(-29)
0
43
-
-
Testing Conducted by the University of Maryland, Department of Engineering
260
-------
greater than 90% and variations were small. Nearly the
entire metal uptake occurred within the top few inches of
the bioretention system. Separate small column labora-
tory experiments with the mulch showed that it has a high
capacity for metals uptake. It is likely that within bioretention
facilities, significant metal uptake occurs within the sur-
face mulch layer. Additionally, examination of the labora-
tory bioretention mulch after several applications of the
synthetic runoff showed elevated metal levels. All results
point to the importance of the mulch layer for metals up-
take in bioretention.
Phosphorus removal appears to have linear correlation
to depth, with better removal resulting as runoff migrates
to deeper levels. Limited additional removal was observed
beyond depths of about 2 to 3 feet.
TKN removal averaged about 60% with significant vari-
ability. The depth of the facility does not appear to signifi-
cantly affect TKN removal rates.
Ammonia removal appears to increase to over 70% at
greater depths. However, due to significant variation in
results, depth correlation is difficult to summarize.
Nitrate removal results were erratic and significant vari-
ability was found. Shallow sampling ports showed nitrate
levels higher than the input, indicating conversion of or-
ganic nitrogen, or ammonia to nitrate during the runoff
event, or washout of nitrate from previously captured ni-
trogen. At depths greater than 30 inches, limited nitrate
removal occurred (15-20%).
Additional research is needed to optimize the pollutant
removal/transformation capabilities of the plant soil com-
plex. It appears possible that the chemical make-up of the
soil complex and the types of plants used for bioretention
can be customized to address unique pollutant runoff prob-
lems and recharge needs for a given land use and water-
shed.
Bioretention Design & Construction
Considerations Based on Monitoring
Results
Laboratory and field studies, along with field observa-
tions over the last five years have provided data neces-
sary to refine bioretention design and construction criteria.
Modifications to the Prince Georges County bioretention
design manual are currently underway based on these find-
ings. Some of the major modifications are listed and briefly
described below:
Optimizing Facility Depth - The original design depth was
set at 4 feet. This was to ensure adequate depth for plant
growth and to maximize pollutant removal. The experiments
performed to date suggest that sufficient removal is
achieved at a 2 -2.5 foot depth. Therefore, design depths
can be reduced without significantly compromising pollut-
ant removal efficiency. This will result in cost savings by
reducing excavation and back-fill material costs.
Underdrains are recommended for all facilities unless
geotechnical reports indicate otherwise.
Minimizing Above Ground Storage Depth-Above ground
storage of runoff is necessary to achieve retention storage
for treatment and management of stormwater runoff. How-
ever, the depth of the runoff should be minimized to allow
quicker infiltration and evapotranspiration. This will ensure
adequate aeration of soils to maintain aerobic conditions
necessary for plant growth and various pollutant transfor-
mation processes. The recommended maximum ponding
depth is 3-6 inches per facility.
Minimizing Hydraulic Loading - The drainage area for
bioretention BMPs should be minimized to reduce hydrau-
lic loading of the facility and avoid excessive saturation of
the soils which could lead to anaerobic conditions. The
maximum recommended drainage area is 1 acre. Ideally,
facilities should be designed to control 0.25 acres or less if
site/soil conditions and land use allow.
Planting Soil and Soil Amendments - Soil amendments
are only necessary for plant viability. The planting soils
should have no greater than 10-15% clay particles. Soil
infiltration rates of greater than 1 inch an hour are prefer-
able if no underdrain system is used. Native plants require
less maintenance.
Facility Sizing and Configuration - Sizing of the facility is
dependent upon the drainage area, land use and site con-
straints. On-lot facilities are limited to space available and
grading. As previously indicated, facilities may have soil
depths of 2 feet and still achieve significant pollutant re-
moval rates. In addition, the side walls do not need to be
straight up and down, but can be shaped in a "dish ar-
rangement" beginning at the ground level around the pe-
rimeter and extending to a the design depth in the center.
All facilities should be "off-line" to avoid pass-through wa-
ters that could cause erosion and flushing of debris from
the facility unless carefully designed to accomodate these
effects.
Settlement and Compaction- During construction, due
to the nature of the planting material and method of instal-
lation, settlement will occur. In the University Maryland
experiment, 15-20% settlement of the soil was observed
after the first influent was applied. Interestingly, after the
initial settlement, only minor settlement was reported. In
construction, settlement is a concern particularly for the
accuracy of final grading and volume storage. Settlement
has implications related not only to construction method,
but cost as well. To help control problems created by dif-
ferential or unexpected settlement, wetting of the planting
soil prior to placement of mulch and plant materials is rec-
ommended. This can be accomplished by hosing down
the area (or if time permits, waiting for a rain event). Com-
paction of the soils must be minimized during construction
activities. Bioretention facilities that are shallow (2-3') can
be constructed and dressed using manual labor and small
machinery.
261
-------
Planting Arrangements- Planting arrangements are im-
portant for aesthetics, soil stability, habitat, pollutant up-
take and the viability of the bioretention facility. Initial plant-
ing arrangements should be designed to provide an im-
mediate dense and perpetual ground cover. For example,
use of ornamental grasses provides thick, quick growing
cover and uptake of pollutants. Facilities located in areas
of high visibility should include plant material that is aes-
thetically pleasing with year-round interest. The use of
native planting arrangements is encouraged. Designers
should work closely with landscape architects and nurs-
erymen to plan an aesthetically pleasing, yet cost conscious
product. Plant materials and landscaping can affect the
cost of the facility dramatically accounting for as much as
60% of the total facility cost.
Designing for Thermal Attenuation - Thermal pollution in
runoff to receiving waters has been correlated to increases
in urban development. Heated runoff temperatures from
impervious surfaces is one of the main causes of elevated
stream temperatures. By directing and capturing heated
runoff into a bioretention facility to allow evaporation, infil-
tration and filtration into the soil, the facility can act as a
heat sink and dissapation. Where protection of cold water
fisheries is an important consideration, bioretention can
be used to reduce thermal impacts of urban runoff.
Designing for Hydraulic Conductivity - When designing
and locating bioretention facilities without underdrains,
special attention must be given to the hydraulic conductiv-
ity of the surrounding in situ soils. They should have a very
high percentage of sand particles (2mm in size or greater)
to ensure adequate infiltration. Soils must be USDA clas-
sification sandy loam or better in the textural triangle. When
performing feasibility analysis for siting bioretention facili-
ties, the hydraulic conductivity of the surrounding soils must
be analyzed sufficiently by geotechnical means such as
soil borings. Soils investigations typical to the requirements
for standard infiltration trenches can be performed to de-
termine the suitability of the in situ soils. Soils having a
significant percentage of small soil particles can lead to
failures. In addition, even where boring samples indicate
small percentages of clay, the soils analysis should deter-
mine the makeup of the soil strata and horizons to ensure
that clay lenses are not present.
Bioretention Applications
Bioretention systems maximize the use of natural physi-
cal, chemical and biological pollutant removal and trans-
formation processes to treat runoff. They are dynamic liv-
ing micro-ecological systems that demonstrate how the
landscape functions to protect the integrity of a watershed's
aquatic and riparian ecosystems. Their designs also dem-
onstrate the interconnections of a wide array of environ-
mental and engineering principles and disciplines includ-
ing: the hydrologic cycle, nonpoint pollutant treatment,
nutrient cycles, resource conservation, habitat creation, soil
chemistry, ecology, horticulture and landscape architec-
ture.
Bioretention systems use upland facultative plants that
are not dependent on a constant source of water such as
conventional stormwater /wetland ponds. Therefore,
bioretention systems can be constructed in the upland ar-
eas avoiding destruction of riparian buffers and streams.
Typically, upland soils are more conducive to the natural
properties that bioretention facilities are attempting to
mimic. Those same soils often are associated with hydro-
logic soil groups "A" and "B", which exemplify well drained
soils.
Widespread use and uniform distribution of bioretention
storage throughout a development can also help to repli-
cate predevelopment watershed hydrologic functions. This
BMP can be used to reproduce similar rainfall storage ca-
pabilities in the developed site that existing prior to devel-
opment. Retention and groundwater recharge functions are
designed to mimic predevelopment runoff characteristics.
Reproducing predevelopment hydrologic functions can be
the most important factor in maintaining the ecological in-
tegrity of receiving waters, small streams and wetland sys-
tems.
To achieve uniform distribution of bioretention practices
for the greatest hydrological benefits and to maximize pol-
lutant removal, it is important to understand how to inte-
grate the practice into the developed landscape. Below is
a listing and brief discussion of the possible ways to apply
bioretention throughout a site.
Incorporated into Parking Lots, Medians and Landscape
Islands - The design must ensure that infiltration and ground
water seepage will not adversely affect the structural in-
tegrity of roadways or buildings. Careful grading, location
and use of underdrains can minimize problems. It is im-
portant to divert overflows to inlets or grass areas in order
to prevent deposits of sediment and debris onto parking
surfaces.
Designed into Existing Meadow or Forested Areas -
These areas can be converted to rain gardens by con-
structing small earthen or stone berms (4 inches or less)
to allow shallow ponding. Care must be taken not to pond
too much for too long, as existing mature vegetation may
be less tolerant to drastic changes in hydrology. Also, ad-
equate measures must be taken to reduce erosion poten-
tial when directing increased volumes and concentrated
flows into existing vegetated areas.
Fringe Forest and Transition Areas - Bioretention can
be used for re-vegetation of forest fringe areas to recreate
a terrestrial forest community and transitional habitat eco-
system. These areas would consist of trees, sub-canopy
under story trees, a shrub layer and herbaceous ground
covers. Plants can be selected fortheir habitat value (food,
shelter and nesting materials). Deep rooted vegetation
would help to promote increased infiltration.
Open Space Meadows - Open space areas not used for
recreation or other purposes can be designed as rain gar-
dens. Where soils and topography allow, wild flower
meadow basins can be constructed. Care must be taken
to prevent erosion and to disperse flows throughout the
bottom of the rain garden basin.
262
-------
Open Swales - Rain gardens must be carefully designed
to be used on-line in an open swale. They may be used as
an appurtenance to a swale in an off-line configuration.
On-line systems are subject to erosion due to high veloci-
ties and concentrated flows without the careful design.
Landscape Trees and Shrubs -A simple application of a
rain garden is to create shallow depression storage areas
around each individual landscape plant. Careful selection
of water tolerant species could allow ponding depths of 2
to 3 inches around each plant where soils allow.
Retrofit Existing Development - Many green spaces and
landscaped areas can be converted to rain gardens. The
applicability of retrofit options will be dependent on a de-
tailed site evaluation.
Sediment Control- Like stormwater management ponds,
bioretention pits may be utilized for sediment control de-
vices for stabilization during construction activities. By us-
ing on-lot bioretention for sediment control compliance, the
need to drain the site runoff to one large holding pond is
eliminated or diminished.
Limitations
Bioretention that relies on infiltration alone fordewater-
ing should not be considered where the watertable is within
6 feet of the ground surface and when the surrounding soil
is unsuitable for infiltration (less than one inch/hour). While
the bioretention concept relies on the natural and physical
properties of infiltration, absorption and evapotransporation,
these processes can have limited capacity under various
conditions such as saturated soil, frozen ground, or high
humidity. The practice is also not recommended for areas
with steep slopes greater than 25% or in areas where ex-
tensive tree removal would be required.
When used on residential lots to fulfill stormwater re-
quirements, property owners must be educated on the need
and routine care of rain garden areas. Maintenance agree-
ments, educational materials and easements are possible
ways to ensure long term use and operation by the prop-
erty owner.
General Cost Comparisons
Bioretention costs are most attractive when compared
to the use of other structural BMPs such as ponds. Cost
savings over conventional BMPs can vary widely depend-
ing on unique site conditions. More efficient use of land
can be achieved when integrated into the typical landscape
features. Savings of 10 to 25% compared to conventional
approaches have been achieved using rain gardens in resi-
dential and commercial sites. Generally, residential rain
gardens will average about $3 to $4 a square foot depend-
ing on soil conditions and density and types of plants. Com-
mercial/industrial sites costs can range between $10 to
$40 a square foot depending on the density/types or plants
and the need for control structures, curbing, storm drains
and underdrains. Planting costs can vary substantially and
can account for a significant portion of the facility cost. In
many cases, bioretention can be an extremely cost-effec-
tive practice for controlling stormwater. Replacing traditional
piping with gardens to convey flow can lead to substantial
savings.
There are additional costs when compared to typical
landscaping treatment due to the increased number of
plantings, additional soil excavation, backfill material and
use of underdrains. The use of a wildflower meadow rain
garden to replace open space turf will have higher site
preparation and planting costs, but long-term maintenance
(mowing) can be reduced to once a year.
Long Term Maintenance Considerations
Rain gardens require routine periodic maintenance (e.g.
mulching, plant replacement, pruning and weeding) typi-
cal of any landscaped area. No special maintenance equip-
ment is needed. Routine maintenance costs will increase
proportionately to the increased number of plants used and
the area planted. To date, there is no solid data on the
longevity of bioretention systems. The use of underdrains
will help to ensure the BMP remains well drained for the
long term health of the plants. Underdrains placed at shal-
low depths (2 to 3 feet) can be easily maintained if clogged
by sediment or roots.
References
Design Manual for Use of Bioretention in Stormwater Man-
agement (PGDER), 1993).
Stormwater Management Design Manual (PGDER, 1991).
Bioretention Design Manual- Final Draft (PGDER, 1997).
Low Impact Development Design Manual (PGDER, 1997).
Native Trees, Shrubs, and Vines for Urban and Rural
America (Hightshoe, 1988).
263
-------
Innovative Stormwater Treatment in Washington State
Stacy Trussler, PE
Northwest Region, Water Quality Program
Washington State Department of Transportation
Seattle, Washington
Bert Bowen
Environmental Affairs Office, Water Quality Program
Washington State Department of Transportation
Olympia, Washington
Introduction
The Washington State Department of Transportation
(WSDOT) has a long history of engineering innovation.
WSDOT built the first floating concrete pontoon bridge in
the world. Other firsts have been achieved in avalanche
control and de-icing treatment.
Not every new technology works as theorized. New sys-
tems often need to be fine tuned before they can become
a standard practice. Innovative Best Management Prac-
tices (BMPs), a continuation of WSDOT's engineering in-
novation tradition, provide design engineers with more ap-
propriate and effective stormwater treatment options. An
innovative BMP is one that has not been approved by the
Washington Department of Ecology, our state environmen-
tal regulatory agency.
WSDOT maintains a Stormwater Management Program
(SWMP) to protect water quality through the National Pol-
lutant Discharge Elimination System (NPDES) municipal
permit requirements. The permit requires WSDOT to:
• Reduce and control discharge of pollutants to the maxi-
mum extent practicable, as required by federal regu-
lations;
• Use all known, available, and reasonable methods of
prevention, control, and treatment.
Innovative BMP research is a critical element of our
Stormwater Management Program. State and local gov-
ernments rely on WSDOT to protect and maintain existing
water quality. WSDOT does this by using BMPs.
Standard BMPs
WSDOT's Highway Runoff Manual (HRM) provides uni-
form technical stormwater management guidelines for our
highway designers and other stormwater professionals.
Stormwater treatment is required when 5,000 square feet
of new impervious surface is added to the highway foot
print. Design guidelines forstandard BMPs are in the HRM
and are approved by the Department of Ecology. They in-
clude: biofiltration swale, wet pond, infiltration pond, wet
vaults, and nutrient control wet ponds.
Innovative BMPs
Oftentimes design and hydraulic engineers cannot se-
lect standard BMPs because:
• Technology is constantly changing and the science of
stormwater treatment is advancing.
• There is a need to comply with more stringent local
requirements.
• Space is limited and expensive.
• Land is not available to install a conventional BMP.
• Land has slope, soil, or light problems.
• A specific water quality need is required which a stan-
dard BMP can not satisfy.
These limitations have created a need for more adap-
tive and effective BMPs. WSDOT has responded by de-
veloping an Innovative BMP Development and Research
Program. After an innovative BMP performance has been
verified and accepted by the Department of Ecology, the
BMP will be added to the HRM — if it performed success-
fully.
Documenting Innovative BMP Performance
Monitoring innovative BMP sites provides valuable per-
formance data. WSDOT's Stormwater Management Pro-
gram defines our stormwater research protocols and pro-
cedures. The SWMP can be downloaded through the
264
-------
Internet from WSDOT's WEB page. WSDOT has deter-
mined that the primary pollutants of concern that are rep-
resentative of highway runoff are:
Solids: Total Suspended Solids
Metals: Cadmium (total)
Copper (total)
Lead (total)
Zinc (total)
Oxygen Biochemical Oxygen Demand
Demand: (5 day)
Nutrients: Phosphorous (total)
Orthophosphates
Total Dissolved Solids
Cadmium (dissolved)
Copper (dissolved)
Lead (dissolved)
Zinc (dissolved)
Chemical Oxygen
Demand
Nitrate-Nitrogen
These stormwater constituents will be analyzed routinely
at our research sites. On a semi-annual basis, WSDOT
performs priority pollutant scans which include polynuclear
aromatic hydrocarbons, ultimate (20-day) biochemical oxy-
gen demand, and effluent toxicity using the Microtox™ tech-
nique.
Vegetated Filter Strip
Research Objective
Vegetative filter strips currently exist along many of our
rural highways; however, they are not considered a stan-
dard BMP by the Department of Ecology. The research
objective is to gather the needed data to demonstrate that
a vegetative filter strip should be accepted as a standard
BMP in our HRM and under what conditions. Three test
filter strips are being evaluated at our SR 8 Black Hills
research site in Thurston County.
Design
Vegetative filter strips are planned for use in rural areas
on state highways. The average daily traffic count at these
sites does not exceed 30,000 vehicles per day. Vegetative
filter strips run parallel to the roadway, and runoff from the
roadway flows off of the roadway, across the shoulder and
then across and into the vegetative filter strip. Typically
filter strips are 15 feet wide and 2 to 4 feet from the edge
of the pavement.
Figure 1 shows a filter strip BMP plan view site layout of
a facility which was built in January 1996. The slot drain
serves as the control for runoff volume and pollutant con-
stituents. Three different soil matrixes are evaluated at the
research facility, including commercially available compost,
organic-rich soil from a local river bottom, and the existing
rock and soil that was excavated (or road-Ex) at the re-
search site. Each filter strip was hand seeded with
Mechlenberg Fescue.
The facility design allows the amount of runoff moving
across and into each filter strip to be monitored, as well as
to determine the pollutant removal performance. Because
each filter strip was lined with a clay liner, a water balance
can be performed. To differentiate between surface over-
flow and flow through the filter strip, perforated drain pipes
were installed at the top and at bottom of each filter strip.
The drain pipes are separated by an impervious layer to
prevent cross-flow.
Preliminary Results
Data have been collected for twelve storm events over
two years. Data will continue to be collected for an addi-
tional year. Washington State University evaluates the data
and prepares a report.
Preliminary results demonstrate that compost fill may
provide stormwater detention as well as treatment. The
compost soil matrix captured and held the stormwater; only
a small proportion of the total storm event flowed through
the compost media. No surface overflow was recorded in
any storm event.. The data also show that composted
material initially releases nitrogen and phosphorus into the
discharge water. This may be a concern in watersheds
with sensitive lakes.
The organic soil matrix allowed about half of the runoff
to sheet flow over the filter strip and the other half to infil-
trate through the soil matrix. The road-Ex soil allowed more
than 80% ofthe runoffto sheet flow overthe filter strip and
less than 20% of the runoff to infiltrate through the filter
strip.
The grass grew best on the compost soil, followed by
the organic-rich top soil. Grass grew the poorest on the
road-Ex. Because of poor seed germination in the organic-
rich soil and road-Ex, grass seed and water was reapplied
on three separate occasions. Runoff and infiltration rates
were directly related to grass growth. Increased infiltration
was observed as grass growth improved.
Difficulties
Accurately recording the discharges through the system
was very difficult. Runoff volumes were too small to get a
measurable head in the smallest (0.4 HS) commercially
available flume. On the other hand, runoff was too large to
contain in a single container. In addition, runoff and infil-
tration ratios varied widely from each filter strip. It was also
very difficult to use commercially available flow-proportion-
ate samplers to collect samples for water quality analysis.
Through many trials and disappointments, we finally were
successful with:
source
slot drains
(control sample)
overflow
underflow
quantity measurement
0.4 HS flume in conjunc-
tion with ISCO 3700
samplers
nutating disk meter
nutating disk meter
quality measurement
flow proportioned
samples using ISCO
3700 samplers
"in-line sampler"
"in-line sampler"
For locations where using automated samplers proved
to be ineffective, an in-line sampler was devised to obtain
flow proportionate samples. The in-line sampler conveys
the flow through a one-inch diameter pipe and across an
irrigation drip valve. Vinyl tubing attached to the drip valve
conveys a small proportion of the total flow to a plastic
bucket. The angle ofthe drip valve controls the amount of
265
-------
o
>,
JO
-------
flow. The vinyl tubing can also be pinched to further re-
duce the flow.
Ecology Embankment
Research Objective
The research objective is to gather the needed data to
see if the ecology embankment can be accepted as a stan-
dard BMP in our HRM. The performance of the ecology
embankment will be compared to that of the bioswale.
Design
The ecology embankment design was developed as part
of a highway improvement project in King County. The
project had limited space to install a standard BMP due to
wetlands, streams, and riparian buffer zones. The ecology
embankment is a modification of the bioswale. The ecol-
ogy embankments will provide water quality treatment of
highway stormwater runoff using the space available in
the side slopes of the highway prism to filter out solids
suspended in highway stormwater runoff.
Figure 2 shows a cross section of the ecology embank-
ment. The design contains an 8-inch PVC underdrain pipe
in a 2-foot-wide trench bedded with gravel at its base. The
underdrain pipe and trench allows conveyance of treated
runoff.
Above the pipe trench, the embankment contains a mini-
mum 1-foot layer of a mixture of soil and soil amendments,
the ecology mix. The ecology mix layer is overlain by a
porous geotextile mat. The mat is crush-resistant, pliable,
resilient, water-permeable, and highly resistant to chemi-
cals and decomposition. The exposed surface of the em-
bankment is seeded, fertilized, and mulched twice.
A slot drain collects control samples for untreated runoff
water quality and runoff volumes. A sampling station vault
houses monitoring equipment including 0.4 HS flumes,
nutating disc meters, deep-cycle marine batteries, and
other equipment.
Difficulties
Accurately recording the discharges through the system
may be difficult. We anticipate that the monitoring tech-
niques we ultimately selected for the filter strip research
site will work here as well.
The maintenance office was very concerned with the
ecology mix and drain pipe associated with this BMP. First,
maintenance states that the ecology mix does not have
the structural integrity to support vehicle loads. They fear
that the travelling public may drive off the road and get
stuck in the embankment. Secondly, because the drain pipe
requires occasional flushing, the maintenance office felt
• See Note 2
Seed, Fertilize, Mulch,
and Second Mulching
Gravel Borrow
Ecology Mix
Synthetic Mat
Gravel Backfill for Drains
8" PVC Underdrain Pipe
Holes as Shown
Construction Geotextile for Separation
Gravel Backfill for Pipe Bedding •
0.65' or as Required for
a Uniform Profile
Anchor 3' on Center in Accordance
with Mfg. Recommendation
3/8" minus Pea Gravel
6 Mil Polyethylene Liner
Cast Spoils to Create a Berm
4" PVC Underdrain Pipe
See Note 3
Figure 2. Modified ecology embankment cross section A-A.
267
-------
that the ecology mix wouldnt support their heavy equip-
ment. The design was modified to allow for flushing of the
drain pipe from the edge of pavement. Clean outs will be
constructed on the roadway shoulder at a considerably
increased cost.
Swirl Concentrator Systems
Research Objective
The Stormceptor™ is a commercially available vault
system that utilizes swirl concentrator technology. The pri-
mary objective of monitoring the Stormceptor™ is to verify
pollutant removal rates independent of manufacturer
claims. The manufacturer reports solids removal rates at
greaterthan 85% for low-flow events, with significantly less
removal capacity at higher-flow events. Since the manu-
facturer emphasizes oil and grease removal as one of its
main features, grab samples will be collected to provide
better estimates of floatable hydrocarbon removal. The
second objective is to collect maintenance requirements
data.
Design
The Stormceptor™ was selected as part of a highway
improvement project in King County. Since the project had
limited space to install a standard BMP due to buildings
and Metro bus stops, the Stormceptor™ units are installed
beneath the pavement in lieu of a catch basin or manhole.
Figure 3 shows the Stormceptor™. It is a dual-level vault
designed for ultra-urban settings to enhance the removal
of sediments and oil. The Stormceptor™ is divided into a
lower storage/separation chamber and upper bypass
chamber.
Normal flows are diverted into the lower treatment cham-
ber where oil and other light non-aqueous phase liquids
rise. They then become trapped and suspended solids
settle to the bottom of the chamber by gravity and centrifu-
gal forces. During high-flow conditions, the bypass cham-
ber conveys water to the downstream storm sewer directly
circumventing the lower chamber. This prevents the
resuspension and scour of settled pollutants.
Figure 3. Stormceptor™.
268
-------
Difficulties
Safety and traffic were major concerns. Placement of
the Stormceptor™ vaults along an urban, unlimited access
highway, coupled with a dedicated High Occupancy Ve-
hicle (HOV) lanes with Metro bus stations made selection
and placement of monitoring equipment difficult. In addi-
tion, change orders during the construction phase forced
us to step back and re-evaluate our sampling plan. The
change orders made it impracticable to use automatic sam-
pling equipment. It was decided that collection of time-
composted grab samples was the only option. Grab
samples will be collected at the catch basin preceding the
BMP and at the outfall to an urban creek.
Enhanced Wet Detention Ponds
Research Objective
WSDOT would like our existing wet ponds to work bet-
ter. In areas, where native soils have a high clay content,
stormwater runoff from construction sites remains turbid
even after being detained and passed through a wet pond.
The discharges often exceed water quality standards for
turbidity and suspended solids. WSDOT is investigating
the use of coagulants to improve turbidity removal.
A group of long-chain polymers called polyacrylamides
(PAM) were selected to test. Only the anionic form of PAM
will be tested. Nonionic and cationic forms of PAM will not
be used in WSDOT's experimental BMP because of toxic-
ity considerations. Two PAM products which will be tested
in WSDOT's innovative coagulation/flocculation BMP -
Cytek Industry's Magnifloc™ 866 A and Magnifloc™ 905N
flocculants.
WSDOT plans to test PAM at a detention pond in south-
ern King County. This area has predominately clay loam
glacial outwash soils. These soils have a history of prob-
lems with mass wasting, erosion control, and water quality
because of high levels of turbidity. Slope stability problems
preclude construction of additional BMPs. The site receives
runoff from approximately 15 acres, 4 of which are cur-
rently paved. Monitoring of grab samples of pond effluent
registered turbidity readings which, on occasion, have ex-
ceeded 200 NTUs.
PAM will be tested at a second pond along SR 5 -
Leverich Park near Vancouver, Washington. The deten-
tion pond which receives SR 5 runoff was constructed in
1978. During large precipitation events, high levels of tur-
bidity have been observed in the Leverich Park detention
pond because of influent highway runoff and resuspension
of sediments residing in the pond. PAM can enhance the
effectiveness of this detention pond by flocculating sus-
pended sediments and preventing resuspension of sedi-
ments that may get discharged to Burnt Bridge Creek.
Design
No special construction provisions are needed for using
PAM to enhance sediment removal in stormwater deten-
tion ponds. Standard construction and design practices for
stormwater detention ponds are adequate.
PAM delivery system will be a "tea bag" device. This can
hold granular PAM and be anchored within the influent pipe
or channel. During precipitation events, stormwater is ex-
pected to infiltrate through the granular PAM tea bag. The
PAM mixed with the sediment-laden stormwater is expected
to flow into the detention pond, where velocities would be
reduced and the flocculation would result.
Preliminary Results
Preliminary results are very promising. Jar tests using
sediment laden construction site runoff demonstrate greater
than 95% turbidity removal in 30 minutes with 2ppm (parts
per million) PAM dosage. The water turned clear, looking
nearly like drinking water and the sediment was captured
in fluffy, quick settling floes. The jar tests also demonstrate
that the floes form at as little as 0.8 ppm PAM dosage.
Aluminum chlorohydrate is an efficient coagulant and acts
as a bridging agent between solids and PAM. Jar tests
demonstrate that a one-to-one ratio of aluminum
chlorohydrate and PAM further improves the settling pro-
cess.
Difficulties
WSDOT has been unable to initiate full-scale field tests
using PAM. The Department of Ecology is very concerned
about the addition of chemicals to the waters of the state.
Standard testing protocols are not available and have de-
layed the testing of this potentially very effective product.
Specifically, short-term and long-term aquatic toxicity test-
ing are required.
Ship Canal BMP Research Facility
Research Objective
The objective is to build a full-scale ultra urban highway
runoff research facility. The facility will be the testing
grounds for new stormwater BMPs designed for limited
space situations. Our goal is to have sufficient performance
data to be able to select and install BMPs at the lowest
possible costs that comply with water quality standards.
Design
WSDOTs most ambitious project is the Ship Canal
Stormwater BMP Facility, located in Seattle. This research
facility will simultaneously evaluate up to six experimental
BMPs treating stormwater runoff from WSDOTs busiest
freeway. The use of controls will minimize bias and im-
prove objective comparative evaluations of BMP perfor-
mance and maintenance requirements. Construction is
currently scheduled for the Summer of 1998.
Figure 4 is a schematic of the Ship Canal Research fa-
cility. The Research facility will contain four test bays. Space
is allowed to add two additional test bays. Commercially
available BMPs or custom-designed BMPs can be evalu-
ated at the site.
Difficulties
A local community action group was concerned about
polluted discharges from the facility, in the event that one
269
-------
o
/(Reversible
Lanes)
Existing 30"
Storm Sewer
Diversion Manhole
(6 mo. storm to
treatment facility)
• To Be
' Decomissioned
Bridge
Columns
(typ)
Outfall
rto Lake Huron
\
Swirl
Concentrator
Figure 4. Schematic of ship canal.
of the units failed or had poor performance. Therefore, the
design calls fora swirl concentrator-type BMP as a polish-
ing step before discharge into Lake Union.
Because the facility will be built in an urban location, the
designer had to work around many issues. Water, cable,
and fiber optic lines will need to be relocated. The facility
also had to work around the bridge columns. Walls will be
constructed because of the large changes in elevation.
The design was first completed in November 1997. The
bids were 42% higher than anticipated. Sufficient funds
were not available to construct the facility. A second, scaled-
back design will be completed during the month of Febru-
ary 1998. We plan to go out to bid during the month of
March.
Challenges
There are many challenges when conducting innovative
BMP performance monitoring. Finding good research sites
takes time and requires thorough research. Major recom-
mendations are:
• Select a safe site for staff and for the traveling public.
Give consideration to highway traffic patterns and ac-
cess. Can sampling staff exit and enter the highway
SAFELY?
Know the hydraulics of the site. In the field there is no
such thing as sheet flow. Dips, gouges, bumps and
irregularities direct and redirect water into or around
the test site. Flow patterns during various storms should
be factored. Sites that include drainage from off-site
contributors should be eliminated to avoid having to
sample all discharges into the site.
Minimize or eliminate confined spaces from the de-
sign. Confined space regulations limit the vertical depth
of our sampling vaults to a maximum of four feet; how-
ever, the vaults should be deeper to allow for the ver-
tical distance needed for water quality monitoring. Our
most recent vault designs allow for both by utilizing a
false floor that the sampling staff can safely stand on.
Sections of the floor can be removed as needed.
Minimize travel time. WSDOT tries to locate research
facilities within 15 to 20 minutes of the office and within
an hour of an environmental laboratory. The longerthe
commute, the more difficult it is to collect or deliver
samples and to pick up spare parts and to make the
repairs necessary to keep the facility operational.
Supervise the construction: as drawn does not mean
as built. Review the plans with the project engineer.
Then check and confirm the construction as shown in
270
-------
the plans. Stay in touch and make sure that all change
orders are approved by the research staff.
• Allow time to listen and understand resource agency
concerns. Balance the need for more effective BMPs
against the need for complete environmental under-
standing and the fate and function of a BMP. If chemi-
cals are added to the water, include short- and long-
term toxicity testing and hazardous waste testing of
sludges as an element of the testing protocols.
• Innovation means change. Change threatens many
people and upsets traditional organizational roles and
responsibilities. The next challenging step is to incor-
porate the research results into how highways are
designed and to have the regulating agencies accept
designs.
Conclusions
The Washington State Department of Transportation has
earned a reputation as an innovative developer of large
public works projects. Constructing and maintaining trans-
portation facilities will always be the primary mission of
WSDOT.
However, the agency accepts responsibility for the po-
tential environmental and social impacts of our facilities. A
major responsibility is to protect our state's water quality
and to preserve our environmental values. WSDOT has
adopted the watershed-need philosophy as the best way
to provide cost-effective water quality treatment.
The Innovative BMP Research program will help pro-
vide clear selection criteria and design parameters for
stormwater BMPs. This, in turn, will provide the best BMP
for each project or outfall site. As the years go by, WSDOT
anticipates a diminishing impact to watersheds from its
transportation facilities.
271
-------
StormTreat™ Technology for Stormwater Treatment
Mark E. Nelson, Director and Scott W. Horsley, President
StormTreat™ Systems, Inc.
Sandwich, Massachusetts
Stormwater treatment standards throughout the nation
have focused on removal of suspended solids, trusting that
other contaminants will be removed as well. Massachu-
setts, for example, requires 80% suspended solids removal.
While this is an important first step, there are a wide range
of other contaminants such as fecal conforms, nutrients,
metals and hydrocarbons that may not be treated by tech-
nologies focusing solely on suspended solids removal.
The StormTreat™ System (STS) was developed in 1994
in response to the need to provide enhanced treatment of
Stormwater, beyond suspended solids removal. The tech-
nology is designed to capture the "first flush" of runoff, and
provide treatment in a 9.5-foot diameter tank through sedi-
mentation, filtration and constructed wetlands uptake. Two
years of independent testing results indicate removal rates
of 97% for fecal coliform bacteria, 90% for phosphorus,
77% for dissolved nitrogen and 99% for total suspended
solids.
StormTreat™ is the first Stormwater technology to be
verified by the Massachusetts Strategic Environmental
Partnership (STEP) Program, a state program designed
to verify the claims and effectiveness of new technologies.
The Massachusetts STEP program is part of a six-state
Partnership for Environmental Technology, including Illi-
nois, California, Pennsylvania, New Jersey and New York.
StormTreat™ has recently developed an optional infiltra-
tion feature to treat Stormwater from as much as one im-
pervious acre per unit. This is accomplished by infiltrating
treated water directly into surrounding and underlying soils.
With this option, treatment costs can be reduced to as low
as $7,000 per acre treated (including installation).
Introduction
Stormwater runoff from streets, parking lots and adja-
cent areas is one of the most significant water pollution
problems today. Lakes, reservoirs, streams, coastal wa-
ters and related wetlands receive "pulses" of oils, metals,
bacteria, nutrients and other pollutants during and follow-
ing each storm event. Where Stormwater is infiltrated or
injected into groundwater, impacts may occur to subsur-
face drinking water supplies.
Chronic petroleum hydrocarbon discharges to the ma-
rine environment from Stormwater runoff far outweigh those
from catastrophic oil spills from tanker ships (such as the
Exxon Valdez). Researchers associated with the EPA-
sponsored Buzzards Bay National Estuary Program have
concluded that stormwater-derived bacterial loadings are
responsible for the majority of shellfish area closures.
Stormwater has also been documented as a major cause
of eutrophication of lakes and ponds nationwide.
Non-point sources of pollution such as Stormwater typi-
cally originate from diffuse areas. Stormwater is generated
from streets, parking lots, rooftops, driveways, lawns, ag-
ricultural fields and forests (Figure 1). Frequently,
Stormwater runoff from several of these various "land uses"
is combined into a Stormwater flow in a drainage ditch or
Stormwater pipe which ultimately discharges to a receiv-
ing water where the impacts are realized. Because of the
diffuse nature of sources, Stormwater management is best
accomplished by a watershed management technique
which treats each area within the watershed independently
as opposed to the more conventional "big pipe" solution
where a large detention/treatment system is constructed
at the bottom of a watershed attempting to catch and treat
all of the Stormwater generated within the watershed.
There is a direct relationship between the traffic volume
on a road and the concentrations of metals found in the
Stormwater. This suggests that Stormwater quality is likely
to degrade over time as urbanization continues and land
use densities increase. A study conducted by Arnold and
Gibbons (1996) has shown a direct relationship between
the percentage of impervious cover and the quality of ad-
jacent surface waters. When impervious surfaces cover
30% of the land area, significant water quality impairment
is evident.
While there is a wide variety of compounds present in
Stormwater, the majority of treatment approaches focus
272
-------
Atmospheric Deposition
I
Pollutants carried away
by wind and traffic
Animal Wastes
Stormwater
Overflow Pipe to
Discharge Area
IT /I
rP 4
1 M0 1 II^J
-«
Catch Basin
^-\_i — ,-,, .1 y
IN. :
in
Figure 1. Stormwater Pollutant Pathways.
on removing suspended solids. The premise is that if sus-
pended solids are removed, a wide range of other pollut-
ants associated with the solids are removed as well. Re-
moval of suspended solids is used as a performance stan-
dard in numerous Stormwater regulations. The most re-
cent example may be the State of Massachusetts which
now requires 80% suspended solids removal fortreatment
systems approved within 100 feet of inland or coastal wet-
lands (MA DEP, 1997).
There are two problems with this thinking. First, the
majority of the suspended solid particles (Table 1) are very
small, with 78% of them less than 44 urn in diameter
(Rexnord, 1984). This presents significant problems with
respect to best management practices (BMPs), which rely
wholly upon physical settling, or separation, of solids. The
settling rates associated with silt and clay-sized particles
are on the order of several hours to days per foot of set-
tling. This has significant implications with regard to the
effectiveness of detention basins and other treatment ap-
proaches in meeting performance standards.
Second, a significant percentage of most pollutants is
associated with the finer particles (Table 2). For example,
approximately 56% of phosphates in Stormwater is asso-
ciated with particles smallerthan 43 urn. If the smaller par-
ticle sizes are not removed, the expected treatment of other
constituents is not obtained.
Table 1. Wet Sieve Analysis of Highway Runoff Composite Samples
Percent of Suspended Solids
Particle Size
(urn)
>250
88-250
44-88
<44
Sacramento
Hwy. 50
1.54
9.07
10.70
78.69
Harrisburg
1-81
6.10
6.70
11.70
75.50
Milwaukee
I -94
14.56
7.00
5.84
72.60
Effland
I-85
3.58
1.30
8.06
87.06
Mean
6.45
6.02
9.08
78.45
Source: Rexnord, Inc., 1984
Generally speaking, a "first flush" effect is observed with
Stormwater quality. The highest pollutant concentrations
are typically observed at the beginning of a storm event.
This is because of the residues which are available on
paved surfaces at the beginning of a rain event. As this
first flush is washed from the paved surface, pollutant con-
centrations typically decline throughout the remainder to
the storm event. The "first flush" principle has been com-
monly observed for total suspended solids. Up to 90% of
the total suspended solids (TSS) are contained within the
first 0.5 inches of runoff (EPA, 1974).
There are exceptions to the first flush principle. For ex-
ample, a long light rain followed by a strong downpour may
exhibit its highest concentrations of pollutants toward the
end of a storm event. However, when one averages all
rainstorms, the majority of annual pollutant loading occurs
during the first flush. Another exception is a large water-
shed which has a long "time of concentration" (the time
required for water to flow from the uppermost part of the
watershed to the final point of discharge). In these cases,
samples taken at the discharge point may integrate the
first flush from the bottom of the watershed but miss the
first pollutant loadings from the upper part of the water-
shed.
The Storm Treat™ System
The StormTreat™ System was designed in 1994 in an
effort to treat Stormwater in a way that would effectively
remove a broad range of pollutants in addition to suspended
solids. It saves space by reducing the need for unsightly
detention basins. It captures and treats the first flush of
runoff which contains 90% of pollutants. An optional infil-
tration feature provides for the treatment of larger quanti-
ties of Stormwater, beyond the first flush, as described later
in this paper.
The system consists of a series of six sedimentation
chambers and a constructed wetland which are contained
273
-------
Table 2. Percent of Street Pollutants in Various Particle-Size Ranges
Particle Size (|im)
Pollutant >2000 840-2000 240-840 104-240 43-104 <43
Total solids
Volatile solids
COD
BOD5
Phosphates
All toxic metals
TKN
All pesticides
PCBs
24.4
11.0
2.4
7.4
0
16.3
9.9
7.6
17.4
4.5
21.1
0.9
17.5
11.6
27.0
66.0
24.6
12.0
13.0
15.7
6.9
14.9
20.0
27.8
16.1
12.4
15.2
6.4
23.5
20.2
9.7
17.9
45.0
17.3
29.6
—
19.6
73.0
34.0
5.9
25.6
22.7
24.3
56.2
27.5
18.7
Source: EPA, 1983
within a modular 9.5-foot diametertank (Table 3). It is con-
structed of recycled polyethylene which connects directly
to existing drainage structures.
Influent is piped into the sedimentation chambers where
larger-diameter solids are removed (Figure 2). The inter-
nal sedimentation chambers contain a series of skimmers
which selectively decant the upper portions of the
stormwater in the sedimentation basins, leaving behind the
more turbid lower waters. The skimmers significantly in-
crease the separation of solids compared with conventional
settling/detention basins. An inverted elbow trap serves to
collect floatables such as oils within one chamber of the
inner tank. After moving through the internal chambers,
the partially treated stormwater passes into the surround-
ing constructed wetland through a series of slotted PVC
pipes.
The wetland is comprised of a gravel substrate planted
with bulrushes and other wetland plants. Unlike most wet-
lands constructed for stormwater treatment, StormTreat™
conveys stormwater into the subsurface of the wetland and
through the root zone, where greater pollutant attenuation
occurs through such processes as filtration, adsorption,
and biochemical reactions.
Precipitation of metals and phosphorus occurs within the
wetland substrate while biochemical reactions, including
microbial decomposition, provide treatment of the
stormwater prior to discharge through the outlet valve. An
outlet control valve provides up to a 5-day holding time
Table 3. StormTreat™ Specifications
StormTreat™ Specifications
Diameter 9.5 ft
Height 4 ft
Storage Capacity 1390 gal
Inflow Pipe Diameter 4 in
Outflow Pipe Diameter 2 in
Average Detention Time 5 days
Average Discharge Rate 0.25 gal/min
Tanks Required per Acre of Impervious Surface 2-5 Tanks*
"The number of tanks depends upon the level of treatment required,
the in-line detention capacity and the use of the optional infiltration
feature.
within the system. The valve can be closed to contain a
hazardous waste spill.
The size and modular configuration of StormTreat™'
makes it adaptable to a wide range of site constraints and
watershed sizes. It can be installed in any type of soil as
the discharge rate is very low (0.25 gal/min) and the gravel
filter/wetland substrate is contained within the system. As
the flow through the system is gravity dependent, the sys-
tem requires an elevation change from the pavement sur-
face to a discharge point of at least 4 feet.
StormTreat™ has been installed in a variety of applica-
tions, including commercial parking lots, industrial sites,
town landings and marinas, transportation facilities and
residential subdivisions. It is an appropriate treatment tech-
nology for both coastal and inland areas, and can be used
throughout the country with only minor system modifica-
tions to fit local conditions.
To date, 315 analyses have been conducted on 33
samples which have been collected over eight indepen-
dent storm events during both winter and summer condi-
tions in New England. Influent stormwater samples were
taken at the entry point to the StormTreat™ tanks at the
catch basin. Effluent samples were taken during the 5 days
following the storm event. The quality of the sampled efflu-
ent was then compared with influent and removal rates
were computed. Test results are summarized in Table 4.
StormTreat™ requires minimal maintenance. Annual in-
spection is recommended to ensure that the system is
operating effectively. At that time the manhole is opened
and the burlap grit screening bag covering the influent line
should be removed and replaced; filters should be re-
moved, cleaned, and reinstalled. Sediment should be re-
moved from the system via suction pump once every 3-5
years, depending on local soil characteristics and catch
basin maintenance practices.
State Verification
StormTreat™ is the first and only stormwater treatment
technology to be verified by the Massachusetts STEP Pro-
gram. The Strategic Environmental Partnership (STEP)
Program is a service provided by the Commonwealth of
Massachusetts to help develop new environmental tech-
nologies and to verify their effectiveness. Both business
management and technological expertise is provided
through the University of Massachusetts. An excerpt from
the executive summary of this assessment reads as fol-
lows:
"It is the conclusion of this assessment that the
(StormTreat™) system, when sized according to recom-
mended criteria, with proper operation and maintenance,
can provide levels of treatment required under Standards
4 and 6, as specified by the Massachusetts DEP
Stormwater Management Handbook. Under special cir-
cumstances, the system may provide as much as 98%
removal of TSS when sized according to design criteria.
The system, when configured for recharge can meet Stan-
274
-------
Constructed
wetland
inverted elbow
Figure 2. Schematic cross section of Storm Treat™ system.
Table 4. Water Quality Sampling Results, Kingston, MA
Pollutant
Average
Stormwater
Influent
Average
Treated
Discharge
Percentage
Removed (%)
Fecal coliform (no./100 ml)
Total suspended solids (mg/l)
Chemical oxygen demand (mg/l)
Total dissolved N (mg/l)
Total Petroleum Hydrocarbons (mg/l)
Lead (mg/l)
Chromium (mg/l)
Phosphorus (mg/l)
Zinc (mg/l)
690
93
95
3569
3.4
6.5
60
300
590
20
1.3
17
520
0.34
1.5
1
26.5
58
97
99
82
77
90
77
98
90
90
Note: Samples were collected by the Jones River Watershed Association in accordance with EPA sampling protocol, and analyzed at state-certified
laboratories (Schueler, T. 1995).
dard 3 and is also likely to meet Standard 5, for land uses
with higher potential pollutant loads, when sized accord-
ing to design criteria."
The Massachusetts STEP is a member of the Six State
Partnership for Environmental Technology which includes
California, Pennsylvania, Illinois, New Jersey, and New
York. Reciprocal certifications in these states may be avail-
able.
Infiltration of Treated Stormwater
An optional infiltration feature enables the StormTreat™
System to process as much as one acre of impervious
surface area per unit. This is accomplished by directing
treated Stormwater into the surrounding and underlying
soils (Figure 3). There are several advantages to this ap-
proach. First is the replenishment of groundwater. Under
natural (undeveloped conditions) a certain percentage of
precipitation infiltrates through the soils and recharges un-
derlying groundwater (this may be as high as 50% of the
annual precipitation resulting in groundwater recharge in
the Northeast US). Commonly, the development of land
results in an increase of impervious surfaces reducing the
recharge rate (in some cases to zero). Less recharge
means less water supply availability, declining watertables
and less baseflow (discharge) to nearby streams and wet-
lands.
A second advantage of this new feature is significantly-
increased treatment and lower costs to customers (as low
as $7,000/acre installed). By allowing treated Stormwater
to infiltrate into the surrounding soils, additional treatment
capacity is achieved. This is accomplished first by infiltrat-
275
-------
Sto-ni
-------
Evaluation of Stormceptor® and Multi-Chamber Treatment Train as Urban
Retrofit Strategies
Steven R. Greb
Wisconsin Department of Natural Resources
Madison, Wisconsin
Steve Corsi and Robert Waschbusch
US Geological Survey
Madison, Wisconsin
Introduction
The installation of water quality best management prac-
tices (BMPs) in developed urban areas is problematic. A
landscape comprised of buildings and pavement presents
little opportunity for placement of new BMPs. To overcome
this obstacle of limited space, a new set of retrofit BMP
technologies are emerging which utilize space under-
ground, thereby avoiding disruption to current above-
ground land uses. This paper evaluates the water quality
benefits of two retrofit BMPs, the Stormceptor and the Multi-
Chambered Treatment Train (MCTT). The installation of
these devices and the subsequent evaluations were co-
operative efforts involving the US Environmental Protec-
tion Agency, US Geological Survey, cities of Milwaukee
and Madison, (Wisconsin), StormceptorXE Corp., Univer-
sity of Alabama-Birmingham, and Wisconsin Department
of Natural Resources.
Study Design
Description of Test BMPs
The Stormceptor consists of a treatment tank and by-
pass chamber (Figure 1). Water initially enters the bypass
chamber from an upstream stormsewer. During periods
when the flow doesn't exceed the unit's capacity, water is
diverted down a drop pipe, where water is discharged tan-
gentially along the chamber's wall. Suspended sediment
falls to the bottom of the tank where after a period of accu-
mulation, it is pumped out and landfilled. Water exits the
treatment chamber at the opposite end through a similar
drop pipe and drains to the downstream stormsewer. Hy-
drocarbons and other lighter-than-water materials are
trapped above the treatment tank's drop pipes. During
periods of surcharge, the portion of the stormwater in ex-
cess of the treatment rate flows directly over the weir in
the bypass chamber and receives no treatment. The unit
chosen for this project was the STC 6000. Its capacity is
6150 gallons and it is designed for a treatment flow capac-
ity of 800 gal/min. The treatment tank is 10 ft in diameter
and 10 ft deep.
The MCTT consists of three components, an inlet area,
a settling chamber, and a filter bed (Figure 2). The largest
grit material accumulates in the bottom of the 4 ft inlet ba-
sin. In addition, water passes over a mesh bag of column
packing balls which enhance aeration and loss of highly
volatile components. The second chamber has inclined
tube settlers which further enhance the settling process.
This chamber also contains sorbent pads which remove
floatable hydrocarbons. Water drains slowly from the sec-
ond chamber into the filter bed chamber via a 0.35-inch
orifice. This final chamber contains a mixed media of sand/
peat/ activated carbon supported by filter fabric and is de-
signed to remove fine particles along with some dissolved
constituents via sorption and ion exchange. The second
and third chambers are constructed from a partitioned con-
crete box (10 ft wide x 19 ft long x 5 ft high). The capacity
of the settling chamber is 750 ft3 although the height of the
orifice results in a "dead storage" capacity of 375 ft3, leav-
ing the actual storm volume capacity at 375 ft3. Once this
capacity is reached, the excess water is bypassed.
Site Descriptions
Both study sites are public works maintenance yards
used forfueling, storage and maintenance of city vehicles,
most of which are heavy trucks. In addition, open storage
of sand, salt and yard wastes can be found at times. The
Stormceptor was installed at the Badger Rd. public works
garage in Madison, Wisconsin. One stormsewer inlet col-
lects the runoff water from the entire facility (4.3 acres)
and the Stormceptor was retrofitted in the existing
stormsewer approximately 300 ft from the inlet. The MCTT
was constructed at the Ruby St. garage in Milwaukee,
Wisconsin. The unit receives waterfrom only a portion (0.2
acres) of the paved area.
Sampling Design
Similarsampling strategies were employed forthe evalu-
ation of both these devices. The evaluation consisted ba-
sically of collecting flow-integrated samples at the influent
and effluent for each BMP. In addition, the Stormceptor's
277
-------
-Manhole Access
Existing
Stormsewer
Bypass Chamber
Effluent
Outlet Sample
Point
Figure 1. Cross section of Stormceptor® device showing placement of sampling points.
Jnflow
Inlet
Sampling
Point
Perforated Pipe
15" Diam.
Concrete
Pipe
Outlet
Sampling
Pipe
Figure 2. Cross section of MCTT device showing placement of sampling points.
bypass water was also collected. The locations of sam-
pling ports are shown in Figures 1 and 2.
Sampling equipment and monitoring instrumentation
were housed in small onsite buildings. In addition, tipping
bucket rain gages were placed on nearby rooftops.
Dataloggers served as the site controller, with a modem
and telephone for external communications. At both sites,
flow was measured either directly, with a doppler or elec-
tromagnetic flow meter, or indirectly, with a pressure trans-
ducer/stage height measurement. The dataloggerwas pro-
grammed to initiate rainfall, stage, and velocity measure-
ments on a variable time scale. Measurements were taken
more frequently during runoff periods and less frequently
during dry periods. The data were recorded using internal
memory of the datalogger and a backup storage module,
and transferred every 24 hours via modem to the USGS
computer in Madison.
Both sites had two refrigerated samplers equipped with
peristaltic pumps and Teflon-lined sample tubing to collect
water quality samples. The automatic samplers were trig-
gered by the datalogger to initiate collection of storm
samples. One composite sample each was collected at
the inlet and the outlet for each storm. Each sample con-
tained between 5 and 40-stormflow volume-weighted
subsamples, which resulted in event-mean concentrations.
Fifteen consecutive storms were monitored at the MCTT
site from April 29, 1996 through September 8, 1996. A to-
tal of 68 constituents were measured in the samples taken
at the MCTT device. At the Stormceptor site, 45 storms
were monitored from the period of August 6, 1996 to May
1,1997. Samples from 15 of the 45 storms were analyzed
for 37 constituents included a variety of solids, nutrient,
metals and polycyclic aromatic hydrocarbons (PAHs). The
remaining 30 storm samples were analyzed only for total
278
-------
suspended solids (TSS), total dissolved solids (IDS) and
total phosphorus (TP). In both studies, loads and removal
efficiencies were calculated (load=storm volume x event-
mean concentration). Two types of removal efficiencies
were calculated: tank efficiencies based on reduction of
load of stormwater passing through the tank, and overall
removal efficiency, which also accounts for load which is
bypassed. In addition, particle-size characterization was
performed on 15 samples at each site. Only a few con-
stituents are presented here to highlight the studies' find-
ings. Complete results can be obtained from the USGS
Water Resource Division in Madison, Wisconsin.
Results
Stormceptor
Precipitation for the 45 monitored storms ranged from
0.02 to 1.31 inches. This rainfall produced runoff amounts
ranging from 120 ft3 to 30,000 ft3. Though for 24% (11 out
of 45 storms) water bypassed the unit, the total water vol-
ume that bypassed equaled only 9%. Stormwater was
observed bypassing the treatment tank at flows greater
than 500 gal/min, which is less than the manufacturer's
specification of 800 gal/min. This difference may have been
caused by the exit sewer pipe being slightly higher in el-
evation than specifications called for, causing a back pres-
sure through the unit and resulting in the unit bypassing
more often.
The influent total suspended solids (TSS) concentrations
found are comparable to parking lot and street runoff con-
centrations observed elsewhere in Wisconsin (Bannerman
et al., 1993) and other locations (Ellis, 1986). The influent
TSS event-mean concentrations for the 45 storms ranged
from 43 to 1236 mg/l with a median value of 251 mg/l. In
general, the highest influent TSS concentrations were ob-
served in the winter months, presumably reflecting the high
activity of sand/salt trucks in the yard area. The influent
TSS load ranged from 0.45 to 224 kg. and the cumulative
influent load for the 45 measured storms was 1670 kg. An
estimated 91 % of this load entered the treatment tank; the
remaining 9% was bypassed.
The TSS concentrations of the water exiting the treat-
ment tank had a median concentration of 151 mg/l and
ranged from 45 to 615 mg/l, approximately half the range
of the influent concentrations. The total load exiting the
treatment tank (1044 kg.) resulted in an overall removal
efficiency for the treatment tank of 26%. The effluent TSS
load totaled 1294 kg. (from 4.3 acres) for the 45 storms,
indicating an overall reduction efficiency (treatment tank +
bypass) of 22%. Because of the hydraulic residence time
of the treatment tank, effluent water can be a mixture of
influent waters from a number of previous storms. There-
fore, caution must be used in interpreting individual storm
efficiency results. In general, the monitored storms had
sufficient runoff water to replace the majority of water in
the treatment tank. Eighty-four percent of the storms ex-
ceeded one tank volume, 62% exceeded two tank volumes.
Figure 3 illustrates the relationship between the individual
effluent load and overall efficiency. There appears to be
considerable variation in efficiency for the smaller storms
and, in general, as the storm loads become larger, the ef-
ficiency decreases.
Total Dissolved Solids (TDS) concentrations were also
quite variable, ranging more than two orders of magnitude
in both the influent and effluent (median influent and efflu-
ent concentrations were 3860 and 4700 mg/l, respectively).
Similar to TSS concentrations, the TDS concentrations
showed a marked increase during the winter season, pre-
sumably due to salt stockpiled onsite and spillage from
trucks. The maximum TDS concentration (114,000 mg/l)
was on 1/26/96. Negative removal efficiencies were found
for both the treatment tank and the overall system (both at
-19%). The fact that efficiencies were negative may be due
to either measurement errors or possible dissolution of
granular salt within the tank. It is interesting to note that
50 100 150
Stormceptor Effluent Load (kg)
Figure 3. Relationship between individual effluent load and overall efficiency at the Stormcepter® site.
200
279
-------
the influent TDS load was more than 20 times the TSS
load.
Total phosphorus (TP), the third constituent measured
in all 45 monitored events had concentrations ranges from
0.08 to 1.3 mg/l in the influent and 0.06 to 0.86 in the efflu-
ent. The tank and overall removal efficiencies (20% and
18%, respectively) were somewhat less than TSS removal
rates.
The remaining 34 constituents were measured in 15
storm samples only. In general, the removal efficiencies of
the total constituent loads were similarto the TSS removal
rates and none had negative efficiencies. Total polycyclic
aromatic hydrocarbons exhibited the highest overall re-
moval rate (32%). Of the four metals quantified, zinc had
the highest concentrations and loading. The overall reduc-
tion efficiency fortotal zinc was 21%. Removal efficiencies
of the dissolved constituents were always less than the
total constituents with the exception of dissolved phospho-
rus, which interestingly was slightly greaterthan total phos-
phorus. Five dissolved constituents had negative efficien-
cies, which may of been a result of load errors, dissolution
in the tank and redox processes in the tank's accumulated
sediment. The increase in chloride mass further suggests
that granular sodium chloride is going into solution after
entering the tank. Given that the device is basically de-
signed for particulate solids removal, the negligible removal
of dissolved constituents was anticipated.
Particle-size analyses of the 15 influent and effluent
samples showed little shift in the size distribution (Figure
4). The small clay-size fraction increased slightly from 3.3
to 3.6%. Silt was the predominant size fraction and in-
creased from 88.9 to 93%; the sand fraction decreased
from 7.8 to 3.5%.
At the end of the study period, the oily surface material,
water, and bottom sediments were pumped from the treat-
ment tank. Approximately 16 ft3 of floating oily material was
captured at the top of the treatment tank. As the tank was
pumped down, the water was subsampled for TSS and
TDS. An increasing gradient in TDS concentration was
observed from top to bottom. At 7 ft from the top of the
tank, the TDS had increased from 51,000 to 138,000 mgl.
Greater water density caused by this high salinity could
have hindered particulate solids settling. In addition, this
sharp increase in density may have resulted in a portion
(30-40%) of the tank's volume being resistant to mixing,
thereby decreasing the treatment ability of the tank. Using
the manufacturer's sizing guidelines, this decreased ef-
fective volume would cause a marked decrease in tank
efficiency. But inconsistent with this hypothesis is the fact
that the first 14 storms (before the saline buildup) exhib-
ited a total removal efficiency of only 5% and the last 14
storms monitored (during the period of high saline condi-
tions) had above average (25%) removal. Therefore, the
impact of the salt on the tank's settling ability is unclear at
this time.
The depth of sediment accumulated on the bottom was
measured and then subsampled and analyzed for dry
weight, particle size, metals and PAHs. As a mass bal-
ance check on the sediment material removed by the treat-
ment tank, the measured mass of sediment deposited in
the tank was compared to the difference in influent and
effluent TSS mass. Based on sediment depth and dry
weight, the mass was calculated to be 536 kg. This value
compares favorably with a TSS load estimate of 473 kg.
These similar mass values lend credence to the method-
ologies used to sample the stormwater flow and concen-
trations.
100
c
(0
20
40
60 80
Particle Size (urn)
100
120
140
Figure 4. Mean particle-size distributions of influent and effluent TSS for both studies.
280
-------
The treatment tank was found to be quite selective with
respect to removal of the larger-sized particles. The siev-
ing of this material found 80% had a particle size of greater
than 250 urn. This finding suggest the influent stormwater
may have had a higher concentration of the sand-size frac-
tion than was found in the suspended solids particle size
analyses. Larger sand-size material, saltating along the
bottom of the stormsewer pipe may have not been ac-
counted for in the TSS loading. This unaccounted material
would further explain why the TSS load (473 kg) is some-
what lower than what was measured in the tank. If the
tank efficiency was based on the mass collected in the
tank (536 kg.) instead of the TSS load difference (473 kg.),
the TSS removal efficiency would increase by 28 to 34%.
The Multi-Chambered Treatment Train
Fifteen storms occurring from April 29,1996 through Sep-
tember 8, 1997 were monitored and sampled. Rainfall
amounts forthese storms ranged from 0.18 to 1.37 inches.
Based on the delineated drainage area (0.2 acres), total
rainfall volumes ranged from 107 to 815 ft3. The actual
quantity of water passing through the MCTT ranged from
60 to 319 ft3 and comprised approximately 60% of the rain-
fall volume. The other 39% of the rainfall volume may have
been lost through cracks in the aged pavement surface or
in joint leaks between the catch basin and the main cham-
ber. An overestimation of the drainage area may also ac-
count for this difference. A consequence of this loss in
stormwater was that the unit never surcharged, even
though the design hydraulics would have suggested 10
out of the 15 storms monitored should have surcharged
and approximately 22% of the total stormwater should have
bypassed. This fact made the calculation of overall effi-
ciency problematic.
Total suspended solids influent concentrations from the
MCTT ranged from 79 to 1050 mg/l, with a median con-
centration of 232 mg/l, values comparable to TSS concen-
trations found in runoff samples at the Stormceptorsite. A
majority (8 out of 14) of the effluent TSS were below the
detection limit. The highest concentration of TSS observed
in the effluent was 18 mg/l. The cumulative influent load of
TSS to the unitforthe 15 consecutive storms was 18.3 kg.
The effluent TSS load was only 0.30 kg, making the unit's
removal efficiency greater than or equal to 98%. Examina-
tion of concentrations at intermediate points, (Pitt et al.
1997) found the majority of the TSS was removed in the
settling chamber.
Though the overall TSS removal efficiency was impos-
sible to directly measure due to the water loss problems
discussed above, an overall TSS removal was estimated
because it is widely used as a key parameter in BMP evalu-
ations. To calculate an overall TSS removal efficiency, it
was assumed that any storm volume in excess of the set-
tling tank's capacity (375 ft3) would bypass the unit. This
method resulted in a calculated overall TSS removal effi-
ciency of 78% for the 15 storms monitored.
Total dissolved solids (TDS) influent samples had a
median value of 652 mg/l with a range from 164 to 5930
mg/l. The major source of the dissolved solids was a store
of road salt located within the drainage area. This salt re-
sulted in a large load of dissolved solids (sodium chloride)
to the unit which was 4.5 times the particulate (suspended
solids) portion of the solids load for the period of study.
The MCTT removed 13% of the TDS load.
Total phosphorus ranged from 0.10 to 0.44 mg/l with a
median value 0.25 mg/l. Effluent concentrations were gen-
erally an order of magnitude less (median= 0.03 mg/l). This
loss signaled a quite high tank removal efficiency of 88%.
Dissolved phosphorus was consistently less than 10% of
the total phosphorus in both the influent and effluent
samples. Dissolved phosphorus removal efficiency (78%),
though somewhat less than total phosphorus removal, was
still substantial.
Of the five metals examined, total zinc concentrations
were consistently the highest (median = 150 mg/l). The
removal efficiency for total zinc was 91%. Avery high level
of removal was observed for all the metals. Because the
majority of the total metal concentrations were in the par-
ticulate form, the physical removal of this material may be
occurring in all three chambers of the unit, although the
bulk of the material (associated with the suspended sol-
ids) is most likely being removed in the settling chamber.
The removal efficiency of dissolved zinc was 68%, which
was somewhat less removal than total zinc. Actual removal
may have been greater because effluent concentrations
were generally at detection limits.
Total polycyclic aromatic hydrocarbon concentrations
(sum of all 16 species) in the influent samples ranged from
2.9 to 23 mg/l. Total fluoranthene and pyrene consistently
had concentrations that were more than double the con-
centrations of other PAH species with median concentra-
tions of 1.8 and 1.4 mg/l, respectively. The total PAH re-
moval efficiency was 94%. The dissolved PAHs averaged
14% of the total concentrations (dissolved PAH conc./total
PAH cone.). The only dissolved PAH species which was
consistently reported above detection was phenanthrene
(median=0.1 mg/l). Because the majority of the PAHs were
found in the particulate fraction, most of the removal prob-
ably occurred in the settling chamber, which is collabo-
rated by (Pitt et al. 1997). Table 1 is a summary of the tank
loads and reduction efficiencies of both BMPs discussed
in this report.
Influent and effluent particle-size distributions for MCTT
site are reported in Figure 4. Particulate material was com-
prised mostly of the silt-size fraction (approximately 88%)
in both the influent and effluent. Somewhat surprising was
the fact that there was no appreciable shift in the particle
size distributions between the influent and effluent. Though
a decrease in overall particle size would be expected, the
size actually increased slightly in the treated water (al-
though the difference was not statistically significant). This
fact may suggest that the unit is not selective in the size of
particles removed. Another possibility is that material from
the filter media, such as sand fines, had escaped around
281
-------
Table 1. Tank Mass Loads and Efficiencies of the Studied BMP's
Constituent
TSS
TDS
Total P
Dis. P
Total Zn
Dis. Zn
Total PAH
Load-in
18.3kg
84.3 kg
19.3g
0.93 g
11.7g
1.03g
0.64 g
MCTT
Load-out
0.30 kg
73.3 kg
2.4 g
0.20 g
1.0g
0.33 g
.039 g
% Eff.
98
13
88
78
91
68
94
Load-in
1 420 kg
37,500 kg
1 .60 kg
0.44 kg
660 g
110g
67 g
Stormceptor®
Load-out
1 040 kg
44,700 kg
1 .29 kg
0.34 kg
520 g
105 g
42 g
% Eff.
26
-19
20
23
21
5
36
or through the filter fabric. (Pitt et al. 1997) also noted a
slight increase in TSS concentrations as the water passed
through the filter tank of their pilot-scale unit. Therefore
the unit may still be removing larger particles that are sub-
sequently replaced by media material, resulting in no net
change in the distributions. It is important to emphasize
that even though there was no appreciable shift in particle
size distribution, there was still a very high removal of all
particulate material (i.e. suspended solids load) in all
storms.
Cost-Effectiveness of the Two BMPs
Though the economics of implementing these BMPs was
not the focus of this evaluation, it is an important issue to
water quality managers, so a brief cost analysis is offered
below. Though actual construction costs forthe MCTT were
$72,000 ($360,000/acre), some of this high cost had to do
with contractor's uncertainties in building an unknown de-
vice, retrofitting around existing sanitary sewer, and addi-
tion of reinforcements for heavy truck traffic. A similar de-
vice was built in Minocqua, Wi. for $95,000 ($38,000/acre)
which is only a tenth of the cost per acre. Table 2 presents
some estimated costs-per-pound of suspended solids re-
moved forthe two management practices. Clearly a num-
ber of assumptions were required to generate these num-
bers. Depending on the location and scale, there can be
considerable variation in capital and maintenance costs
and control efficiencies. For many practices, long-term stud-
ies of removal efficiencies have yet to be conducted. There-
fore caution must be observed when making these simple
comparisons. Even though the MCTT capital and mainte-
nance costs were higher, this was offset by high efficiency
of the unit. Forthe two BMPs studied, the costs per pound
of TSS were of the same magnitude.
Summary
This study evaluates the water quality benefits of two
retrofit BMPs, the Stormceptor and the Multi-Chambered
Treatment Train (MCTT). Both units were placed in public
works maintenance yards where automated sampling
equipment collected event-mean concentration data. The
Stormceptor treated 91% of the total storm volume from
45 storms. Tank reduction efficiencies for the three con-
stituents measured in all 45 storms (TSS, TDS and TP)
were 26%, -19%, and 20% respectively. The extremely high
salt concentrations found in the runoff water at this site
may have compromised the unit's treatment ability. Good
agreement was found between calculated TSS load going
into the tank (from water quality samples) and what was
actually found in the bottom of the tank, although there is
some indication that larger-size particles may have been
missed in the water quality sampling.
The MCTT was designed fora greater level of treatment
and the findings here confirm it. Load reductions of TSS
and TP were 98% and 88% respectively. Even dissolved
constituents such as dissolved phosphorus and zinc had
high removal efficiencies (78% and 68%, respectively).
Somewhat surprising was the little change in particle size
distribution between influent and effluent, which may of
been an artifact of filter media escaping the filter fabric.
Similar costs-per-pound of TSS removed were estimated
for the two BMPs studied.
Table 2. Cost-per-pound of Suspended Solids to Treat One Acre of Parking Lot for Different Management Options.
Practice Efficiency Total Cost Cost/acre/yr1 % Control
1 Capital costs amortized over 50 years plus annual maintenance cost.
2 Based on construction cost at the Minocqua site.
Cost per pound removed
Stormceptor®
Overall
MCTT2
Overall
Tank
Tank
$60,000
$95,000
$479
$1185
26
22
98
78
$1.83
$2.18
$1.21
$1.52
282
-------
References Desbordes, eds. Springer Verlag, Berlin, New York,
Bannerman, R. T., D.W. Owens, R.B. Dodds, and N.J. pp' " '
Hornewer. 1993. Sources of Pollutants in Wisconsin Rjtt, R., B.Robertson, P. Barren, A. Ayyoubi, and S. Clark.
Stormwater. Wat. Sci. Tech. Vol.28, No.3-5, pp. 241- 1997. Stormwater Treatment at Critical Areas. Vol. 1:
259. The Multi-Chambered Treatment Train (MCTT). Re-
port to EPA, Office of Research and Development,
Ellis, J.B. 1986 Pollutional aspects of urban runoff, in Ur- National Risk Management Research Laboratory. Co-
ban Runoff Pollution, H. C. Torno, J. Marsalek, and M. operative Agreement No. CR 819573.
283
-------
Assessing the Effectiveness of Orlando's BMP Strategies
William G. Chamberlin,
City of Orlando
Orlando, Florida
Orlando calls itself "The City Beautiful" for many good
reasons. It is the number one tourist destination in the na-
tion. It has affordable housing, high employment, and an
abundance of cultural and civic amenities. Its natural beauty
has been named in recent polls as one of the greatest
reasons for choosing Orlando as a place of residence and
business. One of the principal elements of its natural beauty
is its lakes. Orlando has 86 lakes, either wholly or partially
located within its borders.
Prior to the mid-1970s there were few regulations which
protected these lakes. For decades, storm sewers were
installed to convey untreated stormwater directly into lakes.
Drainage wells were dug for flood protection. These wells
linked untreated runoff with the surface underground aqui-
fer, a source of drinking water for many. It was not until the
1970s that it became apparent that stormwater runoff en-
tering the lakes was causing problems to the natural sys-
tems.
The lakes, around which some of the most beautiful and
oldest homes in the city were built, were showing stresses
from decades of stormwater pollution. Symptoms as small
as an overabundance of aquatic weeds and as large as
massive fish kills indicated that the lakes were reaching
their limit of assimilating the influx of contaminants found
in stormwater. To address such problems, the city has taken
measures to control stormwater pollution and to provide
lake enhancements, drainage well protection, and shore-
line revegetation.
The Orlando City Code states, in part that, "... the pur-
pose of the Stormwater Utility Bureau and in essence, the
stormwater management program of the city is to 1) im-
prove the public's health, safety and welfare by providing
for the safe and efficient capture and conveyance of
stormwater runoff; 2) authorize the establishment and
implementation of a master plan for storm drainage includ-
ing ... management, operation,... inspection and enforce-
ment; and 3) encourage and facilitate urban water re-
sources management techniques, including ... enhance-
ment of the environment."Orlando's Public Works Depart-
ment, with its Bureaus of Stormwater Utility, Engineering,
Streets, and Drainage, and Project and Construction Man-
agement, work to see that these stormwater management
goals and issues are successfully addressed.
The Public Works Department's Stormwater Priority
Projects List and the Stormwater elements of the city's
Growth Management Plan and the Capital Improvement
Plan reflect projects which address qualitative (environ-
mental) issues as well as quantitative (flooding) issues.
BMPs are the methods by which qualitative issues are
being addressed. A BMP, or a best management practice,
is generally defined as the most effective, practical means
of preventing or reducing the amount of pollution gener-
ated by non-point sources, such as stormwater runoff, to a
level compatible with water quality goals. A BMP should
take into account problem assessment, outcome alterna-
tives, public input, and technological, economic, and insti-
tutional considerations. The city incorporates BMPs rou-
tinely. Some are very basic, while others are more com-
plex. Examples of BMPs utilized by the city include:
Street Sweeping- Street sweeping can be a useful BMP
incorporated by a city. Street sweeping removes pollut-
ants, sediments, leaves, and debris from the street sur-
face before they can be flushed into a receiving water body.
Orlando sweeps over 40,000 curb miles per year. There
are 72 residential routes, seven industrial/commercial
routes, and five downtown/near downtown routes. Over
27,000 cubic yards of material is removed annually by the
street sweeping program. This material is taken to a land-
fill fordisposal instead of being washed off of the streets in
stormwater and deposited into the city's lakes.
Retention/Detention Ponds - In 1984, the City of Or-
lando implemented rules contained in the Orlando Urban
Stormwater Management Manual and became one of the
first Florida municipalities to require both retention and
detention facilities for new development. Underthese rules,
developers must physically separate the retention volume
from the detention volume in what are referred to as "two
pond" systems. All development must include pollution
abatement capabilities to treat either 1/2 in. of runoff or
284
-------
runoff generated by a one-inch rainfall, whichever is great-
est. Treatment is defined as 1) retention with no discharge
and/or 2) detention with an approved filter discharge. A
detention facility for a developed site must be capable of
controlling the runoff volume expected from a 25-year fre-
quency/six-hourduration storm. Detention volume may be
discharged, but only at a rate not to exceed the peak rate
of discharge of the site prior to development.
Lake Revegetation - Twenty-two miles of shoreline in
45 of the city's lakes have been planted with native aquatic
plants. Revegetation returns the lakes to a more natural
state and better enables lakes to assimilate nutrient in-
puts. This also provides valuable wildlife habitat. Lakefront
property owners are also encouraged to plant native aquatic
plants. City staff believe that revegetation enhances lake
aesthetics and water quality. Plantings are required in wet
detention facilities. These ponds are shaped with a littoral
shelf to facilitate plantings.
Pollution Control Devices (PCDs) - Pollution Control
Devices are screening structures that are placed at
stormwater outfall pipes to prevent trash from entering the
lake. The City of Orlando has installed over 250 PCDs.
Cleaning debris from the PCDs after large storms is im-
portant and very labor intensive. If the PCDs are not
cleaned out, the buildup of leaves and debris can cause
the device to be "blown out" by heavy rainfall events,
thereby requiring maintenance and repair.
The most unique PCD the city has is a trash collection
device recently installed at one of its largest outfalls. At
Lake Rowena, a 108" outfall delivers the runoff from a 844
acre drainage basin. Debris that is flushed through this
system is so extensive that the normal screening processes
became ineffective in medium storm events. Therefore, the
city borrowed from the wastewater field and installed flow-
actuated moving bar screen (constructed 20 feet below
the ground) to intercept debris priorto discharge into Lake
Rowena. As flow, and thereby debris, begins to move
through the system, the moving screen begins to operate.
The screen catches debris, lifts it up, and deposits it into a
hopper. The hopper is periodically vacuumed using vactor
truck.
Removal of the trash and debris is only the first line of
defense in the struggle for pollution abatement. To combat
the other deleterious effects of stormwater, runoff has to
be treated. The City of Orlando has incorporated many
innovative BMPs to treat stromwater, most with great suc-
cess.
Notable examples of innovative BMPs used by the city
include alum treatment systems installed at four of the city's
lakes, a Vertical Volume Recovery System, a Packed Bed
Filter Project, an Urban Wetland Systems at Lake Green-
wood and the La Costa canal, and the Lake Wade Per-
iphyton Filter Water Garden.
Lake Dot Alum Treatment System - Lake Dot is a small
(6 acre) lake in downtown Orlando, with a drainage basin
of almost 300 acres. The lake had received untreated run-
off from this highly developed watershed for decades.
Largely as a result of the influx of pollutants during rain
events, the lake had experienced periodic fish kills due to
low dissolved oxygen. In 1988, construction of the Orlando
Arena (home of the NBAs Orlando Magic) within the Lake
Dot drainage basin was being planned. When considering
measures to meet the city's stormwater treatment require-
ments, planners first considered surface retention/deten-
tion ponds. Because the cost of acquiring land needed for
constructing the ponds was prohibitive, an alternative ap-
proach was selected in which alum (aluminum sulfate) is
added to stormwater to tie up nutrients and suspended
solids and precipitate them to the bottom of the lake. Since
95% of all runoff input enters Lake Dot through a single
storm sewer line, the same line that the Arena retention
ponds would have used as a discharge point, it was deter-
mined that the entire amount of stormwater conveyed
through this line could be treated with alum. This system
has been in operation since 1989, and with the exception
of a few operation difficulties, has operated satisfactorily.
Improvements to waterquality have been significant. Prior
to installation of the Alum Treatment System, the lake had
been eutrophic, with a Trophic State Index (TSI) averag-
ing over 60. Since the installation of the system, Lake Dot
has been classified as mesotrophic, with a TSI average of
just over 50.
The success at Lake Dot has encouraged the use of
alum treatment at other locations. Alum treatment systems
have been installed at Lake Lucerne in downtown Orlando,
Lake Holden (as part of a joint project with Orange County),
and Lake Rowena. An alum treatment system is currently
being planned for Clear Lake.
The Vertical Volume Recovery System - The Vertical
Volume Recovery System (VVRS) combines in-pipe stor-
age, a sump device for sediment removal and a sand filter
for fine sediment and dissolved pollutant removal. The
stormwater treatment system was developed as a part of
an inter-basin diversion plan needed to relieve flooding
from Lake Olive, one of our downtown lakes. The system
was designed to treat one inch of run off from the roadway
surfaces within the basin area. It was believed that the
sand filterwould be an effective method of stormwater treat-
ment where land is too expensive for standard retention/
detention. To date, the WR Ssystem has been a mainte-
nance burden. The treatment capabilities of the sand filter
have been disappointing.
Monitoring results showed some removal of dissolved
and particulate pollutants within the sump device. The sump
removed lead at rates between 50 and 78%. Zinc was re-
moved at rates between 40 and 56%. Cadmium was re-
moved at 50 to 60%. No significant removal of other heavy
metals was observed. Total Kjeldahl Nitrogen (TKN) was
removed at rates ranging from 14 to 56%. Total phospho-
rus was removed at rates from 29 to 51 %. Total suspended
solids were removed at rates between 14 to 71%.
Monitoring results also showed very poor removal of dis-
solved and suspended pollutant types by the filter. Removal
285
-------
was 3% for copper, 5% for nitrate-nitrite, 10% fortotal phos-
phorus and 12% for suspended solids. Export of pollut-
ants was observed at 10% for lead, 248% for ammonia,
2% forTKN and 19% for orthophosphate.
The Packed Bed Filter Project - A Packed Bed Filter
Project at Clear Lake has proven to be very successful.
The idea of using vegetated rock media filters to treat
stormwaterwas proposed in response to growing concern
over the water quality of Clear Lake. Clear Lake, once a
lake that was clear, had become eutrophic by the early
1980's, and was one of Orlando's poorest quality lakes.
The Clear Lake drainage basin consists of over three
square miles of highly developed urban area. Stormwater
runoff from this basin is conveyed into Lake Beardall and
Clear Lake. The Packed Bed Filter treats the initial storm
runoff from about 160 acres. Because the project demon-
strated an innovative technology, the city was successful
in obtaining a grant from Florida Department of Environ-
mental Protection (FDEP). The project uses a proven
wastewatertechnology to urban stormwater pollution. Sim-
ply put, the system consists of a packed bed filter, similar
to a trickling filter, and hydroponic aquatic plants.
Ten beds were established, including five crushed con-
crete and five using granite. Four different types of native
aquatic plants were planted in four of the concrete and
four of the granite beds. Two beds were set aside as con-
trols. Assessments were performed on the effectiveness
of the overall packed bed filter system, the performance of
the individual beds, and the best rate at which to operate
the system.
Monitoring results indicated medium to high effective-
ness in treating stormwater. Total phosphorous, total sus-
pended solids, and volatile suspended solids and cadmium
were removed by the overall packed bed system at a rate
exceeding 80%. Total nitrogen, TKN, nitrates, nitrites, chro-
mium, copper, lead, and zinc were removed at rates rang-
ing from 25 to 80%. Removals of ammonia, orthophos-
phate and total dissolved solids were not as effective, with
only 6, 14, and 8% removal, respectively.
The Greenwood Urban Wetlands - What was initially
conceived as a flood protection project, ended as a multi-
purpose, multi-use facility. Florida's first stormwater treat-
ment train was created to provide stormwater quality im-
provement, address the need for additional park area, and
provide flood protection fora 522-acre drainage subbasin.
The project required the expansion of Lake Greenwood to
oversixtimes its original size. It also involved planting over
thirteen acres of wetland and upland vegetation, and con-
structing footpaths, bridges, and stormwater control facili-
ties. In addition to these features, the new man-made wet-
land includes a pumped irrigation system which utilizes
stormwater runoff, instead of potable water to irrigate the
park and a neighboring city-owned cemetery. FDEP pro-
vided a grant to determine the effectiveness of the created
wetland stormwater treatment. The results were very good,
and would likely have been even better had it not been for
a high rate of groundwater inflows. This award winning
project os one of Orlando's most significant stormwater
management success stories.
This Greenwood project has proven to be a success in
all areas it had been intended. It provides flood protection,
stormwater treatment, a much-needed passive park, and
stormwater reuse through irrigation. The city has repeated
this concept, on a slightly smaller scale, with the construc-
tion of another urban wetlands, the La Costa Urban Wet-
lands.
Periphyton Filter Water Garden at Lake Wade - This
joint venture project with the St. John's River Water Man-
agement District will demonstrate the effectiveness of a
managed growth periphyton filtration system. The Periphy-
ton Water Garden is a stair-stepped concrete structure with
a surface that fosters the growth of algae. Water from Lake
Wade is pumped to the top of the structure and flows over
the algae and back to the lake. The algae consume nutri-
ents and traps sediment. Each week, the algae are har-
vested from the water garden and trucked away from the
site. The algae biomass can be used as an environmen-
tally friendly packaging material or be spread on park
grounds or ballfields as a fertilizer.
Periphyton, or attached algae, comprise the most pro-
ductive component of a wetland system. The Periphyton
Filter accelerates natural algae growth processes in a con-
trolled farmed system. A periphyton filter delivers a thin
sheet of untreated water over a diverse community (100-
200 species) of attached algae. Since techniques and de-
vices developed for harvesting the algae leave "roots" in-
tact, rapid regrowth occurs. Algal species diversity and fre-
quent harvesting make periphyton filters adaptable to
changing conditions, for consistently high productivity.
Periphyton Filters work on a wide range and various lev-
els of contaminants. Particularly interesting is their ability
to function at extremely low levels of pollutants, which lends
itself to lake restoration and treatment. The periphyton fil-
ter is a promising approach to reducing non-point source
pollution in Lake Wade.
This project went on-line in December, 1997. The Water
Management District and the City will be conducting re-
search and monitoring during 1998 to determine its effec-
tiveness.
Orlando is a leader in developing innovative approaches
to stormwater treatments for quality and quantity control.
The City of Orlando is very proud of its stormwater man-
agement practices.
Bill Chamberlin
City of Orlando
Phone: (407) 246-2180 or Fax: (407) 246-2152
E-mail: bchamber@ci.orlando.fl.us.
286
-------
Evaluating Public Information Programs:
Experiences with the Florida Yards and Neighborhoods Program
Billie Lofland
Florida Yards and Neighborhoods Program
University of Florida-Hillsborough County Cooperative Extension Service
Seffner, Florida
This report provides an overview of the evaluation meth-
ods applied to the Environmental Landscape Management
(ELM) and Florida Yards and Neighborhoods (FYN) pro-
grams of the University of Florida's Cooperative Extension
Service. Florida Yards and Neighborhoods is a public edu-
cation component of the major state-wide Environmental
Landscape Management program.
The Homeowner Research Questionnaire has been a
useful tool in measuring the level of adoption of environ-
mental landscape management practices by different pro-
gram groups. County-based surveys are used to show
overall program effectiveness and to compare different
program delivery methods. A1997 analysis by the Univer-
sity of Florida comparing three delivery methods used
throughout Florida, indicated that educational seminars are
an essential component in helping people adopt ELM prac-
tices.
Social marketing research conducted in 1997 studied
the perceptions that Tampa Bay homeowners have regard-
ing environmental landscape management practices. This
research included pre- and post-test evaluations of edu-
cational materials used by Florida Yards and Neighbor-
hoods.
In this report, a brief overview of the ELM and FYN pro-
grams will provide the foundation for discussing the evalu-
ation. This overview is followed by a description of two
ELM/FYN evaluation methods. This paper summarizes, in
part, two reports prepared in 1997: Adoption of Landscape
Management Practices by Florida Citizens,11 by Glenn D.
Israel, Janice O. Easton and Gary W. Knox; and Social
Marketing Study of the Perceptions Tampa Bay
Homeowners have Regarding Environmental Landscape
and Protection Practices,2 by Bonnie Salazar. Both are cited
at the end of this paper. The concluding section will look at
lessons learned and future plans for measuring program
effectiveness.
Introduction
Evaluating community-wide environmental education
programs is essential to measuring the effectiveness of
outreach efforts. Determining program effectiveness not
only aids in refining educational tools, it is often essential
for obtaining or maintaining funding. However, measuring
changes in awareness, knowledge and behavior presents
a variety of challenges.
Those implementing community-based educational pro-
grams do not have the advantage of a captive audience
who must come to class and turn in assignments. In fact,
getting the target audience to respond to measurement
techniques is one of the biggest challenges of all, even if it
is a list of questions about theiryard design and care prac-
tices.
Overview of ELM and FYN
Environmental Landscape Management was developed
in the late 1980s to educate professionals and homeowners
on how to create attractive, healthy landscapes by taking
an ecosystem approach to landscape design and mainte-
nance. "Right plant, right spot" is the key tenet. Educa-
tional materials emphasize adopting proper cultural prac-
tices to reduce landscape problems and negative impacts
on the environment.
Florida Yards and Neighborhoods is a component of the
ELM major state program. Developed in partnership with
the Sarasota and Tampa Bay Estuary Programs, FYN in-
creased the emphasis on the problems of stormwater run-
off pollution and the need to protect estuary systems. Since
its implementation in 1992, Florida Yards and Neighbor-
hoods has expanded its scope to include freshwater sys-
tems.
A variety of educational tools have been developed for
the Florida Yards and Neighborhoods Program, many of
them initially funded by the Sarasota Bay and Tampa Bay
National Estuary Programs. As the program continues to
develop, the need for additional tools becomes apparent,
as does the modification of some of the existing ones. A
continuing goal for those involved in the FYN Program is
obtaining the funds to purchase existing materials and
develop new ones.
287
-------
A Guide to Environmentally-friendly Landscaping: The
Florida Yards and Neighborhoods Handbook* (Florida
Cooperative Extension Service, 1994) is in its third print-
ing. The Florida Yardstick Workbook* (Florida Coopera-
tive Extension Service, 1998)4 supports the handbook by
providing a checklist of environmentally friendly landscape
design and care actions that are discussed in the book.
Two slide presentations demonstrate ELM/FYN practices
and discuss the benefits of adopting them: "Creating Your
Florida Yard" and "Maintaining Your Florida Yard." The FYN
message is presented in video form through the half-hour
Reclaiming Paradise: Florida Yards and Neighborhoods.
Various national estuary programs in Florida have pro-
vided funding to implement the FYN program in their coun-
ties. The purpose of this short-term funding was to help
establish a successful program that local government en-
tities would use to help meet their goals of water conser-
vation and reduced stormwater runoff and solid waste. Pro-
gram evaluation plays an important role the continuation
of an FYN Program in a county.
Program Evaluation Approaches
Program evaluation has always been an important com-
ponent of the Environmental Landscape Management Pro-
gram. A self-report survey using the Homeowner Research
Questionnaire is the primary state-wide evaluation tool. In
1997, a social marketing research project provided addi-
tional methods, on a regional basis, for measuring and in-
creasing program effectiveness.
The Homeowner Research Questionnaire
Survey
The Homeowner Research Questionnaire measures the
number of environmental landscape management prac-
tices the respondent is using. Participants complete the
pre-testquestionnaire priorto receiving an educational pro-
gram or educational resources. Ideally, the post-test ques-
tionnaire is completed six months later. Comparisons of
pre- and post-test responses measure changes in behav-
ior.
The pre-test questionnaire contains the following cat-
egories: General Information, Site Analysis, Stormwater
Runoff, Irrigation, Fertilizer, Pest Management, Wildlife,
Information Sources and Waterfront. With the post-test
questionnaire, the Information Source category is dropped
and questions are added under the new category "Views
on FYN Practices." Landscape practices include using
slow-release fertilizers, applying one pound of nitrogen or
less per 1,000 square feet, leaving grass clippings on the
lawn, tolerating some pest damage, watering according to
season and directing downspouts into the lawn or plant
beds.
Each questionnaire is coded according to the type of
program the participant received. Program types can in-
clude the following: neighborhood program, all-day work-
shop, one-hour presentation, exhibit, and Master Gardener
training. In some counties, people also receive publica-
tions by coming into the Extension Service or by calling
and requesting information. This category is referred to as
"Publications only."
Each county enters its own data using a standard tem-
plate. The data is sent by computer disk to be analyzed by
the Program Development and Evaluation Center of the
Institute of Food and Agriculture Services, University of
Florida. Counties use the survey data to evaluate different
delivery methods and in their annual reports to funding
sources. Because the program content and landscaping
practices taught among the counties are consistent, state-
wide comparisons of delivery systems can also be made.
In 1997, the University of Florida Cooperative Extension
Service conducted a state-wide comparisons of three ELM/
FYN program delivery methods: Master Gardener train-
ing, 1-6 hour seminars and publications only.1 A control
group was composed of non-participant residents who
completed the ELM Homeowner Research Questionnaire
in 1993 to establish a baseline. The data for the program
delivery groups was collected immediately priorto and ap-
proximately six months following educational programs
conducted during the 1997 Fiscal Year (July 1, 1996
through June 30, 1997.)
This study looked at average adoption rates in the dif-
ferent groups, as well as the adoption rate for specific land-
scape design and care practices. However, this paper will
discuss only the findings regarding overall adoption rates.
The Master Gardener group data was collected from 134
residents in nine counties. Three hundred and twenty resi-
dents in six counties contributed the data for the 1-6 semi-
nars group. In two counties, data were collected from 72
residents who received publications only. The Master Gar-
deners had the highest response rate, with 63% returning
their surveys, compared to a 42% response rate for semi-
nar participants and 57% for the non-equivalent compari-
son group. This study focused on 39 landscape care and
design practices included in the questionnaire.
A comparison of the pre- and post-test survey results
indicated an increase in the average number of practices
used by participants of all three program delivery meth-
ods, while non-participants remained essentially un-
changed (Figure 1).
The results of an analysis of variance show that while
the type of property, person maintaining the property, pres-
ence of a permanent irrigation system and parcel size did
not affect the adoption of practices, the type of program
did (F=31; p=0.001). A multiple comparison of adjusted
means for net adoption using Fisher's least significant dif-
ference indicated that each program type differed signifi-
cantly from the other groups in the number of practices
adopted (Table 1).
Social Marketing Research
Three grants funded social marketing research on the
Florida Yards and Neighborhoods/Environmental Land-
scape Management Programs in 1997. Funding came from
288
-------
26.7
19.9
Comparison Group Master Gardeners
1-6 hr. Seminars
Pubs only
Pre-program
6 Mo. Follow-up
Figure 1. Mean numbers of Best Management Practices used by program type.
Table 1. Average Number of Environmental Landscape Management PracticesAdopted by Type of Program
Type of Program
Mean Number of
Practices Used
Before the Program
Net Number of
PracticesAdopted
Adjusted Net
Number of Practices
Adopted1
Master Gardener
1-to-6 hr seminar
Publications only
Comparison group
19.4
17.9
17.1
13.3
7.3
4.5
2.8
0.1
6.9
4.3
2.6
0.0
1 Model of R2 of .135 and p=.001 .Adjusted means were generated by the analysis of variance for comparison using Fisher's least significance
differences.
the following: Florida Cooperative Extension Service;
NPDES (National Pollution Discharge Elimination System)
Education Subcommittee of the Florida Department of
Transportation, District 7; and West Coast Regional Water
Supply Authority (WCRWSA). In-kind services, including
co-moderating focus groups, were also supplied by the
Cooperative Extension Service. The research was con-
ducted by Chastain/Skillman, Inc.
Social marketing adapts commercial marketing technolo-
gies to programs designed to influence voluntary behavior
that improves the personal welfare of target audiences and
their society. The purpose of this research project was to
determine the perceptions Tampa Bay homeowners have
regarding environmental landscape and protection prac-
tices. There was the sense in entering this social market-
ing research project that there are different consumer
groups relative to yard care. It was hoped that the research
would help identify those groups by the benefits and draw-
backs they perceived in the ELM program. FYN could then
target those groups that would be the most receptive by
crafting programs to meet their needs.
The research included focus groups, participant-obser-
vation, surveys and the pre- and post-testing of educa-
tional materials. The Florida Cooperative Extension Ser-
vice, especially the Hillsborough County office, coordinated
the effort and provided in-kind services. This paper will
look at the pre- and post-testing of program materials.
The Florida Yardstick Poster was introduced into the
Florida Yards and Neighborhoods Program in 1994. This
colorful poster groups important yard care and design prac-
tices (or actions) into seven categories. Each Florida Yard
action is worth a given number of inches (or credits) and
once a homeowner has accrued 36 inches or more, his or
her yard can be certified as a Florida Yard. The poster is
38 inches tall by 24 inches wide.
The poster was useful in classroom and workshop set-
tings. It graphically showed the relationship between tak-
ing action and creating an environmentally friendly Florida
Yard. However, there was some question as to its useful-
ness as a tool for homeowners. There was a suggestion to
turn the poster into a workbook format—keeping the ac-
tions, but changing the format and adding practical "how-
to" information to augment the handbook's general prin-
ciples.
It was decided to conduct a post-test evaluation of the
workbook during the initial focus groups conducted as part
of the social marketing research. Toward the end of the
289
-------
focus group meeting, participants were asked to look at
the poster and respond to questions about its appearance,
usefulness, intended audience, etc. After the general dis-
cussion, participants were given sections of the poster to
evaluate. They wrote their comments and questions about
the section they received.
In general, participants said the design was friendly and
cheerful. It was noted that for a poster, the print should be
larger. Often, when asked who the intended audience was,
participants said they thought it was for school children.
Some said they would hang the poster in their garage as a
reference. A frequent suggestion regarding format was to
change it into a booklet format. Recommendations regard-
ing text included minimizing the use of technical terms.
Participants indicated that words such as "low-mainte-
nance," "irrigation," "least toxic," and "pervious" were hard
to understand.
The Florida Yardstick Workbook incorporated the col-
ors and graphics used in the poster because of the posi-
tive reaction people had to these elements. The book was
sized to fit into the back flap of the FYN Handbook. The
basic concept of the poster, with the categories and ac-
tions was maintained. A major change was in the use of
terminology. When possible, technical terms were replaced
with more common terms. For example, "water sprinkler"
replaced "irrigation system."
The pre-test evaluation of The Florida Yardstick Work-
bookwas conducted using the first draft of the workbook.
Because of time restrictions, the pre-test was not done in
focus groups. Instead, through one-on-one interviews,
people seeking to enter Master Gardener training were
asked about their perceptions regarding the workbook.
Again, the general response was that the overall appear-
ance is attractive and friendly. Participants found the infor-
mation useful and the writing easy to understand. Many
recommended that a larger print size be used. A common
complaint was that the pages looked cluttered.
In response to these observation during the pre-test
evaluation, a larger print size was used and more white
space was added. Because of cost constraints, informa-
tion was condensed or deleted. However, this process of
honing the content helped define critical elements.
Conclusions
Public information programs are viewed as an essential
component in helping to reduce stormwater runoff pollu-
tion, conserve water resources, protect wildlife habitats,
and engage in other environmental protection actions.
Evaluating environmental information programs is a chal-
lenging but essential process for determining the effec-
tiveness of various strategies, for developing new ap-
proaches and for justifying new, additional or continued
funding.
In 1997, the University of Florida Cooperative Extension
Service used two evaluation methods with its Environmen-
tal Landscape Management Major State-Wide Program,
which also encompasses Florida Yards and Neighbor-
hoods. The ELM/FYN Homeowner Research Survey
proved to be a useful tool for comparing the effectiveness
of three different program delivery methods. Taking a so-
cial marketing approach to evaluating educational publi-
cations was a productive method for making them easier
to read and comprehend.
Of course, it is important to evaluate the evaluation
mechanisms. The ELM team is currently in the process of
refining the ELM/FYN Homeowner Research Question-
naire. Currently, the questionnaire measures changes in
self-reported behavior. An important question is to what
extent does that behavior change result in water conserved,
reduced nutrient runoff from yards, a decrease in the use
of toxic pesticides, or other quantifiable results? There-
fore, one goal in revising the questionnaire is developing
queries that can result in valid, quantifiable answers.
Another reason to have an ongoing process of evaluat-
ing the evaluation techniques is that new research results
regarding stormwater runoff pollution or other environmen-
tal issues may lead to new goals for public action. For ex-
ample, early educational efforts in public information pro-
grams regarding stormwater runoff pollution focused on
having homeowners reduce their use of fertilizers and pes-
ticides and keep their driveways and street gutters clean.
However, with research showing the increasing importance
of atmospheric nitrogen deposition, we may want to focus
on additional behaviors, such as increasing public support
for stormwater retrofit projects, street cleaning, and mass
transportation.
References:
1 Israel, G. D., Easton, J. O., &Knox, G. W (1997). Adop-
tion of Landscape Management Practices by Florida
Citizens. Florida Cooperative Extension Service, In-
stitute of Food and Agricultural Service, University of
Florida.
2 Salazar, B. P. (1997). Social marketing study of the
perceptions Tampa Bay homeowners have regarding
environmental landscape and protection practices.
Report for the Tampa Bay Regional Planning Council,
Saint Petersburg, Florida.
3. Florida Cooperative Extension Service (1994). A Guide
to Environmentally Friendly Landscaping: The Florida
Yards and Neighborhoods Handbook. Institute of Food
and Agricultural Service, University of Florida.
4. Florida Cooperative Extension Service (1998). The
Florida Yardstick Workbook. Institute of Food and Ag-
ricultural Services, University of Florida.
290
-------
Examining the Need for Project Evaluation
Thomas E. Davenport
Water Division - Region 5
United States Environmental Protection Agency
Chicago, Illinois
Abstract
Retrofit project evaluation is as critical as project goals
and objectives. Formulation of a project evaluation should
begin when the project begins; it is an essential part of the
planning process. Without an evaluation system in place,
it is likely that you will waste precious time and funding.
Many assume that evaluations are expensive and require
extensive expertise. Sometimes this is true, but there are
easier and less expensive approaches that can be used, if
you accept the tradeoffs that come with a less complicated
evaluation method. Evaluations make it easier to make a
decision and justify your choices. This article presents the
role of evaluation, describes several types of evaluations
and provides examples.
Introduction
Too often project evaluations are scheduled at the end
of a project, and incompletely determine whetherthe goals
and objectives have been met. Evaluations should instead
be an ongoing part of any retrofit project or program (hence-
forth project means both project and program). The typical
project, designed to accomplish a specific task, is com-
monly evaluated by the completion of just that task. In ad-
dition to accounting for whether specific activities are be-
ing implemented, and resources expended, evaluations
need to be designed so as to determine the impacts over
the whole project. One should also evaluate the broadest
impacts caused by completion of the task. For example, a
recent outreach program to developers and homebuilders
focused on ways to achieve development and still protect
wetlands. The program was evaluated only by the number
of presentations, the amount and type of material distrib-
uted, and the requests for information. The program was
not evaluated against wetland loss prevented, so the
program's entire impact was underestimated.
Most completed nonpoint source project evaluations are
inadequate. Many projects do not even include an evalu-
ation. In a study on river conservation enhancement,
Holmes (1991) found that only five of almost 100 projects
had post-restoration evaluations. The most common ex-
cuses given for not carrying out evaluations were time,
cost, "evaluations not worth it" and "evaluations are diffi-
cult."
Most completed evaluations also fail to determine
whether the project had the long-term impacts it was de-
signed for. For example, most monitoring associated with
nonpoint source control projects focuses on the number of
Best Management Practices (BMPs) implemented, and
stops when implementation funds have been expended.
The Rural Clean Water Program (RCWP) showed the im-
portance of lag time between when BMPs are implemented,
and when water quality changes (USEPA, 1993). Linking
project evaluation to BMP implementation funding creates
an inherent bias toward failure to demonstrate water qual-
ity improvements, because it may take years between BMP
implementation and the realization of water quality improve-
ments.
In response to this lag problem, and the need to have a
national evaluation on the effectiveness of the Section 319
(Nonpoint Source) Program (EPA, 1991), the United States
Environmental Protection Agency (USEPA) created the
Section 319 National Monitoring Program. That program
established frameworks to more carefully document water
quality impacts associated with selected BMP implemen-
tation, especially over a longer period of time than usual.
The National Monitoring Program's objectives arel) to sci-
entifically evaluate the effectiveness of watershed tech-
nologies designed to control nonpoint source pollution, and
2) to improve our understanding of nonpoint source pollu-
tion (USEPA, 1991). For more information on the National
Nonpoint Source Monitoring Program see Osmond, et al.
(1997).
Why Evaluate?
While evaluation is both an art and science, it is critical
to the long-term success of any project. Project evalua-
tions allow managers to build on success and learn from
mistakes. Beech and Drake (1992) say it best; "Would you
invest your own money knowing that you would not get
any feedback on performance of the investment?" Without
an evaluation system for your project, you could be wast-
ing time and money and not know it. The Highland Silver
291
-------
Lake RCWP Project is an excellent example of why evalu-
ations are essential to the success of a project. Designed
to address lake sedimentation, the project's monitoring and
evaluation efforts showed that the lake was instead being
impacted by natric soils (causing high turbidity levels) and
not excessive sediment loading (Davenport & Kelly, 1986).
Evaluation systems can also play an important role in link-
ing your activities to the public and gaining their support.
Evaluation Types
Evaluations can be done at different stages during a
project, or at multiple stages. There are many evaluation
methods, techniques, and approaches designed for spe-
cific purposes and project phases (planning, implementa-
tion, and evaluation). A comprehensive evaluation will uti-
lize a combination of these methods to accurately evalu-
ate a project. The four major types of evaluations are For-
mative, Process, Outcome, and Impact. After discussing
these in the next paragraphs, we will continue with discus-
sion of more modest forms of evaluation.
Formative Evaluation (before the project starts) - A for-
mative evaluation is undertaken to test approaches, mate-
rials, and ideas. Additionally, formative evaluations are used
to understand the target audience before a project is imple-
mented. For projects involving communication, a key fac-
tor is determining the appropriate target audience, based
on project goals and objectives. A second sort of formative
evaluation is a needs assessment. The principal difference
between the traditional formative evaluation and needs as-
sessment involves the decisions that arise from the out-
comes. According to Herman and others (1987), needs
assessments result in better allocation of resources to meet
high priority needs. The formative evaluation is an attempt
to improve implementation prior to getting started. For ex-
ample, school materials about wetlands might be selected
by an advisory group of experienced teachers. Or, land
owners could be surveyed concerning barriers to adopting
conservation practices.
Process Evaluation (during)- While some consider this
part of a formative evaluation, for this paper it is separate
and focuses on tracking activities. Bean-counting is an-
other phrase for this type of evaluation. Using process
evaluation, managers can monitor implementation activi-
ties and provide timely information to their staff and project
sponsors. Process evaluations allow managers to change
project activities in response to ongoing feedback. Major
problems in using process evaluations for nonpoint source
projects include the inability of process evaluations to de-
termine the cause of the problem (before implementation
begins), and the lag time between implementation and re-
porting of accomplishments. Process evaluations can 1)
monitor the program activities by recording activities sys-
temically, 2) keep everyone focused on the big picture,
and 3) provide a database to allow project managers to
evaluate cost-effectiveness at every stage of the project.
Examples of process evaluation are tracking BMP imple-
mentations and requests for technical assistance.
Outcome Evaluation (afterwards) - This is also referred
to as a "summative" evaluation and measures the short-
term results associated with a project. The big question,
"What did I get for my money?" gets answered here. Out-
come evaluations can be used to 1) measure changes in
knowledge, attitudes, awareness, skills, aspirations
(KASAs) or behavior, 2) determine whether the project
worked within the desired time frame, and 3) determine
whether the project goes beyond the desired effects. An
example would be a workshop evaluation that focuses on
participant changes, orthe USEPA Region 1 review of their
Section 314 Clean Lakes Projects (Metcalfe & Eddy, 1992).
Impact Evaluation (much later)- Impact evaluation is the
most difficult type of evaluation to complete. It measures
long-term impacts. This type of evaluation requires dura-
bility of project goals and objectives. It is important to note
that the world keeps changing as your project progresses,
so evaluators must determine whether the project's goals
and objectives have remained constant. Longer-term re-
sults usually vary from short-term results, so impact evalu-
ation is needed to measure the ultimate value of the project.
This type of evaluation is extremely important for pilot and
demonstration projects. Many programs are developed
and implemented based on either "process" or "outcome"
evaluations of pilot or demonstration projects and lack real
impact evaluation.
EPA's Clean Lakes Program (Clean Water Act Section
314) stresses evaluation through all three phases: Diag-
nostic, Implementation, and Post Restoration. Phase 1
evaluations should be the least complex (formative and
process), while Phase 3s are more complex (impact evalu-
ations that provide individual outcomes and future direc-
tion). One Phase 3 Post-Restoration Monitoring study was
conducted five years after the successful completion of a
Phase 2 lake/watershed management program designed
to eliminate watershed sources of pollutants to Lake Le-
Aqua-Na and improve in-lake water quality. Phase 2 of the
project was judged a success after evaluation indicated
the project met its immediate goals and objectives. The
Phase 3 study, five years later, confirmed that the BMPs
were still in place, that the effectiveness of some of the
BMPs had decreased but were still achieving their goal,
and that the overall management program was still having
the desired effect on the lake (Davenport & Kaynor, 1998).
Clean Lakes Projects and Nonpoint Source Control
Projects are similar in that they rely on changes in human
behavior, but the basis of determining which changes are
needed is different for each. Proposed changes in Clean
Lakes Projects are based on stakeholder-involved diag-
nostic/feasibility studies. Nonpoint Source Project changes
are usually prescribed based on preconceived or docu-
mented problems without stakeholder input during the for-
mulation phase. This is an important difference which needs
to be considered in all outcomes or impact evaluations.
In Clean Lakes Projects, landowner/operator education
efforts start during Phase 1, their reactions and KASA lev-
els are designed into the Phase 2 implementation plan.
With Nonpoint Source Projects, landowner/operator reac-
tions and KASA levels are used to modify the implementa-
292
-------
tion plan after it is underway. When a Clean Lakes or
Nonpoint Source Project is promoting BMPs, evaluating
education being used is especially critical to accomplish-
ing long-term behavior changes. Too often NPS projects
judge success only by the number and types of BMPs that
landowners/operators are willing to install. Long-term main-
tenance and operation are seldom considered within
Nonpoint Source Project evaluations, but are a major fo-
cus of Phase 3 Clean Lakes Projects.
Evaluation Components
There are many different evaluation components avail-
able for Nonpoint Source Project evaluation (Table 1).
A brief summary (modified from Beech & Drake, 1992)
of each of the Hierarchy's components follows:
1) Inputs- Project resources that are used to carry out
the work. These resources include, at a minimum,
funds, paid staff, volunteers, office space, and sup-
plies.
2) Activities - Actions that are done, to implement the
work, such as planning.
3) Target Audience - Identifying who the target audi-
ence is for the project, and how they influenced
project design and implementation.
4) Reactions - Reactions of the target audience to the
project activities.
5) KASA Change - The project's activities must change
eitherthe knowledge, attitudes, skills, or aspirations
of of the target audiences (participants) to get imple-
mentation.
6) Changes in Behavior - Participant behavior changes
through the implementation process.
7) End Results- Results compared to the goals and ob-
jectives developed when the activities were planned.
A modification of Bennett's Hierarchy of Evidence for
Program Evaluation is very relevant to Nonpoint Source
Projects (Table 2).
Evidence of real impact becomes stronger as the evalu-
ation ascends the hierarchy. The two lowest levels (inputs
Table 1. Bennett's Hierarchy of Evidence for Program Evaluation
(Beech & Drake, 1992)
Table 2. Modified Bennett's Hierarchy
Level Component
Level
Component
7 (highest)
6
5
4
3
2
1 (lowest)
End Results
Behavior Change
KASA Change
Reactions
Target Audience
Activities
Inputs
7 (highest)
6
5
4
3
2
1 (lowest)
Behavior and Resulting Environmental Change
End Results (linked to funding period)
KASA Change
Reactions
Target Audience
Activities
Inputs
and activities) provide little or no measure of participant
benefit or environmental improvement. Level 3 provides
the first opportunity to get an indication of educational pos-
sibilities. If a focus of the evaluation is to increase project
performance, it is important to apply more evaluation tech-
niques at the lower levels (3 and under). Thorough as-
sessment of project effectiveness requires evaluations to
be conducted at the upper levels of the hierarchy. Evalua-
tions covering Levels 4 (Reactions) and 5 (KASA) provide
an indication of whether or not the activities are working.
KASA changes give an indication of potential BMP adop-
tion. External accountability increases as the level of evalu-
ation increases, while internal accountability follows the
reverse trend. Since the project is likely to change as it
develops, project evaluation techniques selected for higher
levels of the hierarchy must be flexible enough to respond
to change.
The level at which an evaluation is carried out (i.e., 1 to
7) has a tremendous impact on cost, requirements, and
useability. The higher up the hierarchy, the greaterthe prob-
ability that project results will be influenced by external fac-
tors which are more difficult to evaluate. The evaluations
that utilize the higher levels of the hierarchy usually in-
volve more expensive data collection and more time to
obtain results. Also, more expertise is required to design
evaluations, analyze data, and provide feedback. In addi-
tion, evaluation techniques require expertise to separate
external influences from actual project.
Reversing the order of Levels 6 and 7 is more appropri-
ate for Nonpoint Source Projects. The resulting evaluation
then focuses on surrogates for water quality, and complet-
ing the implementation plan. This modification allows evalu-
ators to determine if the BMPs are being maintained and
operated properly after the agencies have moved on to
other priority areas, and whether water quality has im-
proved. It also takes into account the "Hawthorne effect"
where participants respond favorably about implementa-
tion because of attention rather than commitment to change
(when attention has faded they revert to previous behav-
ior). The modification also attempts to link long-term be-
havior change with long-term changes in the environment
that are associated with the behavior in question. The
modified Hierarchy, shifts the focus of the evaluation to
long-term behavior changes after the implementation of
BMPs has been completed.
Table 3 relates the level of evaluation to the type of evalu-
ation. Based on cost and complexity, project managers
293
-------
Table 3. Modified Hierarchy Levels as They Relate to Evaluation Type
Level Component Formative Process Outcome Impact
7
6
5
4
3
2
1
Behavior Changes
End Results
KASA Change
Reactions
Target Audience
Activities
Inputs
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
must decide what type of evaluation is appropriate to docu-
ment their overall project or phase. While not all projects
need the same level of evaluation, the evaluation should
be planned prior to initiation, and not as an afterthought.
The level and type of evaluation utilized must be linked to
program needs. For example, if the program status is
needed, an evaluation that focuses on Levels 1 and 2 of
the hierarchy would be sufficient. If, however, program
managers are contemplating changing the direction of the
program or enhancing specific components, the evalua-
tion must be able to correlate data from Levels 1 and 2
evaluations to target audience (Level 3), reactions (Level
4), KASA (Level 5) and end results (Level 6) for the exist-
ing efforts, and what would be expected with future efforts.
Table 4 shows what level of the modified Hierarchy should
be addressed during each project phase. Using Table 4,
managers can design project evaluations that build from
one phase of the project to the next. Again, the Clean Lakes
Program is an excellent example of this process.
Based on earlier experiences in agricultural nonpoint
source program evaluations (Model Implementation Pro-
gram, Agricultural Conservation Program Special Water
Quality Projects) and the Clean Lakes Program experi-
ence, the third generation of USDAs nonpoint source pro-
gram (RCWP) built evaluation components into the pro-
cess. The premise of RCWP was that agricultural nonpoint
source pollution could be controlled at the farm scale to
cost-effectively improve off-site water quality. The Com-
prehensive Monitoring and Evaluation (CM&E) component
of the RCWP was established to document the farm level
effectiveness of the program, based on a sample of 20%
of the projects funded under RCWP. CM&E Projects were
selected for intensive monitoring and evaluation efforts to
represent the universe of agricultural nonpoint source prob-
lems being addressed. The evaluation framework for
RCWP consists of three tracks: 1) individual project moni-
Table 4. Hierarchy Levels as They Relate to Project Phase
Level Component Planning Implementation Evaluation
7
6
5
4
3
2
1
Behavior Changes
End Results
KASA Change
Reactions
Target Audience
Activities
Inputs
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
toring, 2) CM&E projects and 3) national surveys. The in-
dividual projects were primarily evaluated in terms of in-
puts and activities (Levels 1 & 2) outlined in the implemen-
tation plans. Attempts were made to link these process
evaluations with project outcomes. These evaluations
(tracking activities) were carried out by the project techni-
cal and administrative staff as part of their ongoing duties.
The process evaluations were inexpensive and very
straightforward. In some of the individual projects, process
evaluations were supplemented by Cooperative Extension
Service surveys of the target audience (Level 3) to esti-
mate changes in knowledge (Level 5) and document reac-
tions (Level 4) to the program. CM&E components sepa-
rately fund (usually more than $1,000,000 per project) staff
and work plans; they were integrated as part of the indi-
vidual projects so the project's technical and administra-
tive staffs were involved in the evaluation but not respon-
sible for it. CM&Es integrated all four types of evaluations
in the individual projects over a 10-year period. Formative
evaluations were used with the watershed land owners/
operators to help determine resource allocations between
project components and to establish a baseline (Levels 3
and 4) to document behavior patterns (Level 6) and knowl-
edge levels (Level 5). Process evaluations were used to
determine implementation status and direction (Levels 1-
4 and 7), and a CM&E water quality sampling program
was established to determine the overall impact on water
quality (Level 7). Nationally, the CM&Es and other project
evaluations were supplemented with statistically based
socioeconomic phone surveys of participants and nonpar-
ticipants in the project areas. The combination of these
three evaluations enabled national program managers and
experts 1) to develop a list of recommendations concern-
ing future agricultural nonpoint source control projects
(Farm Bureau, 1992), 2) provide guidance to states to help
them develop their own programs and 3) to determine the
impact of RCWP on watershed level water quality.
A review of the Section 319 Program (which is more re-
cent than RCWP) indicates that states and others have
utilized the lessons learned from RCWP to design the next
generation of agricultural Nonpoint Source Programs. The
RCWP experience shows that ownership of the evalua-
tion 1) increased the useability of the evaluation by the
project implementation team, 2) increased the volume and
improved the quality of material provided to the public and
target audience during project implementation, and 3) de-
creased the response time between a concern or issue
being raised and resolved with project-level data. The
CM&E component proved invaluable in the overall evalu-
ation of the RCWP Program.
Evaluation Barriers
In addition to some barriers already mentioned, there
are a number of factors that may limit the scope, direction,
and success of an evaluation. Leeds and others (1995)
developed a list of factors that may influence an evalua-
tion:
Organizational Politics
Organizational Structure
294
-------
Program Leadership
Professional Influences
Project/Program History
Economics
Social Patterns
Legal Guidelines
Each of these factors has implications for an evaluation.
Examine the purpose of a project evaluation in the context
of these factors. For example, is the evaluation going to
be used to "get" someone (organizational politics)? Know-
ing which of these factors might have an impact on your
evaluation will allow the evaluation team to modify data
collection approaches and methods to account for it.
Closing
1) Evaluation should begin when the project/program
begins and be part of the planning process.
2) Evaluations that are based on valid information are
more useful in decision-making than those based on
assumptions or opinions.
3) Evaluations that focus on average conditions and av-
erage operations may miss the importance of ex-
tremes.
4) Evaluation information that truly reflects the target
audience results in better decision-making.
5) Hard data is generally more comfortable as a basis
of making project decisions. However, hard data is
usually more expensive and time-consuming to ac-
quire.
6) Acceptance of evaluations is more likely when man-
agement is involved throughout the evaluation pro-
cess.
References
Beech, R. &A. Drake. 1992. Designing an Effective Com-
munication Program: A Blueprint for Success. National
Network for Environmental Management Studies Pro-
gram, United States Environmental Protection Agency,
Region 5, Chicago, IL.
Davenport, T. E. and S. Kaynor. 1998. Watershed Man-
agement Works: The Lake Le-Aqua-Na Project. Land
and Water, The Magazine of Natural Resource Man-
agement and Restoration, 42(2): 25-27.
Davenport, T. E. and M. Kelly. 1986. Water Resource Data
for the Highland Silver Lake Monitoring and
Evaaluation Project Madison County, Illinois Phase IV.
IEPA/WPC/86-001. Illinois Environmental Protection
Agency, Division of Water Pollution Control, Spring-
field, IL.
Farm Bureau. 1992. Project Design Checklist for Nonpoint
Source Water Quality Projects on A Watershed Basis.
Farm Bureau, Park Ridge, IL.
Herman, J.L.; L.L. Morris and C.T. Fitz-Gibbon. 1987.
Evaluator's Handbook. Center for Study of Evaluation
University of California, Los Angeles, CA.
Holmes, N. 1991. Post Project Appraisals of Conservation
Enhancement of Flood Defense Works. Research &
Development Report 285/1/A National Rivers Author-
ity, Reading, U.K.
Leeds, C.F., R. Leeds, L.C. Brown and C. Volgstadt. 1995.
Ohio Water Quality Projects Evaluation Workshop.
Ohio State University, Columbus, OH.
Metcalfe & Eddy. 1992. Clean Lakes program Evaluation.
Metcalfe & Eddy, Woburn, MA..
Osmond, D.L., D.E. Line, and J. Spooner. 1997. Section
319 National Monitoring Program: An Overview. NCSU
Water Quality Working Group, Biological and Agricul-
tural Engineering Department, North Carolina State
University, Raleigh, NC.
USEPA. 1991. Watershed Monitoring and Reporting for
Section 319 National Monitoring Program Projects.
USEPA Office of Water, Washington, D.C.
USEPA. 1993. Evaluation of the Experimental Rural Clean
Water Program. EPA-841-R-93-005. USEPA Office of
Wetlands, Oceans, and Watersheds, Washington, D.C.
295
-------
Attendees
S.AMAbbasi
Capital Projects Section
Environmental Resources Department
Prince George's County
Inglewood Center 3, 9400 Peppercorn PI.
Largo, MD 20774
telephone: 301/883-5851
fax: 301/883-7139
e-mail: epaaa@lgcfs003.co.pg.md.us
Mary Abrams
Columbia Slough Watershed Manage-
ment Systems Development Group
Bureau of Environmental Services
City of Portland
1120SW5thAve., Room 400
Portland, OR 97019
telephone: 503/823-7032
fax: 503/823-5344
e-mail:
marya@bessky.gate.bes. portland.onus
Jeff Adam
Department of Public Works - Architecture
and Engineering
Milwaukee County
907 N. 10th St., Annex Room 303
Milwaukee, Wl 53233
telephone: 414/278-4534
fax: 414/223-1366
Thomas R. Adams
Vortechnics, Inc.
41 Evergreen Dr.
Portland, ME 04103
telephone: 207/878-3662
fax: 207/878-8507
e-mail: vortechnics@vortechnics.com
Hans C. Albertsen
Alpha Environmental Research Group,
LLC
25770 Collins Ave.
Chestertown, MD 21620
telephone: 410/778-4858
fax: 410/778-4858 (let ring 6 times)
Jim Alwill
Bureau of Maintenance
Illinois Department of Transportation
201 W Center Ct.
Schaumburg, IL 60196
telephone: 847/705-4171
fax: 847/705-4608
Mary Ambrose
Water Policy and Regulations Division
Texas Natural Resource Conservation
Commission
PO Box 13087
Austin, TX 78711-3087
telephone: 512/239-4813
fax: 512/239-6195
e-mail: mambrose@tnrcc.state.tx.us
Fay Amerson
W270 53565 Oak Knoll
Waukesha, Wl 53188
telephone: 414/549-4577
fax: 414/549-9636
e-mail: famerson@execpc.com
David P. Anderson
Watershed Science Institute
USDA - Natural Resources Conservation
Service
5607 Dogwood Dr.
Lincoln, NE 68516
telephone: 402/437-5178, ext. 46
fax: 402/437-5712
e-mail: danderso@unlinfo.unl.edu
Eric Anderson
Environmental Safety Division, Fire
Department
City of Mountain View
1000 Villa St.
Mountain View, CA 94041
telephone: 650/903-6378
fax: 650/903-6122
e-mail: eric.anderson@ci.mtnview.ca.us
Ronald K. Baba
Planning Department
Oneida Nation
PO Box 365
Oneida, Wl 54155
telephone: 920/869-4527
fax: 920/869-1610
James A. Bachhuber
Water Resources Department
Rust E&l
1210 Fourier Dr.
Madison, Wl 53717
telephone: 608/828-8121
fax: 608/836-9767
e-ma;7:jim_bachhuber@ccmail.rustei.com
TaShara Cornelia Bailey
Executive Office, Public Works Services
City of Grand Rapids
300 Monroe NW
Grand Rapids, Ml 49503
telephone: 616/456-4541
fax: 616/456-4565
Roger T. Bannerman
Fisheries Management and Habitat
Protection Bureau
Wisconsin Department of Natural
Resources
P.O. Box7921
Madison, Wl 53707-7921
telephone: 608/266-9278
fax: 608/267-2800
e-mail: banner@dnr.state.wi.us
Jennifer Barone
County of Summit Engineer
538 E. South St.
Akron, OH 44311
telephone: 330/643-2850
fax: 330/762-7829
296
-------
Cheryl Bartley
Shiawassee District Office, Surface Water
Quality Division
Michigan Department of Environmental
Quality
PO Box 30273
Lansing, Ml 48909-7773
telephone: 517/625-4675
fax: 517/625-5000
e-mail: bartleyc@state.mi.us
Michael Bateman
Stormwater Management Program
Florida Department of Environmental
Protection
2600 Blair Stone Rd.
Tallahassee, FL 32399-2400
telephone: 850/921-5330
fax: 850/921-5217
e-mail: bateman_m@dep.state.fl.us
Julia Bell
Stewardships and Partnerships
National Park Service
U.S. Custom House, 3rd Floor
200 Chestnut St.
Philadelphia, PA 19106
telephone: 215/597-6473
fax: 215/597-0932
e-ma;7:julie_bell@nps.gov
Karen L. Bell
Water Division, Watersheds and Nonpoint
Source Programs Branch (WW-16J)
U.S. Environmental Protection Agency -
Region 5
77 W Jackson Blvd.
Chicago, IL 60604
telephone: 312/353-8640
fax: 312/353-8289
e-mail: bell.karen@epamail.epa.gov
Warren Bell, P.E.
Engineering Department
City of Alexandria
P.O. Box 178, City Hall
Alexandria, VA 22313
telephone: 703/838-4327
fax: 703/838-6438
Vince Berg
Storm ceptor® Corporation
600 Jefferson Plaza, Suite 304
Rockville, MD 20852
telephone: 301/762-8361
fax: 301/762-4190
e-mail: bergvh@erols.com
Melissa Beristain
Operations and Engineering Section;
Bureau of Water Supply, Quality and
Protection
New York City Department of Environ-
mental Protection
465 Columbus Ave.
Valhalla, NY 10595
telephone: 914/773-4447
fax: 914/773-0343
e-mail: mberista@valgis.dep.nyc.nyus
David C. Bier
Futurity, Inc.
5009 N. Hermitage Ave
Chicago, IL 60640
telephone: 773/506-2007
fax: 773/506-8921
e-mail: dcbier@interaccess.com
Dan Bishop
Montgomery Watson
1340 Treat Blvd., Suite 300
Walnut Creek, CA 94596
telephone: 510/274-2238
fax: 510/945-1760
e-mail: dan.bishop@us.mw.com
Ellen Blake
Warren County Soil and Water Conserva-
tion District
777 Columbus Ave.
Lebanon, OH 45036
telephone: 513/933-1336
fax: 513/933-2923
James K. Bland
Integrated Lakes Management
83Ambrogio Dr.
Gurnee, IL 60046
telephone: 847/244-6662
fax: 847/244-0261
Janis A. Bobrin
Drain Commissioner
Washtenaw County
PO Box 8645
Ann Arbor, Ml 48107-8645
telephone: 734/994-2525
fax: 734/994-2459
e-mail: jbobrin@BizServe.com
Thomas H. Boekeloo
Division of Water
New York State Department of Environ-
mental Conservation
50 Wolf Rd.
Albany, NY 12233-3508
telephone: 518/457-9874
fax: 518/485-7786
e-ma;7: thboekel@gw.dec. state, ny us
J. Steven Borroum
Environmental Program
California Department of Transportation
(CALTRANS)
1120 N Street, PO Box 942874 (MS-27)
Sacramento, CA 94274-0001
telephone: 916/653-7396
fax: 916/653-6366
e-mail: jborroum@trmx3.dot.ca.gov
Robert C. Brown
Water Quality Management Division
Manatee County Environmental Manage-
ment Department
P.O. BoxlOOO
Bradenton, FL 34206
telephone: 941/742-5980
fax: 941/742-5996
e-mail: robert.brown@co.manatee.fl. us
Timothy H. Brown
Clean Sites
53 W. Jackson Blvd., Suite 1604
Chicago, IL 60604
telephone: 312/554-0900
fax: 312/554-0193
e-mail: timbrown@igc.org
Whitney E. Brown
Center for Watershed Protection
8391 Main St.
Ellicott City, MD 21043
telephone: 410/461-8323
fax: 410/461-8324
e-mail: mrrunoff@pipeline.com
Dale S. Bryson
Camp Dresser & McKee, Inc.
1425 Briergate Dr.
Naperville, IL 60563
telephone: 630/305-7933
fax: 630/305-8054
e-mail: cleanwater@sprynet.com
"Pete" Ronald R. Cameron
Stormwater Management
FA. Johnson, Inc. / J.K.C Industries, Inc.
4700 Powerline Rd.
Fort Lauderdale, FL 33309
telephone: 954/776-5931
fax: 954/776-5955
e-mail: user605277@aol.com
Mike Campbell
Planning Division
San Jose Department of Planning,
Building and Code Enforcement
801 N. First St., Room 400
San Jose, CA 95110
telephone: 408/277-4576
fax: 408/277-3250
e-mail: mike.campbell@ci.sj.ca.us
297
-------
Bruce K. Carlisle
Coastal Zone Management
Massachusetts Bays Program
100 Cambridge St., Floor 20
Boston, MA 02202
telephone: 617/727-9530
fax: 617/727-5408
e-mail: bruce.carlisle@state.ma.us
Russell Carlson
Land Use
Grand Portage Reservation Tribal Council
PO Box 428
Grand Portage, MN 55605
telephone: 218/475-2442
fax: 218/475-2284
e-mail: gpland@boreal.org
Thomas R. Carpenter
Water Resources Division
Science Applications International
Corporation
1710 Goodridge Dr.
McLean, VA 22102
telephone: 703/748-4297
fax: 703/903-1374
e-mail:
thomas.r.carpenter@cpmx.saic.com
Francesca M. Cava
AbTech Industries
854 Jimeno Rd.
Santa Barbara, CA 93103
telephone: 805/962-8115
fax: 805/966-6312
e-mail: jrobsb@aol.com
William G. Chamberlin, II
Stormwater Utility Bureau
City of Orlando
400 S. Orange Ave.
Orlando, FL 32801
telephone: 407/246-2180
fax: 407/246-2512
e-mail: bchamber@ci.orlando.fl.us
Philip Christopher Champagne
Water Resources Engineering
Dewberry & Davis, Inc.
8401 Arlington Blvd.
Fairfax, VA 22031
telephone: 703/849-0581
fax: 703/849-0103
e-mail: cchampagne@dewberry.com
Karen Chandler
Barr Engineering Co.
8300 Norman Center Dr.
Minneapolis, MN 55437
telephone: 612/832-2600
fax: 612/832-2601
e-mail: kchandler@barr.com
Rudolph A. Chen
Department of Public Works
Milwaukee County
Courthouse Annex, Room 314
907 N. 10th St.
Milwaukee, Wl 53233
telephone: 414/278-4355
fax: 414/223-1853
Janice Cheng
Water Division, Watersheds and Nonpoint
Source Programs Branch (WW-16J)
U.S. Environmental Protection Agency -
Region 5
77 W. Jackson Blvd.
Chicago, IL 60604
telephone: 312/353-6424
fax: 312/886-7804
e-mail: cheng.janice@epamail.epa.gov
John Church
Cooperative Extension Service
University of Illinois
431 S. Phelps, Suite 605
Rockford, IL 61108
telephone: 815/397-7714
fax: 815/397-8620
e-mail: churchj@idea.ag.uiuc.edu
Elizabeth J. Cisar
Great Lakes Office
The Conservation Fund
53 W. Jackson Blvd., Suite 1332
Chicago, IL 60604
telephone: 312/913-9305
fax: 312/913-9523
Richard A. Claytor, Jr. P.E.
Center for Watershed Protection
8391 Main St.
Ellicott City, MD 21043
telephone: 410/461-8323
fax: 410/461-8324
e-mail: mrrunoff@pipeline.com
Alan Cochin
Water Planning Section
Illinois Environmental Protection Agency
1701 S. First Ave., Suite 600
Maywood, IL 60153
telephone: 708/338-7900
fax: 708/338-7930
Larry S. Coffman
Department of Environmental Resources
Prince George's County
Inglewood Center 3, 9400 Peppercorn PI.
Largo, MD 20774
telephone: 301/883-5839
fax: 301/883-9218
e-mail: coffman@ipo.net
Dan Cohen
Department of Public Works, Division of
Wastewater Services
City of Atlanta
55 Trinity Ave., Suite 5800
Atlanta, GA 30335-3029
telephone: 404/330-6899
fax: 404/658-7491
Preston D. Cole
Forestry Division
City of Milwaukee
841 N. Broadway, Room 501
Milwaukee, Wl 53202
telephone: 414/286-3671
fax: 414/286-8097
e-mail: pcole@mpw.net
Javier E. Cruz
USDA - Natural Resources Conservation
Service
16 Professional Park Rd.
Storrs, CT 06268
telephone: 860/487-4034
fax: 860/487-4054
Ellen C. Dailey
Tri-County Regional Planning Commis-
sion
100 N. Main St., Suite 301
EastPeoria, IL 61611-2533
telephone: 309/694-9330
fax: 309/694-9390
e-mail: tcrpc@umtec.com
Michael R. Dailey
Engineering Division, Department of
Public Works
City of Madison
210 Martin Luther King, Jr. Blvd.
City-County Bldg., Rm. 115
Madison, Wl 53710
telephone: 608/266-4058
fax: 608/264-9275
e-mail: mdailey@ci.madison.wi.us
Steve Daut
Midwest Environmental Consultants
6075 Jackson Rd.
Ann Arbor, Ml 48103
telephone: 734/747-9861
fax: 734/747-9865
e-mail: mecofmich@aol.com
Thomas E. Davenport
Water Division, Watersheds and NPS
Branch (WW-16J)
U.S. Environmental Protection Agency -
Region 5
77 W. Jackson Blvd.
Chicago, IL 60604
telephone: 312/886-0209
fax: 312/886-7804
e-mail:
davenport.thomas@epamail.epa.gov
298
-------
Yulya Davidova
Environmental Program
California Department of Transportation
2829 Juan St.
San Diego, CA 92110
telephone: 619/688-0226
fax: 619/688-6655
e-mail: ydavidov@trmx3.dot.ca.gov
David M. Day
Watershed Management Section
Illinois Department of Natural Resources
600 N. Grand Ave. West
Springfield, IL 62701-1787
telephone: 217/785-5907
fax: 217/785-8262
e-mail: dday@dnrmail.state.il.us
Cindy DelPapa
Riverways Program
Massachusetts Department of Fish,
Wildlife and Environment
100 Cambridge St., Room 1901
Boston, MA 02202
te/epftone:617/727-1614, ext. 359
fax: 617/727-2566
e-ma;7:cindy.delpapa@state.ma.us
Thomas A. Deming
Creative Environmental Solutions, Inc.
351 W. Camden St., Suite 100
Baltimore, MD 21201
telephone: 410/625-0300
fax: 410/625-2323
Paul A. DeVito
Oregon Department of Environmental
Quality
2146 NE 4th St., #104
Bend, OR 97701
telephone: 541/388-6146
fax: 541/388-8283
e-mail: devito.paul@deq.state.or.us
John T. Doerfer
Master Planning
Urban Drainage and Flood Control District
2480 W. 26th Ave., Suite 156-B
Denver, CO 80211
telephone: 303/455-6277
fax: 303/455-7880
e-mail: jdoerfer@udfcd.org
Amy Doll
Apogee Research/Hagler Bailly, Inc.
4350 East-West Highway, Suite 600
Bethesda, MD 20814
telephone: 301/657-7504
fax: 301/654-9355
e-mail: doll@apogee-us.com
Dan Donaldson
Lake County Soil and Water Conservation
District
125E. Erie St.
Painesville, OH 44077
telephone: 440/350-2730
fax: 440/350-2601
e-mail: dld@soil.co.lake.oh.us
Leslie Dorworth
Illinois-Indiana Sea Grant/Biological
Sciences
Purdue University Calumet
2200 169th St.
Hammond, IN 46323
telephone: 219/989-2726
fax: 219/989-2130
e-mail: dorworth@calumet.purdue.edu
Dennis W. Dreher
Natural Resources Department
Northeastern Illinois Planning Commis-
sion
222 S. Riverside Plaza, Suite 1800
Chicago, IL 60606
telephone: 312/454-0400
fax: 312/454-0411
e-mail: dreher@nipc.org
Kyle Dreyfuss-Wells
Chagrin River Watershed Partners, Inc.
2705 River Rd.
Willoughby Hills, OH 44094
telephone: 440/975-3870
fax: 440/975-3865
e-mail: drywell@en.com
David Drury
Hydrology and Geology Services Unit
Santa Clara Valley Water District
5750 Almaden Expy.
San Jose, CA 95118-3686
telephone: 408/927-0710
fax: 408/997-9247
Craig W. Dye
Environmental Section, Resource
Projects Department
Southwest Florida Water Management
District
2379 Broad St.
Brooksville, FL 34609
telephone: 352/796-7211
fax: 352/754-6885
J. Marshall Eames
Equipoise, Inc.
2119W. Morse Ave.
Chicago, IL 60645
telephone: 773/761-2431
fax: 773/761-2271
e-mail: equipoz@aol.com
Joseph D. Eigel
Gresham Smith and Partners
410 W. Chestnut St., Suite 300
Louisville, KY 40202
telephone: 502/627-8916
fax: 502/627-8989
e-mail: jeige@gspnet.com
Bill Eisenhauer
Safely Treating Our Pollution (S.T.O.P)
354 9th St., NE
Atlanta, GA 30309
telephone: 404/873-6417
fax: 404/873-6417
Julie Potempa Elliott
Stormwater Division
Louisville and Jefferson County Metropoli-
tan Sewer District
700 W. Liberty St.
Louisville, KY 40203
telephone: 502/540-6112
fax: 502/540-6365
e-mail: elliott@msdlouky.org
Sue Elston
Water Division, Watersheds and Nonpoint
Source Programs Branch (WW-16J)
U.S. Environmental Protection Agency -
Region 5
77 W. Jackson Blvd.
Chicago, IL 60604
telephone: 312/886-6115
fax: 312/886-7804
e-mail: elston.sue@epamail.epa.gov
William R. English
Department of Aquaculture, Fisheries and
Wildlife
Clemson University
G08 Lehotsky Hall
Clemson, SC 29634
telephone: 864/656-2811
fax: 864/656-0678
e-mail: renglsh@clemson.edu
Laura Evans
Superfund Division (SR-6J)
U.S. Environmental Protection Agency
77 W. Jackson Blvd.
Chicago, IL 60604
telephone: 312/886-0851
fax: 312/886-4071
Pamela Failing
USDA - Natural Resources Conservation
Service
16 Professional Park Rd.
Storrs, CT 06268
telephone: 860/487-4026
fax: 860/487-4054
e-mail: pfailing@ct.nrcs.usda.gov
299
-------
Martin Farber
Department of Utilities
City of Sacramento
5770 Freeport Blvd., #100
Sacramento, CA 95822
telephone: 916/433-6318
fax: 916/433-6652
Michael S. Fazzino
Marketing Department
Ruppert Environmental
17701 New Hampshire Ave.
Ashton, MD 20861
telephone: 301/924-7869
fax: 301/774-6840
e-mail: eviron@erols.com
John B. Ferris
Water Resources
Rust E&l
1020 N. Broadway, Suite 400
Milwaukee, Wl 53012
telephone: 414/225-5100
fax: 414/225-5111
e-ma;7:john_ferris@ccmail.rustei.com
Carol Fialkowski
Office of Environmental and Conservation
Programs
The Field Museum
Roosevelt Road at Lake Shore Drive
Chicago, IL 60605
telephone: 312/922-9410, ext. 629
fax: 312/922-1683
e-mail: cfialkowski@fmnh.org
Stuart Finley
Lake Barcroft Watershed Improvement
District
3428 Mansfield Rd.
Falls Church, VA 22041
telephone: 703/820-7700
fax: 703/820-7701 (call first)
Michael Finn
Planning Department
Oneida Nation
PO Box 365
Oneida, Wl 54155
telephone: 920/869-4527
fax: 920/869-1610
Michael Finney
Planning Department
Oneida Nation
PO Box 365
Oneida, Wl 54155
telephone: 920/869-4527
fax: 920/869-1610
Brad Fossum
The Cretex Companies, Inc.
311 Lowell Ave.
Elk River, MN 55330
te/epftone:612/441-2121
fax: 612/441-7626
e-mail: bfossum@cretexinc.com
Gregory T. Fries
Engineering Division
City of Madison
210 Martin Luther King, Jr. Blvd.; Room
115
Madison, Wl 53710
telephone: 608/267-1199
fax: 608/264-9275
e-mail: gfries@ci.madison.wi.us
H. David Gabbard
Division of Engineering
Lexington-Fayette Urban County Govern-
ment
200 E. Main St., 8th Floor
Lexington, KY 40507
telephone: 606/258-3410
fax: 606/258-3458
e-mail: davidg@lfucg.com
John Galli
Department of Environmental Programs
Metropolitan Washington Council of
Governments
777 N. Capitol St., NE, Suite 300
Washington, DC 20002-4239
telephone: 202/962-3348
fax: 202/962-3203
e-mail: jgalli@mwcog.org
Catherine Garra
Watersheds and Nonpoint Source
Programs Branch (WW-16J)
U.S. Environmental Protection Agency -
Region 5
77 W. Jackson Blvd.
Chicago, IL 60604
telephone: 312/886-0241
fax: 312/886-7804
e-mail: garra.catherine@epamail.epa.gov
Tom Girman
TetraTech EM Inc.
330 S. Executive Dr., Suite 203
Brookfield, Wl 53005
telephone: 414/821-5894
fax: 414/821-5946
e-mail: girmant@ttemi.com
Gary Goay
Water Quality Management Division
Louisiana Department of Environmental
Quality
PO Box 82215
Baton Rouge, LA 70884-2215
telephone: 504/765-0511
fax: 504/765-0635
Willie Gonwa
Crispell-Snyder, Inc.
114 W. Court St.
Elkhorn.WI 53121
telephone: 414/723-5600
fax: 414/723-5106
Robert Goo
U.S. Environmental Protection Agency
401 M Street, SW(4503-F)
Washington, DC 20460
telephone: 202/260-7025
fax: 202/260-1977
e-mail: goo. robert@epamail.epa.gov
Timothy W. Good
Planning and Development Department,
Planning Division
Forest Preserve District of Will County
PO Box 1069
Joliet, IL 60434-1069
telephone: 815/727-8700
fax: 815/727-9415
e-mail: larchitect@fpdwc.org
Josh Goode
Independent Contractor - Department of
Public Works
City of Indianapolis
5506 N. Broadway
Indianapolis, IN 46220
telephone: 317/327-4794
fax: 317/327-4954
e-mail: jgood@indygov.org
Leila Gosselink
Environmental Resource Management
Division, Watershed Protection Depart-
ment
City of Austin
POBox1088-WP
Austin, TX 78767
telephone: 512/499-1869
fax: 512/499-2846
e-ma;7:gosselink_l@earth.ci.austin.tx.us
Janine Grauvogl-Graham
Camp Dresser & McKee, Inc.
312 E. Wisconsin Ave., Suite 500
Milwaukee, Wl 53202
te/epftone:414/291-5100
fax: 414/291-2765
e-mail: grahamjl@cdm.com
Steven R. Greb
Environmental Contamination Section
Wisconsin Department of Natural
Resources
1350 Femrite Dr.
Monona, Wl 53716
telephone: 608/221-6362
fax: 608/221-6353
e-mail: grebs@dnr.state.wi.us
300
-------
Randell K. Greer, P.E.
Division of Soil and Water Conservation,
Sediment and Stormwater Program
Delaware Department of Natural Re-
sources and Environmental Control
89 Kings Highway
Dover, DE 19901
telephone: 302/739-4411
fax: 302/739-6724
e-mail: rgreer@state.de.us
Jim Hafner
Minnehaha Creek Watershed District
2500 Shadywood Rd.
Excelsior, MN 55331
telephone: 612/471-0590
fax: 612/471-0682
e-mail: jhafner@minnehahacreek.org
Kendra J. Harmason
Water Quality Management Division
Louisiana Department of Environmental
Quality
PO Box 82215
Baton Rouge, LA 70884-2215
telephone: 504/765-0511
fax: 504/765-0635
e-mail: ke n d ra_h @d eq. state. I a. u s
Harvey H. Harper, P.E.
Environmental Research & Design, Inc.
3419 Trentwood Blvd., Suite 102
Orlando, FL 32812-4863
telephone: 407/855-9465
fax: 407/826-0419
e-mail: hharper@erd.org
Mark Hausner
Product Development
BaySaver, Inc.
1010 Deer Hollow Dr.
Mount Airy, MD 21771
telephone: 301/829-6119
fax: 301/829-3747
e-mail: mbhausner@erols.com
George T. Heard III
USDA - Natural Resources Conservation
Service
175 A Commercial Pkwy.
Madison, MS 39046
telephone: 601/859-4272
fax: 601/859-7091
Theresa B. Heiker
Division of Engineering Services/Public
Works
Leon County
301 S. Monroe St., Room 202
Tallahassee, FL 32301
telephone: 850/488-8003
fax: 850/488-1260
e-ma;7:theresah@mail.co.leon.fl.us
Mark Heinicke
Department of Planning and Development
Forest Preserve District of Cook County
536 N. Harlem Ave.
River Forest, IL 60305
telephone: 708/771-1355
fax: 708/771-1512
Linda Henning
Environmental Planning and Evaluation/
Customer Relations and Environmental
Education
Metropolitan Council/Environmental
Services
230 E. Fifth St.
St. Paul, MN 55101-1633
telephone: 612/602-1279
fax: 612/602-1003
e-mail: linda.henning@metc.state.mn.us
Timothy C. Henry
Water Division
U.S. Environmental Protection Agency -
Region 5
77 W. Jackson Blvd. (W-15J)
Chicago, IL 60604
telephone: 312/353-2147
fax: 312/886-0957
e-mail: henry.timothy@epamail.epa.gov
Steve Hides
H.I.L. Technology, Inc.
94 Hutchins Dr.
Portland, ME 04102
telephone: 207/756-6200
fax: 207/756-6212
Tina Hissong
Lake Maxinkuckee Environmental Council
116 N. Main St.
Culver, IN 46511
telephone: 219/842-3686
fax: 219/842-3704
e-mail: lmec@culcom.net
F. Eric Hjertberg
Environmental Technology Evaluation
Center
1015 15th St. NW, Suite 600
Washington, DC 20005-2605
telephone: 202/842-0555
fax: 202/682-0612
e-mail: ehjertberg@cerf.org
Jim Hodgson
Minnesota Pollution Control Agency
1601 Minnesota Dr.
Brainerd, MN 56401
telephone: 218/828-6065
fax: 218/828-2594
e-mail: James. hodgson@pca. state, mn. us
Robert Holm
Department of Public Works, Environmen-
tal Resources Management Division
City of Indianapolis
2700 S. BelmontAve.
Indianapolis, IN 46221
telephone: 317/327-2234
fax: 317/327-2274
Tom Holmes
Lorain County Soil and Water Conserva-
tion District
42110 Russia Rd.
Elyria, OH 44035
telephone: 216/322-1228
fax: 216/323-0405
Michael C. Houck
Coalition for a Liveable Future
Audubon Society of Portland
5151 NW Cornell Rd.
Portland, OR 97210
telephone: 503/292-6855, ext. 111
fax: 503/292-1021
e-mail: houckm@teleport.com
Jeff H rubes
Beltrami Soil & Water Conservation
District
3217 Bemidji Ave. North
Bemidji, MN 56601
telephone: 218/755-4339
fax: 218/755-4201
e-mail: jmh2@mn.nrcs.usda.gov
Marianne Hubert
Land and Water Quality Division
Maine Department of Environmental
Protection
17 State House Station
Augusta, ME 04989
telephone: 207/287-4140
Holly L. Hudson
Natural Resources Department
Northeastern Illinois Planning Commis-
sion
222 S. Riverside Plaza, Suite 1800
Chicago, IL 60606
telephone: 312/454-0401, ext. 302
fax: 312/454-0411
e-mail: hlhudson@nipc.org
Julia Huffman
Engineering Division
Unified Sewerage Agency
155 N. First Ave., Suite 270, MS 10
Hillsboro, OR 97124-3072
telephone: 503/648-8621
fax: 503/640-3525
301
-------
Joan Hug-Anderson
Summit Soil and Water Conservation
District
2787 Front St., Suite B
Cuyahoga Falls, OH 44221
telephone: 330/929-2871
fax: 330/929-2872
e-mail: summitswcd@aol.com
Nigel Ironside
Environment Division
Auckland Regional Council
BellSouth Centre, 21 Pitt St.
Private Bag 92 012
Auckland, New Zealand
telephone: 011 649 366 2000 ext. 7072
fax: 011 6493662155
e-mail: nironside@arc.govt.nz
Kathryn Jahn
U.S. Fish and Wildlife Service
3817LukerRd.
Cortland, NY 13045
telephone: 607/753-9334
fax: 607/753-9699
e-mail: kathrynjahn@fws.gov
Roger B. James
Water Resources Management
63 Ivy Dr.
Orinda, CA 94563-4228
telephone: 510/631-7950
fax: 510/631-9885
e-mail: roger.james@worldnet.att.net
Anthony J. Janicki
Watershed Studies
Post, Buckley, Schuh & Jernigan
9800 4th St., North, Suite 108
St. Petersburg, FL 33702
telephone: 813/577-6161
fax: 813/576-4313
e-mail: tjanicki@pbsj.com
Barry Johnson
Rouge Program Office
Camp Dresser & McKee, Inc.
220 Bagley, Suite 920
Detroit, Ml 48226
telephone: 313/964-8892
fax: 313/961-1762
Betsy Johnson
Storm Water Services
City of Greensboro
PO Box 3136
Greensboro, NC 27402
telephone: 336/373-2707
fax: 336/373-2988
e-mail: ames121@hotmail.com
Bev Johnson
Planning Department
City of Boulder
PO Box 791
Boulder, CO 80306
telephone: 303/441-3272
fax: 303/441-3241
e-ma;7:johnsonb@ci. boulder, co. us
D. Kent Johnson
Water Quality Section
Metropolitan Council/Environmental
Services
Mears Park Centre, 230 E. Fifth St.
St. Paul, MN 55101
telephone: 612/602-8117
fax: 612/602-8179
e-mail: kent.johnson@metc.state.mn.us
Keshia Johnson
Department of Public Works, Division of
Wastewater Services
City of Atlanta
55 Trinity Ave., Suite 5800
Atlanta, GA 30335-3029
telephone: 404/330-6980
fax: 404/658-7631
Mitch Johnson
Water Resources Group
Bonestroo, Rosene, Anderlik & Associ-
ates
2335 W Highway 36
St. Paul, MN 55113
telephone: 612/604-4806
fax: 612/636-1311
e-mail: mjohnson@bonestroo.com
Karen C. Kabbes
Kabbes Engineering, Inc.
115 W. Coolidge
Barrington, IL 60010
telephone: 847/842-9663
fax: 847/842-9960
e-mail: kckabbes@aol.com
Amy Spies Karhliker
Bureau of Design and Environment
Illinois Department of Transportation
2300 S. Dirksen Pkwy.
Springfield, IL 62764
telephone: 217/785-4614
fax: 217/524-9356
Susan Kenney
Biology Department
University of Illinois at Chicago
3917 W. 70th PI.
Chicago, IL 60629
telephone: 773/735-0281
Jay Kessen
Water Resources Department
Rust E&l
3121 Butterfield Rd.
Oak Brook, IL 60544
telephone: 630/574-2564
fax: 630/574-2007
e-ma;7:james_kessen@ccmail. rustei.com
Andrea S. Kevrick
InSite Design Studio, Inc.
150S. Fifth Ave.
Ann Arbor, Ml 48104
telephone: 313/995-4194
fax: 313/995-3408
Gordon L. King
Industrial Division
The Haskell Company
111 Riverside Ave.
Jacksonville, FL 32202
telephone: 904/791-4583
fax: 904/791-4697
Bruce Kirschner
International Joint Commission
PO Box 32869
Detroit, Ml 48232
telephone: 519/257-6710
fax: 519/257-6740
e-mail: kirschnerb@ijc.wincom.net
Lyn T. Kirschner
Conservation Technology Information
Center
1220 Potter Dr., Room 170
West Lafayette, IN 47906-1383
telephone: 317/494-9555
fax: 317/494-5969
e-mail: kirschner@ctic.purdue.edu
Robert J. Kirschner
Natural Resources Department
Northeastern Illinois Planning Commis-
sion
222 S. Riverside Plaza, Suite 1800
Chicago, IL 60606
telephone: 312/454-0401, ext. 303
fax: 312/454-0411
e-mail: bobkirs@nipc.org
Patricia V. Klein
Environmental Concerns Committee
City of Kalamazoo
2236 Crest Dr.
Kalamazoo, Ml 49008
telephone: 616/343-7965
e-mail: klein@wmich.edu
302
-------
Edward H. Kluitenberg, P.E.
Applied Science, Inc.
660 Plaza Dr., Suite 2000
Detroit, Ml 48226-1207
telephone: 313/963-8300
fax: 313/963-8306
e-mail: appliedscience@mindspring.com
Brandon Koltz
Triad Engineering, Inc.
325 E. Chicago St.
Milwaukee, Wl 53208
telephone: 414/291-8840
fax: 414/291-8841
John A. Kosco, P.E.
Municipal Technology Branch
Office of Wastewater Management (4204)
U.S. Environmental Protection Agency
401 M Street, SW
Washington, DC 20460
telephone: 202/260-6385
fax: 202/260-0116
e-mail: kosco.john@epamail.epa.gov
Michael Kowalski
Sales Department
CSR Masolite
PO Box 1708
Fort Wayne, IN 46801
telephone: 219/432-3568
fax: 219/436-2788
e-mail: mjkowalski@juno.com
Natalie Landry
Water Division
New Hampshire Department of Environ-
mental Services
64 N. Main St.
Concord, NH 03301
telephone: 603/271-5329
fax: 603/271-7894
e-mail: n_landry@des.state.nh.us
James Larsen
Environmental Planning and Evaluation
Department
Metropolitan Council/Environmental
Services
230 E. Fifth St.
St. Paul, MN 55101
telephone: 612/602-1159
fax: 612/602-1130
e-ma;7:jim.larsen@metc.state.mn.us
Jeffrey T. Lee
Environmental Operations
Minneapolis Park and Recreation Board
3800 Bryant Ave., South
Minneapolis, MN 55409-1029
telephone: 612/370-4900
fax: 612/370-4831
e-ma;7:jeffrey.t.lee@ci.minneapolis.mn.us
Stephen Lees
Upper Parramatta River Catchment Trust
Level 1, Macquarie Tower
10 Valentine Ave, Parramatta
PO Box 3720
Sydney, New South Wales 2124, Australia,
telephone: 011 612 9891 4633
fax: 011 61296892537
e-mail: sjlees@ozemail.com.au
Randel J. Lehmoine
Environmental Protection/Stormwater
Management
City of Grand Rapids
1101 Monroe Ave., NW
Grand Rapids, Ml 49503
telephone: 616/456-3253
fax: 616/456-3287
James H. Lenhart, P.E.
Engineering and Research
Stormwater Management, Inc.
2035 NE Columbia Blvd.
Portland, OR 97211
telephone: 503/240-3393
fax: 503/240-9553
e-mail: jiml@stormwatermgt.com
Paul M. Leonard
EDAW, Inc.
3475 Lenox Rd., Suite 100
Atlanta, GA 30326
telephone: 404/365-1110
fax: 404/365-1129
e-mail: leonardp@edaw.com
Mike Liebman
Public Works Division
Foth & Van Dyke
2737 S. Ridge Rd., PO Box 19012
Green Bay, Wl 54307-9012
telephone: 920/496-6765
fax: 920/497-8516
e-mail: mliebman@foth.com
Patrick E. Lindemann
Ingham County Drain Commissioner
707 Buhl, P.O. Box 220
Mason, Ml 48854-0220
telephone: 517/676-8395
fax: 517/676-8364
e-mail: patrickl2@aol.com
Greg Lindsey
Associate Director of the Center for Urban
Policy and Environment, and Associate
Professor in the School of Public and
Environmental Affairs
Indiana University
342 N. Senate Ave., B.S. 4068
Indianapolis, IN 46204-1744
telephone: 317/274-8704
fax: 317/274-7860
e-mail: glindsey@speanet.iupui.edu
Billie Lofland
Hillsborough County Cooperative
Extension Service
University of Florida
5339 S. County Rd. 579
Seffner, FL 33584-3334
telephone: 813/744-5519
fax: 813/744-5776
Kirsteen Macdonald
Waste Water Technology Centre
University of Abertay Dundee
Bell Street
Dundee, Scotland
United Kingdom DD1 1HG
telephone: +441382 308161
fax:+441382 308117
e-mail: k.macdonald@tay.ac.uk
Bill MacElroy
Landscape Architecture Department
University of Washington
348 Gould, Box 355
Seattle, WA 98195
telephone: 206/616-8698
e-mail: macelroy@u.washington.edu
A.E. Machak
Commissioner, Appointed by the Board of
the Illinois Association of Wastewater
Agencies
Northeastern Illinois Planning Commis-
sion
222 S. Riverside Plaza, Suite 1800
Chicago, IL 60606
telephone: 312/454-0400
fax: 312/454-0411
Peter A. Mangarella, P.E.
Woodward-Clyde Consultants
500 12th St., Suite 100
Oakland, CA 94607-4014
telephone: 510/874-3022
fax: 510/874-3268
e-mail: pamangaO@wcc.com
Terese Manning
South Florida Regional Planning Council
3440 Hollywood Blvd., Suite 140
Hollywood, FL 33021
telephone: 954/985-4416
fax: 954/985-4417
e-mail: terryman@sfrpc.com
John D. Mathews
Ohio Department of Natural Resources
1939 Fountain Square Court, Bldg. E-2
Columbus, OH 43224-1336
telephone: 614/265-6685
fax: 614/262-2064
e-ma;7:john.mathews@dnr.state.oh.us
303
-------
Wally Matsunaga
Water Planning Section
Illinois Environmental Protection Agency
1701 S. First Ave., Suite 600
Maywood, IL 60153
telephone: 708/338-7900
fax: 708/338-7930
John Maxted
Division of Water Resources, Watershed
Assessment Branch
Delaware Department of Natural Re-
sources and Environmental Control
29 S. State St.
Dover, DE 19903
telephone: 302/739-4590
fax: 302/739-6140
e-mail: jmaxted@state.de. us
Christopher W. May
Applied Physics Lab
University of Washington
1013 NE 40th St.
Seattle, WA 98105-6698
telephone: 206/543-1300
fax: 206/543-6785
e-mail: may@apl.washington.edu
West Me Adams
South Carolina Sea Grant Extension
Program
Clemson University
259 Meeting St.
Charleston, SC 29401
telephone: 803/722-5940
fax: 803/722-5944
e-mail: mmcdms@clemson.edu
Rick McAndless
North Cook County Soil and Water
Conservation District
PO Box 407
Streamwood, IL 60107
telephone: 847/468-0071
fax: 847/608-8302
Nicole McClain
USDA - Natural Resources Conservation
Service
928 S. Court St., Suite C
Crown Point, IN 46307-4848
telephone: 219/663-0238
fax: 219/663-2547
Robert B. McCleary, P.E.
Division of Highway Operations
Delaware Department of Transportation
P.O. Box 778
Dover, DE 19903
telephone: 302/739-4327
fax: 302/739-2128
e-mail: rmccleary@smtp.dot.state.de.us
Nancy L. McClintock
Environmental Resource Management
Division, Watershed Protection Department
City of Austin
POBox1088-WP
Austin, TX 78767
telephone: 512/499-2652
fax: 512/499-2846
e-mail:
mcclintock_n@earth.ci.austin.tx.us
David McElroy
Warren County Soil and Water Conserva-
tion District
777 Columbus Ave.
Lebanon, OH 45036
telephone: 513/933-1336
fax: 513/933-2923
William A. McKee
Water Quality Control Division
Colorado Department of Public Health
and Environment
4300 Cherry Creek Drive South
Denver, CO 80246-1530
telephone: 303/692-3583
fax: 303/782-0390
e-mail: bill.mckee@state.co.us
Annie McLeod
Office of Ocean and Coastal Resource
Management
South Carolina Department of Health and
Environmental Control
1362 McMillan Ave., Suite 400
Charleston, SC 29405
telephone: 803/744-5838, ext. 132
fax: 803/744-5847
e-mail:
mcleodab@chastn86.dhec.state.se. us
James P. McMahon
Mackinaw River Project
The Nature Conservancy
416 Main St., Suite 1112
Peoria, IL 61602-1103
telephone: 309/673-6689
fax: 309/673-8986
e-mail: mcmahon@cyberdesic.com
James W. Meek
Consultant
708 A Street, SE
Washington, DC 20003
telephone: 202/544-5980
Kimberly W. Merritt
Division of Engineering Services/Public
Works
Leon County
301 S. Monroe St., Room 202
Tallahassee, FL 32301
telephone: 850/488-8003
fax: 850/488-1260
e-mail: merrittk@mail.co.leon.fl.us
Julie Vincentz Middleton
Save Our Streams
The Izaak Walton League of America
707 Conservation Ln.
Gaithersburg, MD 20878-2983
telephone: 301/548-0150
fax: 301/548-0146
e-mail: jvincent@iwla.org
Andy Miller
Water Quality/Watersheds and Planning
South Carolina Department of Health and
Environmental Control
2600 Bull St.
Columbia, SC 29201
telephone: 803/734-9238
fax: 803/734-5355
e-mail:
millerca@columb32. dhec.state.se. us
Gail Miller
G. Miller Consulting, Inc.
2413 Algonquin Rd., Suite #257
Algonquin, IL 60102
telephone: 847/658-8822
fax: 847/658-8862
Davika Misir
ADI Technology Corp.
2345 Crystal Dr., Suite 909
Arlington, VA 22202
telephone: 703/416-0613
fax: 703/416-0182
e-mail: davika.misir@aditech.com
Richard J. Mollahan
Division of Water Pollution Control,
Planning Section, Nonpoint Source Unit
Illinois Environmental Protection Agency
PO Box 19276, 1021 N. Grand Ave. East
Springfield, IL 62794-9276
telephone: 217/782-3362
fax: 217/785-1225
e-mail: epa1184@epa.state.il. us
Michael Morgan
Water Resources Department
Rust E&l
3121 Butterfield Rd.
Oak Brook, IL 60544
telephone: 630/574-3326
fax: 630/574-2007
e-mail: mike_morgan@ccmail.rustei.com
Noel Mullett
Watershed Management Division
Wayne County Department of Environ-
ment
415 Clifford
Detroit, Ml 48226
telephone: 313/964-8868
fax: 313/961-1262
304
-------
Thomas Mumley
San Francisco Bay Region
California Regional Water Quality Control
Board
2101 Webster St., Suite 500
Oakland, CA 94702
telephone: 510/286-0962
fax: 510/286-1380
e-mail: tem@rb2.swrcb.ca.gov
Michael Murphy
Natural Resources Department
Northeastern Illinois Planning Commis-
sion
222 S. Riverside Plaza, Suite 1800
Chicago, IL 60606
telephone: 312/454-0401, ext. 305
fax: 312/454-0411
e-mail: murphy@nipc.org
James E. Murray
Department of Environment
Wayne County
415 Clifford
Detroit, Ml 48226-1515
telephone: 313/224-3632
fax: 313/224-0045
Edwardo Mustata
Surface Water Quality Division
Michigan Department of Environmental
Quality
PO Box 30273
Lansing, Ml 48909-7773
telephone: 517/335-4178
fax: 517/373-9958
e-mail: mustatae@state.mi.us
Nilabh Narayan
Mechanical Research
Tennant, Inc.
PO Box 1452, MD11
Minneapolis, MN 55440
telephone: 612/540-1200
fax: 612/540-1437
e-mail: n1n@tennantco.com
Mark E. Nelson
StormTreat Systems, Inc.
90 Rt. 6A
Sandwich, MA 02563
telephone: 508/833-1033
fax: 508/833-3150
e-mail: mnelson@horsleywitten.com
Peter Neuberger
Department of Public Works
City of Appleton
100N.AppletonSt.
Appleton, Wl 54911
telephone: 920/832-6474
fax: 920/832-6489
Jim Nicita
Huron River Watershed Council
1100 N. Main, Suite 210
Ann Arbor, Ml 48104
telephone: 734/769-5123
fax: 734/998-0163
e-mail: jim_nicita@hotmail.com
Amanda C. Nickel!
Division of Environmental and Financial
Assistance
Ohio Environmental Protection Agency
PO Box 1049
Columbus, OH 43216-1049
telephone: 614/644-3659
fax: 614/644-3687
e-mail: amanda.nickell@epa.state.oh.us
Kirk M. Nixon
Aquifer Studies, Watershed Protection
and Management
San Antonio Water System
PO Box 2449
San Antonio, TX 78298-2449
telephone: 210/704-7392
fax: 210/704-7508
Terry Noonan
Environmental Services
Ramsey County Department of Public
Works
3377 N. Rice St.
St. Paul, MN 55126
telephone: 612/482-5230
fax: 612/482-5232
e-mail: terry. noonan@co.ramsey.mn. us
Rick Noss
Fuller, Mossbarger, Scott & May Engi-
neers, Inc.
6600 Busch Blvd., Suite 100
Columbus, OH 43229
telephone: 614/846-1400
fax: 614/846-9566
e-mail: noss.6@osu.edu
Faruk Oksuz
Water Resources Department
Parsons Engineering Science
1000 Jorie Blvd., Suite 250
Oak Brook, IL 60523
telephone: 630/990-7200
fax: 630/990-7218
e-mail: faruk_oksuz@parsons.com
Stefan L. Olson
Hamilton County Soil and Water Conser-
vation District
29 Triangle Park Dr., #2901
Cincinnati, OH 45246
telephone: 513/772-7645
fax: 513/772-7656
e-mail: stefan.olson@swcd.hamilton-co.org
Sue Olson
Department of Public Works
City of Appleton
100 N. Appleton St.
Appleton, Wl 54911
telephone: 920/832-6474
fax: 920/832-6489
Carla N. Palmer
Division of Surface Water Management
St. Johns River Water Management
District
P.O. Box1429
Platka, FL 32178-1429
telephone: 904/329-4204
fax: 904/329-4315
e-mail:
carla_palmer@district. sjrwmd.state.fi. us
Jeanna M. Paluzzi
Clinton River Watershed Council
1970 E. Auburn Rd.
Rochester Hills, Ml 48307
telephone: 248/853-9580
fax: 248/853-0486
e-mail: clintonriver@compserv.net
Michael J. Paul
Institute of Ecology
University of Georgia
Athens, GA 30602-2202
telephone: 706/542-3414
fax: 706/542-6040
e-mail: mike@sparc.ecology.uga.edu
James Pease
Water Quality Division
Vermont Department of Environmental
Conservation
103 S. Main St., Bldg. 10 North, 2nd Floor
Waterbury, VT 05671
telephone: 802/241-2683
fax: 802/241-3287
e-ma;7:jimp@dec.anr.state.vt.us
Johanna Peltola
Envidata
Finnish Association for Nature Conserva-
tion
Uudenmaankatu 7, FIN-00120
Helsinki, Finland
te/epftone:+35896121 011
fax:+358 9 6957 411
e-mail: Johanna, peltola@envidata.salomaa.fi
Charles A. Peters
Water Resources Division
U.S. Geological Survey
8505 Research Way
Middleton, Wl 53562
telephone: 608/821-3810
e-mail: capeters@usgs.gov
305
-------
Theodore W. Peters
Geneva Lake Environmental Agency
PO Box 200
Fontana, Wl 53125
telephone: 414/275-6310
fax: 414/275-1134
Kristyn Pisor
Cuyahoga Soil and Water Conservation
District
6100 W. Canal Rd.
Valley View, OH 44125
telephone: 216/524-6580
fax: 216/524-6584
e-mail: cswcd@buckeyeweb.com
Jennifer Pondelick
Water Planning Section
Illinois Environmental Protection Agency
1701 S. First Ave., Suite 600
Maywood, IL 60153
telephone: 708/338-7955
fax: 708/338-7930
Josephine A. Powell
Compliance and Public Affairs
Wayne County Department of Environ-
ment
415 Clifford
Detroit, Ml 48226
telephone: 313/224-2652
fax: 313/224-7650
Thomas Price
Natural Resources Department
Northeastern Illinois Planning Commis-
sion
222 S. Riverside Plaza, Suite 1800
Chicago, IL 60606
telephone: 312/454-0401, ext. 304
fax: 312/454-0411
e-mail: tomprice@nipc.org
David Ramsay
Friends of the Chicago River
407 S. Dearborn St., Suite 1580
Chicago, IL 60605
telephone: 312/939-0490
fax: 312/939-0931
e-mail: friends@chicagoriver.org
Krista Reininga
Woodward-Clyde Consultants
111 SW Columbia, Suite 900
Portland, OR 97201
telephone: 503/948-7223
fax: 503/222-4292
e-mail: kxreiniO@wcc.com
Paul Rentschler
Huron River Watershed Council
1100 N. Main St., Suite 210
Ann Arbor, Ml 48104
telephone: 313/769-5123
fax: 313/998-0163
e-mail: p_rentschler@msn.com
Kevin Richards
Water Resources Division/Upper Illinois
NAWQA
U.S. Geological Survey
8505 Research Way
Middleton.WI 53562
telephone: 608/821-3861
fax: 608/821-3817
e-mail: krichard@usgs.gov
James W. Ridgway
Environmental Consulting and Technol-
ogy, Inc.
220 Bagley Ave., Suite 600
Detroit, Ml 48226
telephone: 313/963-6600
fax: 313/963-1707
e-mail: jridgway@ectinc.com
Jay Riggs
Dakota County Soil and Water Conserva-
tion District
4100 220th St. W
Farmington, MN 55024-8087
telephone: 612/460-8004
fax: 612/460-8401
e-mail: jfr@mn.nrcs.usda.gov
Dave Ritter
Summit County Soil and Water Conserva-
tion District
2787 Front St., Suite B
Cuyahoga Falls, OH 44221
telephone: 330/929-2871
fax: 330/929-2872
e-mail: summitswcd@aol.com
John H. Robinson
AbTech Industries
854 Jimeno Rd.
Santa Barbara, CA 93103
telephone: 805/962-8115
fax: 805/966-6312
e-mail: jrobsb@aol.com
Stuart Robinson
Civil Engineering
A. Morton Thomas & Associates, Inc.
12750 Twinbrook Pky., Suite 200
Rockville, MD 20852
telephone: 301/881-2545
fax: 301/881-0814
e-mail: amtengr.intr.net
Donald P. Roseboom
Office of Water Quality Management
Illinois State Water Survey
P.O. Box 697
Peoria, IL 61652-0697
telephone: 309/671-3196
fax: 309/671-3106
e-mail: roseboom@sws.dnr.state.il.us
Jerome K. Rouch
Division of Environmental and Financial
Assistance
Ohio Environmental Protection Agency
PO Box 1049
Columbus, OH 43216-1049
telephone: 614/644-3660
fax: 614/644-3687
e-ma;7:jerry.rouch@epa.state.oh.us
Fred Rozumalski
Barr Engineering Co.
8300 Norman Center Dr.
Minneapolis, MN 55437
telephone: 612/832-2733
fax: 612/832-2601
e-mail: frozumalski@barr.com
Diane Rudin
Mackinaw River Project
The Nature Conservancy
416 Main St., Suite 1112
Peoria, IL 61602-1103
telephone: 309/673-6689
fax: 309/673-8986
e-mail: drudin@cyberdesic.com
Carolyn Rutland
Engineering Division
City of Kalamazoo
415 Stockbridge Ave.
Kalamazoo, Ml 49001
telephone: 616/337-8365
fax: 616/337-8533
e-mail: rutlandc@ci.kalamazoo.mi.us
Jeffrey Sanders
Planning Department
Oneida Nation
PO Box 365
Oneida, Wl 54155
telephone: 920/869-4527
fax: 920/869-1610
Gary C. Schaefer
Hey and Associates
627 N. Second St.
Libertyville, IL 60048
telephone: 847/918-0888
fax: 847/918-0892
306
-------
Tina Schneider
Natural Resources Management
Maryland-National Capital Park and
Planning Commission
1109 Spring St., Suite 802
Silver Spring, MD 20910
telephone: 301/650-4368
fax: 301/650-4371
e-ma;7:schneider@mncppc. state.md.us
Thomas R. Schueler
Center for Watershed Protection
8391 Main St.
Ellicott City, MD 21043
telephone: 410/461-8323
fax: 410/461-8324
e-mail: mrrunoff@pipeline.com
Paul Shadrake
Harza Environmental Services, Inc.
233 S. Wacker Dr.
Chicago, IL 60606
telephone: 312/831-3830
fax: 312/831-3999
e-mail: pshadrake@harza-es.com
Earl Shaver
Environment Division
Auckland Regional Council
BellSouth Centre, 21 Pitt St.
Private Bag 92 012
Auckland, New Zealand
telephone: 011 649 366 2000 ext. 8079
fax: 011 6493662155
e-mail: eshaver@arc.govt.nz
Leslie Shoemaker
Tetra Tech, Inc.
10306 Eaton PI., Suite 340
Fairfax, VA 22030
telephone: 703/385-6000
fax: 703/385-6007
e-mail: shoemle@tetratech-ffx.com
Shelley Shreffler
Natural Resources Program
St. Paul Neighborhood Energy Consor-
tium
475 N. Cleveland Ave. #100
St. Paul, MN 55104
telephone: 612/644-5436
fax: 612/649-3109
e-mail: nec@orbis.net
Pam Sielski
Department of Planning and Development
Forest Preserve District of Cook County
536 N. Harlem Ave.
River Forest, IL 60305
telephone: 708/771-1355
fax: 708/771-1512
Douglas Siglin
Conservation Department
American Rivers
1025 Vermont Ave., NW
Washington, DC 20005
telephone: 202/547-6900, ext. 3103
fax: 202/347-9240
e-mail: dsiglin@amrivers.org
Terry Siviter
Environmental Products Division
Americast
11352 Virginia Precast Rd.
Ashland, VA 23005
telephone: 804/798-6068
fax: 757/498-3586
e-mail: isoilater@aol.com
Rick Smeaton
Kane County Development Department
719 Batavia Ave.
Geneva, IL 60134
telephone: 630/232-3491
fax: 630/232-3411
Bruce E. Smith
Surface Water, Southwest District Office
Ohio Environmental Protection Agency
401 E. Fifth St.
Dayton, OH 45402-2911
telephone: 937/285-6099
fax: 937/285-6249
e-mail: bruce.smith@epa.state.oh.us
Dick Smith
Water Facilities
Florida Department of Environmental
Protection
2600 Blair Stone Rd.
Tallahassee, FL 32308
telephone: 850/488-8163
fax: 850/921-2769
e-mail: smith_r@dep.state.fl. us
Jan Peter Smith
Coastal Zone Management
Massachusetts Bays Program
100 Cambridge St., 20th Floor
Boston, MA 02202
telephone: 617/727-9530, ext. 419
fax: 617/723-5408
e-ma;7:jan.smith@state.ma.us
Kendra Smith
Planning Division
Unified Sewerage Agency
155 N. First Ave., Suite 270, MS 10
Hillsboro, OR 97124-3072
telephone: 503/844-8118
fax: 503/640-3525
e-mail: ksmith@usa-cleanwater.org
Robbin B. Sotir
Robbin B. Sotir & Associates
434 Villa Rica Rd.
Marietta, GA 30064-2732
telephone: 770/424-0719
fax: 770/499-8771
e-mail: rbsotir@aol.com
Kelly S. Standridge
Medina Soil and Water Conservation
District
803 E. Washington St., Suite 160
Medina, OH 44256
telephone: 330/722-2605
fax: 330/725-5829
e-mail:
kelly.standridge@oh.nrcs.usda.gov
Randolph J. Stowe
Natural Areas
10015 Wright Rd.
Harvard, IL 60033
telephone: 815/648-2252
fax: 815/648-2403
e-mail: rstowe@owc.net
Christy Strand
Utility Services Engineering
City of Tacoma
2201 Portland Ave.
Tacoma, WA 98421-2711
telephone: 253/591-5588
fax: 253/502-2107
e-mail: cstrand@ci.tacoma.wa.us
Eric W. Strecker
Woodward-Clyde Consultants
111 SW Columbia Blvd., Suite 900
Portland, OR 97201
telephone: 503/948-7253
fax: 503/222-4292
e-mail: ewstrecO@wcc.com
David C. Strom
Division of Soil Conservation
Iowa Department of Agriculture and Land
Stewardship
Wallace State Office Bldg.
Des Moines, IA 50309
te/epftone:515/281-5142
fax: 515/281-6170
Roger C. Sutherland
Water Resources Engineering
Kurahashi & Associates, Inc.
12600 SW 72nd Ave., Suite 100
Tigard, OR 97223
telephone: 503/968-1605
fax: 503/968-1105
e-mail: kai@spiritone.com
307
-------
Jeff Swano
Enviro Impact Solutions
8738 Washington Ave.
Brookfield, IL 60513
telephone: 708/485-4190
fax: 708/485-0547
e-mail: jswano@concentric. net
Peter Swenson
Water Division (WW-16J)
U.S. Environmental Protection Agency -
Region 5
77 W Jackson Blvd.
Chicago, IL 60604
telephone: 312/886-0236
fax: 312/886-0168
e-mail: swenson.peter@epamail.epa.gov
Sara Synnestvedt
Clinton River Watershed Council
1970 E. Auburn Rd.
Rochester Hills, Ml 48307
telephone: 248/853-9580
fax: 248/853-0486
e-mail: clintonriver@compserv.net
Arthur Talley
Water Quality Division
Texas Natural Resource Conservation
Commission
PO Box 13087
Austin, TX 78711
telephone: 512/239-4546
fax: 512/239-4444
e-ma;7:atalley@tnrcc.state.tx.us
Dan Taphorn
Hamilton County Soil and Water Conser-
vation District
29 Triangle Park Dr., #2901
Cincinnati, OH 45246
telephone: 513/772-7645
fax: 513/772-7656
Susan Tatalovich
School of Civil Engineering
Purdue University
1284 Civil Engineering Bldg.
West Lafayette, IN 47907
telephone: 765/494-0258
fax: 765/496-1210
e-mail: tatalovi@ecn.purdue.edu
Paul Thomas
Water Division (WW-16J)
U.S. Environmental Protection Agency -
Region 5
77 W. Jackson Blvd.
Chicago, IL 60604
telephone: 312/886-7742
fax: 312/886-0168
e-mail: thomas.paul@epamail.epa.gov
Gregg Tichacek
Division of Fisheries
Illinois Department of Natural Resources
600 N. Grand Ave. West, Suite 5
Springfield, IL 62702-2538
telephone: 217/782-6424
fax: 217/785-8262
e-mail: gtichacek@dnrmail.state.il.us
Ginna Tiernan
Parks and Recreation
DeKalb County Government
3681 Chestnut St.
Scottdale, GA 30079
telephone: 404/508-7631
fax: 404/508-7561
Donald L. Tilton
Tilton & Associates, Inc.
204 E. Washington St.
Ann Arbor, Ml 48104
telephone: 313/769-3004
fax: 313/769-1969
e-mail: doctilton@aol.com
Scott Tomkins
Division of Water Pollution Control,
Planning Section, Nonpoint Source Unit
Illinois Environmental Protection Agency
P.O. Box 19276, 1021 N. Grand Ave. East
Springfield, IL 62794-9276
telephone: 217/782-3362
fax: 217/785-1225
e-mail: epal 170@epa.state, il. us
George Townsend
TetraTech, Inc.
10306 Eaton PI., Suite 340
Fairfax, VA 22030
telephone: 703/385-6000
fax: 703/385-6007
e-mail: townsge@tetratech-ffx.com
Michael Troge
Planning Department
Oneida Nation
PO Box 365
Oneida, Wl 54155
telephone: 920/869-4527
fax: 920/869-1610
Stacy Trussler, P.E.
Water Quality Program, Northwest Region
Washington State Department of Trans-
portation
15700 Dayton Ave. North, MS 138
Seattle, WA 98133-9710
telephone: 206/440-4905
fax: 206/440-4805
e-mail: trussler@wsdot.wa.gov
Nichole Vachon
Parks and Recreation
DeKalb County Government
3681 Chestnut St.
Scottdale, GA 30079
telephone: 404/508-7602
fax: 404/508-7561
Paul VanGelder
Clough Harbour & Associates
2001 Rt. 46, Suite 107
Parsippany, NJ 07054-1315
telephone: 973/299-1100
fax: 973/299-1123
e-mail: pvangelder@worldnet.att.net
Doreen M. Vetter
Wetlands Division
U.S. Environmental Protection Agency
401 M Street, SW
Washington, DC 20460
telephone: 202/260-1906
fax: 202/260-8000
e-mail: vetter.doreen@epamail.epa.gov
Laurene von Klan
Friends of the Chicago River
407 S. Dearborn St., Suite 1580
Chicago, IL 60605
telephone: 312/939-0490
fax: 312/939-0931
e-mail: friends@chicagoriver.org
Therese Walch
Public Works Engineering
City of Eugene
858 Pearl St.
Eugene, OR 97401
telephone: 541/682-5291
fax: 541/682-5032
Richard W. Walker
Commonwealth Technology, Inc.
2526 Regency Rd.
Lexington, KY 40503
telephone: 606/276-3091
fax: 606/276-4374
e-mail: richard.walker@ctienv.com
Ruth A. Wallace
Water Pollution Control Program
Missouri Department of Natural Re-
sources
PO Box 176
Jefferson City, MO 65102-0176
telephone: 573/526-7687
fax: 573/526-5797
e-mail: rwallace@mail.state.mo.us
308
-------
Alison Walsh
Office of Environmental Protection (CRI)
U.S. Environmental Protection Agency
JFK Federal Bldg.
Boston, MA 02203
telephone: 617/565-4871
fax: 617/565-9360
e-mail: walsh.alison@epamail.epa.gov
Dave Warburton
Twin Cities Field Office
U.S. Fish & Wildlife Service
4101 E. 80th St.
Bloomington, MN 55425
telephone: 612/725-3548, ext. 203
fax: 612/725-3609
e-mail: dave_warburton@fws.gov
Bill Ward
Tetra Tech, Inc.
200 E. Randolph St., Suite 4700
Chicago, IL 60601
telephone: 312/856-8733
fax: 312/938-0118
e-mail: wardw@ttemi.com
Frank Wash
Robert B. Annis Water Resources
Institute
Grand Valley State University
1 Campus Dr.
Allendale, Ml 49401-9403
telephone: 616/895-3749
fax: 616/895-3864
Margaret Watkins
Water Quality
Grand Portage Environmental Depart-
ment
PO Box 428
Grand Portage, MN 55605
telephone: 218/475-0193
fax: 218/475-2455
Louise Watson
Ramsey-Washington Metro Watershed
District
1902 E. County Road B
Maplewood, MN 55109
telephone: 612/704-2089
fax: 612/704-2092
Matthew C. Weidensee
Land Conservation Department
Walworth County
Courthouse Annex, W3929 Hwy. NN
Elkhorn.WI 53121
telephone: 414/741-2013
fax: 414/741-2886
Patricia Werner
Lake County Stormwater Management
Commission
333-B Peterson Rd.
Libertyville, IL 60048
telephone: 847/918-5269
fax: 847/918-9826
e-mail: lcsmc@ix.netcom.com
Bill White
Office of Realty and Environmental
Planning
Illinois Department of Natural Resources
524 S. Second St.
Springfield, IL 62701-1787
telephone: 217/782-3715
fax: 217/524-4177
e-mail: bwhite@dnrmail.state.il.us
Patricia Payne White
Preservation Fund Inc.
PO Box 15308
Atlanta, GA 30333-0308
telephone: 404/508-7603
Tim White
Erie Soil and Water Conservation District
2900 Columbus Ave.
Sandusky, OH 44870
telephone: 419/626-5211
fax: 419/626-1147
e-mail: tim.white@oh.nrcs.usda.gov
Sean Wiedel
Lake County Stormwater Management
Commission
333-B Peterson Rd.
Libertyville, IL 60048
telephone: 847/918-7695
fax: 847/918-9826
e-mail: lcsmc@ix.netcom.com
Martha Wilczynski
J.F. New & Associates
708 Roosevelt Rd.
Walkerton, IN 46574
telephone: 219/586-3400
fax: 219/586-3446
Wesley A. Wimmer
Stormwater Management Utility
City of Cincinnati
705 Central Ave., Room 400
Cincinnati, OH 45202-1900
telephone: 513/352-5232
fax: 513/352-2407
e-mail: wes.wimmer@cinmsu.rcc.org
Paul Winiarz
G. Miller Consulting, Inc.
2413 Algonquin Rd., Suite #257
Algonquin, IL 60102
telephone: 847/658-8822
fax: 847/658-8862
Jeffrey Witte
Planning Department
Oneida Nation
PO Box 365
Oneida, Wl 54155
telephone: 920/869-4527
fax: 920/869-1610
Christine Woelfel
Seattle Public Utilities
City of Seattle
710 Second Ave., 10th Floor
Seattle, WA 98106
telephone: 206/684-7599
fax: 206/386-9147
e-mail: chris.woelfel@ci.Seattle.wa. us
Dennis M. Wojcik
Drain Commissioner
Washtenaw County
PO Box 8645
Ann Arbor, Ml 48107-8645
telephone: 734/994-2525
fax: 734/994-2459
Chris O. Yoder
Division of Surface Water, Ecological
Assessment Unit
Ohio Environmental Protection Agency
1685WestbeltDr.
Columbus, OH 43228
telephone: 614/728-3382
fax: 614/728-3380
e-mail: chris.yoder@epa.state.oh.us
Jennifer Zielinski
Center for Watershed Protection
8391 Main St.
Ellicott City, MD 21043
telephone: 410/461-8323
fax: 410/461-8324
e-mail: mrrunoff@pipeline.com
309
------- |