![[logo] US EPA](https://webarchive.library.unt.edu/eot2008/20081107122930im_/http://www.epa.gov/epafiles/images/logo_epaseal.gif){kind=logo}
Final Report: Whole Watershed Health and Restoration: Applying the Patuxent and Gwynns Falls Landscape Models to Designing a Sustainable Balance Between Humans and the Rest of Nature
EPA Grant Number: R827169Title: Whole Watershed Health and Restoration: Applying the Patuxent and Gwynns Falls Landscape Models to Designing a Sustainable Balance Between Humans and the Rest of Nature
Investigators: Costanza, Robert , Boumans, Roelof , Maxwell, Thomas , Villa, Ferdinando , Voinov, Alexey , Wainger, Lisa
Institution: University of Maryland
EPA Project Officer: Stelz, Bill
Project Period: March 1, 1999 through February 28, 2001
Project Amount: $699,916
RFA: Water and Watersheds (1998)
Research Category: Water and Watersheds
Description:
Objective:The objectives of this project were to: (1) further develop and test the Patuxent Landscape Model (PLM) as a tool for whole watershed analysis and restoration; specifically, to further develop and test the socioeconomic sectors and their dynamic links to the ecological sectors; (2) apply the approach to the Gwynns Falls watershed and intercompare with the Patuxent; (3) perform detailed scenario analyses, including a range of spatial pattern, management, policy, and climate options; (4) develop (with broad stakeholder participation) and test methods to assess the ecological health of ecosystems and watersheds; (5) develop preferred future states for the watersheds, using the results of (1) and (2) above and broad stakeholder participation; and (6) based on the above, test the degree to which various management policies can restore the ecological health of the Patuxent and Gwynns Falls watersheds and achieve the preferred future states.
Summary/Accomplishments (Outputs/Outcomes):The PLM model and its versions, the Gwynns Falls Landscape Model (GFLM) and the Hunting Creek Watershed Model (HCM), are based upon two software packages developed by our team. These are the integrated environments for high performance, modular, and multiscale spatial modeling called the Spatial Modeling Environment (SME) (Maxwell and Costanza, 1995, 1997a,b), and the Library of Hydro-Ecological Modules (LHEM, Voinov, et al., 2003), which archive a number of models that can be used to study individual processes (hydrology, nutrient cycling, plant growth, etc.) at the local scale, as well as algorithms and methods for spatial dynamics (including hydrologic fluxes, crop rotation, etc.). Using SME, the modules can be integrated into larger and more sophisticated models, such as PLM.
In PLM, by means of the SME, modules from LHEM were integrated into a larger and more sophisticated unit model that included modules for hydrology, nutrient movement and cycling, terrestrial and estuarine primary productivity, animal consumer dynamics, and human system dynamics. The current version of the full unit model contains 21 state variables and over 100 parameters and forcing functions.
The PLM further has been used to analyze a range of scenarios that show the implications of alternative future policies in the watershed on the system (Costanza, et al., 2002). We analyzed 18 scenarios, including: (1) land use in 1650, 1850, 1950, 1972, 1990, and 1997; (2) a "buildout" scenario based on fully developing all the land currently zoned for development; (3) four future development patterns based on an economic land-use conversion model; (4) agricultural "best management practices" that lower fertilizer application; and (5) extreme scenarios of land-use change to analyze the system under strong spatial perturbations.
The Hunting Creek Watershed Model (HCM)
The HCM has been developed in collaboration with the Calvert County Planning
and Zoning Commission as a tool for watershed analysis and comparison of alternative
zoning projections in Calvert County (see http://iee.umces.edu/PLM/HUNT 
).
The HCM showed how the different densities of dwelling units that would be
provisioned
under the alternative zoning plans potentially could result in changes in the
nutrient concentrations in Hunting Creek. The model also has been used to solve
an optimal control problem, wherein the goal function was to maximize the profits
from agricultural production, taking into account the "costs" of
increased nutrient concentrations in the river from fertilizers. The control
variables were the spatial distribution of crops and the amounts of fertilizers
introduced (Seppelt, Voinov, 2002). This exercise also was used to develop
some innovative techniques for spatial landscape optimization.
The HCM was a focal point of yet another project developed to design some
alternative approaches to high school education in landscape ecology and watershed
dynamics. In this project, a team of high school students has been monitoring
three sampling stations on the Hunting Creek watershed and collecting data
on nutrient (N and P) content in the creeks. One of the subwatersheds drained
a predominantly forested catchment; the other one was more agricultural; and
the third one had a larger commercial-residential component. The model then
was implemented for the small subwatersheds to illustrate the link between
the land-use patterns and the water quality in the creeks (see http://iee.umces.edu/AV/EDU/SWN ).
Linked ecological economic models like the PLM are potentially important tools for addressing issues of land-use change at the regional watershed scale. The model integrates our current understanding of ecological and economic processes at the site and landscape scales to give estimates of the effects of spatially explicit land-use or land management changes. The model also highlights areas where knowledge is lacking and where further research should be targeted. Specifically, the PLM model represents advances in the following areas:
(1) The model links topography, hydrology, nutrient dynamics, and vegetation dynamics at a fairly high temporal (1 day) and spatial (200 m) resolution with land-use patterns and the longer-term dynamics of land-use change. As far as we know, it is the most advanced model of its type for application at the regional watershed scale.
(2) The model allows the impacts of the spatial pattern of land use on a large range of ecological indicators to be explicitly assessed, providing decisionmakers and the public with information about the consequences of specific land use patterns.
(3) The model has been calibrated extensively over several time and space scales, a difficult and often ignored operation for models at this scale and complexity. New methods based on multi-criteria decision models were developed for this purpose.
(4) The model operates at several scales simultaneously, including the site (or unit model) scale and the landscape scale, which integrates all the unit models.
(5) The model is process-based, with processes changing in dominance over time. This allows better understanding of the underlying phenomenon occurring on landscapes, and therefore more detailed predictions of the possible results of changing land uses and policies.
(6) Whereas the model is formulated deterministically, extensive sensitivity analysis allows us to understand its complex dynamics without resorting to multiple stochastic replications. In the full spatial mode, when cells change from one land-use type to another, a bifurcation threshold is simulated, and all the parameters in the cell change to those of the new land-use type.
(7) The high data requirements and computational complexities for this type of model mean that development and implementation are relatively slow and expensive. For many of the questions being asked, however, this complexity is necessary. We have tried to find a balance between a simple, general model that minimizes complexity and one that provides enough process-oriented, and spatially and temporally explicit, information to be useful for management purposes (Costanza and Maxwell, 1994).
(8) Spatial data is becoming increasingly available for these types of analyses, and our modeling framework is able effectively to use this data to model and manage the landscape. One also can use the model effectively to estimate the value of specific data collection investments for a particular model, watershed, and set of goals.
One major problem was with the integration between the ecological and economic components of the model. Whereas the ecological part is a dynamic, raster-based, process-based model that runs at daily timesteps, the economic module that we were attempting to integrate with (Bockstael, et al.) was a stochastic Markov chain-type model with a yearly time-step operating over vector polygon-based, spatial coverages. The exchange between the two components was realized on the level of scenarios that were exported from the economic component as land-use maps, and then fed into the ecological component. As a result, there was no feedback from the ecological dynamics to the economic change, except that the two different spatial representations did not merge well, and it was not clear how to bring the polygon-based processes to the raster cells, and vice-versa.
This was one of the major reasons that we have started developing the General Human Model (GHM) that would share the same grid-based formalism and, most importantly, would take into account processes and drivers of economic and social change. In the GHM, we identified, coded, and linked these processes within a nonspatial context to prepare for more tight spatial representations. The knowledge and data requirements for this integration were developed through interactions with stakeholders and scientists of various social and ecological backgrounds. During workshops, we present the integrated model results to the stakeholders, thus informing the stakeholders and potentially modifying their views and priorities.
The GHM is developed to simulate wealth dynamics across space and time for forecasting and analyses of human wellbeing. The complex nature of the multidimensional and nonlinear dynamics embedded within alternative definitions of wealth is equal to those observed in ecosystems. Unlike wealth equated to acquired material goods, the phenomenon that separates wealth from poverty in the GHM is more defined as a multidimensional phenomenon, encompassing the ability to satisfy basic needs. These needs are described as follows: to be in control of resources; to have access to education; to pursue the development and application of skills; to have access to health through proper nutrition, shelter, water, and sanitation; to experience low levels of violence and crime; and to exercise political freedom and voice (Marlin, 1992). Lately, new variables were added to the equation of prosperity documented as ecosystem services (Costanza, et al., 1997). GHM was applied and calibrated as a module within the Global Unified Model of the BiOsphere (GUMBO; Boumans, et al., 2002). The GHM recognizes four different stocks of wealth that contribute to the community and individual wellbeing of humans occupying space on our globe. Wealth is associated with the physical world, or limited within the world of the mind. Wealth in the physical world is called built capital and natural capital; wealth existing within the mind is human capital and social capital. Although the capitals in the physical world are behaving according to the laws of conservation of energy and matter, those within the mind are not. From the beginning, the capitals are linked when they compete for time and resources to be created. Allocation of resources follows the dynamics of human wellbeing, defined as a multidimensional indicator of human values associated with capitals generated.
In built capital, we value the stock of human-made, material resources that can be used to produce a flow of future income. Built capital exists in a wide variety of forms, including buildings, roads, waterworks, tools, cattle and other animals, automobiles, trucks, and tractors, to name just a few forms that built capital can take.
In natural capital, we value the leading of ecosystem functioning to organizations and structures resilient to future disturbances. Natural capital is built by increasing pathways for energy flow (biodiversity) and preservation of stabile nutritional pools (water, nutrients, and carbon). Ecosystems form capital when photosynthetic uptake of carbon exceeds respiration, or when nutrient uptake exceeds mineralization.
In human capital, we value the number of people and the embedded knowledge base. The module that deals with the dynamics for human capital deals with the human population: the growth and declines that result from births, deaths, migrations, and the human production potential; levels of education and training; and health. Investments towards human capital are those that cause increases in births, decreases in deaths, and, in part, costs associated with relocation. Also included are investments made in education and training. Evaluation of human capital is the production potential of those in the population of working age. Important indicators in measuring human capital are the relative levels of income within the labor force.
In social capital, we value the knowledge, understandings, norms, rules, and expectations about patterns of interactions that groups of individuals bring to a recurrent activity. Two stocks are quantified within the social capital module, the social network and "shared norms and rules" (both expressed as indicators between 0 and 1). Investment in social capital increases the social network, which allows for increases in norms and rules. Norms and rules facilitate the efficiency of the investment into the social network. Background rates on declines in network and norms and rules are increased by migrations. Application, verification, and calibration of the regional GHM proceeds parallel to the development of the Human Ecosystem Model, a structural approach to management of socioeconomic data (Machlis, Force, and Burch, 1997). GHM natural capital is linked to the GFLM.
The GHM is an attempt to focus on the conceptual model and qualitative understanding of processes and drivers relevant to the human dimension, even if we are not always sure how the quantitative information about these processes will be obtained. In a way, we are attempting to build a coherent theoretical description of the system, analyze the model, and then try to influence the data collection efforts in the direction that we have found most appropriate in our model analysis. The conceptual model in this case provides an important means of communicating our understanding to and among stakeholders, and helps us formulate the right questions to ask them during workshops and surveys.
Journal Articles on this Report: 11 Displayed | Download in RIS Format
| Other project views: | All 26 publications | 14 publications in selected types | All 12 journal articles |
| Type | Citation | ||
|---|---|---|---|
|
|
Boumans R, Costanza R, Farley J, Wilson MA, Portela R, Rotmans J, Villa F, Grasso M. Modeling the dynamics of the integrated earth system and the value of global ecosystem services using the GUMBO model. Ecological Economics 2002;41(3):529-560 |
R827169 (Final) |
not available |
|
|
Binder C, Boumans RM, Costanza R. Applying the Patuxent landscape unit model to human dominated ecosystems: the case of agriculture. Ecological Modelling 2003;159(2-3):161-177. |
R827169 (Final) |
not available |
|
|
Boumans RM, Villa F, Costanza R, Voinov A, Voinov H, Maxwell T. Non-spatial calibrations of a general unit model for ecosystem simulations. Ecological Modelling 2001;146(1-3):17-32. |
R827169 (Final) R825792 (Final) |
not available |
|
|
Costanza R, Voinov A, Boumans R, Maxwell T, Villa F, Wainger L, Voinov H. Integrated ecological economic modeling of the Patuxent River watershed, Maryland. Ecological Monographs 2002;72(2):203-231 |
R827169 (1999) R827169 (Final) R824766 (1998) R824766 (Final) R825792 (1999) R825792 (2000) R825792 (Final) |
not available |
|
|
Maxwell T. A paris-model approach to modular simulation. Environmental Modelling and Software 1999;14(6):511-517. |
R827169 (Final) |
not available |
|
|
Seppelt R, Voinov A. Optimization methodology for land use patterns using spatial explicit landscape models. Ecological Modelling 2002;151(2-3):125-145. |
R827169 (Final) |
not available |
|
|
Seppelt R, Voinov A. Optimization methodology for land use pattems - evaluation based on multiscale habitat pattern comparison. Ecological Modelling. 2003;168(3):217-231. |
R827169 (Final) |
not available |
|
|
Voinov A, Voinov H, Costanza R. Landscape modeling of surface water flow: 2. Patuxent watershed case study. Ecological Modelling 1999;119(2-3):211-230. |
R827169 (Final) |
not available |
|
|
Voinov A, Costanza R. Watershed management and the Web. Journal of Environmental Management 1999;56(4):231-245. |
R827169 (Final) R824766 (1998) R824766 (Final) |
not available |
|
|
Voinov A, Fitz C, Boumans R, Costanza R. Modular ecosystem modeling. Journal of Ecosystem Modeling and Software. |
R827169 (Final) |
not available |
|
|
Voinov A, Costanza R, Wainger L, Boumans R, Villa F, MaxwellT, Voinov H. Patuxent landscape model: Integrated ecological economic modeling of a watershed. Environmental Modelling and Software 1999;14(5):473-491. |
R827169 (Final) R824766 (1998) R824766 (Final) R825792 (1999) R825792 (2000) R825792 (Final) |
not available |
watersheds, groundwater, land, ecological effects, population, nutrients, ecosystem, indicators, restoration, scaling, terrestrial, aquatic, habitat, integrated assessment, sustainable development, decisionmaking, community-based, socioeconomic, ecology, hydrology, modeling, Mid-Atlantic, Maryland, MD, EPA Region 3. , Ecosystem Protection/Environmental Exposure & Risk, Water, Geographic Area, Scientific Discipline, RFA, Ecosystem/Assessment/Indicators, Water & Watershed, Restoration, Social Science, Aquatic Ecosystem Restoration, Nutrients, Mid-Atlantic, Watersheds, Ecological Effects - Environmental Exposure & Risk, Ecosystem Protection, Monitoring/Modeling, Economics and Business, Urban and Regional Planning, State, stakeholder feedback, water quality, Maryland (MD), aquatic ecosystem, environmental rehabilitation, non-point sources of nutrients, watershed modeling, alternative urbanization scenarios, socioeconomics, watershed restoration, nutrient sensitive ecosystems, ecology assessment models, ecosystem health, landscape characterization, multi-criteria decision analysis, socioeconomic, sustainable development, intergrated watershed model, land use, non-point source pollution, ecological recovery, stakeholder groups, ecosystem restoration, urban landscapes, suburban watersheds, web site development, ecological exposure, Gwynns Falls, aquatic ecosystems, integrated assessment, nutrient transport, community involvement, nutrient loading, watershed sustainablity, Patuxent River watershed, ecosystem, non-point sources, environmentally stable landscape, biodiversity, ecological effects, urban watershed rehabilitation method
Relevant Websites:
http://www.uvm.edu/giee/PLM/
http://www.uvm.edu/giee/LHEM/index.html
http://www.uvm.edu/giee/SME3/
http://www.uvm.edu/giee/PLM/HUNT/
Progress and Final Reports:
1999 Progress Report
Original Abstract