U.S. Geological Survey
Water-Resources Investigations Report 03-4056
Download this report as a PDF
file (18.2 mb)
Download a free copy of
Acrobat Reader
Comparison and Continuous Estimates of Fecal Coliform and
Escherichia Coli Bacteria in Selected Kansas Streams,
May 1999 Through April 2002
By
Patrick P. Rasmussen and Andrew C. Ziegler
The sanitary quality of water and its use as a public-water supply and for recreational
activities, such as swimming, wading, boating, and fishing, can be evaluated on the basis of fecal
coliform and Escherichia coli (E. coli) bacteria densities. This report describes the overall
sanitary quality of surface water in selected Kansas streams, the relation between fecal coliform and
E. coli, the relation between turbidity and bacteria densities, and how continuous bacteria estimates
can be used to evaluate the water-quality conditions in selected Kansas streams.
Samples for fecal coliform and E. coli were collected at 28 surface-water sites in Kansas. Of
the 318 samples collected, 18 percent exceeded the current Kansas Department of Health and
Environment (KDHE) secondary contact recreational, single-sample criterion for fecal coliform
(2,000 colonies per 100 milliliters of water). Of the 219 samples collected during the recreation
months (April 1 through October 31), 21 percent exceeded the current (2003) KDHE single-sample
fecal coliform criterion for secondary contact rec-reation (2,000 colonies per 100 milliliters of
water) and 36 percent exceeded the U.S. Environmental Protection Agency (USEPA)
recommended single-sample primary contact recreational criterion for E. coli (576 colonies per
100 milliliters of water). Comparisons of fecal coliform and E. coli criteria indicated that more
than one-half of the streams sampled could exceed USEPA recommended E. coli criteria
more frequently than the current KDHE fecal coliform criteria. In addition, the ratios of E. coli
to fecal coliform (EC/FC) were smallest for sites with slightly saline water (specific conductance
greater than 1,000 microsiemens per centimeter at 25 degrees Celsius), indicating that E. coli may
not be a good indicator of sanitary quality for those streams. Enterococci bacteria may provide a
more accurate assessment of the potential for swimming-related illnesses in these streams.
Ratios of EC/FC and linear regression models were developed for estimating E. coli densities on
the basis of measured fecal coliform densities for six individual and six groups of surface-water
sites. Regression models developed for the six individual surface-water sites and six groups of
sites explain at least 89 percent of the variability in E. coli densities. The EC/FC ratios and
regression models are site specific and make it possible to convert historic fecal coliform bacteria data to
estimated E. coli densities for the selected sites. The EC/FC ratios can be used to estimate E. coli
for any range of historical fecal coliform densities, and in some cases with less error than the
regression models. The basin- and statewide regression models explained at least 93 percent of
the variance and best represent the sites where a majority of the data used to develop the models
were collected (Kansas and Little Arkansas Basins).
Comparison of the current (2003) KDHE geometric-mean primary contact criterion for fecal
coliform bacteria of 200 col/100 mL to the 2002 USEPA recommended geometric-mean criterion
of 126 col/100 mL for E. coli results in an EC/FC ratio of 0.63. The geometric-mean EC/FC ratio
for all sites except Rattlesnake Creek (site 21) is 0.77, indicating that considerably more than
63 percent of the fecal coliform is E. coli. This potentially could lead to more exceedances of the
recommended E. coli criterion, where the water now meets the current (2003) 200-col/100 mL
fecal coliform criterion.
In this report, turbidity was found to be a reliable estimator of bacteria densities. Regression
models are provided for estimating fecal coliform and E. coli bacteria densities using continuous
turbidity measurements. Prediction intervals also are provided to show the uncertainty associated
with using the regression models. Eighty percent of all measured sample densities and individual
turbidity-based estimates from the regression models were in agreement as exceeding or being
less than the primary and secondary contact recreational criteria. The continuous turbidity
measurements and regression models were used to construct probability curves that can be used to
estimate bacteria concentrations on the basis of measured turbidity values. Duration curves
developed for six sites using the hourly estimates of bacteria density indicate that the current KDHE
(fecal coliform bacteria) and USEPA recommended (E. coli bacteria) primary contact
recreational criteria were exceeded for 21 to 94 and 31 to 97 percent of the spring and summer,
respectively. Estimated bacteria densities most commonly exceeded the current and
recommended criteria in the spring (April through June). Hourly estimates provided in real time
(available on the World Wide Web at
http://ks.water.usgs.gov/rtqw/) allow the
public and water-management agencies to make decisions in regard to whether planned water
activities are appropriate by considering current stream conditions relative to water-quality
criteria.
Annual and seasonal loads and yields were calculated using hourly estimated fecal coliform
and E. coli bacteria densities and streamflow at six surface-water sites for the calendar years 2000
and 2001. Estimated bacteria loads in 2001 were about 2 to 8 times larger than the bacteria loads in
2000 for the Kansas and Little Arkansas Rivers. Data from major point sources upstream from the
surface-water sites in these basins indicate that nonpoint sources accounted for more than 97
percent of the annual loads. Mean daily bacteria loads in 2000 were largest in the winter for five
sites and in the spring for one site. In 2001, mean daily bacteria loads were largest in the spring for
four sites and in the winter for two sites. Annual load differences are caused by varying hydrologic
conditions and higher streamflow caused by over-land runoff. Surface-water sites in the Little
Arkansas River Basin had the largest bacteria yield per acre of watershed.
CONTENTS
- Abstract
- Introduction
- Purpose and Scope
- Methods
- Bacteria Sample Collection and Analysis
- Turbidity Measurements
- Development of Regression Models to Estimate Bacteria Densities
- Measured Bacteria Densities
- Comparison of Fecal Coliform and Escherichia Coli Densities
- Continuously Estimated Bacteria Densities
- Relation Between Turbidity and Fecal Coliform Density
- Probability and Duration of Estimated Fecal Coliform Densities
- Relation Between Turbidity and Escherichia Coli Density
- Probability and Duration of Estimated Escherichia Coli Densities
- Estimated Bacteria Loads and Yields
- Summary
- References
Rasmussen, Patrick 2003, Comparison and Continuous Estimates of Fecal
Coliform and Escherichia Coli Bacteria in Selected Kansas Streams, May 1999
Through April 2002 03-4056, 87 p.
For additional information please visit the Web site at:
http://ks.water.usgs.gov/rtqw/
(best viewed with Internet Explorer 5.0 or greater or Netscape 6.0 or greater)
For additional information contact:
Patrick Rasmussen
U.S. Geological Survey
4821 Quail Crest Place
Lawrence, KS 66049-3839
Telephone: (785) 832-3542
Fax: (785) 832-3500
Email: pras@usgs.gov
|