 |
|
|
||||
| Publications— Scientific Investigations Report |
By Tammy M. Zimmerman
View the report in Portable Document Format (PDF) SIR 2008-5039 (791 KB)
The Lake Erie beaches in Pennsylvania are a valuable recreational resource for Erie County. Concentrations of Escherichia coli (E. coli) at
monitored beaches in Presque Isle State Park in Erie, Pa., occasionally exceed the single-sample bathing-water standard of 235 colonies per 100 milliliters
resulting in potentially unsafe swimming conditions and prompting beach managers to post public advisories or to close beaches to recreation.
To supplement the current method for assessing recreational water quality (E. coli concentrations from the previous day), a predictive regression model
for E. coli concentrations at Presque Isle Beach 2 was developed from data collected during the 2004 and 2005 recreational seasons. Model output
included predicted E. coli concentrations and exceedance probabilities--the probability that E. coli concentrations would exceed the standard. For this
study, E. coli concentrations and other water-quality and environmental data were collected during the 2006 recreational season at Presque Isle Beach 2.
The data from 2006, an independent year, were used to test (validate) the 2004-2005 predictive regression model and compare the model performance to
the current method. Using 2006 data, the 2004-2005 model yielded more correct responses and better predicted exceedances of the standard than the use
of E. coli concentrations from the previous day. The differences were not pronounced, however, and more data are needed. For example, the model
correctly predicted exceedances of the standard 11 percent of the time (1 out of 9 exceedances that occurred in 2006) whereas using the E. coli
concentrations from the previous day did not result in any correctly predicted exceedances.
After validation, new models were developed by adding the 2006 data to the 2004-2005 dataset and by analyzing the data in 2- and 3-year combinations.
Results showed that excluding the 2004 data (using 2005 and 2006 data only) yielded the best model. Explanatory variables in the 2005-2006 model were
log10 turbidity, bird count, and wave height. The 2005-2006 model correctly predicted when the standard would not be exceeded (specificity) with a
response of 95.2 percent (178 out of 187 nonexceedances) and correctly predicted when the standard would be exceeded (sensitivity) with a response of
64.3 percent (9 out of 14 exceedances). In all cases, the results from predictive modeling produced higher percentages of correct predictions than
using E. coli concentrations from the previous day. Additional data collected each year can be used to test and possibly improve the model. The results
of this study will aid beach managers in more rapidly determining when waters are not safe for recreational use and, subsequently, when to close a
beach or post an advisory.
Abstract
Introduction
Methods
Inventory of Water-Use Data
Estimated Unregistered Water-Use Data
Residential
Model Validation
Relation Between Escherichia coli Concentrations and Data Variables
Modeling to Predict Escherichia coli
Summary and Conclusions
Acknowledgments
References Cited
Suggested citation:
Zimmerman, T.M., 2008, Modeling to predict Escherichia coli at Presque Isle Beach 2, City of Erie, Erie County, Pennsylvania: U.S. Geological Survey
Scientific Investigations Report 2008-5039, 13 p.
If you have Adobe® Acrobat® or Adobe® Acrobat®Reader® installed on your computer, you may view and print the PDF version of this report. Acrobat Reader, is a free download from Adobe Systems Incorporated.
View the report in Portable Document Format (PDF) SIR 2008-5039 (791 KB)
For more information about USGS activities in Pennsylvania contact:
Director
USGS Pennsylvania Water Science Center
215 Limekiln Road
New Cumberland, Pennsylvania 17070
Telephone: (717) 730-6960
Fax: (717) 730-6997
or access the USGS Water Resources of Pennsylvania home page at:
http://pa.water.usgs.gov/.
