Search this site:

OWML: Completed Projects

Project Title: Near-real-time assessments of recreational water quality at inland lakes and relation to public health risk.

Project chief:  Donna Francy

Project support:  Chantil Milam and Chris Kephart

Cooperators:  Ohio Water Development Authority, Ohio Department of Natural Resources, and Muskingum Watershed Conservancy District

Project duration:  January 2010 – September 2012

Introduction and problem:
Current Escherichia coli (E. coli) methods used for monitoring recreational water quality and issuing swimming advisories take 18–24 hours to complete. Microbiological water-quality conditions may change during this time, leading to erroneous assessments of public-health risk. Two solutions to improve the timeliness and accuracy of recreational advisories are to use predictive models or rapid analytical methods. Predictive models, developed through statistical techniques such as multiple linear regression (MLR), employ easily measured environmental and water-quality variables to estimate bacterial-indicator concentrations or the probability of exceeding target concentrations. Rapid analytical methods, such as quantitative polymerase chain reaction (QPCR) and immunomagnetic separation/adenosine triphosphate (IMS/ATP), yield results on bacterial indicator levels within 2–3 hours. Although these “rapid assessments” have been tested and applied at coastal beaches in Ohio ( and elsewhere, little work has been done to develop and test these methods in inland lakes and reservoirs.

There is also a paucity of information on the occurrence of pathogens (organisms that cause disease) in inland recreational waters. Pathogens associated with waterborne outbreaks in the United States have included E. coli O157:H7, Shigella sonnei, Cryptosporidium, and enteric viruses. Birds such as gulls and geese, which are commonly found at Ohio beaches, have been known to carry bacterial pathogens that can infect humans; for example, Campylobacter spp. Data on the occurrence of pathogens in inland recreational waters are needed to establish the link between results from predictive models or rapid analytical methods and the density of pathogens and, thus, the human health risk.

Goals and objectives:
The overall goal is to determine whether predictive models and rapid analytical methods can be used to provide near-real-time, accurate assessments of water quality at inland recreational waters and characterize the relations between rapid assessments and pathogen occurrence in inland waters. Specific objectives are to:

  1. Identify the water quality and environmental variables (rainfall, wave height, water turbidity, etc.) that are related to bacterial indicator concentrations at approximately 15 inland recreational waters in Ohio,
  2. Select a subset of 6 sites and develop predictive models to provide reasonably accurate near-real time assessments of recreational water quality,
  3. Determine the occurrence and concentrations of waterborne pathogens at 2 selected sites, and
  4. Collect data to support and validate future implementation of a nowcast system at inland recreational waters.

This collaborative project will include 3 years of data collection, with reporting writing during the 3rd year.

Year 1 of the study (2010) will include collection of routine monitoring data at 15 inland recreational waters. Daily environmental and water-quality data (rainfall, turbidity, pool elevation, wind direction, etc.) will be compiled or measured and concentrations of E. coli by traditional methods will be determined Friday through Monday during each recreational season. Because the U.S. Environmental Protection Agency is considering the use of enterococci and rapid analytical methods as the basis for new recreational water-quality criteria being developed by 2012, analyses for enterococci and for bacterial indicators by IMS/ATP and QPCR will be done at a subset of sites.

In year 2 of the study (2011), daily sampling will continue at 6 sites that will be selected by examining results from year 1; that is, work will continue at 6 sites that are good candidates for predictive models by showing strong relations between E. coli and several variables. At 2 sites during year 2, samples will be collected for Campylobacter spp., Shigella, Cryptosporidium, pathogenic E. coli, and enterovirus and human adenovirus. For the 6 selected sites, data from years 1 and 2 will be used to develop site-specific predictive models. The results from the rapid methods will be tested for use as a stand-alone method or as a variable in a predictive model. The data will also be examined to determine whether model output values are indicative of human-health risk—that is, when the models predict exceedance of the standard, do they also indicate the presence of pathogens in inland recreational waters?

During year 3 (2012), we will continue routine monitoring at the 6 sites and test the accuracy of the predictive models in assessing recreational water quality. Two final reports from this study are planned. A technical paper, published as a journal article or USGS Scientific Investigations Report, will be targeted to the research and public health communities. A fact sheet will be targeted to the general public and will describe how rapid methods and predictive models improve assessments of recreational water quality.