Skip to main content

Predicting Beach Conditions

Public health officials use beach advisories and closings to let the public know when the level of pathogens (disease-causing microorganisms) in the water may be unsafe for swimming, water-skiing or other similar water recreation. These advisories and closings typically occur when monitoring results show that fecal bacteria levels exceed an applicable water quality standard. The presence of fecal bacteria, like E. coli (Escherichia coli), suggests other pathogens may be present in the water column.

Fecal contamination, and its associated microorganisms, can originate from many sources, including:

  • coastal and shoreline development;
  • wastewater collection and treatment facilities;
  • storm sewers and urban runoff;
  • disposal of human waste from boats;
  • bathers themselves;
  • confined animal feeding operations (CAFOs); and
  • natural animal sources such as wildlife and pet wastes.

Fifteen Wisconsin health departments implement monitoring and notification programs at beaches along Lake Michigan and Lake Superior. This involves sampling the water at the beaches 1 to 5 times per week.

  • Advisory signs are posted whenever the water quality criterion of 235 colony-forming units (CFU) of E. coli/100 ml of water is exceeded.
  • Beaches are closed when E. coli levels exceed 1,000 CFU/100 ml, indicating a more serious risk of illness.

Advisories and Closings

The standard beach monitoring and notification method (sometimes called the "persistence model") can take up to 24 hours to complete. As a result, swimming advisories and beach closures are sometimes issued a day or more after a contamination event has occurred. This can expose recreational users to unhealthy levels of pathogens. In addition, actual concentrations of bacteria in the water at the beach can vary over shorter time intervals. This can result in unnecessary advisories or closings, which can have unintended economic impacts.

Potential Solutions

To help address these concerns, researchers are developing and refining rapid detection techniques and "real-time" predictive models, sometimes called "nowcasts." These models use statistical procedures to evaluate the relationships between measured concentrations of bacteria in the water with:

  • meteorological factors (wind direction, precipitation, etc.);
  • onshore conditions (algal mats, waterfowl concentrations, etc.); and
  • near-shore conditions (wave height, turbidity, etc).

Researchers have found nowcast models often produce more accurate estimates of current beach water quality conditions than current notification methods. Nowcast models also provide beach managers with additional information on the relative influence of different factors. This is important because proposed federal legislation could require beach managers to develop and implement source-identification and tracking programs for some beaches. For most local beach managers, however, the staff time and expertise needed to build nowcast models are prohibitive.

In response, U.S. Environmental Protection Agency researchers have developed a prototype decision-support tool called Virtual Beach-Model Builder, which provides a user-friendly interface for local beach managers to build, test and execute predictive models. This tool can use data downloaded from the internet or manual input.

Beach Condition Projects

Currently, nowcast models do not consider bacteria levels in streams with outlets located at or near the beaches. As a result, the models' predictive power may be limited at beaches where polluted runoff is potentially a significant source of bacteria. More importantly, stand-alone statistical models do not provide information on the contribution of polluted runoff from upstream land uses and agricultural practices, or the potential impacts that could be expected in response to altering those land uses or agricultural practices.

To address these limitations, the DNR is working with the Department of Agricultural and Biological Engineering at Purdue University and the Racine Health Department to modify and link the Long-Term Hydrologic Impact Assessment (L-THIA) and Virtual Beach tools. The DNR created this combined tool with funding from the Wisconsin Coastal Management program.

When complete, this tool will allow users to nowcast E. coli levels in coastal streams and at adjacent beaches and predict the potential impact of land-use changes in contributing watersheds. To test this concept, we are developing a pilot model for the lower Root River and North Beach in the city of Racine.

Once completed, DNR staff will evaluate the pilot tool for its transferability to other coastal watersheds in Wisconsin and hope to make the tool available online.

Developing Operational Nowcast Models

Public health officials at only seven of the 540 monitored Great Lakes beaches in the United States currently use nowcast models to help with beach notifications and closures. In Wisconsin, officials at only one of our 124 monitored Great Lakes beaches uses this type of model. With support from the Wisconsin Coastal Management program, we are developing local capacity to build, refine and maintain operational nowcast models. Working with beach managers in Ozaukee County and Racine, we will use U.S. EPA's Virtual Beach 2.0 to build, refine and evaluate models for five additional Lake Michigan beaches. The DNR will provide local partners with individualized, hands-on training on data assembly, model building, evaluation and nowcasting. In addition, efforts in Racine will include the development of an alternative model (using qPCR).

Virtual Beach Pilot Model (2009)

The U.S. EPA, U.S. Geological Survey and National Oceanic and Atmospheric Administration are working with state and local partners to develop, compare and promote tools to provide early warnings of pathogen indicator levels. The City of Racine Health Department and researchers at the University of Wisconsin-Oshkosh tested an early version of Virtual Beach at 14 Great Lakes beaches in Wisconsin during 2007 and 2008. Racine Health Department continues to add data to its pilot models and is examining both current and historic data to gauge their effectiveness.

Wisconsin DNR conducted pilot testing, outreach, and technical assistance in 2008 and 2009 as part of this multi-agency effort. The department's work focused primarily on U.S. EPA's Virtual Beach-Model Builder software and included:

  • testing & evaluating Virtual Beach-Model Builder 1.0;
  • developing step-by-step learning modules;
  • conducting training workshops for beach managers;
  • operational field-testing in Ozaukee County during the summer of 2009; and
  • providing feedback and suggested modifications to the software's developers.

A final report for this project was prepared in 2009.