Forecasting bacteria levels at bathing beaches in Ohio

Abstract

The U.S. Geological Survey developed models for predicting exceedance of the bathing-water standard for Escherichia coli ( E. coli ) at three Lake Erie beaches and one inland lake in Ohio. The statistical models were specific to each beach, and the best model for each beach was based on a unique combination of environmental and water-quality variables as explanatory factors. For the Lake Erie beaches, these factors included wave height, number of birds on the beach at the time of sampling, lake-current direction, rainfall, turbidity, and streamflow of a nearby river. For the inland lake, these factors included date, wind direction and speed, number of birds, and rainfall. The prediction error in the models was too large to accurately estimate concentrations of E. coli ; however, the models can be used like weather forecasts to predict the probability, given a set of input variables, that the Ohio bathing-water standard used to judge swimming safety will be exceeded.