Models of total and presumed wildlife sources of fecal coliform bacteria in coastal ponds

Models that accurately predict fecal coliform bacteria (FCB) concentrations, one of the most widely used measures of estuarine water quality, are needed to improve land use decision-making. Rapidly occurring changes in coastal land uses and the influence on water quality increases the urgency of having improved decision tools. For this study, samples were collected monthly from six coastal ponds, two tidal creeks and four shallow water wells for up to 2 1 2 years. These data were used along with other measures of environmental conditions and land classes within each watershed to construct quantitative relationships between combinations of variables and both total and presumed wildlife sources of FCB. Linear regression, bootstrapping and generalized additive modeling that incorporates both linear and nonlinear terms were used. Results of repeated simultaneous sampling on the same tide stage of ponds and downstream estuarine creeks suggest that most FCB come from wildlife and that the ponds effectively remove these bacteria except immediately following heavy rainfall. Predictive models for concentrations of total and presumed wildlife bacteria are provided along with simple measures to estimate watershed boundaries. It is proposed that these tools can be used to minimize impacts on receiving water body quality. The models can be used to test alternative development approaches within coastal watersheds similar to that found in the southeastern USA coastal zone as well as to evaluate specific proposed landscape alterations.