Predicting Freshwater Microbial Pollution Using a Spatial Model: Transferability between Catchments

Abstract

Freshwater microbial contamination has become a worldwide problem, but fecal indicator organism (FIO) data are lacking in many catchments and large-scale management is expensive. Therefore, a model that can assist in spatial localization to simulate microbial risk maps and Critical Source Areas (CSAs) is needed. This study aims to generate a predicted risk of microbial contamination in Kent and Leven, Northumberland, and East Suffolk based on the ArcMap hydrological tool using the land use parameters in the Wyre and Yealm catchments. Then, this study will compare the value obtained with the E. coli concentration data (observational risk) in order to evaluate whether land cover weightings are transferable between different catchments and provide microbial risk guidelines for ungauged catchments. In the research, the East Suffolk catchment showed strong fitting with actual values in the rainy and dry seasons after using the predictive values weighted by Wyre and Yealm, respectively. Specifically, as for the models with Yealm land cover weightings, the results show that the adjusted R2 in the rainy season for East Suffolk is 0.916 (p < 0.01) while the adjusted R2 values in the dry season is 0.969 (p < 0.01). As for models with Wyre land cover weightings, the adjusted R2 values (rainy season) is 0.872 (p < 0.01), while the adjusted R2 values (dry season) is 0.991 (p < 0.01). This indicates that this spatial model can effectively predict the risk of fecal microbial contamination in the East Suffolk catchment. Second, this research believes that the land cover weightings are more transferable in catchments that have close geographical locations or similar land cover compositions. This paper makes recommendations for future catchment management based on the results obtained.