Converting land use–land cover to E. coli contamination potential classes for improved management of groundwater wells: a case study in Ontario, Canada

Land use-land cover (LULC) types have been used as a proxy for Escherichia Coli (E. coli) sources and transport mechanisms. This study aims to advance the understanding of the relationship between LULC and E. coli presence in wells for the 11 major LULC categories. This represents a novel approach for assessing the broad potential for well contamination and informing groundwater management strategies. The approach combines insights gained from regression analyses conducted using a combination of large datasets with the Intergovernmental Panel on Climate Change (IPCC) method for consistent treatment of uncertainties within literature. Generalized Additive Models for Location, Shape, and Scale (GAMLSS) regression analyses were used to identify and support relationships between a large dataset of E. coli presence in wells and LULC data, identifying potential risk classes. A raster dataset for Ontario, Canada identifying areas of low to very high potential for E. coli presence in wells was created. Notably, the pastoral/agricultural LULC category was found to be in the very high-risk class, urban and aggregate mines in the high-risk class, forest in the moderate risk class, and water and grasslands in the low-risk class. However, gaps in understanding the relationship between some LULC categories and the presence of E. coli in wells remain in the disturbance, bedrock, and scrubland LULCs due to data limitations in both the study area and literature. These results provide private well users, who may lack technical expertise, with an accessible source of information on the potential for E. coli contamination.