Uncovering Impacts of Seasonality and Land Use Land Cover on E. coli Contamination in Private Drinking Water Wells: A Machine Learning Application

In both the global and Canadian contexts, there is a heavy reliance on groundwater as a drinking water source, with approximately 35% and 22% of the respective populations relying solely on groundwater. Further, approximately 4.5 million people living in Canada rely solely on groundwater drawn from an unregulated private source, and are often challenged by limited resources for maintenance, management, and protection [1]. Dependence on private drinking water systems result in an increased health risk due to microbial contamination, such as Escherichia coli (E. coli), some strains of which may result in acute gastrointestinal illness. To mitigate the risk of illness due to groundwater contamination, it is important to understand how and when these microorganisms enter the groundwater system and migrate through it. Recent studies have explored the impacts of geological settings, sampling strategy, seasonality, and land use/land cover on drivers of E. coli concentration in private drinking water wells and have concluded the significant role of the latter three factors. The impact of seasonality is of particular interest within the North American context, as much of the continent has four distinct seasons (winter, spring, summer, autumn), each of which represent nuances when considering E. coli fate and transport in groundwater. Accordingly, the present work adopts the power of data-driven applications (i.e., pattern recognition, stochastic event synchrony, regression analysis) to develop an exploratory model for E. coli contamination in private water wells in Ontario, Canada, by combining a private well water dataset (containing both hydrogeological and microbiological variables) with land use/land cover (LULC; SOLRIS V3) and meteorological data (temperature, rainfall, relative humidity, snow cover; Daymet V3.0). The model developed can be used to advance the current understanding of the drivers of E. coli fate and transport in groundwater and inform effective remediation strategies for contaminated aquifers and management plans for private drinking wells. [1] A. Rivera, The state of ground water in Canada, Ground Water Canada, 2017.