MODELING HEALTH RISKS OF PATHOGENS IN KARST BASINS: IMPLICATIONS OF DATA UNCERTAINTY AND DETECTION LIMITS

Karst aquifers are highly vulnerable to contamination from various sources, including human and animal waste, because of solution-enhanced porosity and permeability. Challenges in addressing contamination by pathogens include early detection of their presence and accurate assessment of human health risks resulting from exposures. Our study focuses on the karstic Royal Spring basin, encompassing both urban and agricultural (pasture and row-crop) land uses in Fayette and Scott counties, Kentucky. Georgetown is the largest municipality in Kentucky that relies primarily on a spring for water supply; as of 2020, the city water utility served 33,075 customers. We aimed to quantify the health risk associated with fecal contamination and, specifically, three pathogens in the Royal Spring basin. We collected 47 water samples (typically weekly) from July 14, 2021, to June 15, 2022, at Royal Spring. Fecal contamination was indicated by the presence of E. coli using the IDEXX Colilert test in split samples at concentrations varying from <1 to 4839.2 most probable number per 100 milliliters (MPN/100 mL) with a mean of 493.4 MPN/100 mL. To quantify pathogen concentrations, each water sample was filtered before DNA extraction followed quantitative polymerase chain reaction. Gene targets included the Shiga toxin gene (stx1) from E. coli O157:H7, beta-giardin from Giardia lamblia, and ntf2 from Cryptosporidium parvum. None of the target pathogens were detected in any of the samples. However, based on previously published literature, risk assessment results could vary based on how data below the detection limit are managed. Therefore, a systematic review of these three pathogens in springs and karst aquifers will be conducted, and a quantitative microbial risk assessment will be performed to model infection risks from different exposure scenarios (i.e., accidental ingestion during recreational activities), exploring concentrations of these pathogens at various levels below the sample limit of detection. Uncertainty and sensitivity analyses will be performed to evaluate the impacts of data uncertainty on modeled health risks.