Microbiological contamination of private water wells in the midwest region of Ireland: investigation of water quality, public awareness and the application of logistic regression in contaminant modelling

Groundwater derived drinking water sources account for a substantial amount of the water infrastructure in Ireland. This study has shown that verocytotoxin producing Escherichia coli coincides with a high dependency on private wells at both a national (44.9% of cases reported in 2012) and a regional level (72.3% of cases in the Midwest in 2012). As private wells are implicated as a major transmission route with regard VTEC epidemiology, this study also assessed the overall quality of private wells in the Midwest region of Ireland to assess risk. In total, 132 households were included in the study and their water supply was sampled three times over the period of a year for the presence of E. coli; a faecal indicator organism. Total contamination was found to be 51.4%; a much higher result than reported groundwater contamination with E. coli in national studies conducted by the EPA. Similar to VTEC notifications, the presence of E. coli had a varied spatial and temporal distribution. The stewardship amongst research participants with regard to the infrastructure of their well and pollution sources was also investigated. Areas of substandard awareness were found to exist with regard to well infrastructure with only 32.6% of participants aware of the presence or absence of a sanitary seal installed to safeguard their water supply mechanism. Similarly, in terms of stewardship, only 32.6% of households had their water tested prior to participation in this study. This study found that overall, private well stewardship and awareness was low, however, the study also found that people were willing to take initiative once presented with contamination information, with 30 (n = 30/132) of the households in this study installing a water treatment system prior to repeat contamination analysis and information. From a contamination perspective, determining the likelihood that groundwater contains faecal coliforms can aid water resource management in facilitating the protection of drinking water supplies. This study predicted groundwater contamination with E. coli through the utilisation of statistic modelling, namely Logistic Regression (LR), which predicted the outcome of the dichotomous dependent variable (E. coli present or absent) based on fifteen potential predictor variables, categorised into three vulnerability factors: Intrinsic, Specific and Infrastructural. Using a geographical information system and questionnaire analysis, the relative hydrogeological and meteorological (Intrinsic), potential contamination sources (specific) and well design and construction parameter (Infrastructural) factors unique to each sampling location were derived. Utilising this information, a logistic regression (LR) model, named ISI-LR (Intrinsic, Specific and Infrastructural Logistic Regression) was used to predict the probability of contamination of private wells with E. coli. In total, the model used eleven predictor variables: soil permeability, aquifer type, presence of Karst features, recharge, rainfall, temperature, number of cattle, number of farms, number of septic tanks (all per Electoral Division), well depth and well type. The full model, containing all predictors, was statistically significant at p= 0.003, indicating that the model distinguished between the independent variables' relationship to the incidence of contamination. The ISI-LR model explained between 59.7% (Cox and Snell R squared) and 81.0% (Nagelkerke R squared) of the variance in contamination and correctly classified 90.7% of cases.