PREDICTING OCCURRENCES OF ARSENIC IN GROUNDWATER IN VIRGINIA AS A TOOL FOR EXPOSURE ASSESSMENT
We are constructing a logistic regression model, using existing datasets of soils, geology, geochemistry, and hydrogeology to predict the probability of As concentrations at or above 10 ppb in VA groundwater. Measured As concentrations in groundwater from the Virginia Department of Environmental Quality will be used as the dependent variable (1: As concentration at or above 10 ppb; 0: otherwise). Geologic units, soil series and texture, land use, groundwater pH and other chemical parameters will be used as explanatory variables in the model. Approximately 20% of data will be retained as the validation data for testing the model and a goodness-of-fit test will be used to evaluate the performance of fitted model. Relationships between explanatory variables (e.g., bedrock geology, soil type, land use) will be evaluated to see under which conditions As is most probable to occur in groundwater. The results will be used to create a risk assessment map in ArcGIS that will identify areas of VA that may have elevated As concentrations in groundwater, thus posing a potential increased health risk for those ingesting the water.