Malaguerra, Flavio1; Albrechtsen, Hans-Jørgen1; Binning, Philip John1
1 Department of Environmental Engineering, Technical University of Denmark2 Urban Water Engineering, Department of Environmental Engineering, Technical University of Denmark3 Water Resources Engineering, Department of Environmental Engineering, Technical University of Denmark
Drinking water wells are often placed near streams because streams often overly permeable sediments and the water table is near the surface in valleys, and so pumping costs are reduced. The lowering of the water table by pumping wells can reverse the natural flow from the groundwater to the stream, inducing infiltration of surface water to groundwater and consequently to the drinking water well. Many attenuation processes can take place in the riparian zone, mainly due to mixing, biodegradation and sorption. However, if the water travel time from the surface water to the pumping well is too short, or if the compounds are poorly degradable, contaminants can reach the drinking water well at high concentrations, jeopardizing drinking water quality. Here we developed a reactive transport model to evaluate the risk of contamination of drinking water wells by surface water pollution. The model was validated using data of a tracer experiment in a riparian zone. Three compounds were considered: an older pesticide MCPP (Mecoprop) which is mobile and persistent, glyphosate (Roundup), a new biodegradable and strongly sorbed pesticide, and its degradation product AMPA. Global sensitivity analysis using the method of Morris was employed to identify the dominant model parameters. Results showed that the presence of an aquitard and its characteristics (degree of fracturing and thickness), pollutant properties and well depth are the crucial factors affecting the risk of drinking water well contamination from surface water. Global sensitivity analysis results were compared with rank correlation statistics between pesticide concentrations and geological parameters derived from a comprehensive database of Danish drinking water wells. Aquitard thickness and well depth are the most critical parameters in both the model and observed data.