Investigating aquitard heterogeneity by inverse groundwater modelling at a drinking water production site

Abstract

This study investigates whether geostatistical properties of aquitards can be determined from data collected at a drinking water well field. A workflow adaptable to any drinking water extraction site with pumping and groundwater head data is developed. Using data from Budel, the Netherlands, a layered groundwater flow model is constructed and calibrated on hydraulic heads. A large number of realizations are generated for the aquitard, considering various heterogeneity parameters. These simulations are upscaled to the grid of the flow model, and their fit to observed heads is evaluated. Results show that optimal geostatistical parameters can be identified, even though many combinations reproduce observed heads. These parameters can be used to parameterize regional groundwater flow models. Particle tracking simulations show heterogeneity decreases contaminant breakthrough times while also decreasing the total flow through the aquitard. These findings emphasize the need to consider aquitard heterogeneity in risk assessments at drinking water production sites.