Impact of Geological Conceptualization on the Calibration of Groundwater Models at the Mannheim Wellfield, Kitchener, Ontario

Abstract

This study investigates the importance of geological data on the calibration of groundwater models using long term pumping/injection and monitoring well records at a wellfield in the Region of Waterloo, Ontario. Four different geological models with homogeneous geological layers are calibrated by coupling HydroGeoSphere (HGS) and the parameter estimation code PEST using water-level variation records collected during municipal well operations. The estimated hydraulic conductivity (𝐾) and specific storage (𝑆𝑆) are consistent to those obtained through previous aquifer tests. The four geological models are well calibrated with assigned initial 𝐾 and 𝑆𝑆 and yield reliable estimates for the upper layers where most data points are collected. However, the 𝐾 and 𝑆𝑆 estimated for lower layers with fewer observation points vary more significantly among the models. The comparison of simulated and observed drawdown for both model calibration and validation reveals that all four groundwater flow models with varying geology can capture the water-level fluctuation pattern quite well. However, these models fail to capture the rapid water- level variations at some wells. This study demonstrates the usefulness of water level fluctuation data resulting from municipal well operations in the calibration of groundwater flow models.