A synthetic study of geophysics-based modelling of groundwater flow in catchments with a buried valley

Groundwater models simulating flow in buried valleys interacting with regional aquifers are often based on hydrogeological models interpreted from dense geophysical datasets and scarce borehole data. For three simple synthetic cases, it is demonstrated that alternative methods of inversion of transient electro-magnetic (TEM) data can lead to very different interpretations of the hydrogeology inside and surrounding a buried valley. The alternative interpreted hydrogeological models are used in numerical modelling of groundwater flow to a pumping well. It is demonstrated that the alternative models result in quite different groundwater-model predictions of capture zone, recharge area, and groundwater age for the pumping well. It is briefly demonstrated that model calibration against hydraulic head data is not likely to improve the predictions or to identify the structural error of the interpreted hydrogeological models. It is therefore concluded that when TEM-based resistivity models are interpreted to construct the hydrogeological framework of a groundwater model, it must be done cautiously with support from deep borehole information. Too much reliance on geophysical mapping can lead to seriously wrong hydrogeological models and correspondingly wrong groundwater-model predictions.