Improving Next-Generation Hydrogeological Models with Geophysics

Abstract

This thesis presents advancements to the disciplines of geophysics and hydrogeological modelling. There are three main scientific contributions in this work. In Chapter 1 I propose a method to couple time-domain electromagnetics (TEM) and surface nuclear magnetic resonance (NMR) datasets with limited drillhole data to provide information on hydrogeological properties in a non-invasive manner, including groundwater salinity and hydraulic conductivity. This method reduces ambiguity in the hydrogeological interpretation of TEM-derived conductivity information by coupling conductivity to porosity values estimated using surface NMR data collected in the same field area. The method was applied to a South Australian River Murray floodplain to investigate the impact of artificial watering on the shallow groundwater salinity. In Chapter 2 I evaluate the impact of incorporating TEM and surface NMR datasets on the prediction error and uncertainty of groundwater models under different hydrogeological conditions using a synthetic approach. A method is presented to couple TEM and surface NMR to derive hydraulic conductivity using the Markov-Chain Monte Carlo method. In Chapter 3 I propose a modelling framework to couple multiple geophysical techniques, including TEM, borehole NMR and audio-frequency magnetotellurics (AMT), with stochastic groundwater modelling through the ensemble-smoother method. This approach allows the uncertainty in geophysical data to be expressed as a prior probability distribution that can be incorporated and accounted for in groundwater model inversion. This framework is applied to a potential in-situ recovery site in Kapunda, South Australia to estimate the regional-scale impact of a hypothetical operation trial in a probabilistic manner. Together, these three novel studies represent a contribution to the coupling between geophysics and hydrogeological modelling.