Improved parameterization of airborne electromagnetic surveys for groundwater salinity mapping using 3D variable-density groundwater flow models

Airborne electromagnetics is a cost-effective and rapid method with which to map the regional distribution of saline groundwater in coastal areas, however the process of transforming observed data into salinity estimates comprises multiple sources of uncertainty. The resulting error primarily relates to two factors: (1) the transformation (or inversion) from airborne observations into physical properties, (2) the availability of lithological information to transform inversion results to salinity estimates. Recent research has shown that this uncertainty can significantly affect the accuracy of resulting groundwater salinity estimates, in particular the location of the fresh-saline interface. Reducing error relating to the two factors is not trivial. Firstly, as the inversion process is non-unique, an infinite number of models can fit the data. Secondly, the availability of lithological information on regional scales is generally low. To highlight potential sources of error and improve parameterization, we investigate the usefulness of combining airborne electromagnetic data with a 3D variable-density groundwater flow and coupled salt transport model. We quantitatively present findings using a synthetic model which was created using an existing large-scale (~100km²) 3D groundwater model based on real data from the Netherlands. The model is created in two steps: (1) the available groundwater model is run until a state of equilibrium is reached with the model boundaries and stress terms and (2), an airborne survey is simulated using standard geophysical forward modelling techniques, resulting in set of observations. The airborne observations are then inverted and used alongside a simulated lithological data acquisition programme, which are finally input as initial conditions to a groundwater model. As the groundwater model is assumed to be in a state of equilibrium, we show the effect of implementing an optimization framework that penalizes the rate of groundwater salinity fluctuations by iteratively changing the input parameters of both the inversion method and the lithological data. Results quantitatively highlight the effectiveness of implementing a simple, inter-disciplinary approach to airborne electromagnetic groundwater mapping.