f Voxel Inversion of AEM Data for Improved Model Integration

the model space is usually linked to the actual observation points. For airborne surveys the spatial discretization of the model space reflects the flight lines (Figure 1). Often airborne surveys are carried out in areas where other ground-based geophysical data are available. The model space of ground-based geophysical inversions is likewise usually referred to the positions of the measurements, e.g. the electrode positions in geoelectrical investigations, and hence does not coincide with the airborne model. Consequently, a model space based on the measuring points is not well suited for jointly inverting airborne and ground-based geophysical data. Furthermore, geological and groundwater models most often refer to a regular voxel grid, not correlated to the geophysical model space. This means that incorporating the geophysical data into the geological and/or hydrological modeling grids involves a spatial relocation of the models, which in itself is a very subtle process where valuable information is easily lost. Also the integration of a priori information, e.g. from boreholes, is difficult when the observation points do not coincide with the position of the prior information. We have developed a geophysical inversion algorithm working directly in a voxel grid disconnected from the actual measuring points (Figure 1b), which then allows for straightforward integration of different data types in a joint inversion, for directly informing geological/hydrogeological models, and for easier incorporation of a priori information. The new voxel model space defines the soil properties (like resistivity) on a set of nodes, and the distribution of the soil properties is computed everywhere by means of an interpolation function (e.g. inverse distance or kriging). Given this definition of the voxel model space, the 1D forward responses are computed as follows: 1) a 1D model subdivision, in terms of model thicknesses is defined for each 1D data set, creating “virtual” layers. 2) the "virtual" 1D models at the sounding positions are finalized by interpolating the soil properties (the resistivity) in the center of the "virtual" layers (Figure 2a). For 2D/3D forward responses the algorithm operates similarly, simply filling the 2D/3D meshes of the forward responses by computing the interpolation values in the centers of the mesh cells (Figure 2b). This definition of the voxel model space decouples the geophysical model from the position of acquired data, allowing for straight-forward integration of different data types in joint inversion and for directly informing (hydro)geological models. We believe that this new approach will facilitate the integration of geophysics, geology, and hydrology for improved groundwater and environmental management. The presented algorithm is a further development of the AarhusInv program package, which manages both large scale AEM surveys (e.g. Auken et al., 2008, Siemon et al., 2009) and ground-based data (e.g. Fiandaca et al., 2013, Behroozmand et al., 2012).