Spatially constrained inversion of area covering datasets

We present a new inversion methodology for large area-covering datasets. Many new data acquisition systems are capable of covering large areas with densely sampled data sets. However, it is not possible to apply 2D or 3D inversion algorithms on a routine basis because of computational problems. We propose a Spatially Constrained Inversion (SCI) inversion scheme using a local 1D model description ensuring fast computation times. Information on the geological variability is included as constraints between model parameters applied on an all-to-all basis. The 1D formulation means that the SCI works best on a sub-horizontal layered geology. To ensure flexibility even for very large surveys we sub-divide the area in hexagons, which are independently inverted and afterwards stitched together. Continuity over hexagon edges is ensured by applying an overlap between neighbouring hexagons. A field example demonstrates that the SCI enables better resolution of otherwise poorly resolved model parameters. Lineations observed in line-inversion are removed.