Lessons worth highlighting
The inversion of multiple data sets can greatly increase the complexity of the models and interpretation. This section presents some the lessons worth sharing for future projects.
Laterally Constrained 1D inversion
We designed a laterally constrained 1D inversion procedure in order to get a more consistent conductivity distribution, as represented schematically in Fig. 391.
Between each 1D iteration, an average conductivity model is interpolated
onto a global 3D mesh and used as a reference for subsequent inversions.
A global data misfit and regularization parameters are used to control
individual 1D inversions, similar to the framework used for 3D algorithms.
Individual 1D inversions are used to populate a large 3D mesh.