Methodology:

Crustal density and susceptibility distribution are inferred from joint inversion of gravity and magnetic data using the 'variation of information' approach. The inversion is carried out with the academic software JIF3D (https://sourceforge.net/projects/jif3d/) developed by Max Moorkamp.

The density and susceptibility inversion models are discretised into meshes with equal horizontal cell sizes of 7.5 km. The vertical cell size is 1 km at the surface and increases with depth by a factor 1.1 for each cell. Increasing vertical cell size with depth is introduced to account for decreasing resolution with increasing distance to the source in potential field applications. A horizontal cell size of 7.5 km is chosen because the input gravity and magnetic data have a resolution of 10 km.

Additionally, a padding area of 20% around the study area is added to avoid edge effects. The resulting inversion mesh contains 244 cells east-west direction, 140 cells in north-south direction and 21 cells in the z direction (60 km below sea level). To constrain the model geometry, the bedrock topography Curie point depths and Moho depths are used as model boundaries.

All cells are set to a starting value of 0 kg/m3 for the density anomaly model and 0 SI for the magnetic model and are iteratively updated during the VI inversion process.

Note that all the elevation information are provided with positive values down.

Data collection:

The data preparation and the inversion set up is performed in python (version 3.10.11). The inversion itself is carried out in the JIF3D software using a Linux terminal.

The gravity and magnetic data were upwardly continued at a constant station height using the Compudrape plug-in for Geosoft Oasis Montaj version 2021.2.

Data quality:

The RMS error between observed and inverted gravity field is 1.5 mGal and 0.7 nT for the magnetic field. The end members of the residual between observed and inverted values are + 37 and -18 mGal with a standard deviation of 3 mGal for the gravity inversion and for the magnetic inversion +341 and -65 nT with a standard deviation of 13 nT.