Three-Dimensional Gravity Inverse Modeling for Basement Depth Estimation Integrating Maximum Difference Reduction (MDR), Trend Surface Analysis (TSA) and Total Variation Regularization

In sedimentary basin studies, gravity data are typically used to estimate the basement topography. Gravity inversion methods are expected to be able to discriminate between continuous and discontinuous sedimentary basins. Most 3D gravity inversion methods require intensive computational resources (computer memory and processing time). MDR3D, a variant of the well-known Bott method, was transformed into the Gauss-Newton inversion approach for extension flexibility. Integration of trend surface analysis (TSA) into the inversion scheme for regional anomaly estimation allows basement depth estimation from the Bouguer anomaly data. The aim of the additional total variation regularization is to stabilize the inversion algorithm and to achieve a geologically feasible model, especially for discontinuous basin types. Evaluation of the proposed method led to satisfactory results both for the synthetic and the field data set. It was found that the regularization parameter can improve the stability of the algorithm and also the depth estimation from noisy data up to ±0.5 mGal.