Optimized Gravity-Magnetic Cross-Gradient Joint Inversion for Characterizing Rift-Sag Geological Structures

The geological analysis of rift-sag structures is of significant importance for understanding crustal tectonic evolution, resource exploration, and seismic activity. However, due to the non-uniqueness of gravity and magnetic data inversion and differences in resolution, a single geophysical method often fails to comprehensively reveal the structural characteristics of complex geological bodies. In this study, we propose an optimized gravity-magnetic cross-gradient joint inversion method, introducing several improvements to the weight calculation process. The model weighting matrix and structural constraint weighting matrix are normalized to simplify the calculation formulas, unify the magnitude of the two matrices, and narrow the range for selecting the weighting factors of the cross-gradient term. Furthermore, a dynamic adjustment mechanism for the cross-term weighting factor is adopted, allowing adaptive parameter adjustment based on the variation of model errors, data fitting, and inversion results during the inversion process. This dynamic parameter optimization enhances the inversion results and achieves real-time correction of parameter values, avoiding the limitations of fixed parameters and improving the reliability of the inversion.

The proposed method was applied to measured gravity and magnetic profiles in the southern margin of the Sichuan Basin. By integrating planar gravity and magnetic anomaly characteristics with vertical boundary identification techniques, the gravity field, magnetic field, structural features, deep characteristics, and sedimentary features of the rift-sag structure were systematically analyzed. The results provide reliable evidence for delineating the spatial distribution of rift-sag boundaries and the internal geological features, offering robust support for geological research and geophysical interpretations in complex tectonic environments.

Acknowledgments: This research was funded by the National Natural Science Foundation of China [grant numbers: 42104092], Fundamental Research Funds Program of Chinese Academy of Geological Sciences [grant numbers: JKYQN202351].