Abstract To mitigate the non-uniqueness problem inherent in separate geophysical data inversions, cooperative inversion of multiphysics data offers an effective approach for delineating subsurface structures and parameters in ore-concentrated areas with minimal reliance on prior geological or borehole information. A cooperative inversion method for gravity and magnetic data is developed with a novel joint structure-guided (JSG) constraint, which incorporates a structure-guided coupling operator into the regularized objective function to enhance structural consistency and extract complementary features between density and magnetization structures. The JSG coupling model adopts a nonlinear guided function based on gradient information and dynamically adjusts the parameter constraint weights to impose different penalties on the edge and non-edge regions. This preserves inherent structural features while integrating complementary information, thereby achieving complementary resolution by exploiting the distinct sensitivities of the two data types. The inversion employs a sequential strategy based on alternating iterative constraints. Additionally, an adaptive scheme for calculating constraint parameters is implemented to ensure stable inversion convergence while alleviating the difficulty in parameter selection. Results from synthetic and field data examples demonstrate that the cooperative inversion exploits the complementary resolution characteristics of gravity and magnetic data, improving lateral resolution from magnetic data and depth resolution from gravity data for multi-scale targets. The method is most effective when density and magnetization are strongly covariant, efficiently transferring complementary structural information, while remaining stable and free of significant spurious structures even when the sources partially overlap spatially. In such cases, the cooperative inversion yields superior resolution compared to separate inversions, while maintaining good data fit. The approach reduces the non-uniqueness inherent in geophysical inversion interpretation to a significant degree, providing more reliable information for refined modeling and metallogenic prediction within ore concentration areas.
Wang et al. (Sun,) studied this question.