3D magnetotelluric constrained inversion integrating realistic structural and rock-physics information: A case study in Tongling ore district

Incorporating realistic geological and geophysical prior information is essential for enhancing the reliability of magnetotelluric (MT) data interpretations, particularly in complex geological settings. However, existing 3D MT inversion methods generally lack the ability to integrate such information. To address this gap, we developed a 3D MT constrained inversion algorithm capable of incorporating detailed structural and rock-physics information using unstructured tetrahedral meshes. First, to ensure accurate computation of electromagnetic response functions for complex subsurface structures with variable topography, we implemented a nested mesh strategy based on unstructured meshes within the forward modeling engine. This nestedmesh approach facilitates precise conductivity mapping between the inversion and forward meshes, effectively eliminating numerical errors arising from parameter mismatches. In constructing the inversion objective function, we supplemented the conventional data fitting and model roughness terms with cross-gradient and guided fuzzy C-means clustering constraints. The cross-gradient constraint integrates interface information derived from geological and geophysical datasets, while the guided fuzzy C-means clustering constraint incorporates rock physics data, promoting sharp boundaries and focused conductivity variations within the inversion model. We utilized the Gauss-Newton method to minimize the objective function, ensuring efficient and rapid convergence of the inversion algorithm. Through a series of synthetic tests, we demonstrated that incorporating prior information significantly enhances MT inversion results. Finally, we applied the algorithm to interpret MT data from the Tongling ore district in Anhui Province, China, constructing a reliable 3D electrical resistivity model of the area. The preferred model suggests that hydrothermal fluids ascend through channels and faults within the deep crust, driving mineralization. This interpretation is in agreement with surface observations, which show that mineral deposits are predominantly located near these fault structures.