Joint interpretation of potential field data using independent Lpq norm inversion and geological modeling: Application to Iron targeting in Central Iran

Abstract

Solving applied geophysical inverse problems using the $L_{\mathrm{pq}}$ norm (mixed $L_p$norm) is a well-established deterministic method, particularly in potential-field inversions. This approach minimizes the objective function by integrating $L_p$ norms ($0\le p\le 2$) for the smallness and smoothness terms in the regularization function, offering high flexibility in controlling the sparsity and smoothness of the recovered models. This study focuses on the computational aspects and parameters of the model structure term that significantly influence the generation of density and susceptibility models. We investigate how different combinations of $L_p$ norms and scaling constants for the smallness and smoothness terms affect the recovery of various geometrical structures and the delineation of iron ore resources. The initial phase of our study involves testing these parameters using a synthetic model designed for gravity and magnetic susceptibility inversion, incorporating a complex dataset. Subsequently, we apply mixed $L_p$ norm inversion with different norm combinations to ground-based potential field data from an iron ore deposit in the Bafq metallogenic belt, central Iran. This region is characterized by a reverse fault that has induced a north-south trend in hematite-magnetite mineralization. A key focus of this study is the impact of employing identical versus non-identical norms in the smoothness term of the model structure. By adjusting the level of compactness to match the target trend, we generate a geologically more interpretable model. Following the inversion, we assess the effectiveness of various combinations of $L_p$ norm in delineating mineralized zones by comparing the inverted models to a lithological model obtained via co-kriging interpolation of borehole data. The results reveal significant hematite mineralization, with high-grade deposits predominantly located in the hanging wall of the fault and lower-grade deposits in the footwall. The magnetite mineralization, while less extensive, exhibits a spatial distribution similar to that of hematite, though it is typically shallower in comparison.