Understanding the recoverability of physical property relationships from geophysical inversions of multiple potential-field data sets

Abstract

Collecting multiple geophysical datasets has been a common practice in resource explorations. How to effectively integrate them and extract useful information has been an active area of study. Among existing approaches, joint inversion followed with geology differentiation has shown promising results. The general workflow is to invert the multiple geophysical sets, whether separately or jointly, to obtain multiple physical property model, which is then followed by geology differentiation, a process of classifying the inverted values into different groups. This process assumes that the recovered physical property relationships are reliable. To understand the recoverability of physical property relationships, we designed four geological scenarios with three causative geological units at different depths with varied physical property values. For each scenario, we performed both separate and joint inversions using both the conventional smoothness-based L2-norm and the novel mixed L12-norm regularization strategies. We examined the distribution of the inverted physical property values and performed geology differentiation for each inversion. We found (1) that depths of the source bodies significantly affect the recoverability of physical property relationships, (2) joint inversion results in better recoverability than separate inversions, and (3) joint sparsity inversion consistently gives rise to the best recoverability.