Unstructured Grid Based Joint Inversion of DC Resistivity and Gravity Data: Adaptive FCM Assisted Model Coupling and Quantitative Similarity Analysis

• Published in: Journal of geophysical research. Machine learning and computation Vol. 2; no. 3
• Main Authors: Verma, Vishnu Kant, Singh, Anand
• Format: Journal Article
• Language: English
• Published: 01.09.2025
• ISSN: 2993-5210; 2993-5210
• DOI: 10.1029/2025JH000647

Many geophysical studies have highlighted the limitations of relying on a single geophysical data set to understand subsurface conditions. Often, individual inversions yield inconsistent geologic models. To address this issue, this study integrates direct current (DC) resistivity and gravity data within a joint inversion (JI) framework. The developed algorithm, implemented in MATLAB, features adaptive fuzzy c‐means‐based model coupling and quantitatively measures the similarity between inverted physical property models. The proposed JI approach incorporates two key parameters: the number of geologic units and the average resistivity and density values for each unit. A comparative analysis was performed between the results of JI and those obtained from separate interpretations of DC resistivity and gravity data sets. The results indicate that the developed approach significantly improves the understanding of the subsurface by producing consistent resistivity and density models—an outcome rarely achieved through individual inversions in complex geological environments. Furthermore, the framework provides a quantitative metric of similarity between the inverted resistivity and density models at each iteration. To validate the proposed approach, tests were conducted on synthetic and real‐field data sets using a triangular mesh. This study addresses a common limitation in geophysical interpretation, where reliance on a single data set often results in inconsistent geologic models. To overcome this issue, a joint inversion (JI) algorithm was developed and implemented in MATLAB. It integrates resistivity and density data using fuzzy c‐means clustering and incorporates two key parameters: the number of geologic units and their average resistivity and density values. A comparison was carried out between the results obtained from the JI and those derived from traditional separate interpretations of DC resistivity and gravity data sets. The results demonstrate that the developed algorithm significantly improves model consistency by generating resistivity and density distributions that are geologically more coherent—a notable enhancement over separate inversions. In addition, it incorporates supplementary information such as the number of geologic units and their average physical property values. The algorithm also computes the similarity between inverted resistivity and density models during each iteration. To validate the proposed approach, tests were conducted on synthetic and field data sets using a triangular discretization. Developed a joint inversion (JI) algorithm that integrates resistivity and density data, enhancing geologic models in geophysical studies The JI method provides consistent models with greater geological insights compared to traditional separate interpretations Validated algorithm's effectiveness on synthetic and real‐field data sets, measuring structural similarity between inverted geologic models