3-D Structurally Constrained Inversion of the Controlled-Source Electromagnetic Data Using Octree Meshes

In this article, we propose a structural constraint method to improve the resolution of 3-D frequency-domain controlled-source electromagnetic (CSEM) data imaging. The subsurface interfaces can be obtained from high-resolution seismic imaging data or reliable geological information. First, we assume that electrical parameters within a given formation are classified, meaning that they exhibit variation around their average value. Then, the structural constraint can be guided by the resistivity averages and ranges obtained from petrophysical measurements. In this way, we can achieve categorical inversion results in known regions or even capture structures that are insensitive to data, thereby enhancing the reliability of the interpretation of CSEM data. In addition, we utilize octree-based nonconforming hexahedral meshes to construct the structurally constrained model to simulate undulating terrain and complex underground interfaces more effectively. We adopt the NLCG algorithm for the inversion of CSEM data. Finally, we test the effectiveness of the proposed structural constraint method using synthetic and field datasets. The inversion results show that our method can constrain the known strata in shallower parts well and significantly improve the resolution of the deeper regions.