Image-guided structure-constrained inversion of electrical resistivity data for improving contaminations characterization

Abstract

Electrical resistivity tomography often has high uncertainty in contamination characterization due to the complex subsurface structure. Utilizing available prior information is crucial for enhancing geological plausibility. We propose an improved structure-constrained method that updates the smooth weights of all eight elements surrounding a boundary element using three different magnitudes. The methodology is tested on seven synthetic models and a field case covering four different types of structural boundaries. Structural constraints are vital to increasing boundary contrast, and the improvement is more significant with increasing depth. With the increase of data points from an optimized array, the influence of structural constraint is reduced. The implementation of accurate structural constraints can enhance the characterization of contamination anomaly zones. However, misplaced structural constraints in the proximity of the contamination can lead to worse depiction. Without knowing the boundary exactly, we propose a method to obtain local constraints from borehole logs, which is verified by both numerical and field data. This method reduces the interference from uncertain structural boundaries and enhances the robustness of the image-guided structure-constrained inversion. In the end, we conclude that correct boundary information even with scattered constraints from borehole logs can increase the identification of the abnormal contamination resistivity.