Trial-and-error modeling of ground-airborne electromagnetic data in the Yishu faulting basin, China

Abstract

Trial-and-error modeling may provide some level of interpretation of the subsurface while sacrificing certainty, and certainty it is a viable alternative for precise three-dimensional (3D) interpretation of real ground-airborne frequency-domain electromagnetic (GAFEM) data. Therefore, a semiautomatic trial-and-error modeling approach has been presented in this study. We first developed a 3D GAFEM forward modeling code. Its accuracy was demonstrated using a 3D synthetic model with a topography and a tilted anomalous body. An initial model was established based on known geological constraints. The code was repeated, and the parameters of the model were renewed semi-automatically based on a predefined geometry-resistivity combination list. The model that could achieve the minimum error between the computed response and collected GAFEM data was selected as the final model. We applied the proposed semi-automatic trial-and-error modeling approach to a geothermal resource survey in the Yishu Faulting Basin, China. The purpose of this survey was to interpret the resistivity structure of the subsurface and evaluate the potential development of geothermal resources in the survey area. The final model obtained by trial-and-error modeling, which was constrained by known geological information and subsurface geoelectric structures inferred from 2D models inverted by the CSAMT and MT data measured at the same location, indicated the existence of geothermal resources. This indication was supported by drilling results from a well site located on the survey line. A comparative analysis was also conducted between the model obtained by trial-and-error modeling and the models obtained by 3D inversion of the GAFEM dataset. The apparent resistivity was calculated using the same data. The results have shown that different approaches can achieve similar subsurface geometries and resistivity distributions for faulting basin structures.