Year: 2026
Type: article
Abstract: Aircraft-based surveying to collect airborne electromagnetic data is a key method to image large swaths of the earth’s surface in pursuit of better knowledge of aquifer systems. Despite many years of advancements, 2D and 3D inversions still pose challenges in terms of computational requirements, regularization selection, hyperparameter tuning, and real-time inversion. A new approach was introduced for the inversion of airborne electromagnetic data that leverages machine learning to overcome the computational burden of traditional inversion methods, which implicitly includes learned regularization and is applicable in real time. The method combined 1D inversion results with geostatistical modeling to create tailored training data sets, enabling the development of a specialized neural network that predicts 2D conductivity models from airborne electromagnetic data. This approach required 3D forward modeling and 1D inversion up front, but no forward modeling during inference. The workflow was applied to the Kaweah Subbasin in California, where it successfully reconstructed conductivity models consistent with real-world data and geologic drill hole information. The results highlighted the method’s capability to deliver fast and accurate subsurface imaging, offering a valuable tool for groundwater exploration and other near-surface applications.
Source: The Leading Edge
Institution: University of British Columbia
Language: English
Subfield: Geophysics
Field: Earth and Planetary Sciences
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7137255906