AI-based geological subsurface reconstruction using sparse convolutional autoencoders

Abstract

Subsurface reconstruction is critical for geological modeling and resource exploration. Conventional spatial interpolation methods are limited by stationarity and spatial isotropy assumptions, while advanced geostatistical techniques require specialized datasets. Deep learning approaches often need large datasets, which is impractical for geoscientific applications. This study presents an AI-based methodology using a sparse convolutional autoencoder for robust subsurface modeling under data constraints and integrating secondary data sources such as Vertical Electrical Sounding (VES) data. A four-stage testing framework was implemented: (1) emulating conventional interpolation for baseline performance; (2) reconstructing subsurface geometries from synthetic data; (3) incorporating geophysical constraints through VES forward modeling; and (4) validating the methodology using a real-world case study from the Huancayo tectonic basin in the Peruvian Andes, using 41 VES measurements across two cross-sections (12 and 14 km long). Results demonstrate that the proposed model effectively emulates kriging interpolation (mean squared error: 1.5 $\times$ 10$-3$ to 1.2 $\times$ 10$-3$ with 100–800 training examples) through transfer learning from an inverse-distance, pre-trained model. In subsurface reconstruction, the model outperforms kriging (37.4–61.7 % improvement across 1–15 % sampling densities) through its ability to adapt to non-stationary conditions. When incorporating synthetic VES data, the model effectively reconstructed subsurface geometries with error reduction from 4.1 $\times$ 10$-1$ to 9.1 $\times$ 10$-3$ as stations increased from 1 to 40, demonstrating diminishing returns beyond this point. Application to the Huancayo basin case study validated the model's practical applicability by successfully identifying previously unmapped features including the contact between basement and sedimentary infill, folds and faults. The methodology demonstrates the AI's capability to enhance geological understanding in complex tectonic settings, revealing subtle features and refining existing assumptions about subsurface architecture.