K-STEMIT: Knowledge-Informed Spatio-Temporal Efficient Multi-Branch Graph Neural Network for Subsurface Stratigraphy Thickness Estimation from Radar Data

TL;DR

K-STEMIT is a novel graph neural network that integrates physical knowledge and adaptive feature fusion to improve subsurface stratigraphy thickness estimation from radar data, outperforming existing methods.

Contribution

It introduces a knowledge-informed, multi-branch spatio-temporal GNN with adaptive feature fusion and physical priors for enhanced accuracy and efficiency.

Findings

K-STEMIT achieves the highest accuracy among compared methods.

Incorporating physical priors reduces RMSE by 21.01%.

The model enables reliable, continuous spatiotemporal assessment of snow accumulation.

Abstract

Subsurface stratigraphy contains important spatio-temporal information about accumulation, deformation, and layer formation in polar ice sheets. In particular, variations in internal ice layer thickness provide valuable constraints for snow mass balance estimation and projections of ice sheet change. Although radar sensors can capture these layered structures as depth-resolved radargrams, convolutional neural networks applied directly to radar images are often sensitive to speckle noise and acquisition artifacts. In addition, purely data-driven methods may underuse physical knowledge, leading to unrealistic thickness estimates under spatial or temporal extrapolation. To address these challenges, we develop K-STEMIT, a novel knowledge-informed, efficient, multi-branch spatio-temporal graph neural network that combines a geometric framework for spatial learning with temporal convolution to capture temporal dynamics, and incorporates physical data synchronized from the Model Atmospheric Regional physical weather model. An adaptive feature fusion strategy is employed to dynamically combine features learned from different branches. Extensive experiments have been conducted to compare K-STEMIT against current state-of-the-art methods in both knowledge-informed and non-knowledge-informed settings, as well as other existing methods. Results show that K-STEMIT consistently achieves the highest accuracy while maintaining near-optimal efficiency. Most notably, incorporating adaptive feature fusion and physical priors reduces the root mean-squared error by 21.01% with negligible additional cost compared to its conventional multi-branch variants. Additionally, our proposed K-STEMIT achieves consistently lower per-year relative MAE, enabling reliable, continuous spatiotemporal assessment of snow accumulation variability across large spatial regions.