GRIT-LP: Graph Transformer with Long-Range Skip Connection and Partitioned Spatial Graphs for Accurate Ice Layer Thickness Prediction

2025-11-23

AI Summary

Graph transformers face challenges of over-smoothing and inadequate long-range dependency modeling in polar ice thickness prediction. To address these, this paper proposes a graph transformer architecture tailored for snow–ice spatiotemporal modeling. Its core contributions are: (1) a partitioned spatial graph construction strategy that explicitly preserves geographical spatial consistency; (2) a long-range skip connection mechanism to mitigate feature homogenization during deep-layer propagation; and (3) a hybrid information flow architecture integrating inductive geometric graph learning with self-attention, leveraging overlapping local neighborhood graphs for efficient spatiotemporal representation. Evaluated on real-world polar observational datasets, the method reduces root mean square error by 24.92% compared to the current state-of-the-art, significantly improving ice thickness prediction accuracy. This advancement provides a more reliable modeling foundation for snow accumulation analysis, paleoclimate reconstruction, and sea-level rise projection.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Reasoning under Uncertainty: Graphical Models

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding

Abstract

Graph transformers have demonstrated remarkable capability on complex spatio-temporal tasks, yet their depth is often limited by oversmoothing and weak long-range dependency modeling. To address these challenges, we introduce GRIT-LP, a graph transformer explicitly designed for polar ice-layer thickness estimation from polar radar imagery. Accurately estimating ice layer thickness is critical for understanding snow accumulation, reconstructing past climate patterns and reducing uncertainties in projections of future ice sheet evolution and sea level rise. GRIT-LP combines an inductive geometric graph learning framework with self-attention mechanism, and introduces two major innovations that jointly address challenges in modeling the spatio-temporal patterns of ice layers: a partitioned spatial graph construction strategy that forms overlapping, fully connected local neighborhoods to preserve spatial coherence and suppress noise from irrelevant long-range links, and a long-range skip connection mechanism within the transformer that improves information flow and mitigates oversmoothing in deeper attention layers. We conducted extensive experiments, demonstrating that GRIT-LP outperforms current state-of-the-art methods with a 24.92% improvement in root mean squared error. These results highlight the effectiveness of graph transformers in modeling spatiotemporal patterns by capturing both localized structural features and long-range dependencies across internal ice layers, and demonstrate their potential to advance data-driven understanding of cryospheric processes.

Problem

• Addresses oversmoothing and weak long-range dependency in graph transformers
• Improves ice layer thickness prediction from polar radar imagery
• Enhances spatiotemporal pattern modeling for cryospheric process understanding

Innovation

• Partitioned spatial graph construction strategy
• Long-range skip connection mechanism
• Inductive geometric graph learning framework