Improving Greenland Bed Topography Mapping with Uncertainty-Aware Graph Learning on Sparse Radar Data

2025-09-10

AI Summary

To address low reconstruction accuracy in Greenland’s subglacial bed topography mapping caused by sparse and irregular radar sounding data, this paper proposes GraphTopoNet—a novel uncertainty-aware graph neural network framework. It constructs a spatial graph structure integrating multi-source remote sensing data, including surface elevation, ice velocity, and mass balance. For the first time in graph neural networks, Monte Carlo Dropout is incorporated to quantify predictive uncertainty. Additionally, a confidence-weighted hybrid loss and a gradient-enhanced graph feature module are designed, coupled with dynamic regularization for robust modeling. Experiments across three Greenland subregions demonstrate that GraphTopoNet reduces root-mean-square error by up to 60% compared to conventional interpolation and state-of-the-art deep learning methods. This significantly improves the fidelity of bed topography reconstruction. The resulting high-resolution bed topography maps have been integrated into climate modeling for sea-level change projections and policy assessment.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• 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: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies
• Web Mining and Content Analysis: Web data integration and cleaning

Abstract

Accurate maps of Greenland's subglacial bed are essential for sea-level projections, but radar observations are sparse and uneven. We introduce GraphTopoNet, a graph-learning framework that fuses heterogeneous supervision and explicitly models uncertainty via Monte Carlo dropout. Spatial graphs built from surface observables (elevation, velocity, mass balance) are augmented with gradient features and polynomial trends to capture both local variability and broad structure. To handle data gaps, we employ a hybrid loss that combines confidence-weighted radar supervision with dynamically balanced regularization. Applied to three Greenland subregions, GraphTopoNet outperforms interpolation, convolutional, and graph-based baselines, reducing error by up to 60 percent while preserving fine-scale glacial features. The resulting bed maps improve reliability for operational modeling, supporting agencies engaged in climate forecasting and policy. More broadly, GraphTopoNet shows how graph machine learning can convert sparse, uncertain geophysical observations into actionable knowledge at continental scale.

Problem

• Mapping Greenland's subglacial topography from sparse radar data
• Handling data gaps and uncertainty in geophysical observations
• Improving accuracy of bed topography for sea-level projections

Innovation

• Graph-learning framework fuses heterogeneous supervision and uncertainty
• Spatial graphs augmented with gradient features and polynomial trends
• Hybrid loss combines confidence-weighted supervision with balanced regularization