VeinCast: Physics-Guided Dynamic Field Graphs with Graph-Conditioned Fusion for Global Medium-Range Weather Forecasting

2026-08-10

AI Summary

This work addresses the challenge of modeling structured, state-dependent interactions among heterogeneous atmospheric fields in global medium-range weather forecasting by proposing the VeinCast framework. VeinCast constructs inter-field relationships through a physics-guided dynamic field graph and introduces a graph-conditioned latent variable fusion mechanism that integrates data-driven learning with physical constraints while preserving field-specific information. Key technical components include Earth-window attention, Top-K residual edges, and node centrality–weighted aggregation. Evaluated on the 1.5° ERA5 benchmark, VeinCast achieves or surpasses the performance of leading models—including Pangu-Weather and GraphCast—across all 69 surface and upper-air meteorological variables throughout the 14-day forecast horizon.

Abstract

Global medium-range weather forecasting requires modeling structured yet state-dependent interactions among heterogeneous atmospheric fields. Existing data-driven models largely learn these interactions implicitly, whereas equation-level physical constraints may inherit approximation and model-form biases. We present VeinCast, a physics-guided dynamic field graph and graph-conditioned fusion framework that jointly forecasts 69 surface and upper-air fields. Within each local window, its Physics-Guided Dynamic Field Graph combines predefined atmospheric relations with state-dependent Top-K residual edges and adapts Earth-window attention using the resulting graph context. Graph-Conditioned Latent Fusion further employs graph context and source-node centrality to guide field-to-latent aggregation, while bounded feedback preserves field-specific information. On the $1.5^\circ$ ERA5 benchmark, VeinCast demonstrates competitive forecasting performance across all 69 meteorological fields at lead times of up to 14 days, compared with representative global weather forecasting models including FuXi, Pangu-Weather, GraphCast, FengWu, and ARROW. Ablations confirm that the two modules provide complementary gains, demonstrating the effectiveness of relational-level physical guidance for data-driven weather forecasting.

Problem

• global medium-range weather forecasting
• heterogeneous atmospheric fields
• state-dependent interactions
• physics-guided modeling
• data-driven weather forecasting

Innovation

• Physics-Guided
• Dynamic Field Graph
• Graph-Conditioned Fusion
• Medium-Range Weather Forecasting
• Relational-Level Physical Guidance