Abstract: Rapidly accumulating, large-scale and long-term meteorological data provide unprecedented opportunities for data-driven meteorological models and fine-grained numerical weather prediction.
Abstract: Graph Neural Networks (GNNs) are widely used across fields, with inductive learning replacing transductive learning as the mainstream training paradigm due to its superior memory efficiency, ...