WeatherGNN: Exploiting Meteo- and Spatial-Dependencies for Local Numerical Weather Prediction Bias-Correction

TL;DR

WeatherGNN is a novel graph neural network-based approach that improves local numerical weather prediction bias correction by capturing complex meteorological and spatial dependencies, leading to state-of-the-art results.

Contribution

It introduces a factor GNN and a hierarchical GNN guided by domain knowledge to effectively model dependencies in weather data.

Findings

WeatherGNN outperforms baselines with 4.75% lower RMSE.

The method effectively captures meteorological dependencies.

The approach demonstrates superior performance on real-world datasets.

Abstract

Due to insufficient local area information, numerical weather prediction (NWP) may yield biases for specific areas. Previous studies correct biases mainly by employing handcrafted features or applying data-driven methods intuitively, overlooking the complicated dependencies between weather factors and between areas. To address this issue, we propose WeatherGNN, a local NWP bias-correction method that utilizes Graph Neural Networks (GNNs) to exploit meteorological dependencies and spatial dependencies under the guidance of domain knowledge. Specifically, we introduce a factor GNN to capture area-specific meteorological dependencies adaptively based on spatial heterogeneity and a fast hierarchical GNN to capture dynamic spatial dependencies efficiently guided by Tobler's first and second laws of geography. Our experimental results on two real-world datasets demonstrate that WeatherGNN achieves the state-of-the-art performance, outperforming the best baseline with an average of 4.75 \% on RMSE.