Improving Predictions of Convective Storm Wind Gusts through Statistical Post-Processing of Neural Weather Models

2025-03-31

AI Summary

Neural weather models (NWMs) suffer from substantial biases and poor extreme-value calibration in convective gust forecasting. To address this, we propose a regionalized, multi-scale statistical–deep learning hybrid post-processing framework tailored to Switzerland’s five major geographical regions, delivering probabilistic hourly gust forecasts for lead times of 0–72 hours. Our approach innovatively embeds Generalized Extreme Value (GEV) distribution constraints into the post-processing pipeline to ensure probabilistic consistency for extreme wind events. Additionally, we design a lightweight CNN module to explicitly capture spatial dependencies while maintaining computational efficiency and high spatial resolution. Evaluated on Pangu-Weather outputs, our method achieves significant improvements over the raw NWM baseline: Brier scores for probabilistic gust forecasting decrease by 18%–26%. This substantially enhances both the reliability and timeliness of early warnings for extreme wind events.

Technology Category

• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Reasoning under Uncertainty: Relational Probabilistic Models
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data
• Systems and Infrastructure for Web, Mobile and WoT: Data management and stream processing for Web, mobile and wireless applications

Abstract

Issuing timely severe weather warnings helps mitigate potentially disastrous consequences. Recent advancements in Neural Weather Models (NWMs) offer a computationally inexpensive and fast approach for forecasting atmospheric environments on a 0.25{deg} global grid. For thunderstorms, these environments can be empirically post-processed to predict wind gust distributions at specific locations. With the Pangu-Weather NWM, we apply a hierarchy of statistical and deep learning post-processing methods to forecast hourly wind gusts up to three days ahead. To ensure statistical robustness, we constrain our probabilistic forecasts using generalised extreme-value distributions across five regions in Switzerland. Using a convolutional neural network to post-process the predicted atmospheric environment's spatial patterns yields the best results, outperforming direct forecasting approaches across lead times and wind gust speeds. Our results confirm the added value of NWMs for extreme wind forecasting, especially for designing more responsive early-warning systems.

Problem

• Improving wind gust predictions using neural weather models
• Applying statistical post-processing for severe storm forecasts
• Enhancing early-warning systems with deep learning methods

Innovation

• Hierarchical statistical and deep learning post-processing
• Convolutional neural network for spatial pattern analysis
• Generalised extreme-value distributions for robust forecasts