Extreme Weather Nowcasting via Local Precipitation Pattern Prediction

2026-02-05

AI Summary

This study addresses key challenges in nowcasting extreme weather, including its highly localized nature, structural complexity, and the limitations of existing methods—either computationally expensive or biased toward moderate precipitation—alongside the absence of balanced datasets encompassing both ordinary and extreme rainfall events. To this end, the authors propose exPreCast, an efficient deterministic framework that integrates a local spatiotemporal attention mechanism, a texture-preserving bicubic dual upsampling decoder, and an adjustable temporal extraction module. Furthermore, they introduce the first balanced radar dataset from the Korea Meteorological Administration (KMA). Evaluated on SEVIR, MeteoNet, and the new KMA dataset, exPreCast achieves state-of-the-art performance, significantly enhancing forecast accuracy and reliability for both ordinary and extreme precipitation while maintaining real-time operational feasibility.

Technology Category

• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Computer Vision: Multi-modal Vision
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Accurate forecasting of extreme weather events such as heavy rainfall or storms is critical for risk management and disaster mitigation. Although high-resolution radar observations have spurred extensive research on nowcasting models, precipitation nowcasting remains particularly challenging due to pronounced spatial locality, intricate fine-scale rainfall structures, and variability in forecasting horizons. While recent diffusion-based generative ensembles show promising results, they are computationally expensive and unsuitable for real-time applications. In contrast, deterministic models are computationally efficient but remain biased toward normal rainfall. Furthermore, the benchmark datasets commonly used in prior studies are themselves skewed--either dominated by ordinary rainfall events or restricted to extreme rainfall episodes--thereby hindering general applicability in real-world settings. In this paper, we propose exPreCast, an efficient deterministic framework for generating finely detailed radar forecasts, and introduce a newly constructed balanced radar dataset from the Korea Meteorological Administration (KMA), which encompasses both ordinary precipitation and extreme events. Our model integrates local spatiotemporal attention, a texture-preserving cubic dual upsampling decoder, and a temporal extractor to flexibly adjust forecasting horizons. Experiments on established benchmarks (SEVIR and MeteoNet) as well as on the balanced KMA dataset demonstrate that our approach achieves state-of-the-art performance, delivering accurate and reliable nowcasts across both normal and extreme rainfall regimes.

Problem

• extreme weather nowcasting
• precipitation forecasting
• spatial locality
• balanced dataset
• real-time application

Innovation

• local spatiotemporal attention
• texture-preserving upsampling
• deterministic nowcasting
• balanced radar dataset
• extreme weather prediction