How Effective Are Time-Series Models for Rainfall Nowcasting? A Comprehensive Benchmark for Rainfall Nowcasting Incorporating PWV Data

2025-09-27

AI Summary

This paper addresses the lack of dedicated benchmarks and effective models for precipitation nowcasting (0–3 hours). To this end, we introduce RainfallBench—the first large-scale, global meteorological time-series benchmark—comprising five years of observations from over 12,000 GNSS stations and integrating six variable types, including precipitable water vapor (PWV). It explicitly models key challenges: zero-inflation, temporal decay, and non-stationarity. We propose the Bi-Focus Precipitation (BFP) module, which synergistically handles sparsity and time-varying dynamics via meteorological priors. We conduct systematic evaluation across 20+ time-series models and six architectural paradigms, employing a novel multi-scale and extreme-precipitation–oriented evaluation protocol. Experiments reveal fundamental limitations of existing methods; BFP achieves substantial accuracy gains—especially for extreme events and longer-horizon forecasts. Both RainfallBench and BFP are practical, scalable, and designed to advance reproducible research in operational precipitation nowcasting.

Technology Category

• Machine Learning: Time-Series/Data Streams
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Rainfall nowcasting, which aims to predict precipitation within the next 0 to 3 hours, is critical for disaster mitigation and real-time response planning. However, most time series forecasting benchmarks in meteorology are evaluated on variables with strong periodicity, such as temperature and humidity, which fail to reflect model capabilities in more complex and practically meteorology scenarios like rainfall nowcasting. To address this gap, we propose RainfallBench, a benchmark designed for rainfall nowcasting, a highly challenging and practically relevant task characterized by zero inflation, temporal decay, and non-stationarity, focused on predicting precipitation within the next 0 to 3 hours. The dataset is derived from five years of meteorological observations, recorded at 15-minute intervals across six essential variables, and collected from more than 12,000 GNSS stations globally. In particular, it incorporates precipitable water vapor (PWV), a crucial indicator of rainfall that is absent in other datasets. We further design specialized evaluation strategies to assess model performance on key meteorological challenges, such as multi-scale prediction and extreme rainfall events, and evaluate over 20 state-of-the-art models across six major architectures on RainfallBench. Additionally, to address the zero-inflation and temporal decay issues overlooked by existing models, we introduce Bi-Focus Precipitation Forecaster (BFPF), a plug-and-play module that incorporates domain-specific priors to enhance rainfall time series forecasting. Statistical analysis and ablation studies validate the comprehensiveness of our dataset as well as the superiority of our methodology. Code and datasets are available at https://anonymous.4open.science/r/RainfallBench-A710.

Problem

• Evaluating time-series models for short-term rainfall prediction
• Addressing rainfall nowcasting challenges with zero inflation and temporal decay
• Incorporating PWV data to improve precipitation forecasting accuracy

Innovation

• Introducing RainfallBench benchmark with PWV data
• Designing specialized evaluation for extreme rainfall events
• Proposing Bi-Focus module for zero-inflation and decay