LangPrecip: Language-Aware Multimodal Precipitation Nowcasting

TL;DR

LangPrecip introduces a novel language-aware multimodal framework for precipitation nowcasting, integrating textual and radar data to improve forecasting accuracy for extreme weather events.

Contribution

It presents a new semantically constrained trajectory generation approach and a large-scale multimodal dataset for improved weather prediction.

Findings

Achieves over 60% improvement in heavy-rainfall CSI at 80-minute lead time.

Demonstrates consistent performance gains on Swedish and MRMS datasets.

Abstract

Short-term precipitation nowcasting is an inherently uncertain and under-constrained spatiotemporal forecasting problem, especially for rapidly evolving and extreme weather events. Existing generative approaches rely primarily on visual conditioning, leaving future motion weakly constrained and ambiguous. We propose a language-aware multimodal nowcasting framework(LangPrecip) that treats meteorological text as a semantic motion constraint on precipitation evolution. By formulating nowcasting as a semantically constrained trajectory generation problem under the Rectified Flow paradigm, our method enables efficient and physically consistent integration of textual and radar information in latent space.We further introduce LangPrecip-160k, a large-scale multimodal dataset with 160k paired radar sequences and motion descriptions. Experiments on Swedish and MRMS datasets show consistent improvements over state-of-the-art methods, achieving over 60 \% and 19\% gains in heavy-rainfall CSI at an 80-minute lead time.