Towards a Spatiotemporal Fusion Approach to Precipitation Nowcasting

2025-05-25

AI Summary

To address the challenge of fusing heterogeneous meteorological data sources (ground stations, ERA5, and GFS) for nowcasting precipitation, this paper proposes a spatiotemporally consistent multi-source observation fusion framework. We introduce the STConvS2S architecture—previously unexplored in this domain—and design a dynamic weighting strategy based on quantified station-network contribution. Furthermore, we establish a novel collaborative inference paradigm integrating GFS numerical weather prediction (NWP) outputs with ground-based observations. Evaluated on 1-hour lead-time heavy precipitation (>25 mm/h) forecasting in the Rio de Janeiro metropolitan area, our method achieves an F1-score of 0.2033, substantially outperforming single-source baselines. Ablation studies confirm complementary gains from the three station systems. Key contributions include: (i) an end-to-end spatiotemporal fusion driven by STConvS2S; (ii) a data-driven dynamic weighting mechanism; and (iii) an NWP-observation joint inference paradigm.

Technology Category

• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

With the increasing availability of meteorological data from various sensors, numerical models and reanalysis products, the need for efficient data integration methods has become paramount for improving weather forecasts and hydrometeorological studies. In this work, we propose a data fusion approach for precipitation nowcasting by integrating data from meteorological and rain gauge stations in Rio de Janeiro metropolitan area with ERA5 reanalysis data and GFS numerical weather prediction. We employ the spatiotemporal deep learning architecture called STConvS2S, leveraging a structured dataset covering a 9 x 11 grid. The study spans from January 2011 to October 2024, and we evaluate the impact of integrating three surface station systems. Among the tested configurations, the fusion-based model achieves an F1-score of 0.2033 for forecasting heavy precipitation events (greater than 25 mm/h) at a one-hour lead time. Additionally, we present an ablation study to assess the contribution of each station network and propose a refined inference strategy for precipitation nowcasting, integrating the GFS numerical weather prediction (NWP) data with in-situ observations.

Problem

• Integrating multi-source meteorological data for precipitation nowcasting
• Improving heavy rainfall prediction accuracy using deep learning
• Assessing station network contributions via ablation study and fusion

Innovation

• Spatiotemporal deep learning for data fusion
• Integrates ERA5 reanalysis and GFS prediction
• Ablation study refines precipitation nowcasting strategy