RainBench: Towards Global Precipitation Forecasting from Satellite Imagery
This work provides a new benchmark dataset and tools for researchers to develop data-driven precipitation forecasting models, which could help mitigate the impact of extreme precipitation events globally, especially in developing regions.
The paper introduces RainBench, a new multi-modal benchmark dataset for global precipitation forecasting, which includes simulated satellite data, ERA5 meteorological data, and IMERG precipitation data. They also release PyRain, a library for efficient processing of large precipitation datasets, and establish baseline results for two medium-range precipitation forecasting tasks.
Extreme precipitation events, such as violent rainfall and hail storms, routinely ravage economies and livelihoods around the developing world. Climate change further aggravates this issue. Data-driven deep learning approaches could widen the access to accurate multi-day forecasts, to mitigate against such events. However, there is currently no benchmark dataset dedicated to the study of global precipitation forecasts. In this paper, we introduce \textbf{RainBench}, a new multi-modal benchmark dataset for data-driven precipitation forecasting. It includes simulated satellite data, a selection of relevant meteorological data from the ERA5 reanalysis product, and IMERG precipitation data. We also release \textbf{PyRain}, a library to process large precipitation datasets efficiently. We present an extensive analysis of our novel dataset and establish baseline results for two benchmark medium-range precipitation forecasting tasks. Finally, we discuss existing data-driven weather forecasting methodologies and suggest future research avenues.