SatRain: A machine-learning-ready benchmark dataset for satellite precipitation estimation

The Satellite Precipitation Retrieval (SatRain) benchmark dataset developed by the International Precipitation Working Group (IPWG) is a benchmark dataset for machine-learning-based satellite precipitation retrievals, i.e., algorithms for detecting and quantifying precipitation from satellite imagery.

Features

• Training, validation, and test splits derived from four years of overpasses of passive-microwave sensors over the conterminous united states (CONUS)
• Collocated satellite observations from visible, infrared, and microwave sensors
• Comprehensive ancillary data covering atmospheric parameters and surface properties
• Independent testing datasets derived from different regions and measurement techniques
• Flexible evaluation framework

]

Retrieved precipitation from three SatRain-trained retrievals applied to observations of Typhoon Khanun during landfall on the Korean peninsula. Each retrieval relies on a different type of input from the SatRain dataset: a single IR window channel (panel d), Himawari-9 observations (panel e), and GMI observations (panel f). The results are compared with reference ground-based radar measurements (panel a) and baseline estimates from ERA5 (panel b) and GPROF V7 (panel c).

Documentation

For instructions on how to get started using the dataset refer to the documentation available here.