Machine Learning Bias Correction for Satellite Precipitation Data Products

  • 9 November 2021
Machine Learning Bias Correction for Satellite Precipitation Data Products

Researchers from Colorado State University and NOAA used machine learning to quantify and correct for the uncertainties associated with satellite precipitation products, especially in the case of rainfall over mountainous regions. Published in the IEEE Transactions on Geoscience and Remote Sensing (Early Access), the study by Haonan Chen, Luyao Sun, Robert Cifelli (NOAA/PSL), and Pingping Xie (NOAA/CPC) compares the precipitation products before and after bias correction using four independent precipitation events over the coastal mountain region in the western United States. Information about rainfall from satellites is limited by how the sensors pick up very light or very heavy precipitation. This project, funded in part by CPO's Climate Observations & Monitoring (COM) program, demonstrated the impact of the mountains on satellite-based precipitation retrievals and found that accuracy of NOAA’s precipitation product, CMORPH, was enhanced after applying the machine learning-based bias correction technique outlined in the study. This machine learning scheme could be used to improve other satellite precipitation products. 

Read the article »




1315 East-West Highway Suite 100
Silver Spring, MD 20910


Americans’ health, security and economic wellbeing are tied to climate and weather. Every day, we see communities grappling with environmental challenges due to unusual or extreme events related to climate and weather.