Evaluation of a New Global Precipitation Analysis at the U.S. Air Force 557th Weather Wing

In summer 2019, the United States Air Force (USAF) upgraded their operational global precipitation analysis to a new algorithm developed by NASA Goddard Space Flight Center. This algorithm employs the Bratseth successive correction scheme to merge global NWP data with rain gauge reports and satellite (IR- and microwave-based) rainfall retrievals. The Bratseth method approximates the classic Optimal Interpolation approach while avoiding computationally expensive direct matrix inversion. The precipitation analysis is coupled with the NASA Land Information System (LIS) run operationally by the 557th Weather Wing, with output available at approximately 10-km resolution across the entire globe. A “reanalysis” has also been generated by NASA from available USAF input data beginning in November 2007. In this study, we evaluate the Bratseth analysis against independent datasets, and compare with the previous Air Force operational product. We also evaluate two USAF satellite retrieval datasets (GEOPRECIP and CMORPH) currently ingested into the Bratseth scheme, and a new candidate (IMERG) produced by NASA using data from the GPM Constellation. Our goal is to illustrate the improvements gained by using Bratseth in operations, and to point toward future improvements by assimilating IMERG.