Numerous algorithms for quantitative precipitation estimation (QPE) from satellites have been developed during the past quarter century. The approaches and data used by these algorithms has varied widely, from manual techniques to artificial intelligence algorithms, relating data from the visible, infrared, and/or microwave portions of the electromagnetic spectrum to rainfall rates at the Earth's surface. However, direct comparison of these algorithms has been hampered by the use of different data sets and different evaluation criteria.

At the Office of Research and Applications (ORA) of the National Environmental Satellite, Data, and Information Service (NESDIS), several operational and experimental satellite QPE algorithms are being compared in near-real time over portions of the continental United States. The initial comparisons, begun 1 April 2001, compared three versions of the Auto-Estimator, two versions of the GOES Multi-Spectral Rainfall Algorithm (GMSRA), and a technique that use microwave-based rain rate estimates to calibrate an IR-based algorithm. Additional algorithms will be added to the comparison as resources and logistics permit. Comparisons are being made between 6-h satellite rainfall estimates and corresponding Stage III radar/raingauge fields, and also between 24-h satellite rainfall estimates and corresponding raingauge data. Initial findings of this study will be presented.