Verification of Rainfall Estimates for the Goddard Profiling Algorithm (GPROF)

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner
Wednesday, 5 February 2014
Hall C3 (The Georgia World Congress Center )
Kathryn Sauter, Virginia Tech, Blacksburg, VA; and J. Bytheway, C. Kummerow, and D. Randel

The Goddard Profiling Algorithm (GPROF) being developed for the Global Precipitation Mission (GPM) uses a Bayesian statistical approach to retrieve surface rainfall and precipitation vertical structure. The latest version of the algorithm, GPROF_2014, has focused on the development of more physical retrievals over land. By making use of all radiometer window channels instead of only the traditional 85 GHz channel, the algorithm attempts to retrieve precipitation directly from the brightness temperatures without the need for empirical screens currently used to distinguish precipitation from radiometerically cold surfaces such as standing water, snow, frozen ground and desserts. This approach is necessary for the GPM mission to homogenize the rain/no rain determination for the diverse radiometers being used by GPM. This project is a first quantitative assessment of the GPROF 2014 output that focuses on the algorithm's ability to first discriminate precipitating from non-precipitating areas and subsequently assess the quantitative precipitation rates. The study focuses on nine days in 2011 from the Special Sensor Microwave Imager/Sounder (SSMIS) on the Defense Meteorological Satellite Program, F16 satellite compared to surface radar observation from the National Mosaic and Multi-Senor Quantitative Precipitation Estimation (NMQ). Results indicate that the algorithm matches the location of liquid precipitation from NMQ quite well, but the intensity of heavy precipitation is significantly underestimated. The algorithm also misses some frozen precipitation. Even if improvements are made to the algorithm before launch, the current study offers a benchmark against which future improvements can be compared.