Refining the Technical Components for the Second Generation CMORPH: LEO-IR Based Precipitation Estimates and Cloud Motion Vector

Abstract:

A prototype system was developed for the second generation CMORPH to produce global analyses of 30-min precipitation on a 0.05^olat/lon grid over the entire globe from pole to pole through integration of information from satellite observations as well as numerical model simulations. The second generation CMORPH is built upon the Kalman Filter based CMORPH algorithm of Joyce and Xie (2011). Inputs to the system include rainfall and snowfall rate retrievals from passive microwave (PMW) measurements aboard all available low earth orbit (LEO) satellites, precipitation estimates derived from infrared (IR) observations of geostationary (GEO) as well as LEO platforms, and precipitation simulations from numerical global models. Key to the success of the 2^nd generation CMORPH, among a couple of other elements, are the development of a LEO-IR based precipitation estimation to fill in the polar gaps and objectively analyzed cloud motion vectors to capture the cloud movements of various spatial scales over the entire globe.

The prototype algorithm for the LEO IR precipitation estimation is refined to achieve improved quantitative accuracy and consistency with PMW retrievals. AVHRR IR TBB data from all LEO satellites are first remapped to a 0.05^olat/lon grid over the entire globe and in a 30-min interval. AVHRR TBB - precipitation relationships are separately established through PDF calibration of the TBB data against temporally/spatially collocated combined PMW retrievals (MWCOMB) and against the CloudSat radar measurements, respectively, then combined using a weighted mean to reflect the strengths of both data sources.

A sub-system is developed to construct analyzed fields of cloud motion vectors from the GEO/LEO IR based precipitation estimates and the CFS Reanalysis (CFSR) precipitation fields. Motion vectors are first derived separately from the satellite IR based precipitation estimates and the CFSR precipitation fields. These individually derived motion vectors are then combined through a 2D-VAR technique to form an analyzed field of cloud motion vectors over the entire globe. Error functions are developed to best reflect the performance of the satellite IR based estimates and the CFSR in capturing the movements of precipitating cloud systems over different regions and for different seasons.