14th Conference on Satellite Meteorology and Oceanography


Improvements of CMORPH resulting from limb adjustments and normalization of AMSU-B rainfall

Robert J. Joyce, RS Information Systems, Inc. and NOAA/NWS/NCEP/CPC, Boca Raton, FL; and R. R. Ferraro

Beam position bias exists in current NESDIS algorithm rainfall rates derived from NOAA spacecraft AMSU-B instruments. Unlike retrievals from the conical scanning SSMI, TMI, and AMSR-E passive microwave instruments, the 89 GHz field of view (FOV) from the cross-tracking AMSU-B scanner increases from ~16 km from near nadir beam positions to ~45 km at the limb. In order to compensate for the increasing earth coverage of limb beam rainfall retrievals relative to those of nadir, some damping of heavy and very light rain rates should occur. However, the relative non-existence of limb retrieval AMSU-B rainfall rates greater than 20 mm/hour does not agree with TRMM 2A-12 TMI rainfall averaged to the earth coverage of the nearest AMSU-B limb estimate. A passive microwave based rainfall algorithm such as CMORPH will inherently suffer problems when combining and propagating rainfall derived from vastly different frequencies, sampling characteristics, and estimation algorithms resulting in pronounced biases. However, the skill of the AMSU-B to estimate rainfall has been shown to be generally better than IR and most likely any other current geostationary satellite sensor. An improved "level 3" AMSU-B rainfall product is derived from matching AMSU-B rainfall with TRMM TMI derived precipitation. Histograms of AMSU-B rainfall are accumulated in small beam position groups, from heaviest to lightest rates, land and ocean separately, for different latitude regions. High resolution TMI rainfall (temporally/spatially coincident with the AMSU-B rainfall) is then "synthetically sampled/averaged" to the exact mapped rainfall region of each beam position within the variable FOV AMSU-B cross-track scan. From these distributions, normalization corrections are derived from frequency matching the histograms dependent on surface type, latitude region, and AMSU-B beam position group. Improvements of CMORPH ingesting the beam position normalized AMSU-B rainfall is then demonstrated as opposed to not utilizing the normalizations.

extended abstract  Extended Abstract (176K)

Poster Session 1, Retrievals and Cloud Products
Monday, 30 January 2006, 2:30 PM-2:30 PM, Exhibit Hall A2

Previous paper  Next paper

Browse or search entire meeting

AMS Home Page