Tropical rainfall-surface temperature relations using Tropical Rainfall Measuring Mission precipitation data

In this study, 9 years (1998-2006) of monthly precipitation data from Tropical Rainfall Measuring Mission (TRMM) are used to examine the relations between tropical rainfall and surface temperature using measurements from both passive and active microwave sensors. These relations are compared to those derived from Global Precipitation Climatology Project (GPCP) analyses. A technique is first developed to adjust the TRMM Precipitation Radar (PR) monthly rainfall data in the tropics (whole ocean and whole land) to account for the effect of the TRMM orbit boost from 350 to 402 km in August 2001. The postboost PR rainfall is adjusted by adding 6.5, 6.0, and 1.0% to the monthly PR rainfall data over the ocean at the estimated surface, the near surface, and the 2 km level, respectively. No adjustment is made for data over land or above the 4 km level. The relationships between the tropical rainfall and surface temperature are then examined with both the TRMM Microwave Imager (TMI) and adjusted PR data. Comparing tropical (25 degrees N-25 degrees S) ocean precipitation to mean sea surface temperature (SST) over the same area, the GPCP and TMI rainfall data have large and similar slopes (similar to 15%/degrees C) against ocean-wide SST anomalies, while the surface monthly rainfall anomalies derived from the TRMM PR exhibit much shallower slopes (similar to 4%/degrees C) against the SST anomalies. At the 4 km level the PR data exhibit a larger slope (12%/degrees C) comparable to the passive microwave value. Over the tropical land, all rainfall data except TRMM PR at 6 km have similar, but negative, slopes against surface temperature anomalies, in contrast to the positive slopes over the ocean. Over the total tropics (ocean and land combined), TRMM TMI and GPCP rainfall data have rather similar smaller positive slopes (6%/degrees C), when compared to ocean plus land surface temperature, but the PR rainfall data slopes are slightly negative, except at the 4 km level (4%/degrees C). Overall, the PR-based surface precipitation-temperature slopes do not confirm slopes based on passive microwave observations. This may be the result of PR retrieval error or inherent passive/active retrieval differences. Further research is needed to advance the use of TRMM data in this regard.