1.3 Analysis of the New Version 7 TMPA Real Time Retrospectively Processed Precipitation Record

Monday, 7 January 2013: 2:15 PM
Ballroom E (Austin Convention Center)
David T. Bolvin, SSAI, Greenbelt, MD; and G. J. Huffman, E. J. Nelkin, R. Adler, and E. Stocker

The TRMM Multi-Satellite Precipitation Analysis (TMPA) is a merged satellite precipitation product produced in both near-real time for precipitation monitoring purposes and two months after real time for research purposes. The research TMPA has the advantage of including gauge analyses and uses the high-quality TRMM Combined Instrument (TCI) estimates as the calibrator. Since neither the gauge nor the TCI is available in real time, we use the TRMM Microwave Imager (TMI) estimates as the calibrator for the real-time TMPA. As well, we apply a two-step monthly climatology calibration to the real-time TMPA to mimic the statistics of the research TMPA to strive for consistency.

Recently, the TMPA was upgraded to Version 7 (V7) to take advantage of newer and better sources of input satellite estimates (and gauge analyses for the research TMPA). The V7 research TMPA record extends from January 1998 – (delayed) present. The V7 real-time TMPA began initial processing in May 2012 with data availability starting on June 1, 2012. Unlike previous versions, in V7 we undertook a retrospective processing of the real-time TMPA record starting in March 2000 and extending through May 2012 to obtain a nearly-complete record of the real-time TMPA estimates. This retrospective processing was performed at the request of users who require a long record of consistent real-time estimates to perform calibrations crucial for monitoring precipitation.

This long record of V7 real-time TMPA gives us an unprecedented opportunity to compare its performance relative to the V7 research TMPA. Previous comparisons of the Version 6 (V6) real-time TMPA with the V6 research TMPA showed that the two-step monthly climatological calibration was neutral with respect to the bias but largely improved the root-mean-square difference statistic. Detailed results of the V7 real-time and V7 research TMPA comparison will be provided to assess the quality of the two-step monthly calibration and to give users a sense of the error associated with the V7 real-time TMPA. As an example, the time series of the statistics varies as particular satellites start recording data, drift in overpass time, and stop recording data.

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner