Extended Abstract (300K)P1.20
Over-ocean rainfall retrieval from TRMM/TMI data during the Typhoon season
Jen-Chi Hu, Chung Cheng Institute of Technology and NDU, Tao-Yuan, Taiwan; and W. J. Chen, G. R. Liu, M. D. Tsai, and M. H. Chang
A new algorithm was developed in this study to estimate quantitative precipitation over the oceans of Taiwan during the Typhoon season for alleviating the disasters caused by the heavy rainfall accompanying with Typhoon. A total of five-year, from 2000 to 2004, TRMM/TMI (Tropical Rainfall Measuring Mission/ TRMM Microwave Imager) and rain gauge data are used to create a rainfall retrieval algorithm using multiple linear regression statistic technique. There are two steps in the rainfall retrieval algorithm. The first one is the rain and no-rain pixel recognition and the SI (scattering Index) technique by Ferraro et al. (1994) was accepted to distinguish rain pixel from no-rain one in this study. The second one is to classify the rain type to be convective or stratiform. Following the step 2 both the convective and stratiform rainfall estimate can be obtained, respectively. The results show that the overall rain pixel recognition rates are 95.0%, 99.6% and 99.1% for 2002-2004, respectively. In validation, the rain gauge is regarded as ground truth. It was shown that the Root-Mean-Square(RMS) error of the satellite rainfall retrieval is 3.75 mm/hr and the coefficient of correlation between them is 0.74. It also shows that the satellite rainfall retrieval is overestimated for weak precipitation system and underestimated for severe precipitation system. The possible reason for this situation is caused by beaming problem of the field of view. Therefore, the high resolution infrared channel data of TRMM/VIRS were used in this study to reduce the RMS error. The standard satellite rainfall retrievals, 2A12, of GPROF (Goddard Profiling Algorithm) , retrieved by physical method and the regression satellite rainfall retrievals by Chen and Li (2000) are also compared with our results and the comparison shows that our algorithm is better than the above two results. The GPROF method was proved suitable for global scale only but not for regional one. In addition, the regression relationship obtained from a certain season data can't be extended to another season. Therefore, it is necessary for us to develop a new relationship between the satellite data and rain gauge for the Typhoon case. AMSR-E(Advanced Microwave Scanning Radiometer - EOS)of AQUA satellite and AMSU(Advanced Microwave Sounding Unit)data of NOAA satellite are planning to include in our algorithm for improving the temporal and spatial resolutions and raise the operational usage of satellite rainfall retrievals.
Key words: quantitative precipitation, rainfall retrieval, microwave channel, linear regression, TMI, VIRS, AMSR-E
Supplementary URL: http://140.132.23.240
Poster Session 1, Retrievals and Cloud Products
Monday, 30 January 2006, 2:30 PM-2:30 PM, Exhibit Hall A2
Previous paper Next paper