Rainfall Retrieval Algorithm over Indian Land and Oceanic Regions Using Trmm Data

The Tropical Rainfall Measuring Mission (TRMM) satellite is a joint US-Japanese mission to explore tropical rainfall and its effects on the earth's energy budget, general circulation, and climate. In the present study, a method has been developed here for rainfall retrieval over Indian land and oceanic regions from remotely sensed microwave (MW) brightness temperature (BT) data obtained from the TRMM Microwave Imager (TMI) and the near-surface rainfall rate from the Precipitation Radar (PR). Artificial Neural Network (ANN) and Multiple Regression (MR) technique both have been applied for rainfall retrieval over Indian Ocean and land regions. It has been found that, over ocean, ANN model with TMI and PR data as inputs and output field vectors works more consistently than MR. On the other hand, MR performs well with selected predictor features like scattering index (SI), polarization correction temperatures (PCT) and the BTs difference (T19V-T37V) over land. The rainfall-rate retrieved from both the techniques is also compared with the TMI surface rain rate based on Goddard Profiling (GPROF) Algorithm [1]. Instantaneous precipitation estimates demonstrated correlations of 0.69 to 0.91 with GPROF rain rate for independent datasets over land while 0.54 to 0.61 over oceanic regions. Developed algorithms have also been successfully applied for various case studies during northeast and southwest monsoon over the area of study.