At Remote Sensing Systems we currently run a physically-based algorithm that uses the passive microwave emission signal to estimate rain rate over the ocean (Wentz and Spencer, 1998). This algorithm is not Bayesian and offers a competing estimate of rain rate to Goddard Profiling Algorithm (Kummerow et al., 2001). We use the same algorithm to estimate rain rate from the SSM/I, TMI, and AMSR sensors. We have recently improved our algorithm by using more realistic rain column heights and by incorporating saturation and footprint resolution effects into our beamfilling correction. We propose three major components of an investigation that will improve our algorithm, improve our understanding of the space-time properties of precipitation variations, and improve climate models through validation of convective parameterizations. We will perform the following work: (1) Intercalibrate the rain estimates from the different passive microwave sensors on the different satellites available to us: 6 SSM/I, 2 AMSR, and TMI. It is important to remove the remaining small differences before moving forward. (2) Study and implement use of scattering information and the 85 GHz channel in the algorithm. Recent work (Hilburn et al., 2006a) has suggested that scattering information may be used to improve light rain versus cloud detection. We will study scattering signals at both 37 GHz and 85 GHz in more detail from both a theoretical standpoint and by examination of data. Also, we will make 85 GHz Polarization Corrected Temperature (PCT) available in our gridded binary data products. (3) Perform a statistical study of the properties of temporal and spatial variability of rainfall in the passive microwave dataset. These results are useful in validating and improving convective parameterizations in climate models.
This work will have the following benefits: (1) An intercalibrated 20-year record of passive microwave rain estimates that we will use for detection of trends in the hydrological cycle. (2a) The availability of 85 GHz PCT in our gridded binary data products will satisfy the requests of many of our data users in the tropical meteorology community. (2b) Improved light rain versus cloud discrimination. (3a) Understanding temporal variability will be crucial for our ability to detect trends in our multi-satellite dataset. (3b) Understanding temporal and spatial variability of passive microwave rain rates complements our previous infrared work (Ricciardulli and Sardeshmukh, 2002) and will help in our assessment and improvement of convective parameterizations in climate models.
