Assessing the Impact of Regime-Dependent Biases on Climate Variability/Trends from a Radiometer Constellation

As the transition is made from TRMM to GPM the focus of algorithm development efforts will change from a single satellite covering the tropics to a global perspective employing a constellation of radiometers on board a diverse set of satellites. In a similar manner we propose to expand/build upon prior research efforts, which have focused on identifying algorithm assumptions leading to regime-dependent rainfall biases by examining differences between rainfall estimates from the TRMM active (PR) and passive microwave (TMI) sensors, to address biases among a diverse constellation of radiometers. This will involve examining issues such as the orbit of the satellite, including the TRMM orbit boost, the characteristics of the sensor, and, as is the case with TRMM, assumptions in the retrieval algorithm. Using the existing constellation of radiometers, including TRMM as the reference satellite, we propose to investigate differences between rainfall estimates from SSM/I on board DMSP F8, F10, F11, F13, F14, and F15, SSMIS on board DMSP F-16, AMSR-E on board EOS Aqua, and WindSat. Multi-year overlaps with TRMM will provide the basis for comparisons, which will take advantage of information from the PR and the precession of the TRMM orbit throughout the diurnal cycle. While the latest version 6 TRMM retrievals both indicate an increase in tropical mean rainfall associated with the 1997/98 El Niņo, there is still a discrepancy in the magnitude of this increase between the PR and TMI estimates as well as between the TMI and SSM/I values. Further research to resolve these differences is important not only to the current rainfall record, but also to ensure the accuracy of long-term merged rainfall datasets through the GPM era.