Consistency check of TRMM rainfall estimates using a radiative transfer model

The difference between TMI and PR estimates is drastically narrowing in Version 6 compared with Version 5. However, certain evidences to support the claim that the TRMM version-6 rain estimates are better than version-5 estimates has not yet be obtained. In this study, consistency between TMI-observed brightness temperatures and those simulated from PR2A25 and TMI2A12 rain profiles are investigated for ITCZ rain systems during the 1998 ENSO event, using a radiative transfer model. We focus on brightness temperature at 10.65 GHz channels which provide total liquid water path estimates.

The better agreement between the observed and simulated brightness temperatures is obtained for PR 2A25 and TMI 2A12 version 6 than those for version 5. Simulated brightness temperature (TBs) from PR2A25 V6 is higher than that from PR2A25 V5, and exhibits better agreement with observed brightness temperature (TBo), especially for higher values (associated heavy rainfall). This is probably explained by the fact that the attenuation due to cloud water that increases as rainfall rate is taken into account in the Version 6. It is nevertheless obvious that TBs from PR2A25 V6 is low. If the effects of cloud water and melting layer that increase brightness temperature are taken into account in the calculation, the better agreement between TBs from PR2A25 V6 and TBo may be achieved. Although TBs from TMI 2A12 V6 also exhibit better agreement with TBo than that from TMI2A12 V5 does, there is an excess of TBs from TMI 2A12 V6 over TBo, especially for the higher range. The excesses of TBs from TMI2A12 V6 over TBo are due to strong ice scattering (85-GHz), which is not strongly correlated to surface rainfall. These results suggest that efforts to improve both PR 2A25 and TMI 1B11 will be required.