Investigation of Rainfall Characteristics Using TRMM PR and Ground Based Radar

Despite relatively good agreement between reflectivity profiles, comparisons of rainfall statistics derived from TRMM Precipitation Radar (PR) deviate from ground-based radar (GR) observations in various field locations across the globe. TRMM PR rain rate probability distribution functions underestimate the occurrence of high rain rates (> 80 mm hr-1) compared with similar ground-based statistics, and similarly, GR distributes the total rain volume over a larger range of rain rates. Analysis of ten years of TRMM data over three field sites has shown that the greatest disagreements occur in the most intense convection, such as over land and during the east and break wind regimes over the Amazon and Australia, respectively. These differences are investigated further in this study. Ten years of TRMM PR data are analyzed in conjunction with data collected during two field experiments involving the NCAR S-Pol radar. S-Pol was deployed in Brazil in the Amazon during TRMM LBA in 1998-1999 and near Mazatlan, Mexico as part of the North American Monsoon Experiment (NAME) in 2004. Additionally, multiple years of data from the Australian Bureau of Meteorology CPOL radar located in Darwin, Australia, are examined to extend the robustness of the GR observations beyond the relatively short field campaigns. Polarimetric data collected by the two radars are used to characterize the differences between TRMM PR and GR observations as a function of bulk hydrometeor type. For example, profiles with significant graupel, as identified by GR, are analyzed to investigate the role of mixed phase in the PR retrievals. The vertical variability of D0 is examined as a function of reflectivity and related to the underlying microphysical conditions using the polarimetric data provided by the GR observations. Spatial variability of D0 is also explored by correlating D0 values derived from GR at different heights. Several significant changes were made to the TRMM processing algorithms in the latest release of the data, version 7 (V7). Among other modifications, the attenuation correction algorithm was adjusted to improve retrievals over land. V6 and V7 reflectivity profiles and rain statistics are compared with ground based radar observations from the three field sites in order to understand the impact of the algorithm changes. Differences in rain rate retrievals between V6 and V7 as a function spatial and temporal regimes, such as land and ocean (NAME and Darwin) and large scale wind regimes (LBA and Darwin), are investigated.