Overlap of Extreme Convective Intensities and Extreme Rain Rates from TRMM and WSR-88D Perspectives

Although extreme rain rates are often associated with convectively intense systems, the TRMM satellite precipitation radar (PR) 2A25 V7 retrieval shows that the overlap fraction between columns that have extreme convective intensity, as judged by the maximum height of the 40 dBZ echo, and those that have extreme low-level rain rate is quite low over the tropics and subtropics in regions of significant rainfall, whether “extreme” is defined as the 95^th (< 35% overlap), 99^th (< 20% overlap), or 99.5^th (< 15% overlap) percentiles of rain rate and maximum 40 dBZ height. However, because TRMM is a Ku-band radar, it suffers from significant attenuation in heavy precipitation, requiring attenuation correction to estimate low level reflectivity, which is used in reflectivity-rain rate (Z-R) relationships to derive rain rate. Additionally, the 2A25 retrieval needs to assume hydrometeor type as a function of altitude and has particular difficulty with situations in which hail is likely contaminating rain rate retrievals. To better understand the potential impacts of these weaknesses, 12 years of June-August TRMM PR retrievals are statistically compared with 1 year of June-August hourly WSR-88D dual-polarimetric S-band radar data for 28 radars over the southeastern United States after matching horizontal resolution and sampling as a function of geographical location between the two datasets and converting TRMM Ku-band measurements to S-band approximations. Additionally, the CSU-HIDRO algorithm is used to derive rain rate from WSR-88D horizontal reflectivity (Z), differential reflectivity (ZDR), specific differential phase (KDP), and hydrometeor identification (HID) retrievals.

WSR-88D confirms that the overlap fraction between extreme convective intensities and extreme rain rates is low, but also shows that this overlap fraction is likely significantly higher than TRMM retrievals indicate, up to ~50% higher in the southeastern United States (e.g., 24% as compared to 16% for the overlap of the 95^th percentiles of maximum 40 dBZ height and rain rate). Low level reflectivity as a function of maximum 40 dBZ echo height is ~2 dBZ higher for WSR-88D than TRMM, while derived rain rate for a given low level reflectivity between 40 and 50 dBZ is 15% higher, caused by the usage of KDP rather than Z in deriving many rain rates for this reflectivity range. These both contribute to a mean WSR-88D rain rate that is nearly double that of TRMM for a given maximum 40 dBZ height above 5-km altitude. For a given low level reflectivity, mean WSR-88D rain rate increases with increasing maximum 40 dBZ height, but TRMM derived rain rates do the opposite, while the low-level reflectivity difference between WSR-88D and TRMM only appears once the maximum 40 dBZ height penetrates into mixed phase and ice regions above 5-km altitude. With the exception of low level reflectivities greater than 55 dBZ, in which hail contamination detected by the WSR-88D HID causes a likely high bias in TRMM-retrieved rain rate, these findings indicate that the TRMM reflectivity attenuation correction process for intense convective systems may produce low-level reflectivities and rain rates that are biased low. Therefore, low-level reflectivity and rain rate profiles retrieved by TRMM in systems with a 40 dBZ echo exceeding 5-km altitude need to be interpreted with caution.