Evaluation of a new polarimetrically-tuned Z-R relation

A new polarimetrically-tuned Z-R estimator of the formZ = aR is evaluated where the coefficient ‘a’ is continuously estimated based on polarimetric radar data of Zdr and Kdp. Six storm events from the TRMM/LBA (Brazil) and TEFLUN-B (Florida) are analyzed using radar data from the NCAR/SPOL radar and a network of gages especially deployed for these two measurement campaigns. For storm total accumulation, the new ‘pol-tuned’ Z-R method gives a normalized bias of 6% and normalized standard error of about 20% (normalized bias is defined as < RZR − Rg > / < Rg > here; normalized standard error is defined as < |RZR − Rg| > / < Rg >; Rg is the gage measurement of rain rate, and RZR is the rain rate estimation from the Z-R estimator; angle brackets denote average values). The method continuously ‘tracks’ the evolution of the drop size distribution and so no classification of rain types is necessary. In contrast, the use of conventional Z-R method (after stratiform-convective rain separation) results in a normalized bias of around 18% for storm-total accumulation and normalized standard error of 24%.