Determining of the ZDR offset and its impact on the hydrometeor classification of the CHUVA Project

The Brazilian space program is planning to launch in 2014 a satellite to measure precipitation, which will be part of the GPM (Global Precipitation Measurement) constellation program. Warm clouds are responsible for a large amount of the precipitation in the tropics, especially in coastal regions. This cloud type is little studied and is not considered in satellite rainfall retrievals. The CHUVA field campaigns will carry out conventional and special observations, in some strategical areas in Brazil, in order to create a meteorological database to give grants for cloud microphysical studies, especially of cloud evolution and its life cycle, to improve different precipitation estimation algorithms, to know better development of thunderstorms and lightning formation and processes in the boundary layer and to develop cloud microphysical modeling. The Polarimetric X band radar (XPOL) is the main instrument to measure precipitation of CHUVA Project. The radar was used in three field campaigns: Fortaleza, Northeast region (April 2011); Belem, North region (June 2011); and Vale do Paraiba, Southeast region (Nov 2011 to March 2012). Since the polarimetric variables depend mainly of difference between horizontal and vertical reflectivity channels, so it is mandatory that channels are well calibrated. However, calibration of these channels is not easy to achieve, especially those values lower than 0.1 dB. One of methodology to procedure of calibration is vertically-pointing, in this case azimuth scanning at 90 degree elevation to measure mainly the differential reflectivity (ZDR). Unfortunately by limitation of radar software, it was possible only to scanning at 89 degree elevation. Thus, the objective of this work is to calculate the offsets of differential reflectivity (ZDR) for those three campaigns to correct them and analyze their impacts on the hydrometeor classification algorithms. The bin resolution is 150 meters and vertically-pointing for each 6 minute. The first three bins were discarded to avoid clutter contamination as well as the bins inside and above melting layer. To ensure the roundness of the raindrops (small), the reflectivity values are limited to range between 5 and 25 dBZ. Once calculated ZDR offset, these were corrected and two hydrometeor classification algorithms, NXRED-Warm and MeteoFrance, both available at the Rainbow software (Selex/Gematronik radar package), were processed and their products are compared with previous results without ZDR offset correction. The results show a ZDR offset for Fortaleza is about -0.29 dB with standard deviation (std) for about 0.11 dB. For Belem, second campaign, the ZDR offset is about -0.56 dB and 0.21 dB for std. The period of Vale do Paraiba campaign was divided in three parts, the ZDR offsets are -0.27 (old radome), -0.33 (new radome) and -0.59 dB (calibration and replacement of some electronic components), respectively. Clearly there was a breakdown of the channel calibrations between Fortaleza campaign and Belem campaign. Maybe one reason is a bad condition of Fortaleza-Belem road, where the radar was transported for about 1.600 km. The top of radome suffered a little damage at Fortaleza campaign and was exchanged by a new one in December 2011, in Sao Paulo. This change is showed in the variation of ZDR offset (-0.27 to -0.33 dB). Due to the oscillation of electric energy and moisture infiltration in waveguides, performance fell more during the second period of the campaign, which led to revision of the electronic components. Even with the electronic calibration, the difference between the channels remained relatively high (-0.59 dB). Despite the difficulties, the methodology was applied and the results show a clear improvement in the hydrometeor classification, especially for rain-hail mixture, locating the exact site of occurrence and its intensity when compared with results without correction. Although the values of ZDR offset are small and perhaps negligible for horizontal reflectivity correction, the corrected ZDR is significant in hydrometeor classification algorithms.