Extended Abstract

3 rd Conference on QPE /QPF and Hydrology. World Meteorological Organization, Nanjing, China, Oct.18-­‐22,2010. Correction of Radar QPE Errors Associated with Partial Bright Band ObservationsNear the GroundYoucun Qi *1,2 , Jian Zhang 1 , Kenneth Howard 1 , David Kingsmill 31 NOAA/National Severe Storms Lab2 Nanjing University of Information Science and Technology3 NOAA/Earth System Research Lab1. IntroductionThe melting of aggregated snow/crystals oftenresults in an enhancement of the reflectivity observed byweather radars, and this is the so-called bright band(BB). The primary cause of the enhancement is a rapidincrease in the dielectric constant of hydrometeors at thetop of the melting layer followed by an increase of fallvelocities of melting snowflakes toward the end of themelting process. The locally high reflectivity causessignificant overestimation in radar quantitativeprecipitation estimates if no appropriate correction isapplied. When the melting layer is very close to theground, it often occurs that only a partial bright band canbe observed by the radar. Since no information isavailable below the BB can be obtained from radarobservations, the partial BB often leads to largeoverestimations in radar QPEs. To correct for theselarge errors, an empirical relationship of BB peak and BBbottom is developed from high-resolution precipitationprofiler data. . The empirical relationship is combinedwith the BB top and peak information from the apparentvertical profile of reflectivity (VPR) and an estimated BBbottom is derived. Radar QPEs are then corrected basedon the estimated BB bottom. The methodology has beentested for two events over the plain area in the UnitedStates. The method is shown to be very effective inreducing the overestimation errors, and the correctedradar rainfall estimates compared much better withgauge observations than those without the correction.*Corresponding author : Youcun Qi, 120 David L. Boren Blvd., Suite 2100 Norman, OK 73072-­‐7304; E-­‐mail: youcun.qi@noaa.gov 2. Methodologya. S-Band Precipitation Profiler Data AnalysisThe profiler data used in this study were obtainedfrom two S-band precipitation profiler radars deployedduring the National Oceanic and AtmosphericAdministration (NOAA) Hydro-meteorological Test-bed(HMT, http://hmt.noaa.gov) in Nov. 2005 to Apr. 2006.The radars were located in Cazadero (CZC) near thewest coast of California and in Alta (ATA) on theCalifornia Sierras. These profiler radars measure timeevolutions of the reflectivity structure along a verticalcolumn with high temporal (1min) and spatial (60 m)resolution (e.g., White et al, 2000, Matrosov et al, 2006).The data used in the current study was 5-min averagesof the 1-min observations.Total of 6336 vertical reflectivity profiles wereanalyzed in the current study. Among them, 2037 (605from CZC and 1432 from ATA) were identified as to havea low but complete bright band. The criteria for having alow and complete bright band are: 1) there was only onepeak below the 0°C height, where the 0°C height wasobtained from a nearby sounding; 2) the peak must bewell-defined in that an inflexion can be found both aboveand below the peak; 3) the inflexion points must be atleast 100 m away from the peak (i.e., the bright bandpeak is at least 200 m wide); 4) the peak reflectivity mustbe at least 6 dBZ higher than the reflectivities of the topand bottom; 5) the bottom of the peak must be less than500m above the radar height. From these profiles, thereflectivites of bright band peak (Z peak ) and bottom (Z bttm )were recorded. The BB bottom reflectivity is subjectively-451-