Recorded presentation6B.3
A Study of Comparison of Reflectivity Using WSR-88D and Disdrometers
Aaron C. Ferrel, Texas A&M University, College Station, TX; and A. Tokay, L. Carey, C. Schumacher, and D. B. Wolff
Radar rainfall estimation has been widely used in hydrological models that are employed in flood forecasting. The multi-sensor precipitation estimate (MPE), which is generated by the NOAA River Forecast Centers, relies heavily on radar (WSR-88D) rainfall estimation. Assuming that the radar is well calibrated and without partial beam filling, the time height ambiguity between the radar measurement aloft and surface rainfall remains one of the major uncertainties in radar rainfall estimation. This study investigates the differences in reflectivity between the radar and disdrometer observations, employing a unique data set where two impact-type disdrometers were located at 19 and 80 km from Houston WSR-88D. The comparisons were made using nine adjacent radar pixels and the average of nine pixels where the center corresponds to the disdrometer site. The disdrometer data, on the other hand, was averaged over three-, five-, and seven-minutes from the radar scanning time. The comparison of the time series of reflectivity between radar and disdrometer provides information on the visual agreement between the two measurements. The scatter diagram and the rainfall statistics such as correlation, bias, absolute bias, and standard deviation quantifies the agreement. We adopted both methods and obtained the statistics for each event invidiually and from combining all the events. We have also determined the behavior of rainfall statistics at different reflectivity ranges. It should be noted that the Houston WSR-88D calibration was evaluated through comparison of TRMM precipitation radar measurements, and we included radar scans where all nine pixels were reporting reflectivity.
Session 6B, Validation of Remotely-Sensed Hydrometeorological Observations—II
Wednesday, 14 January 2009, 10:30 AM-11:45 AM, Room 127C
Previous paper Next paper