Radar data quality control: Evaluation of several algorithms based on accumulated rainfall statistics

The primary function of the TRMM Ground Validation (GV) Program is to create rainfall products that provide a basis for evaluation of satellite-derived precipitation measurements for selected sites within the tropics. A fundamental and extremely important step in creating high-quality GV products is radar data quality control. Quality control (QC) processing of TRMM GV radar data is based on some automated procedures, but the current QC algorithm is not fully operational and requires significant human interaction to assure satisfactory results. Moreover, the TRMM GV QC algorithm, even with continuous manual tuning, still can not completely remove all types of spurious echoes. In an attempt to improve the current operational radar data QC procedures of the TRMM GV effort, an intercomparison of several QC algorithms has been conducted. The results of this effort, however, are valuable beyond the TRMM community and beneficial to any operational efforts dealing with large amounts of radar data.

This presentation will demonstrate how various QC algorithms affect accumulated radar rainfall products. In all, six different QC algorithms will be applied to one month of WSR-88D radar data from Melbourne, Florida, as well as several specific spurious echo case studies from various other sites. Instantaneous, daily, and monthly accumulated radar rainfall statistics will be presented for each quality-controlled data set. The QC algorithms will be evaluated and compared based on their ability to remove spurious echoes without removing significant precipitation. Strengths and weaknesses of each algorithm will be assessed based on their skills in mitigating erroneous rainfall accumulation from spurious echo contamination and removal of true precipitation, respectively. Contamination from individual spurious echo categories will be quantified to assist evaluation of the abilities of each radar QC algorithm. Finally, a cost-benefit analysis will be conducted to determine if a more automated QC algorithm provides a viable alternative to the current, labor-intensive QC algorithm employed by TRMM GV.