Comparison between precipitation estimates of ground-based weather radar composites and GPM's DPR rainfall product over Germany

Autor: Velibor Pejcic, Pablo Saavedra Garfias, Kai Mühlbauer, Silke Trömel, Clemens Simmer
Jazyk: angličtina
Rok vydání: 2020
Předmět:
Zdroj: Meteorologische Zeitschrift, Vol 29, Iss 6, Pp 451-466 (2020)
Druh dokumentu: article
ISSN: 0941-2948
DOI: 10.1127/metz/2020/1039
Popis: We compare more than three years (between 2014 and 2018) of precipitation estimates over Germany from the Dual-frequency Precipitation Radar (DPR) operating on the core satellite of the Global Precipitation Mission (GPM) with the radar-derived precipitation product RADOLAN RY provided by the German national meteorological service (DWD, Deutscher Wetterdienst). Incomplete overlap between the observation volumes due to the different scan geometries and inconsistencies related to the mutually assumed hydrometeor phases lead to large differences, when directly comparing DPR's near surface product (DPRns) with RADOLAN RY. We improve the correspondence between both data sets via two steps. First, we derive an adjusted DPR near surface product (DPRans) extracted from the DPR vertical profiles, that best fits to the scans height and beam width of the surface radar observations. Second, the data is classified into liquid, solid and mixed phases by adjusting hydrometeor phase classification (aHPC) to the RADOLAN scan geometry. With these steps the correlation between both data sets increases from r = 0.49 to r = 0.61 and the RMSD is reduced from 2.94 mm/h to 1.83 mm/h for the commonly observed precipitation, exceeding most of the results found in previous studies. The agreement is best in stratiform precipitation (r = 0.68, RMSD = 1.4 mm/h), for only stratiform and summer season (r = 0.7, RMSD = 1.59 mm/h), and for stratiform with only liquid precipitation (r = 0.67, RMSD = 1.58 mm/h). Unlike the the standard DPRns, the new DPRans product exhibits almost no seasonal differences in the capability of detection; for all seasons the CSI is 0.94 and the FAR/IFAR are 0.04/0.02.
Databáze: Directory of Open Access Journals