WLS Report 48: Analyses and calibration of weather radar measurement values by hyetograph measurements for brief rainfall events, on behalf of the BMLF
Abstract
Release mechanism of events in torrent catchment areas are usually brief heavy rainfalls. Since the spatial density of the precipitation measuring-stations in alpine areas is very small, the analysis of weather radar pictures could offer a supplement to the measuring-stations. It should be clarified, which qualitative and quantitative significance precipitation data - measured by weather radar - of brief strong precipitation events in torrent catchment areas have and which procedures may be used to improve the geostatistical estimation of the spatial precipitation distribution. For this purpose torrent events of the years 1993 ¿ 1997, which are all released by heavy rainfall events, were detected. For the hydrologic description of these rainfall events it is necessary to filter out the characteristic values of 20 precipitation stations. Problems with the weather radar pictures occurred by the rain intensity classes, which are rapidly increasing in class-width to the upper intensity values, as well as the missing accessibility to the original reflection data. For the geostatistical estimation of the spatial precipitation distributions different Kriging methods were used. The estimation of the precipitation heights results in partially realistic values, whereby rough deviations in some cases occur. In future improvements of the results can be expected by more accurate distribution of the weather radar values. In practice weather radar data could be used for the event documentation and accordingly for the advanced warning by reaching prognosticated threshold values.
keywords flood warning system weather radar
Publikationen
Project staff
Johannes Hübl
Univ.Prof. Dipl.-Ing. Dr.nat.techn. Johannes Hübl
johannes.huebl@boku.ac.at
Tel: +43 1 47654-87111
Project Leader
28.12.1998 - 31.05.2001
BOKU partners
External partners
BOKU Vienna, Institute for Mathematics and Statisitcs
none
partner