Abstract : 2E.4
Small-scale variability of orographic precipitation in the Alps: Case studies and idealized simulations for the National Park Berchtesgaden area

Günther Zängl
guenther@meteo.physik.uni-muenchen.de
Meteorological Institute, University of Munich

This study investigates small-scale precipitation variability in the National Park Berchtesgaden area, a high-alpine region with an extraordinarily dense station network. The data analysis is complemented by high-resolution (600 m grid spacing) MM5 simulations for two summertime heavy-precipitation cases (10-12 July and 15-17 August 2005) and idealized flow configurations with various wind directions. Both cases were related to Adriatic cyclones, leading to predominantly northeasterly flow at Alpine crest level (700 hPa). This differs markedly from the wind directions contributing most to climatological precipitation in this region (W, NW and N), and an analysis of the observed precipitation field reveals that the small-scale variability differs from the climatological pattern as well. In particular, there is a valley station that received more than twice as much precipitation as the surrounding stations in both cases (250 and 190 mm/48h, respectively), while the average annual precipitation (about 1600 mm) ranges within the spread of the surrounding stations. The observed rainfall distribution is reproduced quite well by the numerical simulations in both cases. The outstanding values of the above-mentioned station, which look fairly questionable at first sight, turned out to be related to the local precipitation enhancement (seeder-feeder mechanism) of an adjacent mountain range, which is advected towards the valley by the ambient winds. The simulations with idealized flow conditions reveal that the small-scale precipitation variability can be well reproduced with synthetic large-scale fields, provided that wind speed and direction and the height of the freezing level are specified to be in rough agreement with the observed conditions. Changing the wind direction leads to a rapid degradation of the correlation between simulated precipitation fields and station data.