Abstract. Extreme wind events are among the most damaging weather-related hazards in the Czech Republic, forestry is heavily affected. In order to successfully run a landscape model dealing with such effects, spatial distribution of extreme wind speed statistics is needed. The presented method suggests using sector-wise wind field calculations together with extreme value statistics fitted at a reference station. A special algorithm is proposed to provide the data in the form expected by the landscape model, i.e. raster data of annual wind speed maxima.

The method is demonstrated on the area of Bohemian Forest that represents one of largest and most compact forested mountains in Central Europe. The reference meteorological station Churáňov is located within the selected domain. Numerical calculations were based on linear model of WAsP Engineering methodology. Observations were cleaned of inhomogeneity and classified into convective and non-convective cases using index CAPE. Due to disjunct sampling of synoptic data, appropriate corrections were applied to the observed extremes. Finally they were fitted with Gumbel distribution. The output of numerical simulation is presented for the windiest direction sector. Another map shows probability that annual extreme exceeds required threshold.

The method offers a tool for generation of spatially variable annual maxima of wind speed. It assumes a small limited model domain containing a reliable wind measurement. We believe that this is typical setup for applications similar to one presented in the paper.