hide
Free keywords:
-
Abstract:
A high resolution (1 km × 1 km) data set of monthly wind velocities over Germany for the time period 1951-2001 is provided. The data have been reduced to a reference level using a so called 'relative altitude' scheme, interpolated using a simple Inverse Distance Weighting approach and retransformed to the actual topography. However, no parametrization of land use or surface roughness has been integrated in the modelling process of this data and therefore the data is not suitable to serve as criteria for planning wind energy sites. A Cross-Validation scheme applied to this data set yields a mean error of 0.1 m/s for the time period 1951-2001. Regarding the area mean of wind velocities a linear trend of −0.05 m/s is obvious for this period. This negative linear trend changes to a positive one when shorter time scales are considered, e.g. +0.3 m/s for the 1981-2001 period. However, all these temporal trends are not significant. Thus, they could be the result of random features within the dataset and are not further interpreted. These monthly derived wind velocities serve as a reference data set for regional climate model evaluations. The climate models used are two different versions of the hydrostatic regional climate model REMO as well as the nonhydrostatic CLM and MM5 models. All models are capable to reproduce the temporal and spatial variability of the observations to a great extent. Projections of changes in wind velocity have been carried out with these regional climate models. All of these projections show a significant increase in wind velocities over the full model domain, especially over the Baltic Sea and the North Sea, during winter and a decrease during summer. Regarding changes in annual means an increase of up to 1.0 m/s for the Baltic Sea and a decrease in wind velocitites of the same magnitude for the Mediterranean is projected as an average for the 2070-2099 period.
