Wind Power Prediction using Ensembles

The Ensemble project investigated the use of meteorological ensemble fore-casts for the prognosis of uncertainty of the forecasts, and found a good method to make use of ensemble forecasts. This method was then tried based on ensembles from ECMWF in formof a demo application for both the Nysted offshore wind farm and the whole Jutland/Funen area. The utilities used these forecasts for maintenance planning, fuel consumption estimates and over-the-weekend trading on the Leipzig power exchange. Othernotable scientific results include the better accuracy of forecasts made up from a simple superposition of two NWP provider (in our case, DMI and DWD), an investigation of the merits of a parameterisation of the turbulent kinetic energy within thedelivered wind speed forecasts, and the finding that a “naïve” downscaling of each of the coarse ECMWF ensemble members with higher resolution HIRLAM did not improve the error scores or the result space enough to warrant the computational effort.