Validation of KNW atlas with scatterometer winds (Phase 3 of KNW project)

The KNW (KNMI North Sea Wind) atlas is based on the ERA-Interim reanalyses dataset which captures 35 years (1997-2013) of meteorological measurements and generates 3D wind fields consistent with these measurements and the laws of physics. This dataset is “downscaled” using the state-of-the-art weather forecasting model, HARMONIE with a horizontal grid of 2.5 km. The result is a high resolution dataset of 35 years: the KNW dataset. In this report the KNW winds at 10 m above mean sea level are validated againsttwo datasets of scatterometer measurents: the 10 m wind product derived from the ocean surface winds measured by the Seawinds scatterometer instrument on board the polar orbiting QuikSCAT satellite and the ASCAT (Advanced Scatterometer) coastal product of the EUMETSAT OSI-SAF (Ocean and Sea-Ice Satellite Application Facility).

This report deals with the validation of the KNW wind atlas in the horizontal dimension with measurements valid at 10 m above the sea surface. In another report we validate the atlas in the vertical dimension against measurements made on 3 tall masts in the North Sea (Stepek et al, 2015). The reports are identical from the introduction up to and including section 1.1.3. The KNW dataset that is validated horizontally has not been corrected for the model’s underestimation of the vertical shear of the horizontal wind speed (since this has little effect on the wind speed at 10 m) whereas the vertically validated dataset has been corrected. The vertically validated dataset falls just within the area bounded by 50.25-54.75 NB and 1.50-8.25 EL (publicly available from the middle of 2015) whereas the horizontally validated dataset covers the whole of the North Sea. To avoid confusion: the KNW dataset is different from the North Sea wind climatology described in KNMI TR343 (Wijnant et al, 2014). Both the KNW dataset and the climatology of TR343 are based on re-analyses model ERA-Interim, but the “downscaling” procedures used are different. The KNW dataset followed up the wind climatology of TR343 so quickly that the TR343 dataset was never made available to a wider public.

• The probability of 10 m wind speeds of > 10 m/s occurring along the Dutch west coast is correctly modelled by the KNW atlas and more than half the wind energy produced by wind turbines on the North Sea is generatedat these wind speeds. The atlas overestimates the probability by less than 5% in English coastal waters and 3% or less further offshore. These conclusions are based on comparison with both QuikSCAT and ASCAT scatterometer measurements. Comparison with the vertical validation results indicates that the conclusions for the 10 m wind speed may also be valid in general terms at wind turbine hub height.

• The KNW atlas 10 m wind speeds are on average 0.3-0.4 m/s too high for most of the North Sea. For the southern part (including the wind energy areas Borssele and Hollandse Kust) the KNW atlas underestimates the 10 m wind speed by 0.1-0.3 m/s and probably slightly more. The comparison with both scatterometers supports the first conclusion. The second is based on the comparison with the QuikSCAT measurements (as there is only one year of ASCAT measurements and in that area the ASCAT measurements may be contaminated by the influence of the anchor areas). Comparison with the results of the vertical validation against tall measurement masts indicates that it cannot be assumed that the overestimation found at 10 m height means that KNW also overestimates the wind speed at wind turbine hub height.

• The KNW atlas does not make use of HARMONIE at its full potential since the development of small-scale spatial structures, starting from smooth ERA-Interim fields, is still ongoing 6 hours into the forecast and KNW is based on the first 6 hours of the HARMONIE forecasts.