For the derivation of the Hydraulic Boundary Conditions (HBCs), information on extreme
winds over open-water areas is required. To this end, a new method is developed that will
answer the need for a description of not only the storm’s peak impact but also the spatial
and temporal characteristics of extreme storms. The method relies on using high-resolution atmospheric model simulations rather than on using spatial interpolation of sparse point measurements of wind speed. The HARMONIE model, which has a grid spacing of 2.5 km, has been selected to perform the simulations.
The present report is the final report of Work Package 1 (WP1) of the WTI2017 - Wind
Modelling project. WP1 describes the assessment of the high-resolution model, in
particular the spatial and temporal structures of the modelled wind fields. To this end,
HARMONIE simulations of 16 historical storms were performed. Based on a verification
with the available wind measurements, we conclude that the wind fields produced by the
model are realistic in time and space. The temporal correlation between modelled and
observed wind speed is 0.95 over sea. The model represents spatial characteristics of the
storms well. Over sea, modelled wind speeds have a positive bias of about 0.5 m/s; for
most stations the root mean square (rms) difference is between 1.5 and 2.0 m/s. The bias
in wind direction is a few degrees, the rms difference is in the order of 10° for winds of 8
Bft and higher. For these high-wind conditions, modelled wind speeds over Lake IJssel are
generally too high. Overall, the modelled wind speeds over the lake are 5-10% higher than
observed. Over land, extreme wind speeds are underestimated.
The added-value of high-resolution modelling is most pronounced in coastal areas and for
large inland water bodies like Lake IJssel. In those areas, wind and surface-stress values
of models with a coarser resolution become less accurate. This is shown in a comparison
of HARMONIE wind fields with output from the ERA-Interim model dataset, which has a
resolution of approximately 80 km. Storm surge calculations with the WAQUA model
indicate that the water levels predicted from HARMONIE output are consistently closer to
observed water levels than when output from the much coarser resolution ERA-Interim
model is used. To avoid mismatches between drag relations used by the different models,
we use the surface stress to drive the WAQUA model rather than the wind speed.
Overall, we rate the model results as good. As such, we conclude that, although at some
points further evaluation can be done, the model is suitable to be used as it is in Work
Packages 2 and 3 of the Wind Modelling project and that it provides a strong basis for
further use in the WTI program.
Peter Baas. Final Report of WP1 of the WTI2017-HB Wind Modelling project
KNMI number: WR-14-02, Year: 2014, Pages: 72