Towards a concept of deriving wind statistics over inland waters : based on ensemble climate runs

The wind statistics presently used for the design and safety assessment of dikes and dunes in the Netherlands are based on measured time series with a length of approximately 40 years. Extrapolation to return periods of 100.000 years yields considerable statistical uncertainty. The ambition is to use long-term synthetic wind time series to derive wind statistics for the inland water bodies as Lake IJssel. However, the spatial resolution of these long-term timeseries like RACMO (2416 years available, 12 km x 12 km resolution) is too low. On the other hand, high resolution wind models like HARMONIE require too much computational effort to provide these long time series. The goal of the present study is to investigate whether a useful method can be derived to determine wind statistics above inland waters. Presently, for Lake IJssel a translation of the Schiphol wind statistics, based on observations over land, is made. First, we compare DOWA/HARMONIE data to observations during 18 storms to find out that the model performance at Lake IJssel is quite good (+4% bias). HARMONIE gives a slight underestimation for wind speeds below roughly 15 m/s and a slight overestimation for the higher wind speeds. Next, we compare for 28 (other) storms the RACMO and HARMONIE wind with each other. We use the RACMO data at one location in the center of Lake IJssel and compare it with various HARMONIE locations, including the location closest to the RACMO location. The ratio between RACMO and HARMONIE is used as correction factor to translate RACMO wind to HARMONIE quality and to various HARMONIE locations. The correction factors are between 0.9 and 1.1. For the RACMO location the factor is 1/0.932, meaning that the HARMONIE wind is 7% higher than the RACMO wind at that location. The extreme value analysis on 2416 years of original RACMO wind provides wind statistics that turn out to be lower than the actual wind statistics based on the 40 years of observations. Applying the correction factor on RACMO shifts the wind statistics closer to the actual statistics. The use of long-term synthetic wind data to derive wind statistics for hydrodynamic modelling for the safety assessment of dikes is promising.