SWAN uncertainties bases on addition Lake IJssel cases WTI 2017 - Cluster 1: Hydraulic Loads

In general, uncertainties can be divided into inherent uncertainties and knowledge uncertainties. It is the intention that in WTI-2017 besides inherent uncertainties, also knowledge uncertainties are to be implemented, divided into statistical and model uncertainties. Regarding waves on wider rivers, lakes and coasal areas, the model uncertainty concerns SWAN uncertainty. In Deltares (2013b) the model uncertainties are assessed for SWAN in coastal waters and lakes. The present study focusses on the wider rivers and is an extension of Deltares (2013a). Additional cases of autumn 2013 (storms of 27/28 October 2013 and 4/5 December 2013) are included for a better estimation of the model uncertainties. Apart from this, the additional storms are also interesting to include in the growth curve analysis of Deltares (2013a), stretching the reach of observed dimensionless fetches in the direction of more extremes (smaller dimensionless fetches due to larger wind velocities). There are 28 additional momentsfor FL47, 59 for FL48 and 65 for FL49. The additional points show the same trends as the original selection of Deltares (2013a). So - based on scaling with U10 - dimensionless wave heights at the near shore locations FL48 and FL49 are above the standard growth curves and for FL47 the values are rather on top of the growth curves. For the growth curves scaled with u, the additional points come in general closer to the standard Kahma & Calkoen growth curve test. At FL47 the additional higher waves – plotted below the standard growth curve - are likely to be influenced by the limited water depth. Eight cases are added to the twenty SWAN hindeast cases on which the SWAN uncertainties are based. Five of the additional eight cases concern conditions which are more severe than the original cases of FL48 and FL49. Especially case I (October 28th) is very high with an observed significant wave height at FL48 of 1.11 m and at FL49 of 0.87 m. The SWAN performance of the new cases is in line with the other cases, namely underestimation of significant wave height whilst the wave period is simulated quite well.