Abstract
This report describes the results from an uncertainty analysis of the Hydra_k and SWAN models, giving an estimate of the level of accuracy of the Hydraulic Boundary Conditions (HBC) in the Wadden Sea. Specifically, the goal was to quantify the contribution of the SWAN model and Hydra-K model input parameters to the overall uncertainty in the HBC and to determine the quantitative contributions of model inputs and model parameterizations to the uncertainty in the SWAN results. These results can then be used to assist in the improvement of the methodology.
The uncertainty analysis of the HBC was done by using a Monte Carlo sampling technique. Separate Monte Carlo simulations were used for different parts of the calculation associated with different sources of error. The first part involved the SWAN calculation. The uncertainties associated with the SWAN parameters and input were found to be approximately 17% both for the near shore wave height and the wave period at shallow water locations. From comparison with observations (hindcasts), it was concluded that there is an additional uncertainty in the SWAN outcomes as a result of uncertainties in the SWAN. This uncertainty is responsible for a considerable systematic error.
The second part involved the Hydra-K computation of HBC. The uncertainty in three Hydra-K inputs, namely wind speed, water level and results from the SWAN model, contribute to the uncertainty of the Hydra-K output (the HBC). The overall uncertainty in the HBC, expressed as a critical crest level, was found to be between 90 and 100 cm. Although this uncertainty is considerable, it should not lead to different assessment results, because the assessment is based on the expectation value, or best estimate. The results of the uncertainty analysis can, however, be used as guidelines for further improvement of the methodology.
Authors
