Evaluating the applicability of 2DH models for the prediction of mega nourishments

Abstract

In 2003 certain areas of the Dutch coast were found not to meet contemporary safety standards, as a result of erosion that occurred over the last century. One of these 'weak links' stretches from Camperduin to Petten and is known as the Hondsebossche Pettemer Zeewering; a former 'hard' sea dyke with a stone revetment. To fortify this 12 kilometres coastal area construction of a dune, beach and shoreface nourishment began in 2014.

A 2DH model, FINEL2D, was used to predict the development of the zone for twenty years to estimate the maintenance. However, significant deviations in sand volume losses were found when comparing measurements and model predictions. These differences may affect maintenance budgets. Especially in the Nature Zone of the HD, where the contractor must pay a fine if maintenance is needed within the first ten years after construction.

To evaluate the consequences of the deviations on the maintenance prognosis of the coming years, a model is required which predicts the volume losses of the Hondsbossche Dunes with sufficient accuracy. The goal of this research is to evaluate the applicability 2DH models to simulate volume losses of a mega nourishments such as the Hondsbossche Dunes.

An analysis of the data available for the first and second year after construction is performed to provide insight into the causes of the deviations between the measured and modelled volume losses. It focusses on finding the effect of initial processes. Furthermore, it aims to define the relative importance of cross and longshore transport. Afterwards, the model performance is evaluated by means of a hindcast, using brute forcing. Together with the data analysis this information is used to improve future predictions of 2DH models.

From the data analysis it can be concluded that volume losses occurring in the first year are significantly larger than those in the second year. This difference cannot be explained by the difference in occurring wave climate and coastline orientation alone, it can have multiple other explanations; winnowing, settlement of nourished sediment, settlement of the subsoil, sediment losses beyond the dunes, and losses to deeper water. The significant reduction of the volume losses in the second year indicate that initial effects are less pronounced in this year. This indicates that the losses in longshore direction of second year after construction should be predictable. It is also found that aeolian transport is an important process in both years.

The evaluation of the model performance showed higher model accuracy for the second year after construction than for the first year. The modelled volume losses are compared to the losses in longshore direction based on measurements. The second year models the volume losses reasonably well, especially when looking at the Nature Zone, dominated by erosion.

Some aspects that could further improve the model are analysed using a strongly simplified model. Those aspects are assumed to increase the sediment transport and thereby the sediment available for sedimentation in the recreation zone north. Only using a smaller grid size around the bars seems to strongly enhance higher sediment transport. The other aspects considered are the interpolation between SWAN and FINEL2D, the amount of swan calculation per tidal cycle, the change in the cross-shore profile due to lack of 3D processes in 2DH models, and the morphological acceleration factor, none of which significantly affected total sediment transport.

It is found that a 2DH model is applicable in a situation where initial effects are no longer present. In case of a first year after construction, a factor could be applied to account for initial effects. For the second year, the model accuracy is quite good when comparing it to measured losses assumed in longshore direction. Yet an important factor to the losses is aeolian transport which is not accounted for in a 2DH model.

For future projects, it is important to realise that initial effects can significantly affect volume losses, and should be accounted for in some way. Another aspect that is important to take into account when designing a mega nourishment is the contribution of aeolian transport, which is quite constant during the first and second year after construction. A third aspect to keep in mind is the way the model results are interpreted.