P.H.P. Overes1,2*, B.W. Borsje1 , A.P. Luijendijk2,3, S.J.M.H. Hulscher1
1 University of Twente, Netherlands; 2 Deltares, Netherlands, 3 Delft University of Technology, Netherlands
* Corresponding author: p.h.p.overes@utwente.nl
Introduction
Tidal sand waves are found at the bed of sandy seas throughout the world. These large-scale, dynamic bed forms can grow up to 25% of the water depth and migrate up to tens of meters per year. Their migration and shape deformation can lead to significant bed level changes, which pose a threat to offshore constructions, such as windfarms. To facilitate the major increase in offshore activities, due to among others upscaling of green energy production, predictions of future bed levels on decadal timescales are necessary. Currently, these predictions are often based on historic sand wave dynamics from measurement data. However, these methods are not able to account for extreme events and human interventions and rely on the availability of (historic) data. Using numerical models more understanding of these systems can be gained and more processes can be included in the (uncertainty of) bed level predictions. Two attempts at calibrating a numerical model to reproduce sand wave characteristics (Campmans et al., 2022) and dynamics (Krabbendam et al., 2021), have resulted in unrealistic heights and shape deformation of sand waves respectively. To arrive at more accurate model results, an approach based on system understanding is necessary.
Objective and Methods
The goal of this study is to identify the relevance of morphological parameters and processes for in-situ sand wave dynamics in different environments. To this end, two case study models, with contrasting sand wave characteristics and dynamics, are set up in the Dutch North Sea. The 2DV morphological Delft3D FM model, which is based on Overes et al. (2024), is used to hindcast past sand wave evolution. The hydrodynamics are extracted from the Dutch Continental Shelf Model and include tidal and non-tidal currents in the period between the two measurements. No morphological acceleration is used, which is a first in sand wave modelling. The model is used to test the importance of slope induced transport for sand wave characteristics and dynamics, by varying the related input parameter. The resulting sand wave morphology is compared to the measurements and assessed based on three criteria: height, leading slope and migration rate. Using these criteria more detailed driving mechanisms for sand wave dynamics are revealed. Moreover, the use of the RMSE or BSS in evaluating morphological skill, which are often used for model calibration and favor smooth solutions (Bosboom et al., 2018), is circumvented.
Results
Slope induced transport was found to be an important process for sand wave morphology. When overestimated, this effect can lead to major shape deformations through increased diffusion of sediment. Contrary to the generally accepted slope factor (αbs) of 3, the simulations showed that the use of the default value of 1 resulted in better maintenance of sand wave steepness in both areas, while only leading to limited growth of the sand waves. This sand wave growth manifested through a decrease in trough levels. This may indicate that other processes, such as armouring of the trough, which are often excluded from sand wave models, are limiting the growth. In previous studies the role of slope-induced transport may thus have been overestimated.
In this research, for the first time, in-situ sand wave dynamics are studied based on brute-force hydrodynamics, without morphological upscaling. Combining this more realistic set-up with the multicriteria assessment of morphological results, valuable insights are gained into the exact effects of morphological processes in the field. This in turn allows for a knowledge-based improvement of model set-up. By limiting the importance of slope induced transport, more accurate predictions of in-situ sand wave morphodynamics on multi-year timescales, were achieved.
Bed level from measurements and model results with varying importance of slope induced transport at A) a location in the HKZ windfarm zone and B) a location to the west of Texel. A higher bed slope parameter (αbs) indicates larger influence of bed slope induced transport.
References
Bosboom, J., & Reniers, A. (2018). The deceptive simplicity of the Brier skill score. In Handbook of Coastal and Ocean Engineering (pp. 1639-1663). https://doi.org/10.1142/9789813204027_0058
Campmans, G.H.P., van Dijk, T.A.G.P., Roos, P.C., Hulscher, S.J.M.H., 2022. Calibration and validation of two tidal sand wave models: a case study of The Netherlands continental shelf. J. Mar. Sci. Eng. 10 (12), 1902. https://doi.org/10.3390/jmse10121902.
Krabbendam, J.M., Nnafie, A., de Swart, H.E., Borsje, B.W., Perk, L., 2021. Modelling the past and future evolution of tidal sand waves. J. Mar. Sci. Eng. 9 (10), 1071. https://doi.org/10.3390/jmse9101071.
Overes, P. H. P., Borsje, B. W., Luijendijk, A. P., & Hulscher, S. J. M. H. (2024). The importance of time-varying, non-tidal currents in modelling in-situ sand wave dynamics. Coastal Engineering, 104480. https://doi.org/10.1016/j.coastaleng.2024.104480
