PhD Defence Pauline Overes | A Deep Dive into Sand Wave Dynamics | Using Process-Based Modelling and In-Situ Observations

A Deep Dive into Sand Wave Dynamics | Using Process-Based Modelling and In-Situ Observations

Pauline Overes is a PhD student in the department Water systems. (Co)Promotors are prof.dr. S.J.M.H. Hulscher and dr.ir. B.W. Borsje from the Faculty Engineering Technology (ET), University of Twente and dr.ir. A.P. Luijendijk from the Technische Universiteit Delft.

As offshore developments and activities increase at an unprecedented pace, shallow continental shelf areas face increasing pressure. These regions are often covered by dynamic sand waves that can reach 25% of the water depth and migrate up to tens of meters per year. The resulting seabed mobility threatens offshore infrastructure and activities. To reduce risks, decadal bed level predictions are required, which traditionally rely on data-driven approaches. However, these heavily depend on available data and are unsuitable for dealing with changing conditions, due to for example climate change or offshore wind development. Process-based numerical models could provide insights into the processes driving sand wave dynamics and help assess future bed levels, but previous simulations were often schematized and lack validation with in-situ observations. With recent advances in numerical models, we are increasingly able to capture complex processes, like the vertical current structures driving sand wave dynamics, with higher spatial and temporal resolution. This enables us to dive into the details of the processes driving sand wave dynamics in the field. By closing the gap between schematized models and in-situ observations, we can better explain and predict real sand wave dynamics. This thesis aims to increase our understanding of the detailed hydrodynamic, sediment transport and morphodynamic processes driving sand wave dynamics under natural and dredged conditions. To this end we apply the newly developed Delft3D Flexible Mesh (FM) model.

We start by analysing the effects of detailed hydrodynamic forcing on sand wave dynamics. Based on observations we see that sand wave shape and migration rate vary over time. However, available process-based models have not explained these variations, as they include only periodic tidal forcing and steady residual currents. We explore the importance of time-varying, non-tidal currents for sand wave dynamics in the North Sea. Using Delft3D FM we reconstructed these currents on top of periodic tidal forcing and validated the hydrodynamics using in-situ measurements. Compared to tide-only simulations, time-varying, non-tidal currents amplified sedimentation and erosion rates up to 15 times. Additionally, periods of net erosion occurred at locations where tidally forced models predicted only net sedimentation. These findings demonstrate that time-varying, non-tidal currents should be considered when predicting sand wave dynamics in the field.

Next to hydrodynamic forcing, sediment transport processes substantially influences the sand wave morphology. Sand wave models often struggle to preserve sand wave shapes, leading to smoothened crests and slopes and deepened troughs. Using Delft3D FM we demonstrate that the choice of sediment transport formulation significantly affects the stability of sand wave shapes. The widely used Van Rijn 1993 formulation predicts relatively high bed load transport rates, requiring more dominant bed slope-induced transport to limit growth. In contrast, the Van Rijn 2007 formulation predicts lower transport rates and better preserves steep sand wave slopes while limiting growth. The only way to stabilize crest levels is by decreasing the dominance of bed slope-induced transport, but even then trough levels still tend to lower slowly. This indicates that local processes limit sand wave growth and that the importance of slope-induced transport has been overstated in the past. With the adapted setup, the model more accurately represents sand wave evolution over multiyear timescales.

We continue with a challenging full-3D case: sand wave recovery and trench infill after dredging. Such dredged trenches are created to prevent cable exposure, but their evolution cannot be predicted using traditional data-driven methods. We apply a Delft3D FM model to reproduce measured trench recovery and analyse 3D interactions between the dredged area and surrounding sand waves. Results show that sand wave regrowth is driven by tide-averaged vertical current circulation, which is remarkably still present in the absence of sand waves in the trench due to the influence of adjacent non-dredged sand waves. Trench infill, on the other hand, is governed by slope-induced bedload transport, at these water depths. The infill volume rate depends on side-slope steepness, which varies over time and space with changing bathymetry. The highest infill volume rates are found between the surrounding crests, thereby also contributing to sand wave regrowth. The trench width was found to have negligible effect on infill rates. Tidal spring-neap variability primarily controls infill rates and storm-induced currents introduce asymmetry in the infill but do not significantly accelerate infill. Our uncalibrated simulations captured multiyear infill volumes well when compared to measurements, demonstrating their predictive capability for assessing recovery timescales and design choices.

Finally, we explore the ability to estimate sand wave migration based on insight in driving processes from large-scale models. Leveraging an extensive dataset of observed sand wave migration rates, with over 300,000 verified datapoints (around 17,000 km of crest length) across the Dutch Continental Shelf, environmental data and hydrodynamic models, we explore the ability of hydrodynamic drivers and sediment mobility to explain and quantify migration rates. Results demonstrate that tidal current asymmetry, quantified through the mean peak velocity difference, can provide a first-order estimate of migration rates, since these variables show a high linear correlation. The inclusion of non-tidal currents improves predictions, particularly for less mobile sand waves, by either dampening or amplifying the tidal asymmetry. To better represent the physical processes, we incorporate sediment mobility through excess velocity and excess bed shear stress indicators, which account for the critical threshold of motion. These measures reveal clearer non-linear relations with migration rates which show greater consistency across different regions of the continental shelf. The squared excess velocity emerges as the most promising predictor, showing a clear 2nd order fit with the migration rate, which is consistent over the areas. Nevertheless, the shear stress-based indicators show greater robustness in dealing with uncertainty in sediment grain size. These process-based relations enable first-order migration rate estimates in data-scarce areas and can provide insights into how changing environmental conditions may affect future sand wave dynamics.

In conclusion, this thesis demonstrates the importance of detailed hydrodynamic, sediment transport and morphological processes for understanding and reproducing sand wave dynamics. With new details in forcing conditions and sediment transport processes as well as 3D processes in the model, we are better able to understand, explain and predict sand wave dynamics in the field. In addition, the validation efforts have greatly increased our confidence in the numerical model. Finally robust empirical relations allow us to use the model to gain understanding of the local driving forces and make first-order estimation of migration rates. Together these steps make that we are better prepared to face the challenges related to seabed morphodynamics that lie in the expansion of our offshore activities.