Ebook: Dune erosion during storm surges

A large scale dune erosion experiment has been conducted in a flume to examine the effect of the wave period on dune erosion and to perform detailed measurements of inner surf water pressure, flow velocities and sediment concentrations. Profile measurements reveal that a 50% increase in the wave period results in 25% larger dune erosion volumes for storm surge conditions that are representative for the Dutch coast. Analysis of the detailed measurements shows that both short wind waves and long waves are important to inner surf hydrodynamics. The mean flows are directed offshore and increase towards the shoreline whereas mean sediment concentrations rise sharply towards the dune face (up to 50 g/l near the bed). The sediment transport is dominated by the mean offshore directed flow and is partly compensated by shoreward sediment transports that take place above the wave trough or are associated with intra wave processes. The effect of the wave period on dune erosion is mainly caused by an O(60%) increase of the time and depth averaged near dune sediment concentrations whereas the offshore directed mean flows are comparable, yielding a larger offshore directed transport capacity. This increase in transport capacity is only partly compensated by a concurrent increase of the wave related sediment transports (in landward direction).

The interaction of dune face and swash zone is studied in more detail for the large scale dune erosion experiment. First an algorithm is proposed that can make three dimensional reconstructions of the dune face and beach from collected image pairs with two synchronized cameras. Next, available stereo video reconstructions and profile measurements are used to confirm a linear relation according to Fisher et al. (1986) between the average wave impact force on the dune face and the erosion rate. It is found that initially a different and more effective dune erosion mechanism is present in which waves run-up the dune face and steepen it by scouring the face. When the dune face is sufficiently steep waves start to impact it after which the steepening continues until a critical slope is reached and the dune face collapses. Slumping characteristics change over a storm surge and the time interval between successive slumps increases whereas the average volume associated with a slump does hardly change as a surge progresses.

Inner surf zone hydrodynamics have been simulated with a surf beat model for the large scale dune experiment and the field. It is shown that the observed shift in variance towards lower frequencies in the inner surf zone during the experiment is associated with the generation and interaction of long waves with short wave groups. Considering dissipative conditions in the laboratory and field, near shore hydrodynamics can be accurately reproduced with the surf beat model whereas for reflective conditions this is less the case since short (wind) waves are also present near the shoreline. In shallow water (H_rms,lf ~ h) long waves contribute to the mean offshore directed flow.

Sediment concentration measurements have been analyzed in more detail and it is found that the mean near dune sediment concentration correlates much better with the maximum wave surface slope than with flow drag. The maximum surface slope is associated with the intensity of wave breaking and if the diffusion of turbulence from the water surface towards the bed is taken into account the correlation with the mean sediment concentrations improves.

The O(100%) increase in the near-bed sediment concentrations for a larger wave period correlates well with an increase in the intensity of wave breaking whereas the wave averaged turbulence production is comparable for the range of wave periods studied. For this reason it is hypothesized that the near-bed turbulence energy varies over the wave cycle since breaking induced turbulence is generated at the wave front and is injected in the water column over a short period (the bore interval). In addition it is presumed that sediment suspensions respond nonlinearly to the near-bed turbulence intensity and will increase for more intense breaking waves.

The effect of near-bed turbulence on sediment concentrations is examined with a 1DV suspension model. Simulations with constant turbulence energy over the wave cycle are compared with simulations with wave varying turbulence energy. It is found that in cases with a sufficient short duration of the bore interval (T/T_bore > 7) the wave averaged sediment suspensions are substantially higher (one order of magnitude) when the near-bed turbulence intensity varies over the wave cycle. For larger bore intervals sediment suspensions have the same order of magnitude and the effect of wave breaking induced turbulence on sediment suspensions is expected to be small since this turbulence does not reach the bed. The 1DV model results are aggregated in a wave averaged equilibrium sediment concentration formulation that in addition to flow drag is a function of the bore averaged turbulence intensity.

Obtained insights in dune erosion physics are coupled within a 2DH morphodynamic model XBeach (Roelvink et al., 2007). The model is extended with an adapted wave dissipation model, an equilibrium sediment concentration formulation that depends on the bore averaged turbulence energy and a wave shape model from which the bore interval is estimated. The wave shape model is also utilized to estimate intra wave sediment transports and the dissipation rate in bores that develop after wave breaking.

After optimization the XBeach model is applied to simulate:

1. The large scale dune erosion experiment described in this thesis. The model physics over the developing foreshore are favorably compared with detailed measurements during a dune erosion test. In addition the effects of the wave period and spectral shape on dune erosion are correctly simulated. Profile evolution during a small dune breach looks reasonable however the amount of erosion is overestimated at the end of the test.

2. The effect of a dune revetment on foreshore evolution during a storm surge. It is found that at this stage the model lacks the right physics to simulate the development of a scour hole and simulated long wave run-up is too small to erode sand above a revetment of medium height.

3. Profile evolution during moderate and calm wave conditions. The model tends to erode the beach near the waterline; however the amount of erosion is small.

4. Impact of the 1953 storm surge on the Delfland coast in The Netherlands. An erosion volume of 73 m³/m is predicted, which is within the range of estimated erosion volumes (55 m³/m – 155 m³/m).

In addition, the sensitivity of simulated dune erosion to short waves and long waves is examined. It is found that dune erosion rates during the start of a test are determined by the sediment supply from the dunes rather than by the offshore transport capacity of the near dune hydrodynamics. Considering only short (wind) waves still a reasonable estimation of the amount of dune erosion can be made and the erosion volume is underestimated with about 30%. Wave group generated long waves contribute to the amount of dune erosion (about 30%) and are effective in releasing sand from the dunes.

Dune face erosion is simulated with a simple avalanching algorithm. Robustness of this algorithm is tested for an instable dry bank and for a dune erosion test by performing simulations on various grids. Next, the parameters associated with avalanching are varied showing that the sediment supply from the dunes influences the dune erosion volume during a storm surge.

2DH simulations are conducted with a time varying surge (representative for the Dutch coast) and with directionally spread incoming waves. First a uniform coast is considered and 2DH model results are compared to 1D results showing that dune erosion rates are comparable. Next, a simulation with an alongshore varying dune height and uniform bathymetry is performed. A uniform foreshore develops and sand from areas with higher dunes is transported to cross-shore transects with lower dunes. Simulation results for an alongshore uniform dune height and varying bathymetry (including an offshore bar, an intertidal bar intersected by rips and beach cusps) reveal highly non-uniform flows and sediment concentrations. However, the foreshore develops quite uniform in alongshore direction and dune face retreat varies only slightly along the coastline. Finally, the transition of a (non-erodible) dike that protrudes seaward of a sandy dune system is modeled. In case a longshore flow is present dune erosion volumes are significantly larger downstream of the dike, which is caused by longshore sediment transport gradients related to the presence of the dike.