Formation and evolution of nearshore sandbars

Abstract

The aim of this study is to understand whether hydrodynamic processes or geometrical characteristics play a dominant role in the response of the nearshore sandbar system to hydrodynamic conditions. To that end a depth-averaged (2DH) process-based model has been used to compute the morphological evolution of nearshore sandbars. The morphological evolution was computed for an initially alongshore uniform beach profile with two bars with an alongshore length of 7 km, forced with constant hydrodynamic conditions over a period of two weeks. The computations aimed to investigate the evolution of the system. It was found that an identical initial cross-shore profile responds distinctly different to different constant hydrodynamic conditions, showing the role of the hydrodynamic conditions. The length scales of the bars (corresponding to rip channel distances) increased with increasing alongshore velocities and increasing depths of the bar crests. The length scales ranged from 300-700 m for the inner and 600-2000 m for the outer bar. The response time of the system was in the order of days and depends linearly on the local wave height, the alongshore current, the steepness of the bar and inversely on the active volume of the bar. Bars with a smaller volume were found to respond quicker. To speed up the morphological computations, the initial alongshore uniform bathymetries were perturbed with a random seed in the order of cm. Different seedings resulted in different locations of the evolving features, while maintaining the length scales corresponding to the forcing condition. The role of the antecedent morphology was further investigated with computations with an increasing level of initial morphological variability. A high level of variability is for example formed by deep rip channels. With deeply imprinted bathymetrical patterns, the resulting hydrodynamical patterns prohibited the evolution of new patterns. This prohibited the adaptation toward length scales that would match the concurrent forcing conditions if the initial bathymetry would have been alongshore uniform. Only if the level of variability is small (smaller than O 0.5 m), the patterns adjusted toward the expected length scales. This was found for both evolutions with increasing as well as decreasing energy levels. This explains why observed nearshore bar patterns rarely match the concurrent conditions. The antecedent level of variability is often high, which inhibits complete adaptation. Further, the forcing conditions rarely persist for periods of time that are long enough for a system to evolve towards the corresponding length scales even if the initial variability would have been minimal. A hindcast was performed of an observed morphological evolution at Palm Beach, New South Wales, Australia, during a ten day period including a storm event. Palm Beach is a pocket beach of about 2 km length. During the event, the wave energy increased from moderate to storm levels, subsequently decreasing again to moderate conditions. The observed morphological variability changed from a single barred beach with rip channels toward a reset morphology (no alongshore variability) during the storm with subsequently newly evolving rip channels during the quieter post-storm conditions. The initial bathymetries used for the model computations were inferred from video-observations of the dissipation patterns. The effect of wave groups, wave asymmetry, long wave induced sediment stirring, the amount of turbulence and the rate of morphological change was tested in creating and hindcasting the observed patterns. It was found that these processes affect the magnitude and pace of morphological evolution. With optimal settings, the model including all mentioned processes forecasted a morphological evolution with decreasing variability during the storm event -similar to the observations. However, the observed amount of increase in bathymetrical variability after the storm event could not be matched in magnitude by the model. In general, best matches with observations were obtained for computations with a duration of up to three days. Within this period the different process settings only clearly changed the morphological evolution when the storm event was included within this period. Excluding wave groups resulted in the evolution of slightly shorter length scales. During the storm event an offshore bar formed and subsequent evolution was small and occurred both near the shore (wave groups have a diffusive effect in shallow water) and in deeper water. Computations starting after the storm event showed very little difference in morphological evolution whether wave groups were included or not. Excluding wave asymmetry resulted in shoreward migration of the shoreline and very little morphological evolution after all initial features had been erased. Long wave induced sediment stirring has a large diffusive effect on the evolving morphological features. Excluding this stirring resulted in the evolution of extreme shore-attached features. When the turbulent diffusion in the model was decreased, similar types of features evolved, though at slightly different rates and moments. Increasing the rate of morphological change resulted in the evolution of increased, mainly shore-attached, morphological variability. It was found that obtaining the correct pace and magnitude of morphological evolution is crucial for the level of success. If an event would encompass only an increasing or decreasing level of energy, it could be modelled. However, maintaining the correct pace and magnitude of evolution throughout a storm event has not been achieved with the currently tested model formulations and settings. This indicates that the model formulations need improvement. It is suggested that improving the description of the diffusion, including improving the turbulence description, can improve the model's capabilities. This does not only require improved model formulations, but also increased knowledge of the turbulence processes in the nearshore zone. The location of evolving features was found to be highly sensitive to the location and depth of imprint of features in the initial bathymetry. As this is rarely available to the required degree of accuracy, it is not expected that the exact location of features will be predicted correctly. However, the length scales and level of variability could be hindcasted if optimal settings are found, process descriptions improved (e.g. the diffusion) and if the model is morphologically calibrated both for evolutions with increasing and decreasing wave energy. In conclusion, morphological evolution of nearshore sandbar patterns is found to be influenced by the initial morphology in two ways. First, if the initial variability is low, local hydrodynamic forcings -determined by the off-shore conditions and the local geometry- and their duration will determine the length scale. The location of initially small perturbations -order of cm- influences the location of rip channels. Second, if the initial morphology has a high level of variability, the bathymetry will remain the same due to the occurring hydrodynamic circulations which are reinforced by the incoming waves. Only an event with extreme energy may cause changes in the morphology in this case. The hydrodynamics seem to have a rather small role in changing the patterns of a bathymetry in case there is a significant level of variability: they reinforce existing patterns and are only capable of drastically changing existing patterns when the energy level is extremely high. They do affect the length scales and the response time in case of small initial variability. Using a morphological model requires accurate calibration of both the hydrodynamics as well as of the morphodynamics. Different nearshore processes have different effects with various magnitudes at different locations in the nearshore zone. Obtaining the correct balance between processes which amplify or damp existing patterns throughout an event with varying energy levels is currently a challenge. However, hindcasting either an up-state or a down-state morphological evolution is possible.