Hydrography and cohesive sediment modelling: application to the Rømø Dyb tidal area

Abstract

Estuaries act as sinks for fine-grained sediments and because of the cohesive properties of these sediments, heavy metals and nutrients tend to accumulate in estuaries. In order to quantify the erosion, transport and deposition of these pollutants, modelling of cohesive sediment dynamics is a very useful tool. However, the description of cohesive sediment dynamics through numerical analysis is a difficult task since the physical properties of cohesive sediments are of complex nature.

In this paper, setup and calibration of a cohesive sediment transport model covering the period from October 20, 1999 to December 13, 1999 are described and an interpretation of the results is carried out. Further, a comparison with measured suspended sediment concentrations and bed level measurements from the modelling period is presented.

A detailed bathymetry is used. This is necessary in order to describe the water movements in a realistic way. A bottom description is created in a way so that differences in erodibility of the sediment can be described. Further, a spatial differentiated description of critical shear stress, both for erosion and deposition of the cohesive sediment bottom, is made in order to describe the processes of settling and scour lag.

The hydrodynamic simulation is shown to be very reliable, and therefore, it has been possible to extract key values for the area. Thus, the maximum current velocities for the Lister Dyb area are modelled to 1.2 and 0.93 m s⁻¹ for the flood and ebb period, respectively. The tidal prism has likewise been computed to 620×10⁶ m³.

The cohesive sediment transport modelling has shown that the highest sediment concentrations at a given site appear when onshore winds are prevailing. Further, it can be recognized in the results that an inward sediment transport direction is prevailing, especially after a windy period with waves has mobilized considerable amounts of sediment.

A detailed investigation of the cohesive sediment's settling velocities collected in the area is used to give a site-specific description. Since the description of the settling velocity changes with temperature, several simulations using different descriptions have been carried out. These simulations have shown that this parameter is very influential for the net deposition result. Thus, an increase of the settling velocity will increase the deposition rates considerably.

The modelling results and the comparisons to measured data presented in this paper show that it has been possible to calibrate the hydrodynamic model in accordance with observed values. In continuation of this, the deformation of the tidal wave has also been modelled satisfactorily. Modelling of the cohesive sediment dynamics has been more complicated and the modelling results show divergence from the measured results. However, the levels of the sediment concentrations and the overall net sedimentation pattern show accordance with observed values.

Introduction

The processes of flocculation, settling and scour lag, and the asymmetry of the tidal currents make cohesive sediment transport in estuaries difficult to forecast Dyer, 1989, Teisson, 1991, van Leussen, 1994, Parker, 1997, Van der Lee, 2000. Because of the cohesive properties of the fine sediments, nutrients, heavy metals and other pollutive substances tend to bind to the sediment's surface (Rae, 1997). Consequently, pollutants can be concentrated in the estuaries, thus being of great environmental interest. In addition, the mudflats occurring in estuaries are important biotopes for a large number of micro- and macro-faunal species and act as feeding places for a number of birds (Eisma, 1998). This makes forecast of erosion, transport and deposition of cohesive sediment of great interest in estuaries.

Because of this, modelling of the dispersion of cohesive sediments will be of major importance in the future, and numerical simulation models will have a great potential when it comes to estimation of the effects of dredging of ports, building of dams, discharge of environmental pollutants and the global eustatic sea level rise. These simulation models can make it possible to quantify the sediment erosion, transport and deposition over a large area during a given period. The results can be investigated at time scales down to e.g. 5-min intervals. This combination of features can hardly be obtained by the use of traditional in situ measurements. Still, the fieldwork is the basis for the development of the model and fieldwork is required in order to calibrate and validate the model. The use of fieldwork and modelling can therefore be an extremely profitable combination.

The now-existing cohesive sediment transport models all use an advection–dispersion equation to simulate the cohesive sediment transport in the water column (e.g. Teisson, 1997). The advection–dispersion equation requires current velocity components and water levels that are normally provided from a decoupled hydrodynamic model. The models have further incorporated a variety of equations that describe the cohesive sediment erosion, flocculation and deposition processes in different ways (e.g. Mehta et al., 1989, Teisson, 1991). However, most papers that describe setups of cohesive sediment transport models (e.g. Cancino and Neves, 1999, Le Normant, 2000) only present very short time series (hours or few days). If a realistic extrapolation of modelling results in time and space should be carried out, longer time series with a high degree of truthfulness needs to be obtained. This first requires a very well-calibrated hydrodynamic model. Hereafter, a profound calibration of the cohesive sediment transport model going thoroughly through each of the calibration parameters must be carried out.

In the future, there will be a need for comprehensible sediment transport models and the demand for reliability will further increase.

In this research project, it was attempted to set up and calibrate a numerical model so that the sediment transport is described satisfactorily, and further, to model cohesive sediment transport in a Danish estuary over a period of several weeks.

This paper describes the setup of a state-of-the-art sediment transport model (MIKE 21 MT) for the Rømø Dyb tidal area in the Danish Wadden Sea. The model is used for hindcast of sediment transport. The paper includes an analysis of the results and a comparison to previously published research from the area and theoretical aspects of cohesive sediment transport.