Coastal Engineering Today, Gainesville, Florida, October 2003
MORPHOLOGICAL MODELLING: A TOOL FOR OPTIMISATION OF COASTAL STRUCTURES Ida Brøker1, Julio Zyserman2, Erik Østergaard Madsen3, Karsten Mangor4, John Jensen5 This paper discusses the problems of sedimentation in a fishery port and its impact on the morphology of a sandy, very exposed coastline. An improved layout of the main breakwaters has been developed. The results obtained from testing the new layout in a morphological modelling complex are discussed. The new layout is expected to lead to improved bypass, decreased sedimentation and coastal impact, to a greater natural depth at the entrance, and to provide safer navigation conditions. Abstract:
INTRODUCTION AND BACKGROUND
The West Coast of Denmark is exposed to the severe North Sea wave climate. The west coast, from the Wadden Sea in the south to the Skaw in the north, is several hundred kilometres long. The net sediment transport along the coastline ranges from 500,000 to 1,000,000 m3/year. Natural shoreline retreats of up to 5-8 m/year take place at some stretches. The sandy coastline is undergoing constant reshaping mainly due to gradients in the longshore sediment transport. This coastal landscape was basically formed following the latest glacial period. An equilibrium shape has not
1
Ida Brøker, M.Sc., PhD. Head of Coastal and Estuarine Dynamics, DHI Water & Environment, Agern Allé 5, 2970 Hørsholm, Denmark,
[email protected] 2 Julio Zyserman, PhD. Chief Engineer, Coastal and Estuarine Dynamics, DHI Water & Environment, Agern Allé 5, 2970 Hørsholm, Denmark,
[email protected] 3 Erik Østergaard Madsen, M.Sc., PhD. Research Engineer, Coastal and Estuarine Dynamics, DHI Water & Environment, Agern Allé 5, 2970 Hørsholm, Denmark,
[email protected] 4 Karsten Mangor, M.Sc. Chief Engineer, Coastal and Estuarine Dynamics, DHI Water & Environment, Agern Allé 5, 2970 Hørsholm, Denmark,
[email protected] 5 John Jensen, M.Sc. Senior coastal engineer, Danish Coastal Authority, Højbovej 1, 7620 Lemvig , Denmark,
[email protected]
1
Brøker, Zyserman, Madsen, Mangor & Jensen
Coastal Engineering Today, Gainesville, Florida, October 2003
been reached and even though till cliffs, forming semi-hard points, are present, the entire coastline is slowly retreating in its natural condition. Thorsminde fishery port is located at a tidal inlet on this coastline, on one of the narrow barriers which divides coastal lagoons from the sea. The port is located at the entrance to the coastal lagoon. Sluices regulate the water exchange between the lagoon and the sea. Figure 1 shows a location map and a close-up of Thorsminde.
Fig. 1 Location map and close-up of Thorsminde Port. Up until the 1980's, i.e. for about 100 years, critical parts of the coast were protected by traditional structures such as groynes. A comprehensive nourishment scheme was established in the mid-eighties. Nourishment of about 3 million m3/year along a stretch of approximately 115 km has now stabilised the beach at critical stretches, while other stretches have been left to retreat as part of an overall shoreline management plan prepared and controlled by The Danish Coastal Authority. The left panel in Figure 2 shows the central part of the coastal stretch, where the most critical retreat takes place. It also shows the shoreline retreat in two periods, before and after nourishment, and indicates the distribution of nourishment. It can be clearly seen how the shoreline retreat has been alleviated by intensive nourishment, Laustrup et al. (1998). The littoral drift and corresponding shoreline retreat have been simulated by DHI using the model complex, LITPACK, see Kerper et al. (2002) for details. The middle and right panels of Figure 2 show the simulated yearly littoral drift and shoreline retreat based on the time series of wave parameters from 1991-1996. The littoral drift was calculated along the entire coastline. Blocking structures, variability in wave conditions, current conditions, sediment properties and the shape of the coastal profiles were taken into account. Calculations of the littoral drift were carried out every hour.
2
Brøker, Zyserman, Madsen, Mangor & Jensen
Coastal Engineering Today, Gainesville, Florida, October 2003
LITPACK is a complex of modules for the simulation of wave transformation, longshore wave-driven currents, longshore and cross-shore sediment transport, shoreline evolution and coastal profile evolution. The bed contours are assumed to be quasi-uniform in the longshore direction and the waves and currents are considered to be quasi-stationary. These two basic assumptions limit the use of the tool to cases of long and uniform sandy beaches and cases where the shoreline evolution is the result of the overall gradients in the longshore sediment transport capacity. However, total or partial blockage of the littoral drift by structures as well as the effect of revetments is included in the shoreline evolution model. Due to the assumptions described above, long coastal stretches can be investigated over long time spans. Various elements in LITPACK are described in Deigaard et al. (1986), Deigaard et al. (1991), Deigaard et al. (1993) and Elfrink et al. (1996).
Fig. 2 Left: Observed erosion, 1977-1986. Middle and right: simulated yearly littoral drift and shoreline retreat. Thorsminde Port is located in the central part of this very exposed stretch, where the net littoral drift is southward with an order of magnitude of 0.4 million m3/year, but where the gross transport is several times larger. The sedimentation and shoaling problems at present affecting the harbour entrance are illustrated in Figure 3. The upper and lower panels show the results of the
3
Brøker, Zyserman, Madsen, Mangor & Jensen
Coastal Engineering Today, Gainesville, Florida, October 2003
simulation of nearshore waves and wave-driven currents for northwesterly waves and southwesterly waves, respectively.
Fig. 3 Examples of wave and flow fields around the existing harbour Upper Hs = 3.0 m, 315°° , Lower Hs = 2.5 m, 235°°
4
Brøker, Zyserman, Madsen, Mangor & Jensen
Coastal Engineering Today, Gainesville, Florida, October 2003
The simulations were performed using a spectral wind wave model and a depthintegrated hydrodynamic model. Both modules are included in the morphological modelling system, MIKE 21 CAMS, which will be discussed below. For northwesterly waves, it is clearly seen how the wave-driven currents converge at the end of the northernmost groyne and diverge south of the inlet jetty. The southward sediment transported along the coast will thus be pushed around the tip of the groyne and will settle in the large eddy south of the inlet jetty. For southwesterly waves, the northward littoral transport will be pushed directly into the harbour entrance. The natural (equilibrium) water depth at the entrance to the harbour is 2-3 m, if no maintenance dredging is carried out. The harbour entrance is at present dredged to 3.5-4 m. Sedimentation of the entrance typically occurs during conditions with wind speeds between 8 and 15 m/s. The harbour is “small” compared to the width of the littoral zone and there is a need for maintenance dredging after almost every storm, partly due to sedimentation in the entrance channel and partly due to shoaling in front of the entrance area. On average, 0.10 million m3 are dredged every year. The downdrift coastline suffers from erosion as some of the sediment migrates past the harbour and continues to the south at some distance from the shore. The transport capacity in the inner nearshore zone immediately south of the harbour is, therefore, bigger than the sediment supply, which leads to the erosion of the beach. At present, this southern beach is protected by beach breakwaters and some of the dredged material is placed artificially on the eroding beach. The goal of a new harbour layout can be summarised as follows: •
To increase the natural depth in front of the entrance, thereby decreasing the downtime for access to the port and decreasing maintenance dredging
•
To improve navigation conditions to the port, especially during storms from SW, where opposing waves and currents co-exist in the present layout
It may be a positive side effect of a new layout that the downdrift erosion is reduced. EXAMPLES OF BYPASS HARBOURS
The concept of bypass harbours has been used at several locations in Denmark. Figure 4 shows Hanstholm harbour, a large fishery and ferry port located 80 km north of Thorsminde, see Figure 1. This harbour was built in the 60’s at a critical location with about 0.4 million m3/year net northward transport and a gross transport of around 1.5 million m3/year. The symmetrical and streamlined layout creates a convergence of the flow past the harbour entrance and has resulted in a very small sedimentation, localised in the
5
Brøker, Zyserman, Madsen, Mangor & Jensen
Coastal Engineering Today, Gainesville, Florida, October 2003
outer harbour immediately inside the entrance, and a natural depth in the entrance area of 9 m. Flows around this harbour are mainly driven by meteorological forcing, variations in wind and pressure, and, to a less extent, by wave breaking.
Fig. 4 Hansholm Harbour, west coast of Denmark Figure 5 shows a small harbour on the northern coast of the Danish island of Sealand, see Figure 1. The old fishery port now constitutes the inner basin. The updrift deposition is clearly seen to the right of the photograph. The fishery port is more than 100 years old.
Fig. 5 Hornbæk harbour, north coast of Sealand, Denmark
6
Brøker, Zyserman, Madsen, Mangor & Jensen
Coastal Engineering Today, Gainesville, Florida, October 2003
The main updrift breakwater was extended several times following its construction until the nineties, when the downdrift breakwater was added. The symmetrical and streamlined layout has reduced the sedimentation problem by a factor of three. It can even be seen from the photo how the natural bypass occurred and how, in years to come, it will feed the downdrift coast. The expected development of the shoreline adjacent to a bypass harbour after construction is sketched in Figure 6.
Fig. 6 Sketch of bypass harbour, from Mangor (2001) IMPROVED LAYOUT FOR THORSMINDE
A new layout has been developed for Thorsminde harbour using the principles of natural bypass. The new layout includes a downdrift breakwater, a streamlining of the entrance by a small extension of the existing main breakwater to the southwest and a shortening of the updrift groyne. It is expected that the contraction of the wave-driven currents around the harbour entrance will be enough to maintain an equilibrium depth in front of the harbour, which will be suitable for navigation. Strong, but well-defined currents will be present in front of the entrance during storms. The large outer harbour basin makes this current pattern acceptable for navigation in rough weather. The present sedimentation problem related to waves from a southwesterly direction will be
7
Brøker, Zyserman, Madsen, Mangor & Jensen
Coastal Engineering Today, Gainesville, Florida, October 2003
alleviated. The naturally bypassed sediment from north to south, will, with time, develop a bypass shoal and start to feed the downdrift beach. The new layout can be seen in Figure 9. Analysis method: the Coastal Area Modelling System, MIKE 21 CAMS
The critical parameter for the new harbour layout is the equilibrium depth at its entrance. This equilibrium depth is reached when the sediment transport capacity in front of the harbour is similar to the updrift littoral transport. The equilibrium depth has been evaluated using the numerical modelling system, MIKE 21 CAMS. MIKE 21 CAMS consists of modules for the simulation of waves, currents, sediment transport and bed level changes, with continuous updating of the bed levels and the subsequent re-calculation of waves, currents and sediment transport. The use of a (2D) morphological area model paves the way for the possibility of following the bed development with time during storm conditions. This is a large step forward in modelling procedure. Previously, coastal studies often relied on time series of sediment transport on a fixed bed. The use of a morphological model allows for a better comparison of the effectiveness of structures of different layouts. The applied modules are described briefly in the following. Wave modules. Two wave modules have been used for this study (i) MIKE 21 PMS and (ii) MIKE 21 NSW. MIKE 21 PMS is based on the parabolic approximation to the mild-slope equation (Kirby, 1986) and accounts for the effects of shoaling, refraction, diffraction, wave breaking, directional spreading, forward scattering and bed friction. MIKE 21 NSW is a spectral wind-wave model, which describes the propagation, growth and decay of short-period waves in nearshore areas by solving the equations for the conservation of wave action (Holthuijsen et al., 1989). The model includes the effects of refraction and shoaling, wave generation due to wind, and energy dissipation due to bottom friction and wave breaking. The effects of current on these phenomena may be included. In both wave modules, the dissipation of wave energy due to breaking is calculated according to the model of Battjes and Janssen (1978). The hydrodynamic module, MIKE 21 HD, calculates the flow field from the solution of the depth-integrated continuity and momentum equations, Abbott (1979). In addition to wind and tide, the forcing terms may include the gradients in the radiation stress field as calculated by the wave module. The currents and the mean water level are calculated on a bed evolving at a rate equal to ∂z/∂t, as calculated by the sediment transport module. The non-cohesive sediment transport module, MIKE 21 ST, is used to calculate the transport rates of graded sediment and the rates of bed level change ∂z/∂t due to the combined action of waves and current. MIKE 21 ST uses DHI’s deterministic intra-wave sediment transport model STP to calculate the total (bed load + suspended load) transport rates of non-cohesive sediment.
8
Brøker, Zyserman, Madsen, Mangor & Jensen
Coastal Engineering Today, Gainesville, Florida, October 2003
The sediment transport model, STP, is identical in LITPACK and MIKE 21 ST. The model has been described in detail in a series of papers, see e.g. Fredsøe (1984) and Deigaard et al. (1986). STP includes a quasi-3-dimensional description of the flow and the sediment transport, as described in Elfrink et al. (1996, 2000). Use of this approach allows simultaneous calculation of net sediment transport rates both in the longshore and cross-shore directions. The MIKE 21 CAMS, Coastal Area Modelling System, Johnson et al. (1994), is a complex in which the wave, flow and sediment transport fields and rates of bed level changes are simulated in sequence and which includes full feedback from the developing bathymetry to all modules. The updating technique is described in detail in Johnson and Zyserman (2002). The modelling system is sketched in Figure 7. Model calibration and validation
The morphological modelling complex has been calibrated and validated against measured pre-storm and post-storm bathymetries of the harbour entrance for two storms from the northwest and southwest respectively. The Danish Coastal Authority maintains a directional wave meter 3 km northwest of the site. The water level is recorded every 15 minutes inside the harbour. These recordings have been used as boundary conditions for the morphological modelling complex. General surveys are available for each summer, but detailed surveys are carried out regularly, especially after a severe event has taken place around the harbour entrance. The pre-storm and post-storm bathymetries measured around the entrance, as well as the model bathymetry prior to the storm (lower left panel), are shown in Figure 8. The grid spacing of the model bathymetry is 6 m. The bathymetry outside the area where pre-storm measurements are available is constructed from the general surveys. The lower right panel shows the modelled post-storm bathymetry. It can be seen that the sand shoal off the main breakwater was pushed to the south during the storm and that the model is able to reproduce this process. The calibration of the morphological modelling complex comprised the tuning of bed roughness, wave breaking parameters, the selection of a wave model (parabolic model versus spectral wind wave model) and the testing of the influence of the start bathymetry outside the area, where pre-storm measurements are available. The sediment properties are constant all over the model area. The median grain size is 0.3 mm is this area and the sediment is well sorted. The best calibration is obtained with the following parameters/choices: • • •
breaking parameter γ2 = 1.0 bed roughness: M = 48 m1/3 /s the spectral wind wave model
9
Brøker, Zyserman, Madsen, Mangor & Jensen
Coastal Engineering Today, Gainesville, Florida, October 2003
Fig. 7 Flow diagram of the morphological modelling system MIKE 21 CAMS The adopted breaking parameters result in a narrower surf zone compared to the standard value of γ2 = 0.8 in Battjes and Janssen(1978). The forward scatter of waves on the structures has shown, in this case, to lead to too high waves, which prevents the deposition of sand in the entrance when using the parabolic mild slope model. More realistic morphological changes were obtained using the spectral wind wave model. The initial bathymetry was changed to include a larger sand bar stretching 100 m to the north and with the same shape as the observed pre-storm bar. The simulated post-storm bathymetry was not sensitive to this change. The simulation of a historical storm from the southwest using the above-mentioned model setting showed acceptable results.
10
Brøker, Zyserman, Madsen, Mangor & Jensen
Coastal Engineering Today, Gainesville, Florida, October 2003
Fig. 8 Boundary conditions, observed pre-storm and post-storm bathymetries, simulated pre-storm and post-storm bathymetries, 16 and 27 October 1997
11
Brøker, Zyserman, Madsen, Mangor & Jensen
Coastal Engineering Today, Gainesville, Florida, October 2003
TESTING AND DOCUMENTATION OF THE NEW LAYOUT
The proposed new layout is seen in Figure 9, which shows the result of a repetition of the simulation of the October 1997 storm with the new breakwater, including the redesigned existing structures. The left panel is the initial model bathymetry, the right panel is the bathymetry after the storm. It appears that a water depth of about 3.5 m can be maintained in front of the entrance after the storm with the modified layout. Furthermore, no sedimentation seems to have taken place in the entrance area. The “equilibrium” depth in front of the harbour was investigated through six thirty-day morphological simulations with persistent wave conditions. The bathymetries after 15 and 30 days of constant wave conditions, Hs = 3.5 m, from the northwest applied at the offshore boundary of the model area, are shown in Figure 10 together with the initial bathymetry. The initial bathymetry included in all six cases a deep area in front of the harbour, down to –6 m. The simulation results show that this deep area is migrating to the south and is filling in, see also Figure 11. After about 20 days of constant wave action, the water depth at the entrance stabilises at approximately 3.3 m. The lower right panel shows the bed level changes with time at two points, A and B, just off the existing main breakwater and at the entrance. The locations of points A and B are shown in Figure 10. It appears that large-scale bed forms migrate towards the south in the morphological model and that the “equilibrium bed level” at the entrance is dynamic, with a minimum depth of around 3.2 m.
Fig. 9 New layout. Morphological modelling of the October 1997 storm
12
Brøker, Zyserman, Madsen, Mangor & Jensen
Coastal Engineering Today, Gainesville, Florida, October 2003
The results seem to indicate that the 3 m depth contour never reaches the tip of the northern breakwater. A shoal develops immediately downdrift of the harbour, in the area where the contracted current expands after flowing across the entrance. This shoal keeps growing until the depth decreases to a level where the sediment transport capacity, due to wave breaking and wave-driven currents, corresponds to the amount which bypasses the harbour. The simulation indicates that the present coastal erosion on the downdrift side will decrease with time following the construction of the new structures. This is illustrated further in Figure 11, which shows the sediment transport field after 15 and 30 days of morphological simulation. The figure indicates that longshore gradients in transport capacity decrease with time as the downdrift shoal develops. CONCLUSIONS
Maintenance of a minimum depth in front of the entrance to a small fishery port on an exposed sandy coast is the key issue for the success of an improved harbour layout. A new streamlined layout, which optimises natural sand bypass, has been developed and investigated. A morphological modelling complex, which simulates waves, currents, sediment transport and corresponding bed level changes, has proven to be a useful tool in supporting the understanding of the processes around the harbour and in estimating the minimum depth at the entrance during and after storm events.
13
Brøker, Zyserman, Madsen, Mangor & Jensen
Coastal Engineering Today, Gainesville, Florida, October 2003
Fig. 10 Modelled morphological evolution for constant waves, Hs = 3.5 m and 315°°. Bathymetries after 0 days, 15 days, 30 days. Time series of bed level in points A and B.
14
Brøker, Zyserman, Madsen, Mangor & Jensen
Coastal Engineering Today, Gainesville, Florida, October 2003
Fig. 11 Sediment transport capacity after 15 and 30 days, respectively, with constant waves, Hs = 3.5 m and 315°° REFERENCES
Abbott M.B. 1979. Computational hydraulics, elements of the theory of free surface flows. Pitman, London. Battjes, J.A. and Janssen J.P.F.M. 1978. Energy loss and Set-up due to breaking of random Waves. Procs. of the 16th Int. Conf. On Coastal engineering, ASCE, 569587. Deigaard, R. 1993. A note on the 3-dimensional shear stress distribution in a surf zone”. Coastal Engineering, 20, 157-171. Deigaard, R., Fredsøe, J. and Brøker Hedegaard, I. 1986. Suspended sediment in the surf zone. J. Waterway, Port, Coastal and Ocean Engineering, 112 (1), ASCE, 115-128. Deigaard, R., Justesen, P. and Fredsøe, J. 1991. Modelling of the undertow by a one equation turbulence model. Coastal Engineering, 15, 431-458. Elfrink, B., Brøker, I. and Deigaard, R. 2000. Beach Profile Evolution due to Oblique Wave Attack. Procs. of the 27th Int. Conf. on Coastal Eng., ASCE, 3021-3034. Elfrink, B., Brøker, I., Deigaard, R., Hansen, E.A. and Justesen, P. 1996. Modelling of 3D sediment transport in the surf zone, Procs. of the 25th Int. Conf. on Coastal Eng., ASCE, 3805-3817. Fredsøe J. 1984. The turbulent boundary layer in combined wave-current motion. Journal of Hydr. Eng., 110(8), ASCE, 1103-1120.
15
Brøker, Zyserman, Madsen, Mangor & Jensen
Coastal Engineering Today, Gainesville, Florida, October 2003
Holthuijsen, L.H., Booij, N. and Herbers, T.H.C. 1989. A prediction model for stationary, short-crested waves in shallow water with ambient currents. Coastal Engineering, 13, 23-54. Johnson, H.K. and Zyserman, J.A. 2002. Controlling spatial oscillations in bed level update schemes. Coastal Engineering, 46, 109-126. Kerper R.D., Brøker I., Damgaard Christensen, E. and Zyserman J.A. 2002. Application of coastal modelling systems in support of integrated coastal zone management. Proceeding of Coastal Disasters conference, San Diego, California. Kirby, J.T. 1986. Rational approximations in the parabolic equation method for water waves. Coastal Engineering, 10, 355-378. Laustrup, C. and Toxvig, H. 1998. Evaluation of the Effect of 20 Years of Nourishment, Procs. of the 26th Int. Conf. on Coastal Eng., ASCE, 3074-3085. Mangor, K. 2001. Shoreline Management Guidelines, DHI Water & Environment, 232p.
16
Brøker, Zyserman, Madsen, Mangor & Jensen
