Skillbed XBeach 1.22.4687 Kingsday Release

XBeach skillbed report, revi-
sion 4867
status update trunk default
Revision: 4867
October 21, 2015

XBeach skillbed report, revision 4867
Published and printed by:
Deltares
Rotterdamseweg 185
p.o. box 177
2600 MH Delft
The Netherlands
telephone: +31 88 335 85 85

fax: +31 88 335 85 82
e-mail: [email protected]
www: http://www.deltares.nl
For support contact:
telephone: +31 88 335 85 55

fax: +31 88 335 81 11
e-mail: [email protected]
www: http://www.xbeach.org/
Copyright 2015 Deltares

All rights reserved. No part of this document may be reproduced in any form by print,
photo print, photo copy, microfilm or any other means, without written permission from the
publisher: Deltares.
XBeach skillbed report, revision 4867 Contents October 2015
4867
Contents
1 Introduction 1
1.1 Introduction to the XBeach model . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Model approach and innovations . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 XBeach skillbed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2 Hydrodynamic tests 5
2.1 Long wave propagation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 1D wave runup (analytical solution) . . . . . . . . . . . . . . . . . . . . . . . 6
2.3 2D wave runup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.4 High- and low-frequency wave transformation . . . . . . . . . . . . . . . . . . 10
2.5 Field experiment: DELILAH . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.6 Field experiment: Ningaloo reef . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.7 Longcrested refraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.8 Tide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.8.1 Blankenberge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3 Morphological laboratory tests 23

3.1 Scale relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.1.1 Small scale tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.1.2 Additional small scale tests . . . . . . . . . . . . . . . . . . . . . . . . 26
3.1.3 Large scale tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.2 Large scale tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.2.1 Revetments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.2.2 Extreme conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.2.3 Bar evolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.2.4 Influence of wave period . . . . . . . . . . . . . . . . . . . . . . . . . . 42
Deltares iii
October 2015 Contents XBeach skillbed report, revision 4867
4867
4 Morphological field tests 61

4.1 1953 storm surge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
4.2 1976 storm surge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
4.3 Erosion and overwash of Asseteague Island . . . . . . . . . . . . . . . . . . . 64
4.4 Breach growth at Zwin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.5 MICORE field experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
4.5.1 Lido di Dante, Italy . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
4.5.2 Praia de Faro, Portugal . . . . . . . . . . . . . . . . . . . . . . . . . . 69
4.5.3 Cadiz Urban Beach, Spain . . . . . . . . . . . . . . . . . . . . . . . . . 70
4.5.4 Dziwnow Spit, Poland . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
4.5.5 Kamchia Shkorpilovtsi Beach, Bulgaria . . . . . . . . . . . . . . . . . 72
5 Comparisons with other models 75

5.1 Field applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
5.1.1 Retreat distances JARKUS . . . . . . . . . . . . . . . . . . . . . . . . 75
6 Specific functionalities 77
6.1 River outflow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
6.2 Drifters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
6.3 Multiple sediment fractions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
6.4 Curvilinear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
7 References 83
A Model Performance Statistics 89

A.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
A.2 MPS parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
A.3 Mean Error & Standard Deviation . . . . . . . . . . . . . . . . . . . . . . . . 90
A.4 Correlation coefficient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
A.5 Relative Bias . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
A.6 Scatter Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
A.7 Brier Skill Score . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
A.8 Brier Skill Score (Murphy and Epstein, 1989) . . . . . . . . . . . . . . . . . . 91
B Overview 93
iv Deltares
XBeach skillbed report, revision 4867 Introduction October 2015
4867
Chapter 1
Introduction
1.1 Introduction to the XBeach model
The devastating effects of hurricanes on low-lying sandy coasts, especially during the 2004
and 2005 seasons have pointed at an urgent need to be able to assess the vulnerability
of coastal areas and (re-)design coastal protection for future events, and also to evaluate
the performance of existing coastal protection projects compared to do-nothing scenarios.
In view of this the Morphos-3D project was initiated by USACE-ERDC, bringing together
models, modelers and data on hurricane winds, storm surges, wave generation and nearshore
processes. As part of this initiative an open-source program, XBeach for eXtreme Beach
behaviour, has been developed to model the nearshore response to hurricane impacts. The
model includes wave breaking, surf and swash zone processes, dune erosion, overwashing and
breaching (Roelvink et al., 2009).
Existing tools to assess dune erosion under extreme storm conditions assume alongshore
uniform conditions and have been applied successfully along relatively undisturbed coasts
(Vellinga, 1986; Steetzel, 1993; Nishi and Kraus, 1996; Larson et al., 2004), but are inadequate
to assess the more complex situation where the coast has significant alongshore variability.
This variability may result from anthropogenic causes, such as the presence of artificial inlets,
sea walls, and revetments, but also from natural causes, such as the variation in dune height
along the coast or the presence of rip channels and shoals on the shoreface (Thornton et al.,
2007). A particularly complex situation is found when barrier islands protect storm impact on
the main land coast. In that case the elevation, width and length of the barrier island, as well
as the hydrodynamic conditions (surge level) of the back bay should be taken into account
to assess the coastal response. Therefore, the assessment of storm impact in these more
complex situations requires a two-dimensional process-based prediction tool, which contains
the essential physics of dune erosion and overwash, avalanching, swash motions, infragravity
waves and wave groups.
With regard to dune erosion, the development of a scarp andepisodic slumping after under-
cutting is a dominant process (Van Gent et al., 2008). This supplies sand to the swash and
surf zone that is transported seaward by the backwash motion and by the undertow; without
it the upper beach scours down and the dune erosion process slows down considerably. One-
dimensional (cross-shore) models such as DUROSTA (Steetzel, 1993) focus on the underwater
offshore transport and obtain the supply of sand by extrapolating these transports to the dry
dune. Overton and Fisher (1988), Nishi and Kraus (1996) focus on the supply of sand by the
dune based on the concept of wave impact. Both approaches rely on heuristic estimates of
Deltares 1
October 2015 Introduction XBeach skillbed report, revision 4867
4867
the runup and are well suited for 1D application but difficult to apply in a horizontally 2D
setting. Hence, a more comprehensive modelling of the swash motions is called for.
Swash motions are up to a large degree a result from wave-group forcing of infragravity
waves (Tucker, 1954). Depending on the beach configuration and directional properties of
the incident wave spectrum both leaky and trapped infragravity waves contribute to the swash
spectrum (Huntley et al., 1981). Raubenheimer and Guza (1996) show that incident band
swash is saturated, infragravity swash is not, therefore infragravity swash is dominant in storm
conditions. Models range from empirical formulations (e.g. Stockdon et al., 2006) through
analytical approaches (Schaeffer, 1994; Erikson et al., 2005) to numerical models in 1D (e.g.
List, 1992; Roelvink, 1993b) and 2DH (e.g. Van Dongeren et al., 2003; Reniers et al., 2004a,
2006). 2DH wavegroup resolving models are well capable of describing low-frequency motions.
However, for such a model to be applied for swash, a robust drying/flooding formulation is
required.
1.2 Model approach and innovations
Our aim is to model processes in different regimes as described by Sallenger (2000). He defines
an Impact Level to denote different regimes of impact on barrier islands by hurricanes, which
are the 1) swash regime, 2) collision regime, 3) overwash regime and 4) inundation regime.
The approach we follow to model the processes in these regimes is described below.
To resolve the swash dynamics the model employs a novel 2DH description of the wave groups
and accompanying infragravity waves over an arbitrary bathymetry (thus including bound,
free and refractively trapped infragravity waves). The wave-group forcing is derived from
the time-varying wave-action balance (e.g. Phillips, 1977) with a dissipation model for use
in combination with wave groups (Roelvink, 1993a). A roller model (Svendsen, 1984; Nairn
et al., 1990; Stive and De Vriend, 1994) is used to represent momentum stored in surface
rollers which leads to a shoreward shift in wave forcing.
The wave-group forcing drives infragravity motions and both longshore and cross-shore cur-
rents. Wave-current interaction within the wave boundary layer results in an increased wave-
averaged bed shear stress acting on the infragravity waves and currents (e.g. Soulsby et al.,
1993, and references therein). To account for the randomness of the incident waves the de-
scription by Feddersen et al. (2000) is applied which showed good skill for longshore current
predictions using a constant drag coefficient (Ruessink et al., 2001).
During the swash and collision regime the mass flux carried by the waves and rollers returns
offshore as a return flow or a rip-current. These offshore directed flows keep the erosion
process going by removing sand from the slumping dune face. Various models have been
proposed for the vertical profile of these currents (see Reniers et al., 2004b, for a review).
However, the vertical variation is not very strong during extreme conditions and has been
neglected for the moment.
Surf and swash zone sediment transport processes are very complex, with sediment stirring by
a combination of short-wave and long-wave orbital motion, currents and breaker-induced tur-
bulence. However, intra-wave sediment transports due to wave asymmetry and wave skewness
are expected to be relatively minor compared to long-wave and mean current contributions
(Van Thiel de Vries et al., 2008). This allows for a relatively simple and transparent formula-
tion according to Soulsby & Van Rijn (Soulsby, 1997) in a shortwave averaged but wave-group
resolving model of surf zone processes. This formulation has been applied successfully in de-
scribing the generation of rip channels (Damgaard et al., 2002; Reniers et al., 2004a) and
2 Deltares
XBeach skillbed report, revision 4867 Introduction October 2015
4867
barrier breaching (Roelvink et al., 2003).

In the collision regime, the transport of sediment from the dry dune face to the wet swash, i.e.
slumping or avalanching, is modeled with an avalanching model accounting for the fact that
saturated sand moves more easily than dry sand, by introducing both a critical wet slope and
dry slope. As a result slumping is predominantly triggered by a combination of infragravity
swash runup on the previously dry dune face and the (smaller) critical wet slope.
During the overwash regime the flow is dominated by lowfrequency motions on the time scale
of wave groups, carrying water over the dunes. This onshore flux of water is an important
landward transport process where dune sand is being deposited on the island and within
the shallow inshore bay as overwash fans (e.g. Leatherman et al., 1977; Wang and Horwitz,
2007). To account for this landward transport some heuristic approaches exist in 1D, e.g. in
the SBeach overwash module (Larson et al., 2004) which cannot be readily applied in 2D.
Here, the overwash morphodynamics are taken into account with the wave-group forcing of
low-frequency motions in combination with a robust momentum-conserving drying/flooding
formulation (Stelling and Duinmeijer, 2003) and concurrent sediment transport and bed-
elevation changes.
Breaching of barrier islands occurs during the inundation regime, where a new channel is
formed cutting through the island. Visser (1998) presents a semi-empirical approach for
breach evolution based on a schematic uniform cross-section. Here a generic description is
used where the evolution of the channel is calculated from the sediment transports induced
by the dynamic channel flow in combination with avalanche-triggered bank erosion.
1.3 XBeach skillbed
The XBeach code and related functionalities develop fast. As a result there is a need from
modelers and code developers to develop a tool that gives insight in the effect of code devel-
opments on model performnace. The XBeach skillbed tries to fulfill this need by running a
range of tests including analytical solutions, laboratory tests and practical field cases every
week with the latest code.
Deltares 3
October 2015 Introduction XBeach skillbed report, revision 4867
4867
4 Deltares
XBeach skillbed report, revision 4867 Hydrodynamic tests October 2015
4867
Chapter 2
Hydrodynamic tests
Morphodynamics start with hydrodynamics. In this chapter the hydrodynamic results of

XBeach are discussed. All tests are run without the morphological module and the analsyis
is focussed on the wave propagation and transformation computed by XBeach.
First, two analytical solutions are reproduced by XBeach. Subsequently, two laboratory
experiments are discussed and finally a field experiment.
2.1 Long wave propagation
The purpose of the this test is to check if the NSWE numerical scheme is not too dissipative
and that it does not create large errors in propagation speed.
A long wave with a small amplitude of 0.01m and period of 80s was sent into a domain of
5m depth, grid size of 5m and a length of 1km. At the end,
a fully reflecting wall is imposed.
The wave length in p this case should
p be g d T = 9.81 5 80 = 560m. The velocity
amplitude should be g/h A = 9.81/5 0.01 = 0.014m. After the wave has reached the wall,
a standing wave with double amplitude should be created.
Deltares 5
October 2015 Hydrodynamic tests XBeach skillbed report, revision 4867
4867
Water level at t = 0 to 160s with 20s interval

0.04
0.02
[m] 0
0.02
0.04
0 200 400 600 800 1000
x [m]
Water velocity at t = 0 to 160s with 20s interval
0.04
0.02
u [m/s]
0.02
0.04
0 200 400 600 800 1000
x [m]
Figure 2.1: Water levels and velocities from the start of the experiment until the wave
just reaches the end of the flume
Water level at t = 0 to 1200s with 20s interval

0.04
0.02
[m]
0.02
0.04
0 200 400 600 800 1000
x [m]
Water velocity at t = 0 to 1200s with 20s interval
0.04
0.02
u [m/s]
0.02
0.04
0 200 400 600 800 1000
x [m]
Figure 2.2: Snapshots of water levels and velocities showing a standing wave pattern
2.2 1D wave runup (analytical solution)
The purpose of this test is to check the ability of the model to represent runup and rundown of
non-breaking long waves. To that end, a comparison was made with the analytical solution
of the NSWE by Carrier and Greenspan (1958), which describes the motion of harmonic,
non-breaking long waves on a plane sloping beach without friction.
A free long wave with a wave period of 32 seconds and wave amplitude of half the wave
breaking amplitude (ain = 0.5 abr ) propagates over a beach with constant slope equal to 1/25.
6 Deltares
4867

The wave breaking amplitude is computed as abr = 1/ 1283 s2.5 T 2.5 g 1.25 h00.25 = 0.0307m,
where s is the beach slope, T is the wave period and h0 is the still water depth at the seaward
boundary. The grid is non uniform and consists of 160 grid points. The grid sizex is
decreasing in shoreward direction and is proportional to the (free) long wave celerity ( g h).
The minimum grid size in shallow water was set at x = 0.1m.
To compare XBeach output to the analytical solution of Carrier and Greenspan (1958), the
first are non-dimensionalized with the beach slope s, the acceleration of gravity g, the wave
period T , a horizontal length scale Lx and the vertical excursion of the
p swash motion A. The
horizontal length scale Lx is related to the wave period viaqT = Lx/gs and the vertical

excursion of the swash motion A is expressed as: A = ain / 0.125sT g/h0
Water level variation in time
5.2
water level [m]
5.1
4.9
4.8
50 60 70 80 90 100 110 120 130 140 150

crossshore distance [m]
Water velocity variation in time
1
flow velocity [m/s]
0.5
0.5
50 60 70 80 90 100 110 120 130 140 150

crossshore distance [m]
Figure 2.3: Snapshots of water level and velocity
2.3 2D wave runup
The verification cases so far considered solely the cross-shore dimension and assumed a long-
shore uniform coast. In the following case the potential of the model to predict coastal and
dune erosion in situations that include the two horizontal dimensions is further examined.
A first step towards a 2DH response is to verify that the 2DH forcing by surge run-up and
run-down is accurately modelled by testing not against Zelt (1986), but actually Ozkan-Haller
and Kirby (1997). The reason is that Zelt modeled the NSW equations including some dis-
persive and dissipative terms, which the present model does not include. For that reason, we
compared the model to the results of Ozkan-Haller and Kirby (1997) who modeled the NSW
equations using a Fourier-Chebyshev Collocation method, which does not have any numerical
dissipation or dispersion errors. They use a moving, adapting grid with a fixed y (which
is equal to the present models y in this comparison) but with a spatially and temporally
varying x so that the grid spacing in x near the shoreline is very small. In the present
model x is set equal to y, which means that we can expect to have less resolution at the
shoreline than Ozkan-Haller and Kirby (1997).
Deltares 7
4867
Figure 2.4: Definition scetch of concave beach bathymetry
Figure 2.4 shows the definition sketch of the concave beach bathymetry in the present coor-
dinate system, converted from the original system by Zelt (1986). The bathymetry consists
of a flat bottom part and a beach part with a sinusoidally varying slope. For Zelt (1986)s
fixed parameter choice of = Lhsy = 104
, the bathymetry is given by
h={
hs , x Ls
0.4(xL
s )
hs y
, x > Ls
3cos L
y
where hs is the shelf depth, Ls is the length of the shelf in the modeled domain and Ly is
the length scale of the longshore variation of the beach. This results in a beach slope of
1
hx = 10 in the center of the bay and of hx = 15 normal to the headlands. In the following
we chose Ly = 8 m, which determines hs = 1.0182 m. We set Ls = Ly . Different values for
Ls only cause phase shifts in the results, but no qualitative difference, so this parameter is
not important in this problem. Also indicated in the figure are the five stations where the
vertical run-up (the surface elevation at the shoreline) will be measured.
At the offshore (x = 0) boundary we specify an incoming solitary wave, which in dimensional
form reads
q
3g
i (t) = hs sech2 4hs (1 + )(t to )
8 Deltares
4867
which is similar to Zelt (1986)s Eq. (5.3.7). The phase shift to is chosen such that the surface
elevation of the solitary wave at t = 0 is 1% of the maximum amplitude. The only parameter
yet to be chosen is . We will compare our model to Zelts case of = hHs = 0.02, where
H is the offshore wave height. Zelt found that the wave broke for a value of = 0.03, so
the present test should involve no breaking, but has a large enough nonlinearity to exhibit a
pronounced two-dimensional run-up.
Any outgoing waves will be absorbed at the offshore boundary by the absorbing-generating
boundary condition. At the lateral boundaries y = 0 and y = 2 Ly we specify a no-flux
(wall) boundary condition following Zelt (1986). The model equations used in this test are
the nonlinear shallow water equations without forcing or friction. The numerical parameters
are x = y = 18 m with a Courant number = 0.7.
5
4
0
2
0
/H
/H
2
0
4
0
5 6
5 0 5 10 15 0 0.5 1 1.5 2
t/T y/Ly
Figure 2.5: Normalized vertical runup in time (panel 1) and maximum and minimum
values (panel 2)
The first panel in Figure 2.5 shows the vertical runup

normalized with the offshore wave height
H as a function of time, which is normalized by g hs /Ly at the 5 cross-sections indicated in
Figure 2.4. The solid lines represent the present model results, while the dashed lines denotes
Ozkan-Haller and Kirby (1997)s numerical results. The second panel in Figure 2.5 shows the
maximum vertical run-up and run-down, normalized by H, versus the alongshore coordinate
y. Table 2.1 presents error statistics of the model run with respect to the measurements.
Table 2.1
R2 Sci Rel. bias BSS

Timeseries (min) 0.10 0.23 -0.15 -5.14
Timeseries (max) 0.97 2.27 0.09 0.94
Max. runup 0.99 0.09 -0.09 0.99
Deltares 9
4867
2.4 High- and low-frequency wave transformation
Boers (1996) performed experiments with irregular waves in the physical wave flume at Delft
University of Technology with a length of 40 meters and a width of 0.8 m. The flume is
equipped with a hydraulically driven, piston type wave generator with second-order wave
generation and Active Reflection Compensation. Boers ran waves over a concrete bar-trough
beach, which was modelled after the Delta Flume experiments. He ran three different irregular
wave conditions, but in this report we will focus on case 1C, a Jonswap spectrum with Hm,0 =
0.1m and Tp = 3.3s. The surface elevation was measured in 70 locations shown in Figure 2.6.
Figure 2.6: Locations of surface eleveation measurements
The comparison between the model and the data for the wave height transformation of the
short waves and the long waves (defined as waves with a frequency greater than fp /2 and less
than fp /2, respectively) is shown in Figure 2.7.
The red dashed line and triangles indicate the short wave height transformation. The blue
line and circles indicate the mean (steady) set-up. The dotted red line and upside-down
traiangles indicate the total (incoming and reflected) low frequency wave.
The observational data is separated into incoming and reflected long wave components using
an array of wave gauges (Bakkenes, 2002) and the numerical data has been separated into
two components using co-located surface elevation and velocity information.
10 Deltares
4867
wave heights and water levels

0.1
wave height (HF,measured)
wave height (LF,measured)
wave height (measured)
0.08 setup (measured)
wave height (HF)
wave height (LF)
wave height
setup
0.06
height [m]
0.04
0.02
0.02
0 5 10 15 20 25 30
distance [m]
Figure 2.7: Wave height transformations during Boers 1C experiment
2.5 Field experiment: DELILAH
In order to verify the 2DH hydrodynamics of XBeach when forced by directionally-spread

short waves, a simulation is set up to compare model results to field measurements. In this
case the DELILAH field experiment at Duck, North Carolina is selected as a suitable test
location. The period that is modeled is October 13th 1990, which was a stormy day, between
16:00 and 17:00 hours. The significant wave height at 8 m water depth was 1.81 m, with a peak
period of 10.8 s and a mean angle of incidence of -16o relative to the shoreward normal. This
period is selected because the wave conditions are energetic enough to generate a significant
infragravity wave component and the incident wave spectrum is sufficiently narrow-banded to
justify the assumptions in the model boundary conditions. The model is forced with the wave
spectrum measured at 8 m water depth (Birkemeier et al., 1997). A measured tidal signal is
imposed on the model boundaries of which the mean level is 0.69 m above datum. The slope
of the wave front in the roller model is set to 0.05, which is found to be a slight improvement
over the value of 0.10 used in the previous sections. A constant grid size of 5 m in cross shore
and 10 m in longshore direction is used. The resolution of the wave model in directional
space is 15o . The model is set to generate output at the location of the primary cross shore
measurement array, gauge numbers 10, 20, 30, 40, 50, 60, 70, 80 and 90 (Figure 2.8).
Deltares 11
4867
Figure 2.8: Bathymetry and measurement locations
The modeled time-averaged wave heights of the short waves are compared to the time-
averaged wave heights measured at the gauges. These results are shown in the first panel of
Figure 2.9. Unfortunately, no data exist for gauge number 60.
The infragravity wave height is calculated as follows (Van Dongeren et al., 2003):
q R
0.05Hz
Hrms,low = 8 0.005Hz Sdf
Figure 2.9 shows the infragravity wave height. The measured and modelled time-averaged
longshore current are shown in the second panel of Figure 2.9. The correlation coefficient,
scatter index, relative bias and Brier Skill Score for the simulation are shown in Table 2.2.
12 Deltares
4867

2
1.5 wave height (measured)
height [m]
wave height (HF)
1 wave height (LF)
wave height
0.5
0
0 100 200 300 400 500 600 700 800 900
distance [m]
flow velocities
0
flow velocity (v,mean,measured)
flow velocity (v,mean)
velocity [m/s]
0.5
1.5
0 100 200 300 400 500 600 700 800 900
distance [m]
Figure 2.9: Wave height transformation and flow velocities during the DELILAH field
experiment 1990
Table 2.2

Hrmsl f 0.70 0.13 -0.12 0.50
Hrmsh f 0.94 0.15 -0.12 0.76
vmean 0.69 0.46 0.28 0.43
The modeled and measured sea surface elevation spectra at four gauge locations are shown
in Figure 2.10. Note that the modeled surface elevation spectra only contain low frequency
components associated with wave groups.
Deltares 13
4867
1 2
0.4 0.8
0.3 0.6
S [m2/Hz]
S [m2/Hz]
0.2 0.4
0.1 0.2
0 0
0 0.1 0.2 0.3 0 0.1 0.2 0.3
f [Hz] f [Hz]
3 4
0.5 0.8
0.4
0.6
S [m2/Hz]
S [m2/Hz]
0.3
0.4
0.2
0.2
0.1
0 0
0 0.1 0.2 0.3 0 0.1 0.2 0.3
f [Hz] f [Hz]
Figure 2.10: Modelled surface elevation spectra for four locations in the primary cross
shore array during the DELILAH field experiment 1990
2.6 Field experiment: Ningaloo reef
Ningaloo Reef is a large fringing reef on the northwest coast of Australia and consists of a
series of reef-channel cells, exposed to tropical cyclones and Southern Ocean swells. A field
data set of wave transformation on a shore-normal transect (Figure 2.11) taken at Sandy Bay
in June 2009 is described in (Pomeroy et al., 2012). The transect is composed of a steep fore
reef, a shallow reef flat ( 1 2m depth) that is separated from the shore by a slightly deeper
lagoon ( 2 3m average depth). Instrument C1 was deployed on the forereef slope, C3 and
C4 were located on the reef flat, while C5 and C6 were located inside the lagoon behind the
reef. The wave field exhibits a dramatic decay in the incident swell band on the fore reef
section with a transfer of part of the energy to infragravity waves which are dissipated due
to bottom friction over the reef flat and lagoon.
The XBeach model formulations have been extended with a friction dissipation term in the
wave action equation (Dongeren et al. (2013), which also describes the following case). For
this site, optimum friction coefficient values (fw = 0.6, cf = 0.1) were determined for a 1D
version of the model based on conditions at the peak of the swell event. These settings were
subsequently used to simulate the entire swell event from June 14 12:00 hours to June 19
00:00 hours (109 hours in total) when wave conditions varied significantly. Good agreement
was generally observed throughout the simulation and at all sites (Figure 2.12; Figure 5 in
Dongeren et al. (2013)). The model reproduced the spatial variability in wave heights across
the reef, as well as temporal changes in the response to the varying offshore wave conditions
and tidal variations. The short wave height predictions matched the data reasonably well
(Figure 2.12 a-e), except for a small positive bias of a few centimeters. The IG wave heights
were slightly under predicted (negative bias) at C1, but were generally in very good agreement
for sites on the reef (Figure 2.12 f-j). The time series of the predicted mean water level
residuals (the time-averaged difference between the observed water level on the reef and
the observation at C1, zs = zs zs,C1 , thus describing wave setup) followed the observed
residuals reasonably well, albeit that the model over predicts the observations by about 0.1
14 Deltares
4867
m. (Figure 2.12 l-o) Note that at C1 the observed and predicted water levels rather than the
residuals are shown (Figure 2.12 k). A summary of the model skill (bias and the RMS error)
for the short wave heights, IG heights and mean water level is shown in Figure 2.13 (Figure
6 in Dongeren et al. (2013)).
Figure 2.11: Cross-shore profile of the bathymetry along the main measurement transect
with instrument locations shown.
Deltares 15
4867
Hrms.sw (m) Hrms.IG (m) zs (m)

4 0.5 1
(a) (f) (k)
2 0.3 0
0.1
0 0 1
06/1506/1606/1706/1806/19 06/1506/1606/1706/1806/19 06/1506/1606/1706/1806/19
0.4 0.5 1
(b) (g) (l)
0.2 0.3 0
0.1
0 0 1
06/1506/1606/1706/1806/19 06/1506/1606/1706/1806/19 06/1506/1606/1706/1806/19
0.2 0.5 1
(c) (h) (m)
0.1 0.3 0
0.1
0 0 1
06/1506/1606/1706/1806/19 06/1506/1606/1706/1806/19 06/1506/1606/1706/1806/19
0.2 0.5 1
(d) (i) (n)
0.1 0.3 0
0.1
0 0 1
06/1506/1606/1706/1806/19 06/1506/1606/1706/1806/19 06/1506/1606/1706/1806/19
0.2 0.5 1
(e) (j) (o)
0.1 0.3 0
0.1
0 0 1
06/1506/1606/1706/1806/19 06/1506/1606/1706/1806/19 06/1506/1606/1706/1806/19
Figure 2.12: Comparison between the 1D model results (blue) and measured data (red)
for the duration of the 5 day swell event. (a,f,k) are for instrument C1, (b,g.l) C3, (c,h,m)
C4, (d,i,n) C5 and (e,j,o) C6. The peak of the storm is indicated by the red vertical line.
Note the large reduction in vertical scale between C1 and the reef sites C3-C6.
0.15 0.15
(a) (b)
0.1
0.05 0.1
RMSE (m)
Bias (m)
0.05 0.05
Swell
0.1 IG
zs
0.15 0
C1 C3 C4 C5 C6 C1 C3 C4 C5 C6
Instruments Instruments
Figure 2.13: Bias and RMS error from the 1D swell duration ( 5 day) model results
compared with the measured data at each site, based on the short wave heights, IG wave
heights, and mean water levels.
2.7 Longcrested refraction
Longcrested waves are supported in XBeach by using a value s larger than 1,000. The
directional distribution in that case is a spike corresponding to the mean wave direction. Due
to the discretization in the wave directional dimension, numerical diffusion in this dimension
may occur. In case of a narrow directional distribution, the numerical diffusion may lead to
a loss of wave energy through the boundaries of the directional grid.
To limit the numerical directional diffusion, the directional grid may be chosen with large grid
cells and thus decreasing the gradients between adjacent grid cells. Directional advection may
be turned of entirely when using a single directional grid cell, since refraction is not possible
with a single grid cell at all. However, when using oblique longcrested waves in a 2DH model,
none of these solutions suffice.
16 Deltares
4867
This test shows the numerical directional diffusion in the comparison of a few runs where the
angle of wave incidence and the number of directional bins vary. The bathymetry is a linear
sloping beach and no morphological change is computed in these runs.
0o 10o 20o
1 1
1 bin 0.5 0.5
0 0
0 0.5 1 0 0.5 1
3 bin
5 bin
Figure 2.14: Wave height
0o 10o 20o
1 1
1 bin
0.5 0.5
0 0
0 0.5 1 0 0.5 1
3 bin
5 bin
Figure 2.15: Mean wave direction
Deltares 17
4867
0o 10o 20o
1 1
1 bin
0.5 0.5
0 0
0 0.5 1 0 0.5 1
3 bin
5 bin
Figure 2.16: Time-averaged wave forcing in lateral direction
This test is to check the refraction behaviour of loncrested waves, which are simulated using
instat=stat and m=1000. Four cases are generated and compared: directional bins of 2.5, 5
and 10 degrees and a single bin. In all cases the incident wave angle is 60 degrees and theta
grid runs from -10 to 90 degrees. The results are compared among themselves and, for the
wave direction, with Snels law and with mean direction computed by integrating c /cg,x over
x. The results all agree within approx. one degree, whch is considered satisfactory. Note
that the longshore current does not exhibit any negative velocities, as was noticed when the
refraction is not computed correctly.
bathymetry [m]
10
20
0 100 200 300 400 500 600 700 800 900 1000
direction [o]
250
200
0 100 200 300 400 500 600 700 800 900 1000
H [m]
0
0 100 200 300 400 500 600 700 800 900 1000
0.2
ue [m]
0.2
0 100 200 300 400 500 600 700 800 900 1000
2
o
ve [m]
1 10
2p5o
0 5o
0 100 snellius
200 300 400 500 600 700 800 900 1000
distance [m]
Figure 2.17: Wave energy, direction and flow velocities for different sized of theta bins.
18 Deltares
4867
10o 2p5o
1000 1000
15000 15000
800 800
distance [m]
600 10000 600 10000
400 400
5000 5000
200 200
0 0
0 0
0 20 40 60 80 0 20 40 60 80
5o snellius
1000 1000
15000 15000
800 800
distance [m]
600 10000 600 10000
400 400
5000 5000
200 200
0 0
0 0
0 20 40 60 80 0 20 40 60 80
direction [o] direction [o]
Figure 2.18: Refraction with different sizes of theta bins. The red dotted line indicates
the solution according to Snel law.
2.8 Tide
This test was set up to verify the capability of XBeach to model tidal elevation and longshore
currents in a small model, based on water level time series at the seaward corner points,
which have a small time shift. The current pattern should spin up in a longshore uniform
pattern and remain longshore uniform at all times. In the figure velocity patterns at six
times during a one-day simulation are shown. Also the input timeseries of water levels at the
corners (dots) can be compared with the model water
level time series, which should closely follow the input but may have a small delay depending
on the type of boundary condition chosen (abs 1d, abs 2d or waterlevel) and on settings of
epsi (-1 in this case) and cats (5). In this case without wave forcing, the abs 2d boundary
type does not perform well; water levels do not match the input. This is probably due to the
absence of waves. With abs 1d the results as in the reference figure are shown.
Deltares 19
4867
t = 12hr t = 14hr t = 16hr

500
y [m]
500
1000
t = 18hr t = 20hr t = 22hr

500
0
y [m]
500
1000
0 1000 2000 0 1000 2000 0 1000 2000

x [m] x [m] x [m]
6
0.4
4
zs [m]
0.3
0.2
2
0.1
velocity [m/s] 0
0 5 10 15 20
time [hr]
Figure 2.19: The lower-left panel shows the tidal timeseries imposed to the offshore corners
of the model (dots) and the generated waterlevels at these locations (lines). For six
moments indicated with the black vertical lines, the flow field is shown in the upper
panels.
2.8.1 Blankenberge
This represents a realistic test model of the coast and port of Blankenberge on the Belgian
coast, in a model set up by IMDC and Flanders Hydraulics. The model experienced bad
tidal currents; after fixing this problem the accompanying figure show smooth and realistic
current patterns with a nearshore current dominated by wave-driven currents (instationary,
JONSWAP-type, with waves from 270 degrees, so a large angle to the coast. In this case the
abs 2d boundary option does not pose a problem to the tidal forcing. In the present setup
we run it with theta bins of 20 degrees.
20 Deltares
4867
5 t = 6hr t = 8hr t = 10hr

x 10
3.74
3.73
y [m]
3.72
3.71
5 t = 12hr t = 14hr t = 16hr

x 10
3.74
3.73
y [m]
3.72
3.71
6000 4000 2000 6000 4000 2000 6000 4000 2000

x [m] x [m] x [m]
1
1
zs [m]
0 0.5
1
velocity [m/s] 0
0 5 10 15 20
time [hr]
Figure 2.20: The lower-left panel shows the tidal timeseries imposed to the offshore corners
of the model (dots) and the generated waterlevels at these locations (lines). For six
moments indicated with the black vertical lines, the flow field is shown in the upper
panels.
Deltares 21
4867
22 Deltares
XBeach skillbed report, revision 4867 Morphological laboratory tests October 2015
4867
Chapter 3
Morphological laboratory tests
In this chapter, the performance of XBeach is compared to results obtained from physical
model tests performed in a variety of laboratory facilities. Many of those tests are part of
fundamental research into dune erosion and other morhpological processes. Research took
place on different scales, mainly depending on the size of the facility used. A key issue in
laboratory research is the translation of the laboratory results to the prototype situation,
which often requires scale relations.
The first section of this chapter covers an important measurement campaign commissioned
by the Dutch Ministry of Public Works in order to derive these scale relations. The follow-
ing sections cover in-depth analysis of other laboratory tests and research to bar evolution,
interactions with structures and other specific processes.
3.1 Scale relations
During the 1953 storm surge, large areas in the western part of the Netherlands were inun-
dated. This catastrophic event urged the Dutch Government to initiate fundamental research
to sea defences in general and dunes in particular. In the Netherlands, a major part of the sea
defence consists of dunes. The aim of the government was to make all sea defences withstand
a storm surge with a once in 10,000 years frequency of occurance.
In order to maintain this newly introduced norm, fundamental research to dune erosion was
necessary. A key issue in this field of research was the translation from laboratory results
to prototype situations. Therefore, the Ministry of Public Works commisioned a research
campaign in order to derive scale relations that facilitate this translation. Many experiments
with a variety of scales, sediments and hydraulic conditions are performed in this context
between 1974 and 1981. The experiments are presented in three parts: exploring experiments
on small scale, additional experiments on small scale and verification experiments on a large
scale. The three parts are discussed in the following subsections.
In contrast to models that are currently used for the assessment of dunes in The Netherlands,
XBeach is a process based model. If a process based model describes the relevant processes
of dune erosion (on different scales) accurately, it should be capable of reproducing the entire
series of tests in this research campaign. In order to verify whether XBeach is capable hereof,
the measurement results are compared to the XBeach results in the following subsections.
During all experiments described in this section, the reference profile for the Holland coast
Deltares 23
October 2015 Morphological laboratory tests XBeach skillbed report, revision 4867
4867
is used. This profile is a schematized profile that is considered representative for the Holland
coast.
3.1.1 Small scale tests
From 1974 to 1975, 58 model experiments for the derivation of scale relations are performed
in the Wind Flume of Laboratory De Voorst in the Netherlands (Van de Graaff, 1976). Four
different depth scales are used in these experiments: 150, 84, 47 and 26. Also, two different
sediment diameters are used: 225m and 150m. During all experiments, the dune is exposed
to a significant wave height of 7.6m and a constant maximum surge level of 5m+NAP on
prototype scale.
Table 3.1: Overview of experiments
Experiment Scale Profile Sediment Water Wave Wave

contraction diameter depth height period
BT13 84 9 150 0.461 0.091 1.31
CT14 84 9 225 0.461 0.091 1.31
BT15 84 7 150 0.461 0.091 1.31
CT16 84 7 225 0.461 0.091 1.31
BT17 84 5 150 0.461 0.091 1.31
CT18 84 5 225 0.461 0.091 1.31
BT23 47 7 150 0.585 0.163 1.76
CT24 47 7 225 0.585 0.163 1.76
BT25 47 5 150 0.585 0.163 1.76
CT26 47 5 225 0.585 0.163 1.76
BT27 47 3 150 0.585 0.163 1.76
CT28 47 3 225 0.585 0.163 1.76
AT33 26 3 150 0.806 0.292 2.35
DT34 26 3 225 0.806 0.292 2.35
AT35 26 1.5 150 0.806 0.292 2.35
DT36 26 1.5 225 0.806 0.292 2.35
AT37 26 5 150 0.806 0.292 2.35
DT38 26 5 225 0.806 0.292 2.35
AT61 84 5 150 0.461 0.091 1.31
BT62 84 5 150 0.461 0.091 1.31
CT63 84 7 225 0.461 0.091 1.31
DT64 84 5 225 0.461 0.091 1.31
AT71 26 3 150 0.806 0.292 2.35
BT71 17 2.6 150 0.806 0.292 2.35
CT71 26 3 225 0.806 0.292 2.35
DT71 26 3 225 0.806 0.292 2.35
AT91 47 3 150 0.806 0.292 1.76
BT92 26 3 150 0.806 0.292 1.76
CT93 26 3 225 0.806 0.292 1.76
DT94 26 4 225 0.806 0.292 1.76
AT95 47 3 150 0.806 0.163 2.35
BT96 47 4 150 0.806 0.163 2.35
CT97 26 3 225 0.806 0.163 2.35
DT98 47 3 225 0.806 0.163 2.35
24 Deltares
4867
The flume had a length of approximately 100m and a width of 4m. The flume was subdivided
in four parts with a width of 1m each. Using this configuration, it was possible to perform
four experiments at a time. All experiments are therefore considered to be one-dimensional.
Among the 58 experiments available, 6 experiments were calibration experiments, another 6
experiments are performed on a very small scale (150), in 4 experiments the profile develop-
ment was disturbed and in another 8 experiments a variable surge comparable with the 1953
storm surge was used. These experiments are excluded from the skillbed as for now. The 34
experiments left are presented in Table 3.1.
The profile development computed by XBeach is compared to the measurements for all 34
experiments. One of these comparisons is shown in Figure 3.2. For all other experiments,
only the resulting Brier Skill Score is presented in Figure 3.3.
profiles Windflume_M1263_I CT26

0.85
initial
measured
0.8 XBeach (current) BSS=0.80
0.75
0.7
0.65
height [m]
0.6
0.55
0.5
0.45
0.4
0.35
75.5 76 76.5 77 77.5 78 78.5
distance [m]
Figure 3.1: CT26
Deltares 25
4867
bed level change

0.2
0.1
height [m]
0
0.1
computed
0.2
0 10 20 30 40 50 60 70 80 90
distance [m]
erosion volume
0.15
volume [m3/m]
0.1
0.05
0
computed
0.05
0 1 2 3 4 5 6 7
time [h]
retreat distance
0.8
distance [m]
0.6
0.4
0.2
computed
0
0 1 2 3 4 5 6 7
time [h]
Figure 3.2: CT26
DT98
DT94
DT74
DT64
DT48
DT34
CT97
CT93
CT73
CT63
CT46
CT28
CT26
CT24
CT18
CT16
CT14
BT96
BT92
BT72
BT62
BT45
BT27
BT25
BT23
BT17
BT15
BT13
AT95
AT91
AT71
AT61
AT47
AT33
2 1.5 1 0.5 0 0.5 1
Figure 3.3: Overview of Brier Skill Scores
3.1.2 Additional small scale tests
The experiments presented in the previous section resulted in a set of scale relations (Van de
Graaff, 1976). Moreover, the experiments indicated that the process of dune erosion scaled
with the dimensionless parameter THw . In order to verify these findings, a series of additional
small scale tests is performed in 1976 and 1977 (Vellinga, 1981). Again, the Wind Flume of
Laboratory De Voorst in The Netherlands was used.
During the additional experiments, only the depth scales 84, 47 and 26 are used. The sediment
diameter is varied between 95m and 225m. Tests with constant and varying waterlevels
are performed. Here, only the tests with constant water levels are considered. XBeach is
compared to all tests performed on scale 26 plus the tests on the scales 84 and 47 for which
26 Deltares
4867
a sediment diameter of 225m was used. These tests are summarized in Table 3.2.

111 84 4.0 225 0.461 0.091 1.31
115 84 2.9 225 0.461 0.091 1.31
101 47 3.4 225 0.585 0.163 1.76
105 47 2.5 225 0.585 0.163 1.76
121 26 3.0 225 0.806 0.292 2.35
122 26 2.0 225 0.806 0.292 2.35
123 26 2.2 150 0.806 0.292 2.35
124 26 1.5 150 0.806 0.292 2.35
125 26 1.6 130 0.806 0.292 2.35
126 26 1.1 130 0.806 0.292 2.35
127 26 1.3 95 0.806 0.292 2.35
128 26 1.0 95 0.806 0.292 2.35
Figure 3.5 shows the profile comparison of one of the tests from Table 3.2. For the other
tests, the Brier Skill Score is presented in Figure 3.6.
profiles Windflume_M1263_II 121

1.3
initial
measured
1.2 XBeach (current) BSS=0.90
1.1
0.9
height [m]
0.8
0.7
0.6
0.5
0.4
92 93 94 95 96 97
distance [m]
Figure 3.4: 121
Deltares 27
4867
bed level change

0.4
0.2
height [m]
0
0.2
computed
0.4
0 10 20 30 40 50 60 70 80 90 100
distance [m]
erosion volume
0.6
volume [m3/m]
0.4
0.2
0
computed
0.2
0 1 2 3 4 5 6 7
time [h]
retreat distance
1.5
distance [m]
0.5
computed
0
0 1 2 3 4 5 6 7
time [h]
Figure 3.5: 121
128
127
126
125
124
123
122
121
115
111
105
101
0.2 0 0.2 0.4 0.6 0.8 1 1.2
3.1.3 Large scale tests
The experiments presented in the previous sections (Van de Graaff, 1976; Vellinga, 1981),
resulted in a final theory on scale relations and the translation of small scale results to
prototype situations (Vellinga, 1986). In order to verify this theory, large scale model tests
are performed in 1980 and 1981 in the Deltaflume of Delft Hydraulics, currently known as
Deltares (Vellinga, 1984). The Deltaflume is approximately 230m long, 5m wide and 7 to 9m
deep.
Five experiments are performed, as presented in Table 3.3. Tests 1, 2 and 5 had a constant
surge level, while tests 3 and 4 had a variable surge level with a course depicted in Figure 3.7
and Figure 3.8 respectively.
28 Deltares
4867

1 5 1.91 225 4.2 1.50 5.4
2 5 1.27 225 4.2 1.50 5.4
3 5 1.27 225 4.2 1.50 5.4
4 3.27 1.91 225 4.2 1.85 5.0
5 1 1 225 5.0 2.00 7.6
4.5
water level [m]
3.5
3
0 1 2 3 4 5 6 7 8
time [s] 4
x 10
2
wave height [m]
1.5
1
0 1 2 3 4 5 6 7 8
time [s] 4
x 10
5.5
peak wave period [s]
4.5
0 1 2 3 4 5 6 7 8
time [s] 4
x 10
Figure 3.7: Boundary conditions for test 3
4.5
water level [m]
3.5
3
0 1 2 3 4 5 6 7
time [s] 4
x 10
2
wave height [m]
1.8
1.6
1.4
0 1 2 3 4 5 6 7
time [s] 4
x 10
6
peak wave period [s]
5.5
4.5
0 1 2 3 4 5 6 7
time [s] 4
x 10
Figure 3.8: Boundary conditions for test 4
Deltares 29
4867
As in the previous sections, the profile developments are compared to the measurements in
Figure 3.9 to Figure 3.13. Moreover, erosion patterns and volumes and retreat distances are
analyzed in Figure 3.15 to Figure 3.19. Figure 3.14 provides an overview of the Brier Skill
Scores of the profile development for the different tests.
profiles Deltaflume_M1263_III Test1

6.5
initial
measured
6 XBeach (current) BSS=0.95
5.5
4.5
height [m]
3.5
2.5
2
160 170 180 190 200 210
distance [m]
Figure 3.9: Profile development during test 1

6.5
initial
measured
5.5
4.5
height [m]
3.5
2.5
2
160 165 170 175 180 185 190 195 200 205 210
distance [m]
30 Deltares
4867

6.5
initial
measured
5.5
4.5
height [m]
3.5
2.5
2
150 160 170 180 190 200 210
distance [m]

6.5
initial
measured
6
XBeach (current) BSS=0.54
5.5
4.5
height [m]
3.5
2.5
1.5
140 150 160 170 180 190 200 210
distance [m]
Deltares 31
4867

9
initial
measured
6
height [m]
1
120 130 140 150 160 170 180 190 200 210
distance [m]
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
Test1 Test2 Test3 Test4 Test5
32 Deltares
4867
bed level change

1
height [m]
1
2
computed
3
20 40 60 80 100 120 140 160 180 200 220
distance
erosion [m]
volume
30
volume [m /m]
20
3
10
computed
0
0 2 4 6 8 10 12
timedistance
retreat [h]
15
distance [m]
10
5
computed
0
0 2 4 6 8 10 12
time [h]
Figure 3.15: Development of erosion volumes during test 1
bed level change

1
0
height [m]
2
computed
3
20 40 60 80 100 120 140 160 180 200 220
distance
erosion [m]
volume
20
volume [m /m]
15
3
10
5
computed
0
0 2 4 6 8 10 12
timedistance
retreat [h]
15
distance [m]
10
5
computed
0
0 2 4 6 8 10 12
time [h]
Deltares 33
4867
bed level change

1
height [m]
1
2
computed
3
0 50 100 150 200 250
distance
erosion [m]
volume
20
volume [m /m]
15
3
10
5
computed
0
0 2 4 6 8 10 12 14 16 18 20
timedistance
retreat [h]
15
distance [m]
10
5
computed
0
0 2 4 6 8 10 12 14 16 18 20
time [h]
bed level change

1
0
height [m]
2
computed
3
100 50 0 50 100 150 200 250
distance
erosion [m]
volume
15
volume [m /m]
10
3
5
computed
0
0 2 4 6 8 10 12 14 16 18
timedistance
retreat [h]
10
distance [m]
computed
0
0 2 4 6 8 10 12 14 16 18
time [h]
34 Deltares
4867
bed level change

2
height [m]
2
4
computed
6
100 50 0 50 100 150 200 250
distance
erosion [m]
volume
80
volume [m /m] 60
3
40
20
computed
0
0 1 2 3 4 5 6 7
timedistance
retreat [h]
30
distance [m]
20
10
computed
0
0 1 2 3 4 5 6 7
time [h]
3.2 Large scale tests
This section discusses a selection of large scale laboratory experiments. The experiments
focus on specific situations and processes that XBeach is assumed to model well.
3.2.1 Revetments
Revetments or other hard structures influence the dune erosion process. Scour holes and
wave runup on structures can occur, both important phenomena in the assessment of water
defences. Scour holes are mainly caused due to the obstructed transport of sediments in
seaward direction from the dune. A transport gradient causes a scour hole to occur. This is
also valid for other hard elements present in or near a dune. (Short) wave runup can cause
overtopping of the revetment, which can cause erosion at the rear-side of the revetment.
3.2.1.1 Influence of reventment height
Steetzel (1987) describes a series of large scale experiments with revetments of different
heights. The experiments are performed in the Deltaflume of Delft Hydraulics, now known
as Deltares. A depthscale nd = 5 is used for all experiments (Vellinga, 1986) and the initial
profile in the flume correponds to the reference profile for the Holland coast. At the dune
foot, which was located at 193m from the wave board and 3.80m above the flumes floor, a
concrete revetment is applied that covers almost the whole dune face with a slope of 1:1.8.
The lower end of the revetment is located at 2.5m above the flume floor. The location of the
top of the revetment varied in each experiment. The tests were conducted with a constant
water level of 4.2m and wave conditions that correspond to a Pierson-Moskowitz spectrum
with Hm0 = 1.50m and Tp = 4.20s. The sediment applied in the test had a median grain
diameter of approximately D50 = 210m. An overview of the tests is given in Table 3.4.
Deltares 35
4867
Experiment Revetment height w.r.t. flume floor

T1 6.20m
T2 5.40m
T3 4.80m
The profile developments as measured and computed by XBeach are presented in Figure 3.20
to Figure 3.22.
Figure 3.20: Profile development during test T1
36 Deltares
4867
3.2.2 Extreme conditions
This model test, described in Arcilla et al. (1994) (2E), concerns extreme conditions with a
raised water level at 4.6m above the flume bottom, a significant wave height, Hm0 , of 1.4m
and peak period, Tp , of 5s. Bed material consisted of sand with a D50 of approximately
0.2mm. During the test substantial dune erosion took place.
Based on the integral wave parameters Hm0 and Tp and a standard Jonswap spectral shape,
Deltares 37
4867
time series of wave energy were generated and imposed as boundary condition. Since the
flume tests were carried out with first-order wave generation (no imposed super-harmonics
and sub-harmonics), the hindcast runs were carried out with the incoming bound long waves
set to zero as well. Active wave reflection compensation (ARC) was applied in the physical
model, which has a result similar to the weakly reflective boundary condition in XBeach,
namely to prevent re-reflecting of outgoing waves at the wave paddle (offshore boundary).
A grid resolution of 1m was applied and the sediment transport settings were set at default
values. For the morphodynamic testing the model was run for 0.8 hours of hydrodynamic
time with a morphological factor of 10, effectively representing a morphological simulation
time of 8 hours.
In Figure 3.23 the hydrodynamic results are shown for first order wave generation (as in the
flume tests).

1.5
1
height [m]
setup (measured)
wave height (HF)
wave height (LF)
0.5 wave height
setup
0
0 20 40 60 80 100 120 140 160 180
distance [m]
flow velocities
0.8
flow velocity (RMS,LF,measured)
flow velocity (RMS,measured)
0.6 flow velocity (RMS,LF)
velocity [m/s]
flow velocity (RMS)

0.4
0.2
0
0 20 40 60 80 100 120 140 160 180
distance [m]
Figure 3.23: Hydrodynamics during test 2E
In Figure 3.24 the horizontal distribution of sedimentation and erosion after 8 hours is shown,
and the evolution in time of the erosion volume and the dune retreat. Noteworthy is the
episodic behaviour of the dune erosion, both in measurements and model. An important
conclusion for physical model tests is that for dune erosion it does make a difference whether
first-order or second-order wave steering is applied.
38 Deltares
4867
bed level change

0.5
height [m]
0.5
measured
1
computed
1.5
0 20 40 60 80 100 120 140 160 180 200
distance
erosion [m]
volume
10
volume [m3/m] 5
0 measured
computed
5
0 1 2 3 4 5 6 7 8
timedistance
retreat [h]
10
distance [m]
5
measured
computed
0
0 1 2 3 4 5 6 7 8
time [h]
Figure 3.24: Morphodynamics during test 2E
A key element in the modelling is the avalanching algorithm; even though surfbeat waves
running up and down the upper beach are fully resolved by the model, without a mechanism
to transport sand from the dry dune face to the beach the dune face erosion rate is substan-
tially underestimated. The relatively simple avalanching algorithm implemented, whereby an
underwater critical slope of 0.3 and a critical slope above water of 1.0 are applied, proves to
be quite successful in representing the retreat of the upper beach and dune face. In ?? the
measured and modelled bed evolution are shown.
profiles DeltaflumeLIP11D 2E
6
initial
measured
5.5
4.5
height [m]
3.5
2.5
130 140 150 160 170 180 190
distance [m]
Figure 3.25: Profile development during test 2E
Deltares 39
4867
3.2.3 Bar evolution
In Arcilla et al. (1994) two types of profiles and three different dynamic states are investigated:
near-equillibrium, erosive and accretive. Also, a test with extreme erosive conditions is
performed. In this section the erosive and accretive tests for the first profile are discussed:
test 1B and 1C. The two tests are intended to simulate the erosive effects during a storm and
the restoring effects afterwards. A clear dune front is absent in the profile used, only a gently
sloping beach is present. In Table 3.5 the conditions of these two tests are summarized.
Experiment Hm0 Tp Water level Duration

1B 1.4m 5s 4.1m 18h
1C 0.6m 8s 4.1m 13h
In Figure 3.26 and Figure 3.27 the computed and measured hydrodynamics are shown.

1
0.5
height [m]
setup (measured)
wave height (HF)
wave height (LF)
0 wave height
setup
0.5
0 20 40 60 80 100 120 140 160 180 200
distance [m]
flow velocities
0.8
velocity [m/s]
flow velocity (RMS)

0.4
0.2
0
0 20 40 60 80 100 120 140 160 180 200
distance [m]
Figure 3.26: Hydrodynamics during test 1B
40 Deltares
4867

0.6
0.4 wave height (measured)
height [m]
setup (measured)
wave height (HF)
0.2
wave height (LF)
wave height
0 setup
0.2
0 20 40 60 80 100 120 140 160 180
distance [m]
flow velocities
0.8
velocity [m/s]
flow velocity (RMS)

0.4
0.2
0
0 20 40 60 80 100 120 140 160 180
distance [m]
Figure 3.27: Hydrodynamics during test 1C
The profile development is shown in Figure 3.28 and Figure 3.29.
profiles DeltaflumeLIP11D 1B
1
initial
measured
0.5
0
height [m]
0.5
1.5
130 140 150 160 170 180 190 200
distance [m]
Figure 3.28: Profile development during test 1B
Deltares 41
4867
profiles DeltaflumeLIP11D 1C
1
initial
measured
0.5
0
height [m]
0.5
1.5
130 140 150 160 170 180 190 200
distance [m]
Figure 3.29: Profile development during test 1C
3.2.4 Influence of wave period
In Van Gent et al. (2008) and Van Thiel de Vries et al. (2008) describe large-scale laboratory
experiments on the influence of the wave period on the dune erosion process. It was concluded
that not only short waves, but also (wave group generated) long waves are important in the
dune erosion process. Initially, about 30% of the dune erosion is due to long wave energy,
but this amount increases with the development of an erosion profile. Moreover, an increase
of the wave period increases the dune erosion volumes.
XBeach, being a wave group solving model, is expected to be able to reproduce the observed
influence of the wave period well. The following sections provide an in-depth comparison
of the first experiment from Van Gent et al. (2008). It continue with a comparison of the
influence of the wave period and spectrum shape between the measurements obtained and
the XBeach results.
All experiments are performed in the Deltaflume of Delft Hydraulics, currently known as
Deltares, using the reference profile for the Holland coast. This is a schematized profile that
is considered representative for the Holland coast. Furthermore, a significant wave height
1.50m and a water depth of 4.50m is used. The test programme is given in Table 3.6. During
test T04 a profile with a double dune row is used.
Experiment Tp Tm1,0 Spectrum

T01 4.90 4.45 Pierson-Moskowitz
DP01 6.12 3.91 double-peaked
DP02 7.35 5.61 double-peaked
42 Deltares
4867
3.2.4.1 Detailed analysis
In this section, a detailed comparison between simulated physics over an evolving bathymetry
and the measurements obtained during the Deltaflume experiment in 2006 (Van Gent et al.,
2008) is made. For brevity this comparison is performed only for test T01 (this test corre-
sponds best to the Dutch normative conditions). The simulations are performed on a regular
grid with dx = 1m and input to the model are time series of short wave varying energy (low
pass filtered on the wave group time scale) and incoming (bound) long waves. The time series
are constructed from pressure and flow measurements at x = 41m from the wave board. The
short wave group velocity (associated with advection of wave action) is based on the Tm1,0
wave period. Other model settings can be found in Van Thiel de Vries (2009).
Wave height transformation and wave setup (Figure 3.45) are favourably reproduced with the
model. The long wave height is slightly underestimated whereas the wave setup is slightly
overestimated. The correlation between measured short wave variance and long wave water
surface elevations (Figure 3.31) corresponds reasonably well with the measurements. Towards
the shoreline this correlation increases (Abdelrahman and Thornton, 1987; Roelvink and
Stive, 1989) meaning the highest short waves travel on top of long waves, which likely causes
that more short wave energy gets closer to the dune face.
Short wave skewness and asymmetry are reasonably predicted with the extended Rienecker
Fenton model (Figure 3.32, panel 1). However, in the inner surf zone both wave skewness
and asymmetry are overestimated. Possible explanations are wave breaking, which limits the
steepness and height of waves and the presence of free harmonics in the flume. Both these
effects are not included in the wave shape model but indeed are present in the flume test
(Van Thiel de Vries, 2009). From simulated skewness and asymmetry it follows that the total
nonlinearity of a short wave is overestimated close to the dune face (Figure 3.32, panel 2).
The phase is favourably simulated with the model but is underestimated further offshore.

1.5
1 wave height (measured)
height [m]
setup (measured)
wave height (HF)
0.5
wave height (LF)
wave height
0 setup
0.5
40 60 80 100 120 140 160 180 200 220
distance [m]
flow velocities
1
flow velocity (RMS,HF,measured)
velocity [m/s]
0.5
flow velocity (u,mean,measured)
flow velocity (RMS,HF)
flow velocity (RMS,LF)
0 flow velocity (RMS)
flow velocity (u,mean)
0.5
40 60 80 100 120 140 160 180 200 220
distance [m]
Figure 3.30: Wave height transformations and flow velocities
Deltares 43
4867
correlations
0.8
correlation HF variance/LF elevation (measured)
correlation HF variance/LF elevation
0.6
0.4
0.2
correlation []
0.2
0.4
0.6
0.8
40 60 80 100 120 140 160 180 200 220
distance [m]
Figure 3.31: Correlation between short wave variance and long wave surface elevations
wave shape
1
wave skewness (measured)
skewness & asymmetry []
0.5 wave asymmetry (measured)

wave skewness
0
wave asymmetry
0.5
1.5
2
40 60 80 100 120 140 160 180 200 220
distance [m]
wave nonlinearity
2
wave nonlinearity (measured)
nonlinearity & phase []
wave phase (measured)

1 wave nonlinearity
wave phase
0
2
40 60 80 100 120 140 160 180 200 220
distance [m]
Figure 3.32: Wave shape and nonlinearity
The simulated test and depth averaged flow velocity shows the same trend as in the mea-
surements and increases towards the shoreline (Figure 3.33). However, in the simulation the
cross-shore range with a high offshore mean flow is smaller and extends less far seaward than
in the measurements. This is possibly explained by differences in profile development (??)
or inaccurate measurements. In addition, another explanation may be found in the incorrect
modeling of the roller energy dissipation. Simulations (not shown) with a smaller roller dis-
sipation rate revealed that roller energy in the inner surf increases, leading to higher return
flow over a broader cross-shore range.
Long waves contribute to the time and depth averaged flow close to the shoreline. The
contribution of long waves to the mean flow is explained by on average larger water depths
44 Deltares
4867
during the interval associated with shoreward flow velocities in relation to the interval with
offshore flow velocities. Considering continuity and a uniform vertical structure of the long
wave flow this means a time and depth averaged offshore directed flow should be present.
Nonlinear waves may cause onshore sediment transport presuming non-uniform sediment
stirring over the wave cycle and a positive correlation between sediment suspension and the
intra wave flow. In order to include the wave averaged effect of nonlinear waves on the
sediment transport a mean flow uA is computed, which is added to the mean (Eulerian) flow
Um (Van Thiel de Vries, 2009). The simulated time averaged flow associated with nonlinear
waves shows a comparable evolution as in the measurements but is overestimated especially
closer to the dune face. Near the shoreline the wave skewness related sediment transport
vanishes (Figure 3.32, panel 1) since waves develop towards fully saw tooth shaped bores
that have negligible skewness.
The orbital flow velocity (Figure 3.45, panel 2) is favourably predicted by the model. The
short wave orbital flow velocity is slightly overestimated whereas the long wave orbital flow
is underestimated. The underestimation of the simulated long wave orbital flow corresponds
well to the slight underestimation of the observed long wave water surface variance.
wave orbital velocity and undertow

0.9
0.8 flow velocity (RMS,measured)
undertow velocity (measured)
flow velocity (RMS)
undertow velocity
0.6
velocity [m/s]
0.5
0.4
0.3
0.2
0.1
0
40 60 80 100 120 140 160 180 200 220
distance [m]
Figure 3.33: Wave motion velocities and undertow
The simulated test and depth averaged sediment concentration increases towards the shoreline
but is underestimated, especially in deeper water where the modeled sediment concentration
is smaller (Figure 3.34). In the proximity of the dune face the simulated mean sediment
concentration is within a factor two with the measurements. Further offshore the discrep-
ancy between simulations and measurements is larger. The sharp rise in the near dune
sediment concentration compares well with the bore averaged near-bed turbulence intensity
(Figure 3.35) that also increases towards the shoreline. This increase in turbulence intensity
through the inner surf is explained by more intensive wave breaking (turbulence pro-duction
at the water surface increases) and by decreasing water depth (generated turbulence at the
water surface is more effective in reaching the bed).
The simulated time averaged sediment transport compares well with the measured sediment
transport computed from profile changes (Figure 3.36, panel 1). Sediment is eroded from the
Deltares 45
4867
dune face via avalanching and as a result the sediment transport associated with avalanching
is dominant over the dune face and in the swash zone. From the swash zone seaward, the
flow based sediment transport becomes more important. At 205m from the wave board, in
a water depth that varies between 0.1m and 0.2m, the flow related sediment transport is
dominant.
The simulated flow related sediment transport is separated in sediment transports as-ociated
with nonlinear waves (SW ), long waves (SL ) and the short wave driven under-tow (SR )
(Figure 3.36, panel 2):
SW = uA ch
SL = uL ch
SR = (uE uL )ch
The offshore sediment transport results from the short wave and roller driven under-tow (SR )
combined with the transport associated with the long waves (SL ). The transport that follows
from the short wave undertow is dominant in the present simu-lation but the long wave
related sediment transport cannot be neglected (about 30% at the location of the maximum
offshore transport). The wave related sediment transport (SW ) is onshore and suppresses the
offshore sediment transport with some 30%.
Profile evolution and dune erosion volumes are favourably predicted with the model during
test T01 (Figure 3.39 and Figure 3.42, panel 2). Between t = 2.04 and 6.0 hours (interval
E) the dune erosion rate is slightly underestimated. At the offshore edge of the developing
foreshore, the model seems not capable to reproduce the steep transition from the original
(unaffected) profile towards the newly developed foreshore. A bar type feature is observed at
this transition that is hypothesized to be related to (partly) plunging breakers that generate
a water jet, which penetrates in the water column and causes additional sediment stirring
when it reaches the bed. Though the effect of wave breaking induced turbulence on sediment
suspension is included in the simulation, the applied model only considers spilling breakers,
which are expected to be less efficient than plunging breakers in stirring up sand.
sediment concentration
30
concentration (measured)
concentration
25
20
concentration [g/l]
15
10
0
40 60 80 100 120 140 160 180 200 220
distance [m]
Figure 3.34: Sediment concentrations
46 Deltares
4867
tubulence
1.4
wave averaged tubulence
bore averaged tubulence
bore averaged nearbed turbulence
1.2
tubulence [m /s ]
2 2
0.8
0.6
0.4
0.2
0
40 60 80 100 120 140 160 180 200 220
distance [m]
Figure 3.35: Wave induced turbulence
4
x 10 sediment transport (morphodynamics)
4
bed level changes (measured)
bed level changes
transport [m3/m/s]
2 avalanching
suspended
0 bedload
4
40 60 80 100 120 140 160 180 200 220
distance [m]
4
x 10 sediment transport (hydrodynamics)
2
total hydrodynamics
1 long waves
transport [m3/m/s]
short wave undertow

0 wave asymmetry
3
40 60 80 100 120 140 160 180 200 220
distance [m]
Figure 3.36: Sediment transports by transport type and forcing
Deltares 47
4867
profiles Deltaflume2006 T01

2
initial
measured
1.5
0.5
height [m]
0.5
1.5
185 190 195 200 205 210 215
distance [m]
Figure 3.37: Profile development
bed level change

1
0.5
0
height [m]
0.5
1.5 measured
computed
2
0 50 100 150 200 250
distance [m]
erosion volume
10
8
volume [m3/m]
2 measured
computed
0
0 1 2 3 4 5 6 7
time [h]
Figure 3.38: Bed level changes and erosion volumes
It is concluded that profile evolution and dune erosion during test T01 are favourably simu-
lated. Also simulated wave heights, flows, sediment concentrations and sediment transports
compare reasonably well with measurements. However,looking at the results in more detail
some discrepancies are found:
1. The long wave height and especially associated long wave orbital flows are un-derestimated.
2. The test and depth averaged flow between x = 170 m and x = 200 m is underestimated.
Close to the shoreline no reliable measurements are available to verify the model results.
3. The simulated sediment concentration compares well with measurements close to the
48 Deltares
4867
dune face. However, for smaller sediment concentrations in deeper water the simulated
concentration is underestimated.
4. The offshore sediment transport is mainly driven by the short wave and roller induced
undertow (O(70%) at the location of the maximum offshore transport) whereas the
offshore directed long wave related sediment transport cancels out with the onshore
sediment transport due to nonlinear short waves.
It is remarked that shoreward of the maximum offshore sediment transport, the importance
of the long wave related transport increases and eventually becomes dominant in relation to
the transport associated with short wave and roller driven undertow. Considering the mainly
long wave associated sediment transport in proximity of the dune face and the importance
of long wave run-up for avalanching it is expected that long waves are mainly responsible for
the swash zone sediment transport.
An overview of the Model Performance Statistics for test T01 with respect to the measure-
ments is given in ??.
Table 3.7

zb 0.99 0.03 -0.01 0.94
Hrmsl f NaN NaN NaN NaN
Hrmsh f NaN NaN NaN NaN
rho 0.84 0.61 -0.44 0.70
SK NaN NaN NaN NaN
AS NaN NaN NaN NaN
B NaN NaN NaN NaN
beta NaN NaN NaN NaN
urmsl f NaN NaN NaN NaN
urmsh f NaN NaN NaN NaN
umean NaN NaN NaN NaN
c NaN NaN NaN NaN
dz 0.02 5002.69 73.47 -25989440.45
3.2.4.2 Wave period
During test T01, T02 and T03 of the 2006 Deltaflume experiment (Van Gent et al., 2008) the
effect of the wave period on dune erosion was examined. It was found that the dune erosion
volume increases for a larger wave period, which is caused by a larger flow related sediment
transport. In addition, the increase in flow related sediment transport is mainly a result of a
higher mean sediment concentration whereas the time and depth averaged flow velocity has
the same order of magnitude within the range of wave periods examined.
The aim of this section is to examine the processes, which cause the wave period effect in
the model and to what extent these processes are in line with the mechanisms observed in
the measurements. To this end test T01, T02 and T03 of the Deltaflume experiment are
simulated. The simulations are performed on a regular grid with dx = 1 m and input to
the model are time series of short wave varying energy (low pass filtered on the wave group
time scale) and incoming (bound) long waves. The time series are constructed from pressure
and flow measurements at x = 41 m from the wave board. The short wave group velocity
Deltares 49
4867
(associated with advection of wave action) is based on the Tm1,0 wave period. Settings for
the wave energy dissipation model are found in Table 3.8 and other setting are defaults as
listed in in Appendix B in Van Thiel de Vries (2009).
Table 3.8: Wave dissipation parameter settings
Experiment n
T01 1.0 0.50 10
T02 1.0 0.48 10
T03 1.0 0.45 10
In Figure 3.39 to Figure 3.41 the simulated profile evolution during the tests compares well
with the measured profile evolution. However, the bar type feature at the offshore edge of
the developing foreshore is not reproduced in the simulations. Comparing measured and
simulated dune erosion volumes in Figure 3.42 to Figure 3.44 it is seen that for test T03 the
erosion volume after 1 hour and 2.04 hours of waves is underestimated whereas for test T01
and T02 simulated erosion volumes are reasonably well predicted over all test intervals.

2
initial
measured
1.5
0.5
height [m]
0.5
1.5
185 190 195 200 205 210 215
distance [m]
50 Deltares
4867

2
initial
measured
1.5
height [m] 0.5
0.5
1.5
180 185 190 195 200 205 210 215
distance [m]

2
initial
measured
1.5
0.5
height [m]
0.5
1.5
180 185 190 195 200 205 210 215
distance [m]
Deltares 51
4867
bed level change

1
0.5
height [m] 0.5
1.5 measured
computed
2
0 50 100 150 200 250
distance [m]
erosion volume
10
8
volume [m3/m]
2 measured
computed
0
0 1 2 3 4 5 6 7
time [h]
Figure 3.42: Erosion pattern and volumes and retreat distance during test T01
bed level change

1
0.5
0
height [m]
0.5
1.5 measured
computed
2
0 50 100 150 200 250
distance [m]
erosion volume
12
10
volume [m3/m]
2 measured
computed
0
0 1 2 3 4 5 6 7
time [h]
52 Deltares
4867
bed level change

1
0.5
height [m]
0.5
1.5 measured
computed
2
0 50 100 150 200 250
distance [m]
erosion volume
12
10
volume [m3/m]
2 measured
computed
0
0 1 2 3 4 5 6 7
time [h]
Simulated time and depth averaged flows are compared in Figure 3.45 to Figure 3.47, which
shows that the mean flow slightly increases with a larger wave period (6% between x = 170 m
and x = 205 m). This tends to be a slightly different trend than observed in the measurements
that show a 3% increase in the same cross-shore area. It is remarked though that any firm
conclusions cannot be made since the measured time and depth averaged flow velocities are
based on limited sensors over depth (and for that reason it was concluded that the mean
flows for test T01 and T03 are comparable).
The increase in the simulated offshore directed mean flow is caused by an increase of the
short wave related mass flux whereas the roller related mass flux is only slightly different.
The maximum long wave related mean flow is larger during test T01 than during test T03.

1.5
height [m]
setup (measured)
wave height (HF)
0.5
wave height (LF)
wave height
0 setup
0.5
40 60 80 100 120 140 160 180 200 220
distance [m]
flow velocities
1
velocity [m/s]
0.5
0.5
40 60 80 100 120 140 160 180 200 220
distance [m]
Figure 3.45: Hydrodynamics during test T01
Deltares 53
4867

1.5
height [m]
setup (measured)
wave height (HF)
0.5
wave height (LF)
wave height
0 setup
0.5
40 60 80 100 120 140 160 180 200 220
distance [m]
flow velocities
1
velocity [m/s]
0.5
0.5
40 60 80 100 120 140 160 180 200 220
distance [m]

1.5
height [m]
setup (measured)
wave height (HF)
0.5
wave height (LF)
wave height
0 setup
0.5
40 60 80 100 120 140 160 180 200 220
distance [m]
flow velocities
1
velocity [m/s]
0.5
0.5
40 60 80 100 120 140 160 180 200 220
distance [m]
It is concluded that the effect of the wave period on dune erosion and dune face retreat is
favourably predicted with the XBeach model. The simulated increase in dune erosion with
a larger wave period is mainly caused by an increase of the mean sediment concentration
of O(60%), which is comparable to the measurements. The near dune return flow slightly
increases with approximately 6% for a larger wave period. However, the accuracy of measured
time and depth averaged flows is insufficient to verify this increase.
54 Deltares
4867
3.2.4.3 Spectral shape
Test DP01 and DP02 were conducted with a double-peaked wave spectrum to investigate
what (spectral) wave period is best qualified to describe dune erosion (Van Gent et al.,
2008). In this document the tests are discussed to obtain further insight in the capability of
the model to simulate dune erosion for various wave spectra. The simulations are performed
on a regular grid with dx = 1m and input to the model are time series of short wave varying
energy (low pass filtered on the wave group time scale) and incoming (bound) long waves.
The time series are constructed from pressure and flow measurements at x = 41 m from the
wave board. The short wave group velocity (associated with advection of wave action) is
based on the Tm1,0 wave period. Settings for the wave energy dissipation model are found
in Table 3.9.
Table 3.9: Wave dissipation parameter settings
Experiment n
DP01 1.0 0.50 10
DP02 1.0 0.48 10
Simulated and measured profile evolution and dune erosion volumes for test DP01 and DP02
are compared in Figure 3.48 to Figure 3.51 respectively. For test DP01 the profile evolution
is accurately reproduced and results for test DP02 are reasonable even though the erosion
rate during the last interval is overestimated.
profiles Deltaflume2006 DP01

2
initial
measured
1.5
0.5
height [m]
0.5
1.5
185 190 195 200 205 210 215
distance [m]
Figure 3.48: Profile development during test DP01
Deltares 55
4867
profiles Deltaflume2006 DP02

2
initial
measured
1.5
0.5
height [m]
0.5
1.5
180 185 190 195 200 205 210 215
distance [m]
Figure 3.49: Profile development during test DP02
bed level change

1
0.5
0
height [m]
0.5
1.5 measured
computed
2
0 50 100 150 200 250
distance [m]
erosion volume
10
8
volume [m3/m]
2 measured
computed
0
0 1 2 3 4 5 6 7
time [h]
Figure 3.50: Erosion pattern and volumes and retreat distance during test DP01
56 Deltares
4867
bed level change

1
0.5
height [m]
0.5
1.5 measured
computed
2
0 50 100 150 200 250
distance [m]
erosion volume
15
volume [m3/m]
10
5
measured
computed
0
0 1 2 3 4 5 6 7
time [h]
Figure 3.51: Erosion pattern and volumes and retreat distance during test DP02
The imposed double-peaked wave spectra affect the time scale and amplitude of the simulated
wave groups. Consequently, the interaction of simulated long waves with the short wave
groups is different and hydrodynamics in front of the dune face are expected to have other
characteristics. In Figure 3.52 and Figure 3.53 the simulated wave transformation, flows and
sediment concentrations are favourably compared with measurements obtained during test
DP01. It seems the model is capable to take into account the effect of various wave spectra
on near dune hydrodynamics and sediment transports.

1
0.5
height [m]
setup (measured)
wave height (HF)
wave height (LF)
0 wave height
setup
0.5
40 60 80 100 120 140 160 180 200 220
distance [m]
flow velocities
1
velocity [m/s]
0.5
0.5
40 60 80 100 120 140 160 180 200 220
distance [m]
Figure 3.52: Hydrodynamics during test DP01
Deltares 57
4867

1
0.5
height [m]
setup (measured)
wave height (HF)
wave height (LF)
0 wave height
setup
0.5
40 60 80 100 120 140 160 180 200 220
distance [m]
flow velocities
1
velocity [m/s]
0.5
0.5
40 60 80 100 120 140 160 180 200 220
distance [m]
Figure 3.53: Hydrodynamics during test DP02
It is concluded that the effect of the wave spectral shape on dune erosion and dune face retreat
is favourably predicted with the model. The time averaged simulated wave transformation,
flow and sediment concentration compare well with the mobile frame measurements obtained
during test DP01.
In Van Gent et al. (2008) the spectral mean wave period Tm1,0 is argued to be more qualified
to describe dune erosion than the peak wave period Tp . The simulations presented in this
subsection are performed with the Tm1,0 wave period and show satisfying results suggesting
the spectral mean wave period proposed by Van Gent et al. (2008) is indeed a good measure
to describe dune erosion. It is remarked though that any firm conclusion would require extra
simulations in which the peak wave period Tp is applied instead of the Tm1,0 wave period. In
addition this would demand for a new model optimization and most likely different settings
for the wave dissipation model.
3.2.4.4 Double dune system
Test T04 of the Deltaflume experiment 2006 (Van Gent et al., 2008) is carried out with an
initial profile that contains a small dune in front of a larger volume dune that collapses after
approximately one hour of waves (interval C and D). After breaching of the small dune,
the foreshore is already very efficient in reducing wave impacts on the dune face resulting in
small erosion rates over the remaining test intervals. In this test it is examined to what extent
the dune breach can be reproduced with the XBeach model and whether the (substantially
smaller) erosion rate at the end of a storm is correctly predicted. The simulations is performed
on a regular grid with dx = 1 m and input to the model are time series of short wave varying
energy (low pass filtered on the wave group time scale) and incoming (bound) long waves.
The time series are constructed from pressure and flow measurements at x = 41 m from the
wave board. The short wave group velocity (associated with advection of wave action) is
based on the Tm1,0 wave period. Settings for the wave energy dissipation model are equal
to the settings for the T03 test.
Simulated profile evolution and dune erosion volumes are shown in Figure 3.54 and Figure 3.55
58 Deltares
4867
respectively. The simulated hydrodynamics are presented in Figure 3.56. For the first three
intervals (the small dune breaches in interval 3) the dune erosion rate is slightly overestimated
but the profile evolution compares favourably with the measured profiles. Considering the
last two intervals, erosion rates and dune face retreat are too large.
Breaching of a small dune in front of a larger volume dune causes that suddenly the foreshore
is significantly closer to equilibrium with the storm surge conditions. As a result near shore
hydrodynamics, near shore sediment transports and wave impacts on the dune face are less
severe. It is concluded that the feedback between profile evolution and near dune processes
is not sufficiently well included in the model at the end of test T04, which is representative
for conditions at the end of a storm. More insight in the model performance at the end of a
storm could possibly be obtained by comparing the evolution of simulated driving processes
(undertow, sediment concentrations and avalanching) with measurements.

2
initial
measured
1
0
height [m]
4
60 80 100 120 140 160 180 200 220
distance [m]
Deltares 59
4867
bed level change

0.5
height [m]
0.5
1.5 measured
computed
2
0 50 100 150 200 250
distance [m]
erosion volume
8
6
volume [m3/m]
2
measured
computed
0
0 1 2 3 4 5 6 7
time [h]

1.5
height [m]
setup (measured)
wave height (HF)
0.5
wave height (LF)
wave height
0 setup
0.5
40 60 80 100 120 140 160 180 200 220
distance [m]
flow velocities
1
velocity [m/s]
0.5
0.5
40 60 80 100 120 140 160 180 200 220
distance [m]
60 Deltares
XBeach skillbed report, revision 4867 Morphological field tests October 2015
4867
Chapter 4
Morphological field tests
In this chapter, XBeach is compared to a variety of measurement results obtained from the
field. This involves major storm events from the past as well as more moderate situations
where morphological behavior is monitored. In general, these tests can be seen as example
applications of the XBeach model.
4.1 1953 storm surge
In order to test the model performance in prototype conditions, this test studies the impact
of the 1953 storm surge on the Dutch coast at Delfland. The initial profile for the simulation
is obtained from test T4 of the M1263-III experiment conducted in the Deltaflume (Vellinga,
1984) and is scaled-up to prototype. The profile is representative for the coast at Delfland.
The applied grid is uniform with dx = 4.56 m and the applied hydrodynamic conditions vary
over the storm (see Figure 4.1). Simulation settings are default except for the maximum
erosion rate dzmax, which is scaled-up to 0.17m3 /ms applying the scale relation for the
erosion volume (Vellinga, 1986).
5
4
3
[m]
2
1
0
0 5 10 15 20 25 30
time [h]
10
Hrms [m]
0
0 5 10 15 20 25 30
time [h]
12
11
Tp [s]
10
8
0 5 10 15 20 25 30
time [h]
Figure 4.1: Surge and wave conditions during the 1953 storm surge
Deltares 61
October 2015 Morphological field tests XBeach skillbed report, revision 4867
4867
The simulated profile evolution is shown in Figure 4.2. Evolution of the simulated dune
erosion volume is shown in Figure 4.3.
profiles 1953_storm_surge default

30
initial
XBeach (current)
25
height [m]
20
15
1100 1150 1200 1250 1300 1350
distance [m]
bed level change

2
0
height [m]
4
computed
6
0 500 1000 1500
distance [m]
erosion volume
150
volume [m3/m]
100
50
computed
0
0 5 10 15 20 25 30 35
time [h]
retreat distance
20
distance [m]
15
10
5
computed
0
0 5 10 15 20 25 30 35
time [h]
Figure 4.3: Morphological response
4.2 1976 storm surge
This test studies the impact of the 1976 storm surge on the Dutch coast at Holland. The
surge during the 1976 storm was the highest since the 1953 storm. In the night between
January 2nd and 3rd wind speeds reached a maximum near the Dutch coast. In the central
and eastern parts of the North Sea the wind speed reached 51 m/s from the Northwest. The
62 Deltares
4867
strongest gale on the Dutch mainland was recorded at Vlissingen: southwestern wind with a
speed of 26 m/s. This caused a significant storm surge with a maximum water level in the
afternoon of January 3rd (Van der Werf et al., 2011).
Between October 1975 (pre-storm) and February (post-storm) 1976 30 cross-shore profile
measurements were performed at eight locations (among others) along the North-Holland
coast. The observed profiles generally extend from the low water line to above the dune
foot (+3 m NAP). We have extended these with the latest JARKUS data (yearly profile
measurement from approximately -8 m NAP to the first dune row), measured in spring 1975.
1D XBeach models were setup for these 30 transects, of which 3 are used in this test.
We base the XBeach model settings on an extensive sensitivity analysis and on recent XBeach
studies (Van der Werf and Van Santen, 2010). Furthermore, we take a morphological accel-
eration factor of 1, set the offshore boundary at a depth of 20 m, and use measured instead
of computed water levels as input.
As an example, Figure 4.4 shows the initial cross-shore profile with the pre-storm measure-
ment, the post-storm measurement and the final profile predicted by XBeach for one of the
30 transects. ?? shows the Brier Skill Score of all computed transects.
profiles 1976_storm_surge raai3400

20
initial
measured
15
10
height [m]
10
400 300 200 100 0 100
distance [m]
Figure 4.4: raai3400
Deltares 63
4867
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
raai3400 raai568 raai6050
4.3 Erosion and overwash of Asseteague Island
This test concerns the morphodynamic response of sandy dunes to extreme storm impacts
at Assateague Island, Maryland, USA, which was analyzed before by Jimenez et al. (2006).
Two consecutive North Easters attacked the barrier island during late January and early
February, 1998. The bathymetry was measured using LIDAR in September 1997 and again
February 9th and 10th, 1998 after the two storms had subsided.
Three types of dunes were identified by Jimenez et al. (2006), shown in Figure 4.6. Profile A
(upper left panel) is initially characterized by a steep faced dune, where the maximum run-up
exceeded the dune crest height and the mildly sloped back of the dune. The morphological
response is characterised by profile lowering, decrease of the beach face slope and landward
barrier displacement, while retaining barrier width.
Profile type B is a double-peaked dune profile and has two different shapes. Profile B1 (upper
right panel) is initially characterized by a primary and secondary dune, both of which are
lower than the maximum run-up height and which are separated by a valley. Profile B2
(bottom left panel) initially has two peaks of which the seaward one is lower. The backside
of the barrier of either type is therefore either characterized by a secondary dune line (profile
B1) or a taller crest of the dune (profile B2) which prevents the eroded sand from being
transported to the backside of the dune. The main morphological response for these profile
types is a decrease of the beach face slope, outer shoreline retreat and narrowing of the
barrier.
The height of the dune crest of profile C (lower right panel) exceeds the maximum run-up
height and so little overwash is observed. The morphological response of this type of profile
is crest lowering due to slumping, decrease of the beach face slope and retreat of the outer
shoreline. The width of the barrier is seen to decrease.
The storm impact of the two North Easters on Assateague Island were modelled with XBeach
for the four profiles described by Jimenez et al. (2006). The profiles were extended with
64 Deltares
4867
a shallow foreshore and a 1:100 slope in seaward direction till a water depth of 9 m below
NAVD88. As XBeach has not been shown to accurately simulate morphological change during
very long storm durations, the simulations were run for a total of 20 hours. The measured
wave and surge conditions were parameterized for each storm by a constant surge level and
a constant wave spectrum (Pierson-Moskowitz) (see Table 4.1). This approach assumes that
two 72 hour storms with varying surge and wave conditions can be approximated by two
10 hour simulations with constant maximum surge and wave conditions following a similar
approach as Vellinga (1986). This approach also facilitates further sensitivity studies into
the effect of varying hydraulic forcing conditions. The calculation grid size varies from 18 m
at the offshore boundary to 2 m on the islands. A morphological acceleration factor of 5 is
applied. The final simulated bed profiles are shown in Figure 4.6.
Table 4.1: Hydrodynamic boundary conditions XBeach simulations
Storm 1 Storm 2
Surge level [m +NAVD] 0.8 1.0
Hs [m] 4.1 3.9
Tp [s] 8.5 8.5
ProfileA ProfileB1
3 3
Elevation (m +NAVD88)
2 2
1 1
0 0
1 1
2 2
3 3
800 1000 1200 1400 800 1000 1200 1400
x (m) x (m)
ProfileB2 ProfileC
3 3
2 2
1 1
0 0
1 1
2 2
3 3
800 1000 1200 1400 800 1000 1200 1400
x (m) x (m)
Figure 4.6: Pre-storm profiles (black dotted line), measured post-storm profiles (black
solid line) and modelled post-storm profiles (red solid line). Upper left panel: profile A.
Upper right panel: profile B1. Lower left panel: profile B2. Lower right panel: profile C.
The seaward side is on the left in all panels. Note that the measured post-storm profiles
contain only the sea surface and emerged topography and no submerged topography.
4.4 Breach growth at Zwin
This test is absed on the Zwin breach growth experiment, as reported by Visser (1998). In
the mouth of the Zwin, a tidal inlet located at the border between the Netherlands and
Belgium, an artificial dam was constructed with a crest height of 3.3 m +N.A.P. (Dutch
datum, approx. MSL), crest width 8 m, inner slope 1:3 outer slope 1:1.6 and length 250 m.
Deltares 65
4867
An initial depression of 0.8 m was made in the middle of the dam having a width of 1 m
and a side slope of 1:1.6 to ensure that the breach initiated at this location. The level of the
surrounding sea bed was about 0.7 m + N.A.P. The mean tidal prism of the Zwin is about
350,000 m3. The polder area Ap as a function of the water level behind the dam is given by:
Ap = 170.000m zs 100.000m2 , 0.6m < zs < 2.3m + N AP
Ap = 2.100.000m zs 4.540.000m2 , zs 2.3m + N AP
At t = 0, about 10 minutes prior to high water, the water level at the seaside was 2.72m+NAP.
At t = 10 minutes a water level of 2.75m+NAP. was reached. For the remainder of the test,
which had a total duration of 1 hour, the water level marginally decreased. After 1 hour the
breach growth became nil, as the water level of the polder area behind the breached equaled
the sea level. The wave height near the dam was negligible during the experiment. The wind
speed was about 2 m/s.
Until t = 6.5 minutes the breach depth grew whereas the breach width remained constant.
At t = 6.5 minutes the original dike structure had nearly completely disappeared over the
initial depression width of 1 m. Near t = 6.5 minutes the onset of lateral breach growth was
observed. The scour hole developed further down to a depth of -1.6m+NAP. (4.9 m below
the original dam crest level). The rate of lateral breach growth was about 2 cm/s. After
approximately 40 minutes the process slowed down considerably and after approximately 1
hour the water levels at both sides were equal.
A schematized representation of the Zwin test was created in XBeach, with at the sea side
a uniform bed level at 0.7m+NAP, and inside the basin a prismatic profile with the deepest
point at 0.7m+NAP and sloping sides, such that the polder area as a function of the water
level was in accordance with the equations above. The grid is non-equidistant with grid sizes
gradually varying from 0.5 m near the breach to approx. 50 m far away from it. The median
grain diameter D50 of the bed material was set to 0.3 mm in accordance with the prototype
test conditions for the artificial dam. The applied critical slopes for avalanching are the same
as in other tests and standard settings were applied for the transport formulations. Waves
were negligible in the test and were set to zero. The model was run with a CFL of 0.5 and
remained smooth and stable despite the steep slopes and supercritical flows.
In Figure 4.7 a sequence of 3D images is shown depicting the various stages in the breaching
process: the initial overflowing, the cutting back of the breach, the deepening and finally the
widening of the breach.
66 Deltares
4867
Figure 4.7: Growth of breach width
In Figure 4.8 a comparison is given between measured and simulated water levels, flow veloc-
ities and development of the breach width in time. Observation point MS2 is 30 m upstream
of the centre point of the breach and MS4 is 30 m downstream of it. In MS4 there was some
ambiguity in the measured initial water level, which explains the initial discrepancy between
measurements and simulations.
3
water level (m)
2
MS2 obs
MS2 comp
1
MS4 obs
MS4 comp
0
0 10 20 30 40 50 60 70
time (min.
6
MS4 obs
velocity (m/s)
MS4 comp
4
0
0 10 20 30 40 50 60 70
time (min.
60
breach width (m)
40
20 B obs
B comp
0
0 10 20 30 40 50 60 70
time (min.
Figure 4.8: Breach width related to water level and flow velocities
4.5 MICORE field experiments
Within the European Communitys Seventh Framework Programme, the MICORE project
facilitated the validation of XBeach for 9 European coastal sites and compared the XBeach
Deltares 67
4867
results to results obtained from models locally used (Deltares, 2011).

The validation of XBeach started off with a measurement campaign at each of the 9 sites.
During the measurements, bathymetric and hydrodynamic data is collected, which is subse-
quently used as input for XBeach and the other models. The following subsections describe
the results for 6 of the 9 field sites.
4.5.1 Lido di Dante, Italy
The study site is the Lido di Dante-Lido di Classe area, an 8 km stretch of sandy beaches along
the Emilia-Romagna coastline in northern Italy on the Adriatic Sea. The site is a mixture of
urbanised (approximately 40% of the total area) and relatively pristine (approximately 60%)
coastal environments. The seaside towns of Lido di Dante and Lido di Classe are located at
the sites northern and southern boundaries respectively. In these regions the beaches are
protected by offshore breakwaters and groins and backed by moderate coastal development in
the form of beach huts, holiday accommodation and paved roads. Between these two towns
is a natural park consisting of natural vegetated dunes and no coastal protection. Three river
mouths are located at the site: one at Lido di Dante (Fiumi Uniti); one at Lido di Classe
(Fiume Savio); and one in the centre of the natural park (Torrente Bevano).
The submerged beach is generally composed of fine sand, while the beachface is made up of
fine to medium sands (D50 = 0.03 mm). The intertidal beach slope varies significantly along
the 8 kms of coastline, from mild (2.5%) to steep (14%). Steep values are representative of
areas adjacent to coastal defence structures (i.e. groins) while the area inside the natural park
is characterised by lower gradients. The mean submerged beach slope is 3%. According to
the morphodynamic classification of Wright and Short (1984), the beaches are considered as
having intermediate beach states. Low tide terraces are often observed both in the protected
and natural areas. Submerged longshore bars meanwhile are only present in the areas outside
of the offshore structures.
The wave climate of this region is generally small, with 91% of significant wave heights below
1.25 m. The prevalent wave direction is from the east, while the most intense storms are from
the ENE (known as the Bora wind). The Bora wind is a strong, cold, gusty wind that blows
intermittently but mainly during the winter months. It not only has a strong influence on
the wave climate of this region, but of the general circulation patterns of the entire Adriatic
Sea. South-easterly waves meanwhile are much less significant, since SE winds are sheltered
to some degree by the Conero Headland approximately 120 km south of the site.
In regards to water level variations, the area is microtidal with a mean neap tidal range
of 30-40 cm and a mean spring tidal range of 80-90 cm. The tidal signal has both diurnal
and semidiurnal components. Tidal anomalies of up to double the maximum tidal elevation
can occur as a result of surge. This is particularly the case during SE wind conditions,
where, considering the SE-NE orientation of the Adriatic Sea, there is the greatest fetch for
wind-driven surge.
Within the 2008-2010 MICORE monitoring period, three storms were selected for calibration
of the off-the-shelf and XBeach models. Each storm has distinct properties that encompass
the range of storm conditions typical for the Emilia-Romagna region. Here the modelling is
restricted to profile MN15 for the 1-3 December 2008 storm.
68 Deltares
4867
profiles
4
initial
measured
XBeach BSS=0.16
3
1
height [m]
4
4300 4350 4400 4450 4500 4550 4600 4650
distance [m]
Figure 4.9
4.5.2 Praia de Faro, Portugal
The study area is Ancao Peninsula constituting the westernmost barrier of the Ria Formosa
barrier island system. It is a NW-SE oriented sandy barrier that is attached to mainland by
its western terminus.
The area is mesotidal, with an average tidal range of about 2 m that can reach up to 3.5 m
during spring tides. Analysis of two years of records from a tidal gauge on the Algarve Coast
showed a maximum observed storm surge level of +0.75 m (Gama et al., 1994). The return
period of a sea level 2.23 m above Mean Sea Level (MSL) is 10 years (Gama et al., 1994).
The offshore wave climate is dominated by west-southwest waves (71% of occurrences). SE
waves that consist of short period waves generated by regional winds (locally called Levante)
are also frequent (about 23%). Wave energy is moderate with an average annual significant
offshore wave height of 1.0 m and average peak period of 8.2 s (Costa et al., 2001).
Storm events in the region are considered when the significant offshore wave height exceeds 3
m (Pessanha and Pires, 1981) and typically correspond to less than 1% of the offshore wave
climate (Costa et al., 2001). A 5.0 m significant wave height for a SE storm has a return
period of 50 years, whilst a 5.7 m SW storm is expected every 5 years (Pires, 1998). Due to
its northwest-southeast orientation it is directly exposed to west-southwesterly waves, and
is relatively protected from SE waves.
Several storm events have been recorded during the MICORE campaign among which the
two most important are discussed in the MICORE report. Here a group of several individual
WSW storms is simulated that took place at Faro beach from 18/12/2009 until 5/1/2010.
The significant wave height reached 4 m and the peak period up to 20 sec, while given the
long duration intense wave conditions coincided with both spring and neap tides. The event
had a significant impact on the coast, as overtopping and dune erosion occurred at several
sections.
Deltares 69
4867
profiles
10
initial
measured
other BSS=0.30
8 XBeach BSS=0.44
4
height [m]
4
1200 1250 1300 1350 1400 1450 1500 1550 1600
distance [m]
Figure 4.10
4.5.3 Cadiz Urban Beach, Spain
The field site is located around Cadiz town, in south-western Spain, facing the Atlantic Ocean.
It is constituted by two different beaches extending along 10 km, providing the opportunity
for studying the effects of storms on different types of coastal environments.
The study area is a mesotidal coast with a mean tidal range of 3.2 m and 1.1 m during
springs and neaps tides, respectively. Dominant winds blow from ESE (19.6% of annual
occurrence) and WNW (12.8%), which together with coastline orientation makes sea and swell
waves approach generally from the third and fourth quadrants. According to this, prevailing
longshore drift is directed south-eastwards. Significant wave height is usually lower than 1
m, with waves over 4 m high being uncommon and occurring only during the most important
storms, which usually take place between November and March and approach from the third
quadrant. In fact, waves greater than 1.5 m are considered storm waves, so the area can be
classified as a low-energy one.
The storm event that was selected is a moderate storm event with a return period of about
1 year. The maximum significant wave (Hs) height during the peak of the event was 3.7m
with a spectral period (Tp) of 8.7sec. The total duration of the storm was 46 hours (light
grey shaded area). The tidal conditions over that period were from springs to neaps with an
average tidal range of 2.27m.
70 Deltares
4867
Figure 4.11
4.5.4 Dziwnow Spit, Poland
The Polish study site is the 14 km long Dziwnow Spit of barrier type built of Holocene deposits
(mainly sands) with dunes 3.5 to 10 meters high. Behind the spit there are relatively wide
lowlands of glacial or glaciofluvial origin, in most cases filled with peat. Their surface is 1 to
3 m above sea level. In the lowlands alos the Kaminski Lagoon is found with a rather small
depth (maximum 2-3 m). At the middle of the spit there is a connection between the lagoon
and sea (Dziwna). The mean beach width calculated for the pilot area is 33 m.
The average tide range in the Baltic is very small and is less then 10 cm. This is due to the
small area of the Baltic, its geographical situation and the presence of the Danish Straits,
which prevent propagation of North Sea tides into the Baltic. Thus, surface waves (wind
waves and swell) are the most important factor of the Baltic coastal zone hydrodynamics.
The wave climate in Poland is highly diversified because of the wealth of fetches and wind
speeds occurring throughout the year.
Since 1 June 2008, (from the beginning of observations taken within MICORE project) 1
extreme storm (12.10.2009) was noticed which caused significant morphological changes at
the shore. The storm return period was about 4 years and was simulated with the XBeach
model.
The storm occurred on 12.10.2009 and lasted for almost 4 days (93hours). The highest sea
level observed on tide gauge located in the Dziwna (Dziwnow Port Authority area) was 0,76
m. above mean sea level. The maximum significant wave height (Hs) reached 3.2 m and teh
maximum peak period (Tp) was 11.17 sec.
The XBeach model was run for profile 386 in 1D mode. The profile was interpolated to a
cross-shore varying grid with a minimum cell size of 3 m. Offshore wave data timeseries from
the WAM model were inputted to XBeach using a Jonswap spectrum and setting instat=41.
Wave direction values were changed to 270 degrees, which means that incoming waves are
shore normal. Surge input data were taken as the hourly mean sea level.
Deltares 71
4867
profiles
8
initial
measured
6 XBeach BSS=0.22
2
height [m]
8
0 100 200 300 400 500
distance [m]
Figure 4.12
4.5.5 Kamchia Shkorpilovtsi Beach, Bulgaria
The study area, called Kamchia-Shkorpilovtsi, is situated in the western Black Sea, and
spreads from cape Paletsa to cape Cherni Nos (Figure 9.1), located 25 and 40 km to the
south of Varna city, respectively. It comprises the longest and the largest sandy beach along
the Bulgarian Black Sea coast, with well-developed dunes and the two rivers mouths, these
of the Kamchia River and the Fundakliiska River. In the middle of the site, near the mouth
of the Fundakliiska River a scientific pier is built perpendicularly to the shoreline, reaching
4.5 m water depth. The beach is formed as a result of accumulation of erosive and fluvial
sediments. The main morphological feature of the study area is its rectilinear shoreline with
almost parallel isobaths. The bottom slope is covered with sands of different size. In its upper
part down to 2.5 m depth, over 95% of bottom sediments consist of coarse and medium sand
fractions. As the depth grows, the content of these fractions decreases and at 8-10 m over
90% of the sediment grain size is less than 0.25 mm.
The beach is open to waves of the eastern half. In the case of severe storms the wind speed
magnitude can reach 35-40 m/s and 9 m height of maximum significance wave at depths of
about 1000 m. The large seasonal variability is one of the most marked features of the wave
climate. The winter storms are much more frequent than the summer ones. In the western
Black Sea the most frequent are the winds from northeast and east, which trigger the most
severe storms.
In the beginning of March a short but very intense storm occurred in the western Black sea.
This event was distinguished with all features of the severe storms known from the historical
overview well defined phases growth took place on 08.03, peak 09.03 and decay 09.03 - 10.03.
Wind and wave direction were quite stable turning from ESE to ENE. Maximum SWH
reached almost 4.20 m.
Model results for this storm are shown in Figure 4.13.
72 Deltares
4867
Figure 4.13
Deltares 73
4867
74 Deltares
XBeach skillbed report, revision 4867 Comparisons with other models October 2015
4867
Chapter 5
Comparisons with other models
In this chapter, XBeach is compared to results obtained from other models. The comparisons
are currently focussed on the DUROS+ and D++ (Vellinga, 1986; Delft Hydraulics, 2006;
Deltares, 2010) models that are used for the assessment of dunes along the Dutch coast.
Comparisons with other models like DurosTA (Steetzel, 1993) are made throughout the report
and are not discussed specifically in this chapter at the moment.
5.1 Field applications
In this section, DUROS+ (Vellinga, 1986; Delft Hydraulics, 2006) and D++ are (Deltares,
2010) compared with XBeach based on field applications. Comparisons are made based on
erosion volumes and retreat distances, since these are the main parameters of interest in dune
safety assessment.
5.1.1 Retreat distances JARKUS
In this test, retreat distances obtained from DUROS+ and XBeach using a selection of
JARKUS profiles characteristic for the Dutch coast are compared (Den Heijer et al., 2011).
The comparison is presented in Figure 5.1. The retreat distance is defined as the horizontal
distance between the NAP+5m contour and the erosion point. The erosion point is defined
as the first diversion point between the pre-storm and post-storm profile, when going from
the land side in seaward direction. Diversion is considered as a vertical difference of more
than 5cm.
Deltares 75
October 2015 Comparisons with other models XBeach skillbed report, revision 4867
4867
XBeach vs. DUROS+

140 XBeach vs. D++
1:1
DUROS+ resp. D++ retreat distance [m] 120
100
80
60
40
20
20 40 60 80 100 120 140

XBeach retreat distance [m]
Figure 5.1: Scatter plot of retreat distances obtained from XBeach and DUROS+ (blue
circle) and XBeach and D++ (green cross).
76 Deltares
XBeach skillbed report, revision 4867 Specific functionalities October 2015
4867
Chapter 6
Specific functionalities
In the previous chapters, the core functionalities are discussed: hydro- and morphodynamics
on different scales. However, XBeach is equipped with a variety of additional functionalities
that facilitate more complex computations. One example already discussed is hard elements
in the profiles. This chapter discusses a variety of other functionalities, like discharges, drifters
and multiple sediment fractions.
6.1 River outflow
The river outflow case is meant to test XBeach for the combined effects of a river outflow
and a steady wave-driven longshore current on the sediment transport and the morphological
evolution. Though purely hypothetical, this case contains many salient features of real-life
applications, such as longshore currents through open side-boundaries and exchange of water
and sand through a gap in a closed boundary. Thus, the formulation of open boundary
conditions is also tested here.
The initial topography consists of a plane beach (slope 1:50), which is interrupted by a 75 m
wide river mouth with a water outflow of 150 m3 /s. The bottom contours are straight and
parallel to the shoreline, except for a shallow submerged channel in line with the river.
The computational grid is rectangular, with 56 nodes in the x-direction (cross-shore) and
111 nodes in the y-direction (longshore), with a uniform grid spacing of 15 m. The waves
are irregular and long-crested, with a root-mean-square height of 2m at a water depth of
13.5m. The direction of wave incidence is 30 degrees with respect to the shore-normal. The
peak wave period is 8s. The bed material is uniform sand of 250, with a settling velocity of
0.031m/s.
In this figure the bathymetry is shown after approxiately 4 days; arrows indicate the sediment
transport vectors. plotted for every cross-shore cell and every third longshore cell. When
functioning correctly, we see a channel that has turned towards the north and striaght contour
lines downstream of the channel.
Deltares 77
October 2015 Specific functionalities XBeach skillbed report, revision 4867
4867
0 2
5
10
1600
1400 4
1200
1000
800 6
600 800
400 600
400
200 200
0 8
0
10
Figure 6.1: Final bathymetry
6.2 Drifters
This purely hypothetical test combines three specific functionalities of XBeach: curvilinear
grids, discharges and drifters. The model is basicly an odd-shaped bathtub with a single
discharge location at the left boundary. The bathtub is filled by the discharged water, in-
dicated by the increasing water levels (colored background). The bathtub is approximately
70m x 70m. Considering that, the discharge of 50m3 /s is rather large. Every 10 seconds,
two drifters are released just before the discharge opening. One of the two drifters is released
at the upper boundary of the opening, the other at the lower boundary. The entire path a
single drifter has followed at a certain moment in time is plotted in Figure 6.2. Two large
eddies driven by the large discharge are revealed by these paths.
78 Deltares
4867
Figure 6.2: Drifter paths after 50, 100, 150 and 450 seconds
6.3 Multiple sediment fractions
The purpose of this simulation is to ensure the multiple sediment fractions model in XBeach
performs as expected. In this test, the Deltaflume 2006 T01 test is recreated with two types
of sand with different colours, red and blue. The sand is initially placed in a zebra-stripe
pattern in the profile. The properties of both types of sand such as the grain size and mobility
are the same as the sand used in the Deltaflume experiment. For the test to be successful,
the following conditions should be met:
1. The simulated final profile should be the same as the final profile in the original
Deltaflume 2006 T01 test.
2. The two sediment types should mix and form layers over each other (yellow).
The figures Figure 6.3 and Figure 6.4 show the initial and final distribution of red and blue
sediment in the profile near the dune face. The red and blue lines in the same figure show
the sediment concentration of each sediment type in the water column. If the simulation is
successful, the red and blue sediment will be well mixed on the foreshore and fresh blue sedi-
ment will be deposited over the red sediment at the dune foot as the dune face retreats. The
concentration of blue sediment in the water column should be higher than the concentration
of red sediment in the water in areas where only blue sediment is available in the top layer
of the bed.
The red line in Figure 6.5 shows the predicted dune face retreat and bed level change in
the XBeach multiple-sediment model. The black line in the same figure represents the cor-
responding measured final profile. If the simulation has been successful, the red and black
line will align reasonably well. The results of this simulation should be compared to the
Deltaflume 2006 T01 test described earlier in this report.
Deltares 79
4867
Grain distribution and sediment concentrations

2 1
0.9
1
height [m] and concentration (g/l) 0.8
0 0.7
0.6
1
0.5
2 0.4
0.3
3
0.2
Initial bed level

Concentration red 0.1
4
Concentration blue
0
60 80 100 120 140 160 180 200 220
distance [m]
Figure 6.3: Initial bathymetry and sediment distribution
Grain distribution and sediment concentrations

2 1
0.9
1
0.8
height [m] and concentration (g/l)
0 0.7
0.6
1
0.5
2 0.4
0.3
3
0.2
Initial bed level

Concentration red 0.1
4
Concentration blue
0
60 80 100 120 140 160 180 200 220
distance [m]
Figure 6.4: Final bathymetry and sediment distribution
80 Deltares
4867
profiles Deltaflume2006 T01_zebra

2
initial
measured
1.5
height [m] 0.5
0.5
1.5
185 190 195 200 205 210 215
distance [m]
6.4 Curvilinear
In this test a curvi-linear grid is applied to simulate waves and hyrdodynamics around a
virtual island. The island is circular with a linear sloping profile. The lower half of the
offshore boundary generates bichromatic waves that partly refract towards the island, but
also passes the island. The waves passing the island leave the model through the upper half
of the offshore boundary. Reflections at the offshore boundary are minimal due to the 2D
weakly-reflective boundary condition used (Van Dongeren and Svendsen, 1997).
Figure 6.7 shows the RMS wave height (color) and the flow velocity field (arrows) around
the island in the last time step. Figure 6.7 shows the same parameters, but in this case the
overall mean values are presented.
Deltares 81
4867
Figure 6.6: Snapshot of the RMS wave height and the flow velocity field in the last time
step
Figure 6.7: Mean values of the RMS wave height and the flow velocity field
82 Deltares
XBeach skillbed report, revision 4867 References October 2015
4867
Chapter 7
References
Abdelrahman, S. M. and Thornton, E. B. (1987). Changes in the short wave amplitude and
wavenumber due to presence of infragravity waves. In Proceedings Speciality Conference
on Coastal Hydrodynamics, pages 458478. 43
Arcilla, A. S., Roelvink, J. A., OConnor, B. A., Reniers, A., and Jimenez, J. A. (1994). The
delta flume 93 experiment. In Coastal Dynamics, pages 488502. 37, 40
Bakkenes, H. J. (2002). Observation and seperation of bound and free low-frequency waves
in the nearshore zone. Masters thesis, Delft University of Technology. 10
Birkemeier, W. A., Donoghue, C., Long, C. E., Hathaway, K. K., and Baron, C. F. (1997).
1990 delilah nearshore experiment: Summary report. Technical Report CHL-97-4-24, U.S.
Army Corps of Engineers. Field Research Facility. 11
Boers, M. (1996). Simulation of a surf zone with barred beach, part 1: Wave heights and
wave breaking. Communications on Hydraulic and Geotechnical Engineering 69-5, Delft
University of Technology. 116 p. 10
Carrier, G. F. and Greenspan, H. P. (1958). Water waves of finite amplitude on a sloping

beach. Journal of Fluid Mechanics, 4:97109. 6, 7
Costa, M., Silva, R., and Vitorino, J. (2001). Contribuio para o estudo do clima de agitao
martima na costa portuguesa. In 2as Jornadas Portuguesas de Engenharia Costeira e
Porturia, page 20p, Sines. in Portuguese. 69
Damgaard, J., Dodd, N., Hall, L., and Chesher, R. (2002). Morphodynamic modelling of rip
channel growth. Coastal Engineering, 45:199221. 2
Delft Hydraulics (2006). Dune erosion product 2: Large-scale model tests and dune erosion
prediction method. Report H4357, Delft Hydraulics. 75
Deltares (2010). Ontwikkeling detailtoets duinen 2011. Interim report 1202124-003, Deltares.
in Dutch. 75
Deltares (2011). Validation of dune impact models using european field data. Report 1002266,
Deltares. 68
Den Heijer, C., Walstra, D. J. R., Van Thiel de Vries, J. S. M., Huisman, B. J. A., Hoonhout,
B. M., Diermanse, F. L. M., and Van Gelder, P. H. A. J. M. (2011). Importance of
dune erosion influencing processes. Journal of Coastal Research, SI 64(1):283287. ISSN
0749-0208. 75
Deltares 83
October 2015 References XBeach skillbed report, revision 4867
4867
Dongeren, A. V., Lowe, R., Pomeroy, A., Trang, D. M., Roelvink, D., Symonds, G., and
Ranasinghe, R. (2013). Numerical modeling of low-frequency wave dynamics over a fringing
coral reef. Coastal Engineering, 73(0):178 190. 14, 15
Erikson, L., Larson, M., and Hanson, H. (2005). Prediction of swash motion and run-up
including the effects of swash interaction. Coastal Engineering, 52:285302. 2
Feddersen, F., Guza, R. T., Elgar, S., and Herbers, T. C. (2000). Velocity moments in along-
shore bottom shear stress parameterizations. Journal of Geophysical Research, 105:8673
8688. 2
Gama, C., Dias, J. A., Ferreira, O., and Taborda, R. (1994). Analysis of storm surge in
portugal, between june 1986 and may 1988. In Proceedings Second International Sym-
posium on Coastal Zone Research - Management and Planning, pages 381387, Lisbon.
EUROCOAST. 69
Huntley, D. A., Guza, R. T., and Thornton, E. B. (1981). Field observations of surf beats,
1, progressive edge waves. Journal of Geophysical Research, 86:64516466. 2
Jimenez, J. A., Sallenger, A. H., and Fauver, L. (2006). Sediment transport and barrier island
changes during massive overwash events. In 30th International Conference on Coastal
Engineering, San Diego, California, USA. ASCE. 64
Larson, M., Erikson, L., and Hanson, H. (2004). An analytical model to predict dune erosion
due to wave impact. Coastal Engineering, 51(89):675696. 1, 3
Leatherman, S. P., Williams, A. T., and Fisher, J. S. (1977). Overwash sedimentation asso-
ciated with a large-scale northeaster. Marine Geology, 24:109121. 3
List, J. H. (1992). A model for two-dimensional surfbeat. Journal of Geophysical Research,

97:56235635. 2
Murphy, A. H. and Epstein, E. S. (1989). Skill scores and correlation coefficients in model
verification. Monthly Whether Review, 117:572581. iv, 89, 91, 92
Nairn, R. B., Roelvink, J. A., and Southgate, H. N. (1990). Transition zone width and impli-
cations for modelling surfzone hydrodynamics. In Edge, B. L., editor, 22th International
Conference on Coastal Engineering, pages 6881, Reston, Virginia, USA. ASCE. 2
Nishi, R. and Kraus, N. C. (1996). Mechanism and calculation of sand dune erosion of
storms. In 25th International Conference on Coastal Engineering, pages 30343047, Or-
lando, Florida, USA. ASCE. 1
Overton, M. F. and Fisher, J. S. (1988). Laboratory investigation of dune erosion. Journal

of Waterway, Port, Coastal and Ocean Engineering, 114(3):367373. 1
Ozkan-Haller, H. T. and Kirby, J. T. (1997). A fourier-chebyshev collocation method for the

shallow water equations including shoreline runup. Applied Ocean Research, 19:2134. 7, 9
Phillips, O. M. (1977). The Dynamics of the Upper Ocean. Cambridge University Press, New
York, 2nd edition. 2
Pires, H. O. (1998). Project india, preliminary report on wave climate at faro. Technical
report, Instituto de Meteorologia, IST, Lisbon. 69
84 Deltares
4867
Pomeroy, A., Lowe, R., Symonds, G., Dongeren, A. V., and Moore, C. (2012). The dynamics
of infragravity wave transformation over a fringing reef. Journal of Geophysical Research,
117(C11022):1 17. 14
Raubenheimer, B. and Guza, R. T. (1996). Observations and predictions of run-up. Journal

of Geophysical Research, 101(C10):25,57525,587. 2
Reniers, A. J. H. M., MacMahan, J., Thornton, E. B., and Stanton, T. P. (2006). Modelling
infragravity motions on a rip-channel beach. Coastal Engineering, 53:209222. 2
Reniers, A. J. H. M., Roelvink, J. A., and Thornton, E. B. (2004a). Morphodynamic mod-

elling of an embayed beach under wave group forcing. Journal of Geophysical Research,
109(C01030). 2
Reniers, A. J. H. M., Thomas, E. B., Stanton, T. P., and Roelvink, J. A. (2004b). Vertical flow
structure during sandy duck: observations and modeling. Coastal Engineering, 51(3):237
260. 2
Roelvink, J. A. (1993a). Dissipation in random wave groups incident on a beach. Coastal

Engineering, 19:127150. 2
Roelvink, J. A. (1993b). Surf beat and its effect on cross-shore profiles. PhD thesis, Delft
University of Technology. 2
Roelvink, J. A., Reniers, A. J. H. M., Van Dongeren, A. R., Van Thiel de Vries, J. S. M.,
McCall, R. T., and Lescinski, J. M. (2009). Modelling storm impacts on beaches, dunes
and barrier islands. Coastal Engineering, 56(11-12):11331152. 1
Roelvink, J. A. and Stive, M. J. F. (1989). Bar-generating cross-shore flow mechanisms on a

beach. Journal of Geophysical Research, 94(C4):47854800. 43
Roelvink, J. A., Van Kessel, T., Alfageme, S., and Canizares, R. (2003). Modelling of barrier
island response to storms. In Coastal Sediments 03, Clearwater, Florida, USA. 3
Ruessink, B. G., Miles, J. R., Feddersen, F., Guza, R. T., and Elgar, S. (2001). Modeling the
alongshore current on barred beaches. Journal of Geophysical Research, 106:22,45122,462.
2
Sallenger, A. H. (2000). Storm impact scale for barrier islands. Journal of Coastal Research,
16(3):860895. 2
Schaeffer, H. A. (1994). Edge waves forced by short-wave groups. Journal of Fluid Mechanics,
259:125148. 2
Soulsby, R. L. (1997). Dynamics of Marine Sands. Thomas Telford, London. 2
Soulsby, R. L., Hamm, L., Klopman, G., Myrhaug, D., Simons, R. R., and Thomas, G. P.
(1993). Wavecurrent interaction within and outside the bottom boundary layer. Coastal
Engineering, 21:4169. 2
Steetzel, H. J. (1987). Systematic research on the effectiveness of dune toe revetments - large
scale model investigation. Technical Report H298-I, Delft Hydraulics. 35
Steetzel, H. J. (1993). Cross-shore transport during storm surges. PhD thesis, Delft University
of Technology. 1, 75
Deltares 85
4867
Stelling, G. S. and Duinmeijer, S. P. A. (2003). A staggered conservative scheme for every

froude number in rapidly varied shallow water flows. International Journal for Numerical
Methods in Fluids, 43:13291354. 3
Stive, M. J. F. and De Vriend, H. J. (1994). Shear stresses and mean flow in shoaling
and breaking waves. In Edge, B. L., editor, 24th International Conference on Coastal
Engineering, pages 594608, Reston, Virginia, USA. ASCE. 2
Stockdon, H. F., Holman, R. A., Howd, P. A., and Sallenger, A. H. (2006). Empirical
parameterization of setup, swash, and runup. Coastal Engineering, 53:573588. 2
Sutherland, J., Peet, A. H., and Soulsby, R. L. (2004). Evaluating the performance of mor-
phological models. Coastal Engineering, 51(89):917939. 89
Svendsen, I. A. (1984). Wave heights and set-up in a surf-zone. Coastal Engineering, 8:303
329. 2
Thornton, E. B., MacMahan, J., and Sallenger, A. H. (2007). Rip currents, mega-cusps, and
eroding dunes. Marine Geology, 240(14):151167. 1
Tucker, M. J. (1954). Surfbeats: sea waves of 1 to 5minutes period. In Proceedings of Royal
Society, number A in 202, pages 565576, London. 2
Van de Graaff, J. (1976). Scale series dune erosion. Technical Report M1263 part I, Delft
Hydraulics. in Dutch. 24, 26, 28
Van der Werf, J., Van Santen, R., Van Ormondt, M., Briere, C., and Van Dongeren, A. (2011).
Operational model to simulate storm impact along the holland coast. In Proceedings Coastal
Sediments 2011, Miami, Florida, USA. 63
Van der Werf, J. J. and Van Santen, R. (2010). Operational model to simulate storm im-
pact along the holland coast: system development and test application. Technical report,
Deltares, Alkyon and Arcadis. 63
Van Dongeren, A. R., Reniers, A. J. H. M., Battjes, J. A., and Svendsen, I. A. (2003).
Numerical modeling of infragravity wave response during delilah. Journal of Geophysical
Research, 108(C9):3288. 2, 12
Van Dongeren, A. R. and Svendsen, I. A. (1997). Absorbing-generating boundary condition
for shallow water models. Journal of Waterway, Port, Coastal and Ocean Engineering,
pages 303313. 81
Van Gent, M. R. A., Van Thiel de Vries, J. S. M., Coeveld, E. M., De Vroeg, J. H., and
Van de Graaff, J. (2008). Large-scale dune erosion tests to study the influence of wave
periods. Coastal Engineering, 55(12). 1, 42, 43, 49, 55, 58
Van Rijn, L. C., Walstra, D. J. R., Grasmeijer, B., Sutherland, J., Pan, S., and Sierra, J. P.
(2003). The predictability of cross-shore bed evolution of sandy beaches at the time scale of
storms and seasons using process-based profile models. Coastal Engineering, 47(3):295327.
ISSN 0378-3839. 91, 92
Van Thiel de Vries, J. S. M. (2009). Dune erosion during storm surges. PhD thesis, Delft
University of Technology. 43, 45, 50
Van Thiel de Vries, J. S. M., Van Gent, M. R. A., Walstra, D. J. R., and Reniers, A. J.
H. M. (2008). Analysis of dune erosion processes in large-scale flume experiments. Coastal
Engineering, 55(12). 2, 42
86 Deltares
4867
Vellinga, P. (1981). Scale series dune erosion. Technical Report M1263 part II, Delft Hy-
draulics. in Dutch. 26, 28
Vellinga, P. (1984). Scale series dune erosion: Large scale tests in deltaflume. Technical
Report M1263 part III, Delft Hydraulics. in Dutch. 28, 61
Vellinga, P. (1986). Beach and Dune Erosion during Storm Surges. PhD thesis, Delft Uni-
versity of Technology. 1, 28, 35, 61, 65, 75
Visser, P. J. (1998). Breach growth in sand dikes. PhD thesis, Delft University of Technology.
3, 65
Wang, P. and Horwitz, M. H. (2007). Erosional and depositional characteristics of regional

overwash deposits caused by multiple hurricanes. Sedimentology, 54:545564. 3
Zelt, J. A. (1986). Tsunamis: the response of harbours with sloping boundaries to long
wave excitation. PhD thesis, W.M. Keck Laboratory of Hydraulics and Water Resources,
Division of Engineering and Applied Science, California Institute of Technology. 318 p. 7,
8, 9
Deltares 87
4867
88 Deltares
XBeach skillbed report, revision 4867 Model Performance Statistics October 2015
4867
Appendix A
Model Performance Statistics
A.1 Introduction
In this Appendix the theory behind the Model Performance Statistics (MPS) used in the
XBeach skillbed is explained. The MPS are used to quantify the performance of model
results based on a comparison with measurement data. Different MPS parameters are used
as each parameter has its own characteristics.
First an overview is given of the MPS parameters used in the XBeach skillbed, summarized
in table form including some basic characteristics. Consequently, each MPS parameters listed
in the overview table is further explained in separate sections.
A.2 MPS parameters
An overview of the MPS parameters used in the XBeach skillbed is given in Table A.1.
Table A.1: MPS parameters
Parameter Description Ranges

ME & STD Mean Error & Standard Deviation 0: perfect prediction
R Correlation coefficient (range: [0 1]) 1: perfect correlation
Rel. bias Systematic error relative to the mean low value: good performance
Sci Scatter Index low values: performance
BSS Brier Skill Score (Sutherland et al., 2004) see below
BSS Brier Skill Score (Murphy and Epstein, 1989) see below
Each parameter listed in the table is further explained in the following paragraphs.
Deltares 89
October 2015 Model Performance Statistics XBeach skillbed report, revision 4867
4867
A.3 Mean Error & Standard Deviation
The Mean Error (ME) and the Standard Deviation (STD) of the error of a time series are a
useful measure to quantify model performance for parameters such as wave heights or water
levels. The SD is in general not so useful when applied to morphological parameters such as
the bed level evolution.
N
1 X
ME = (fcomp.,i fmeas.,i ) (A.1)
N
i=1
v
u N
u 1 X
ST D = t (fcomp.,i fmeas.,i M E)2 (A.2)
N 1
i=2
A.4 Correlation coefficient
The Correlation Coefficient R is a measure quantifying the correlation of the measurements

and simulation results, but does not indicate significance because the distributions of the
series are not taken into account.
A.5 Relative Bias
The Relative Bias (Rel. Bias) is the systematic error relative to the mean. Relative low
values of the mean can cause high vales of the Rel. Bias.
PN
i=1 (fcomp.,i fmeas.,i )
Rel.Bias = PN (A.3)
i=1 fmeas.
A.6 Scatter Index
The Scatter index (Sci) is the standard deviation relative to the mean value of the measured
signal. Relative low values of the mean can cause high vales of the Sci.
q PN
1
N 1 i=2 (fcomp.,i fmeas.,i M E)2
Sci = (A.4)
fmeas.
A.7 Brier Skill Score
The Brier Skill Score (BSS) calculates the performance of the performance relative to a
baseline prediction. The BSS calculates the mean square difference between the prediction
90 Deltares
XBeach skillbed report, revision 4867 Model Performance Statistics October 2015
4867
and observation with the mean square difference between baseline prediction and observation.
1 PN
N i=1 (zb,c zb,m )2
BSS = 1 1 PN (A.5)
N i=1 (zb,0 zb,m )2
where zb,c is the computed bottom, zb,m is the measured bottom and zb,0 is the initial bottom
(variables taken at each cross-shore coordinate i).
Perfect agreement gives a Brier score of 1, whereas modelling the baseline condition gives a
score of 0. If the model prediction is further away from the final measured condition than the
baseline prediction, the skill score is negative. Van Rijn et al. (2003) proposed a classification
for the Brier Skill Score as shown in Table A.2.
The BSS is very suitable for the prediction of bed evolution. The baseline prediction for
morphodynamic modelling will usually be that the initial bed remains unaltered. In other
words, the initial bathymetry is used as the baseline prediction for the final bathymetry. A
limitation of the BSS is that it cannot account for the migration direction of a bar; it just
evaluates whether the computed bed level (at time t) is closer to the measured bed level (at
time t) than the initial bed level. If the computed bar migration is in the wrong direction,
but relatively small; this may result in a higher BSS compared to the situation with bar
migration in the right direction, but much too large. The BSS will even be negative, if the
bed profile in the latter situation is further away from the measured profile than the initial
profile. The limitation shown here is that position and amplitude errors are included in the
BSS. Distinguishing position errors from amplitude errors, requires a visual inspection of
measured and modelled profiles or the calculation of further statistics (Murphy and Epstein,
1989). The BSS can be extremely sensitive to small changes when the denominator is low,
in common with other non-dimensional skill scores derived from the ratio of two numbers.
Table A.2: Brier Skill Score quantification (Van Rijn et al., 2003)
Qualification Brier Skill Score

Excellent 1.0 - 0.8
Good 0.8 - 0.6
Reasonable fair 0.6 - 0.3
Poor 0.3 - 0.0
Bad <0.0
A.8 Brier Skill Score (Murphy and Epstein, 1989)
Murphy and Epstein (1989) decomposed the BSS, leading to contributions due to errors in
predicting the amplitude (), the phase () and the mean () as presented in Table A.3.
The decomposition facilitates linking performance quantifications to model processes and
accordingly bringing the model performance to a higher level.
+
BSS = (A.6)
1+
Deltares 91
October 2015 Model Performance Statistics XBeach skillbed report, revision 4867
4867
Y 0 2 <Y 0 > <X 0 > 2 <X 0 > 2

= rY2 0 X 0 ; = (rY 0 X 0 ) ; = ( ) ; = (A.7)
X 0 X 0 X 0
Table A.3: Brier Skill Score decomposition factors (Murphy and Epstein, 1989)
Factor Indication Perfect modelling

phase error () transport locations =1
amplitude error () transport volumes =0
map mean error () - =0
normalization term () - -
Van Rijn et al. (2003) also proposed a classification for the decomposed Brier Skill Score as
shown in Table A.4.
Table A.4: Brier Skill Score (Murphy and Epstein, 1989) quantification (Van Rijn et al.,
2003)
Qualification Brier Skill Score

Excellent 1.0 - 0.5
Good 0.5 - 0.2
Reasonable fair 0.2 - 0.1
Poor 0.1 - 0.0
Bad <0.0
92 Deltares
XBeach skillbed report, revision 4867 Overview October 2015
4867
Appendix B
Overview
In the table below, the statuses of all tests found in the skillbed are summarized. The
first three colomns identify the test with a combination of a binary, test and run name.
The following three columns provide information on the status of that specific run. These
columns indicate whether the model ran, the analysis ended succesfully and whether the run
used default or custom settings. An empty status means the run or analysis is ignored, a
cross indicates failure and a checkmark indicates success. The other columns in the table
provide information on the model configuration.
Tests can be run multiple times using different settings. Different runs are identified by a
run name, which follows after the test name. If a test runs once only, it is common use to
name the run default. This is not the same as running a test with default settings, which is
indicated in the Default settings column.
The table provides all tests ran in a single skillbed run. This report not necessarily uses all
available tests. Therefore, not all tests in the table are necessarily found in the continuation
of this report.
Table B.1: Status overview skillbed tests Groundwater flow

Default settings
Water levels**
Configuration
Morphology
Hard layers
Fractions
Waves*
Matlab
Status
Test Run
1953 storm surge default 3 3 1D WG V 1 3
1976 storm surge raai3400 3 3 1D WG V 1 3
Assateague Island profA 3 3 1D ST C 1 3
Assateague Island profB1 3 3 1D ST C 1 3
Assateague Island profB2 3 3 1D ST C 1 3
Assateague Island profC 3 3 1D ST C 1 3
Boers 1C custom 3 3 1D WG C 1
Boers 1C default 3 3 1D WG C 1
CarrierGreenspan default 3 3 1D ST C 1
Curvi Island default 3 3 2D WG C 1
Deltares 93
October 2015 Overview XBeach skillbed report, revision 4867
4867
Table B.1: Status overview skillbed tests
Groundwater flow
Default settings
Water levels**
Configuration
Morphology
Hard layers
Fractions
Waves*
Matlab
Status
Test Run
DUROS refere... 3 3 1D WG C 1 3
DUROS transe... 3 3 1D WG C 1 3
Delilah 199010131000 default 3 3 2D WG C 1
Deltaflume2006 DP01 3 3 1D WG C 2 3
Deltaflume2006 DP02 3 3 1D WG C 2 3
Deltaflume2006 T01 3 3 1D WG C 2 3
Deltaflume2006 T01 zebra 3 3 1D WG C 2 3
DeltaflumeH298 T1 3 3 1D WG C 1 3 3
DeltaflumeLIP11D 1B 3 3 1D WG C 1 3
DeltaflumeLIP11D 1C 3 3 1D WG C 1 3
DeltaflumeLIP11D 2E 3 3 1D WG C 1 3
Deltaflume H4731 T11 3 3 1D WG C 1 3 3
94 Deltares
4867
Groundwater flow
Default settings
Water levels**
Configuration
Morphology
Hard layers
Fractions
Waves*
Matlab
Status
Test Run
Deltaflume M1263 III Test-1 3 3 1D WG C 1 3
Deltaflume M1797 T01 3 3 1D WG V 1 3
Deltaflume M1797 T02 3 3 1D WG V 1 3
Drifters default 3 3 2D ? C 1
GWK86 T01 3 3 1D WG V 1 3
GWK86 T02 3 3 1D WG V 1 3
GWK86 T03 3 3 1D WG V 1 3
GWK86 T04 3 3 1D WG V 1 3
GWK86 T05 3 3 1D WG V 1 3
GWK86 T06 3 3 1D WG V 1 3
GWK98 A9 3 3 1D WG C 1 3
GWK98 B2 3 3 1D WG C 1 3
GWK98 C2 3 3 1D WG C 1 3
GWK98 F1 3 3 1D WG C 1 3
GWK98 H2 3 3 1D WG C 1 3
MICORE Cadiz Urb... north 7 7 1D WG V 1 3
MICORE Dziwnow Spit 386 3 3 1D WG V 1 3
MICORE Kamchia S... pr04 7 7 1D WG V 1 3
MICORE Lido di Dante mar201... 3 3 1D WG V 1 3
MICORE Mariakerk... s116a 3 3 1D WG V 1 3 3
MICORE Praia de Faro event9 3 3 1D WG V 1 3
Netcdf default 7 7 1D ST C 1
Ningaloo reef default 3 3 1D WG V 1 3
River Outflow default 3 3 2D ST C 1 3
Scheldtflume H4265 T01 3 3 1D WG C 1 3
Scheldtflume H4265 T02a 3 3 1D WG C 1 3
Scheldtflume M1819 I T01 3 3 1D WG V 1 3
Deltares 95
4867
Groundwater flow
Default settings
Water levels**
Configuration
Morphology
Hard layers
Fractions
Waves*
Matlab
Status
Test Run
Scheldtflume M18... T01 3 3 1D WG V 1 3 3
Windflume M1263 I AT33 3 3 1D WG C 1 3
Windflume M1263 I BT13 3 3 1D WG C 1 3
Windflume M1263 I CT14 3 3 1D WG C 1 3
96 Deltares
4867
Groundwater flow
Default settings
Water levels**
Configuration
Morphology
Hard layers
Fractions
Waves*
Matlab
Status
Test Run
Windflume M1263 I DT34 3 3 1D WG C 1 3
Windflume M1263 II 101 3 3 1D WG C 1 3
Zelt Case1 default 3 3 2D WG C 1
Zwin T01 default 3 3 2D ST V 1 3
long wave propag... default 3 3 1D ST C 1
longcrested refr... 1bin a... 3 3 2D WG C 1
longcrested refr... 10deg 3 3 1D ST C 1
longcrested refr... 2p5deg 3 3 1D ST C 1
longcrested refr... 5deg 3 3 1D ST C 1
longcrested refr... snellius 3 3 1D ST C 1
tideonly blanke... 3 3 2D ST V 1
tideonly default 3 3 2D ST V 1
Deltares 97
4867
Groundwater flow
Default settings
Water levels**
Configuration
Morphology
Hard layers
Fractions
Waves*
Matlab
Status
Test Run
* ST = stationary, WG = wave groups, NH = non-hydrostatic
** C = constant, V = varying
98 Deltares

Skillbed XBeach 1.22.4687 Kingsday Release

Uploaded by

Copyright:

Available Formats

Skillbed XBeach 1.22.4687 Kingsday Release

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Skillbed XBeach 1.22.4687 Kingsday Release

Uploaded by

Copyright:

Available Formats

XBeach skillbed report, revi-

status update trunk default

October 21, 2015

Published and printed by:

telephone: +31 88 335 85 85

For support contact:

telephone: +31 88 335 85 55

Copyright 2015 Deltares

3 Morphological laboratory tests 23

4 Morphological field tests 61

5 Comparisons with other models 75

A Model Performance Statistics 89

1.1 Introduction to the XBeach model

1.2 Model approach and innovations

barrier breaching (Roelvink et al., 2003).

1.3 XBeach skillbed

Morphodynamics start with hydrodynamics. In this chapter the hydrodynamic results of

2.1 Long wave propagation

Water level at t = 0 to 160s with 20s interval

Water level at t = 0 to 1200s with 20s interval

2.2 1D wave runup (analytical solution)

Water level variation in time

50 60 70 80 90 100 110 120 130 140 150

50 60 70 80 90 100 110 120 130 140 150

Figure 2.3: Snapshots of water level and velocity

2.3 2D wave runup

Figure 2.4: Definition scetch of concave beach bathymetry

The first panel in Figure 2.5 shows the vertical runup

R2 Sci Rel. bias BSS

2.4 High- and low-frequency wave transformation

Figure 2.6: Locations of surface eleveation measurements

wave heights and water levels

Figure 2.7: Wave height transformations during Boers 1C experiment

2.5 Field experiment: DELILAH

In order to verify the 2DH hydrodynamics of XBeach when forced by directionally-spread

Figure 2.8: Bathymetry and measurement locations

wave heights and water levels

R2 Sci Rel. bias BSS

2.6 Field experiment: Ningaloo reef

Hrms.sw (m) Hrms.IG (m) zs (m)

2.7 Longcrested refraction

1 bin 0.5 0.5

Figure 2.14: Wave height

Figure 2.15: Mean wave direction

Figure 2.16: Time-averaged wave forcing in lateral direction

600 10000 600 10000

600 10000 600 10000

t = 12hr t = 14hr t = 16hr

t = 18hr t = 20hr t = 22hr

0 1000 2000 0 1000 2000 0 1000 2000

5 t = 6hr t = 8hr t = 10hr

5 t = 12hr t = 14hr t = 16hr

6000 4000 2000 6000 4000 2000 6000 4000 2000

Morphological laboratory tests

3.1 Scale relations

3.1.1 Small scale tests

Table 3.1: Overview of experiments

Experiment Scale Profile Sediment Water Wave Wave

profiles Windflume_M1263_I CT26

Figure 3.1: CT26