XBeach Model Description and

Manual
1200116/1002266
Report
Unesco-IHE Institute for Water Education, Deltares and Delft
University of Technology
J une 21, 2010 version 6
XBeach Model Description and
Manual
Dano Roelvink (P.I.), Ad Reniers, Ap van Dongeren, J aap van
Thiel de Vries, Jamie Lescinski, Robert McCall
Report
J une 21, 2010
Unesco-IHE Institute for Water Education, Deltares and Delft
University of Technology
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
i
Contents
1 Introduction...................................................................................................... 11
1.1 Motivation.............................................................................................11
1.2 Objective...............................................................................................12
1.3 Context..................................................................................................12
1.4 Model approach.....................................................................................12
2 Model formulations .......................................................................................... 21
2.1 Coordinate system..................................................................................21
2.2 Grid Setup..............................................................................................22
2.3 Wave action equation solver...................................................................23
2.3.1 Wave boundary conditions.........................................................26
2.4 Roller energy equation solver...............................................................213
2.5 Shallow water equations solver ............................................................215
2.5.1 Depth-averaged equations........................................................215
2.5.2 Quasi-3D equations (advanced option) ....................................216
2.5.3 Flow boundary conditions........................................................216
2.6 Sediment transport...............................................................................224
2.6.1 Wave asymmetry.....................................................................226
2.7 Morphological updating.......................................................................226
2.7.1 Avalanching............................................................................226
2.7.2 Morfac options........................................................................228
2.8 Multiple sediment fractions (advanced option).....................................228
2.9 Hard Layers (advanced option) ............................................................230
2.10 Groundwater flow (advanced option)...................................................230
2.10.1 Physical and numerical principles............................................230
2.10.2 Determining groundwater head................................................230
2.10.3 Momentum balance.................................................................231
2.10.4 Determining vertical flow........................................................231
2.10.5 Mass balance...........................................................................233
2.10.6 Groundwater boundary conditions...........................................235
2.11 Drifters (advanced option) ...................................................................236
2.12 Discharge boundaries (river input) (advanced option)...........................238
2.13 Dam break (advanced option)...............................................................240
3 Descripton of program structure...................................................................... 31
3.1 Single domain setup...............................................................................31
3.2 Implementation of parallel computing using MPI ...................................32
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
i i
4 Compiling the model......................................................................................... 41
5 Running the model............................................................................................ 51
5.1 Input file structure..................................................................................51
5.2 Physical processes..................................................................................51
5.3 Specifying grid and depth.......................................................................52
5.4 Physical constants..................................................................................53
5.5 Time management..................................................................................53
5.6 Wave input.............................................................................................54
5.6.1 Action balance...........................................................................54
5.6.2 Wave dissipation model .............................................................55
5.6.3 Roller model..............................................................................55
5.7 Wave boundary conditions.....................................................................56
5.7.1 User input..................................................................................56
5.7.2 Instat =0 to 3. wave boundary condition parameters for non-
spectral input.............................................................................57
5.7.3 Instat =4, spectral parameter input ............................................59
5.7.4 Instat =5, SWAN spectrum input.............................................511
5.7.5 Instat =6, Use of formatted variance density spectrum.............513
5.7.6 Instat =7, reusing existing boundary condition files..................515
5.7.7 Instat =8, boundary conditions for non-hydrostatic model .......515
5.7.8 Instat =9, no boundary condition.............................................515
5.7.9 Instat =40, sequence of stationary sea states............................515
5.7.10 Instat =41, sequence of sea states to make time-varying wave
groups.....................................................................................516
Notes.......................................................................................516
5.7.11 on generation of boundary conditions......................................516
5.8 Flow input............................................................................................518
5.9 Flow boundary conditions....................................................................518
5.9.1 Time-varying tide/surge...........................................................520
5.10 Wind....................................................................................................521
5.11 Limiters...............................................................................................521
5.12 Sediment transport...............................................................................521
5.13 Multiple sediment fractions and hard layers..........................................522
5.14 Morphological updating and avalanching.............................................523
5.15 Groundwater........................................................................................524
5.16 Drifters................................................................................................525
5.17 Discharge boundaries (river input)........................................................525
5.18 Nonhydrostatic.....................................................................................526
5.19 Data types............................................................................................527
5.19.1 Regular spatial output..............................................................527
5.19.2 Time-averaged spatial output...................................................528
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
i i i
5.20 Selecting output data types...................................................................528
5.20.1 Regular spatial output..............................................................528
5.20.2 Time-averaged spatial output...................................................529
5.20.3 Point output.............................................................................529
5.20.4 Cross-sections.........................................................................531
5.20.5 Default values..........................................................................531
5.21 Choosing output times..........................................................................531
5.21.1 Output at fixed intervals..........................................................531
5.21.2 Output times defined by external file.......................................532
5.21.3 Combinations of fixed interval and external files.....................533
5.21.4 Default values..........................................................................533
5.22 NetCDF CF output............................................................................533
5.23 Reading output files in Matlab..............................................................534
5.24 Output variable keywords.....................................................................537
5.25 Example params.txt file........................................................................541
6 References ......................................................................................................... 61
Appendices
A Description of gen.ezs file (instat+2,3)............................................................. A1
B Numerical implementation.............................................................................. B1
B.1 Wave action equation............................................................................ B1
B.2 Shallow water equations........................................................................ B2
B.3 Advection-diffusion equations............................................................... B5
B.4 Bed update............................................................................................ B6
C Multiple sediment fractions proof of concept .............................................. C1
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
i
Abstract
XBeach is a numerical model of nearshore processes intended as a tool to compute the
natural coastal response during time-varying storm and hurricane conditions, including dune
erosion, overwash and breaching.
The model consists of formulations for short wave envelope propagation, nonstationary
shallow water equations, sediment transport and bed update. Innovations include a newly-
developed time-dependent wave action balance solver, which solves the wave refraction and
allows variation of wave action in x, y, time and over the directional space, and can be used
to simulate the propagation and dissipation of wave groups. An added advantage to this set-
up, compared to the existing surfbeat (infragravity wave) model, is that a separate wave
model is not needed to predict the mean wave direction, and it allows different wave groups
to travel in different directions. Wave-current interaction in the short wave propagation is
included. Various wave breaking dissipation model are implemented for use in the
nonstationary and stationary wave energy balance (in other words, when the wave energy
varies on the wave group timescale).
The Generalised Lagrangean Mean (GLM) approach was implemented to represent the
depth-averaged undertow and its effect on bed shear stresses and sediment transport, cf.
Reniers et al. (2004). Quasi 3D formulations are included as well as ground water flow
through a porous medium.
Soulsby Van Rijn transport formulations have been included, which solves the 2DH
advection-diffusion equation and produces total transport vectors, which can be used to
update the bathymetry. The pickup function follows Reniers et al (2004) was implemented.
An avalanching routine was implemented with separate criteria for critical slope at wet or
dry points providing a smooth and robust solution for slumping of sand during dune erosion.
Since length scales are short in terms of wave lengths and supercritical flow frequently
occurs, the numerical implementation is mainly first order upwind, which in combination
with a staggered grid makes the model robust. The momentum-conserving form of Stelling
and Duinmeijer (2004) is applied which improves long-wave runup and backwash on the
beach. The model scheme utilizes explicit schemes with an automatic time step based on
Courant criterion, with output at fixed or user defined time intervals, which keeps the code
simple and makes coupling and parallellization easier, while increasing stability.
The model has been validated with a series of analytical, laboratory and field test cases. The
model performs well in different situations including dune erosion, overwash and breaching
and these cases are all modelled using a standard set of parameter settings.
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
1 1
1 Introduction
1.1 Motivation
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, but 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. Concurrently, the Dutch Dune Safety Assessment required that advanced models
be developed which can be used to assess the protection provided by dunes against flooding
of the hinterland for situations where empirical model assumptions do not hold anymore. In
the Netherlands, most of the central Holland Coast is protected by dunes. For these reasons
an open-source program, XBeach for eXtreme Beach behavior, has been developed to model
the nearshore response to hurricane impacts and storms.
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 and episodic slumping after
undercutting 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 the runup and are well suited for 1D application but difficult to
apply in a horizontally 2D setting. Hence, a more comprehensive modeling of the swash
motions is called for.
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
1 2
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 wave group 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 Objective
The main objective of the XBeach model is to provide a robust and flexible environment in
which to test morphological modeling concepts for the case of dune erosion, overwashing
and breaching. The top priority is to provide numerical stability and robustness, while still
providing accurate results in a reasonable computational time.
1.3 Context
The XBeach model can be used as stand-alone model for small-scale (project-scale) coastal
applications, but will also be used within larger (shelf sea) systems, where it will be driven
by boundary conditions provided by the wind, wave and surge models and its main output to
be transferred back will be the time-varying bathymetry and possibly discharges over
breached barrier island sections.
1.4 Model approach
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.
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
1 3
The wave-group forcing drives infragravity motions and both longshore and cross-shore
currents. 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
description 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 return flow or 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
turbulence. 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 formulation according to Soulsby Van Rijn (Soulsby, 1997) in a short-wave
averaged but wave-group resolving model of surf zone processes. This formulation has been
applied successfully in describing the generation of rip channels (Damgaard et al., 2002
Reniers et al., 2004a) and 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 run-up on the previously dry dune face and the (smaller) critical wet-
slope.
During the overwash regime the flow is dominated by low-frequency 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, 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 momumtum-
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.
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
1 4
To this end, the code has the following functionalities:
Flow
- Depth-averaged shallow water equations including time-varying wave forcing
terms; combination of sub- and supercritical flows;
- Numerical scheme in line with Stelling and Duinmeijer method, to improve long-
wave runup and backwash on the beach. The momentum-conserving form is
applied, while retaining the simple first-order approach.
- Generalised Lagrangean Mean (GLM) approach to represent the depth-averaged
undertow and its effect on bed shear stresses and sediment transport, cf. Reniers et
al. (2004)
- Smagorinsky viscosity formulation
- Drifter (passive particle) option
- White-Colebrook roughness
- Quasi 3D formulation
- Automatic time step based on Courant criterion, with output at fixed or user-defined
time intervals.
- Ground water flow
- Discharge boundaries
- Non-hydrostatic formulation
- MPI (Message Passing Interface) implementation with automatic domain
decomposition for parallel (multi-processor) computing
Waves
- Time-varying wave action balance including refraction, shoaling, current refraction
and wave breaking;
- Roller model, including breaker delay
- Wave amplitude effects on wave celerity;
- Wave-current interaction
- Roelvink (1993) wave dissipation model for use in the nonstationary wave energy
balance (in other words, when the wave energy varies on the wave group timescale)
- Baldock et al. (1998) wave dissipation formulation for stationary wave energy
balance.
Sediment transport and bed updating
- Depth-averaged advection-diffusion equation to solve suspended transport;
- Bed updating algorithm including possibility of avalanching;
- Soulsby Van Rijn transport formulations, cf Reniers et al (2004).
- Multiple sediment fractions and bed layer bookkeeping.
- Intra-wave sediment transport
- Avalanching mechanism, with separate criteria for critical slope at wet or dry points.
- Hard structures
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 1
2 Model formulations
The model solves coupled 2D horizontal equations for wave propagation, flow, sediment
transport and bottom changes, for varying (spectral) wave and flow boundary conditions.
Because the model takes into account the variation in wave height in time (long known to
surfers) it resolves the long wave motions created by this variation. This so-called surf beat
is responsible for most of the swash waves that actually hit the dune front or overtop it. With
this innovation the XBeach model is better able to model the development of the dune
erosion profile, to predict when a dune or barrier island will start overwashing and
breaching and to model the developments throughout these phases.
2.1 Coordinate system
XBeach uses a coordinate system where the computational x-axis is always oriented towards
the coast, approximately perpendicular to the coastline, and the y-axis is alongshore, see
Figure 2.1 . This coordinate system is defined relative to world coordinates (x
w
,y
w
) through
the origin (x
ori
,y
ori
) and the orientation alfa, defined counter-clockwise w.r.t. the x
w
-axis
(East). The grid size in x- and y-direction may be variable but the grid must be rectilinear.
land
sea
x
y
(xori,yori)
xw
yw
alfa
Figure 2.1 Coordinate system
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 2
2.2 Grid Setup
The grid applied is a staggered grid, where the bed levels, water levels, water depths and
concentrations are defined in cell centers, and velocities and sediment transports are defined
in u- and v-points, viz. at the cell interfaces. In the wave energy balance, the energy, roller
energy and radiation stress are defined at the cell centers, whereas the radiation stress
gradients are defined at u- and v-points, see Figure 2.2.
Velocities at the u- and v-points are denoted by uu and vv respectively; velocities u and v at
the cell centers are obtained by interpolation and are for output purpose only. The water
level, zs, and the bed level, zb, are both defined positive upward. uv and vu are the u-
velocity at the v-grid point and the v-velocity at the u-grid point respectively. These are
obtained by interpolation of the values of the velocities at the four surrounding grid points.
The model solves coupled 2D horizontal equations for wave propagation, flow, sediment
transport and bottom changes, for varying (spectral) wave and flow boundary conditions.
Because the model takes into account the variation in wave height in time (long known to
surfers) it resolves the long wave motions created by this variation. This so-called surf beat
is responsible for most of the swash waves that actually hit the dune front or overtop it. With
this innovation the XBeach model is better able to model the development of the dune
erosion profile, to predict when a dune or barrier island will start overwashing and
breaching and to model the developments throughout these phases.
1,1 2,1 3,1 nx-1,1 nx,1 nx+1,1
1,2 2,2 3,2 nx-1,2 nx,2 nx+1,2
1,3 2,3 3,3 nx-1,3 nx,3 nx+1,3
1,ny+1 2,ny+1 3,ny+1 nx-1,ny+1 nx,ny+1 nx+1,ny+1
1,ny 2,ny 3,ny nx-1,ny nx,ny nx+1,ny
1,ny-1 2,ny-1 3,ny-1 nx-1,ny-1 nx,ny-1 nx+1,ny-1
uu,vu
vv,uv
zs,zb,u,v
Figure 2.2 Staggered grid
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 3
2.3 Wave action equation solver
The wave forcing in the shallow water momentum equation is obtained from a time
dependent version of the wave action balance equation. Similar to Delft Universitys
(stationary) HISWA model (Holthuijsen et al., 1989), the directional distribution of the
action density is taken into account whereas the frequency spectrum is represented by a
frequency, best represented by the spectral parameter f
m,-1,0
.. The wave action balance is then
given by:
y
x w
c A
c A c A D A
t x y
u
u o
c
c c c
+ + + =
c c c c
(2.1)
with the wave action:
( , , , )
( , , , )
( , , )
w
S x y t
A x y t
x y t
u
u
o
= (2.2)
where u represents the angle of incidence with respect to the x-axis,
w
S represents the
wave energy density in each directional bin and o the intrinsic wave frequency. The wave
action propagation speeds in x- and y-direction are given by:
( , , , ) cos( )
( , , , ) sin( )
L
x g
L
y g
c x y t c u
c x y t c v
u u
u u
= +
= +
(2.3)
With u
L
and v
L
the cross-shore and alongshore depth-averaged Lagrangian velocities
respectively (defined below), and the group velocityc
g
obtained from linear theory. If wave-
current interaction is turned off (wci=0) then the last term in either equation is not taken into
account.
The propagation speed in u -space is obtained from:
( , , , ) sin cos cos sin cos
sinh2
sin sin cos
h h u u
c x y t
kh x y x y
v v
x y
u
o
u u u u u u
u u u
| | | | c c c c
= + +
| |
c c c c
\ . \ .
| | c c
+
|
c c
\ .
(2.4)
taking into account bottom refraction (first term on the RHS) and current refraction (last two
terms on the RHS) and h is the total water depth. If wave-current interaction is turned off
(wci=0) then the last two terms are not taken into account.
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 4
The wave number k is obtained from the eikonal equations:
0
0
x
y
k
t x
k
t y
e
e
c c
+ =
c c
c
c
+ =
c c
(2.5)
where the subscripts refer to the direction of the wave vector components and e represents
the absolute radial frequency. The wave number is then obtained from:
2 2
x y
k k k = + (2.6)
The absolute radial frequency is given by:
L L
x y
k u k v e o = + + (2.7)
and the intrinsic frequency is obtained from the linear dispersion relation. If wave-current
interaction is turned off (wci=0) then the last two terms are not taken into account.
The total wave energy dissipation, i.e. directionally integrated, due to wave breaking is
modelled according to Roelvink (1993a), which is coded asbreak=1:
rms
max
max
2
H 8 tanh
=1-exp - , ,
w b w
rep
n
w
b
D Q E
T
E kh
Q H H
H g k
o

=
| |
| |
| = =
|
|
\ .
\ .
(2.8)
with (1) O o = , the water density, the breaker index (a free parameter) and the total
wave energy is given by:
2
0
( , , ) ( , , , )
w w
E x y t S x y t d
t
u u =
}
. (2.9)
In a variation of the above, one could also state (break = 3)
2
rms
w b w
rep
H
D Q E
T h
o
= (2.10)
Finally, Roelvink and Daly (break = 4) developed a breaking formulation based on (2.10)
where the fraction of breaking waves is modelled as
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 5
2
1
0
b rms
b rms
Q if H h
Q if H h

= >

= <

(2.11)
which states that waves are fully breaking is the wave height exceeds a threshold and stop
breaking if the wave height fall below another threshold. Details are in Daly and Roelvink
(ICCE 2010).
In the stationary case, we apply Baldock et al. (1998) (break = 2) which states
,
2 2
1
4
b rep b rms
D Q gf H H o = + (2.12)
With (1) O o = and
rep
f representing a representative intrinsic frequency. The fraction of
breaking waves is given by:
2
2
exp
b
b
rms
H
Q
H
( | |
=
( |
\ .
(2.13)
Where the breaking wave height is:
0.88
tanh
0.88
b
kh
H
k

(
=
(

(2.14)
And is a calibration parameter.
In either instationary or stationary case the total wave dissipation, D, is distributed
proportionally over the wave directions:
( , , , )
( , , , ) ( , , )
( , , )
w
w w
w
S x y t
D x y t D x y t
E x y t
u
u = (2.15)
The bed friction dissipation is modelled as
3
2
3 sinh
f w
rep
H
D f
T kh
t
t
| |
=
|
|
\ .
(2.16)
This closes the set of equations for the wave action balance. Given the spatial distribution of
the wave action and therefore wave energy the radiation stresses can be evaluated (using
linear wave theory):
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 6
,
,
2
,
, ,
2
,
1
( , , ) 1 cos
2
( , , ) sin cos
1
( , , ) 1 sin
2
g
xx w w
g
xy w yx w w
g
yy w w
c
S x y t S d
c
c
S x y t S S d
c
c
S x y t S d
c
u u
u u u
u u
| |
= +
|
\ .
| |
= =
|
\ .
| |
= +
|
\ .
}
}
}
(2.17)
Instationary vs. stationary wave solver
For situations where infragravity motions play an important role the wave and roller
equations must be solved in instationary, wave-group resolving mode (instat=1, by default).
For some applications which do not focus on swash motions and where surfbeats are small,
we have implemented an option (instat=0) to solve the stationary problem directly using a
forward marching technique, where the equations are solved grid row by grid row in an
iterative fashion. The wave module is then called everywavint seconds rather than each time
step, which often means a large reduction in computation time.
2.3.1 Wave boundary conditions
Offshore wave boundary conditions
At present, a number of wave boundary conditions can be specified at the offshore
boundary. These are numbered instat=0 through instat=41. The overview is as follows:
instat abbreviated
name
description
0 stat stationary wave boundary condition (sea state)
1 bichrom bichromatic (two wave component) waves
2 ts_1 first-order timeseries of waves (generated outside XBeach)
3 ts_2 second-order timeseries of waves (generated outside
XBeach)
4 jons wave groups generated using a parametric (J onswap)
spectrum
5 swan wave groups generated using a SWAN 2D output file
6 vardens wave groups generated using a formatted file
7 reuse reuse of wave conditions
8 nonh boundary conditions for nonhydrostatic option
9 off no wave boundary condition
40 stat_table a sequence of stationary conditions (sea states)
41 jons_table a sequence of time-varying wave groups
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 7
In the params.txt file the boundary conditions options can be invoked using either the
number (instat=3 for instance) or by using the name instat=ts_2. The name must be
lower case and match exactly.
Stationary wave boundary conditions (instat = 0).
In this case a uniform, constant wave energy distribution is set, based on given values of
H
rms
, T
m01
, direction and power of directional distribution function.
,
,
max
min
0
2
cos
( ) , / 2
cos
1
8
m
m
mean m
m
m
mean rms
e E
E gH
0
0
0 0
0 0 0 t
0 0 0

= <
A
=

(2.18)
Wave energy varying periodically in time (instat = 1):
In this case regular wave groups (i.e. bichromatic waves) are specified.
,
,
max
min
0
cos 1
( ) * 1 cos 2 ,
2
cos
sin( )
m
m
mean
m
long long
m
g long
long
m
t y
e E
T L
c T
L
0
0
0 0
0 t
0 0 0
0
| | | | | |

| = + | |
| | |
\ . \ . \ . A
=

(2.19)
First-order longcrested, irregular wave groups (instat = 2)
In this case E is read in as a function of time; the timeseries is shifted along the y-axis to
account for the oblique incidence.
,
,
max
min
0 ( )
cos
( , ) ,
cos
sin( )
( )
m
m
t y
m
m
m
g
e y E
y
y
c
t 0
0
0 0
0
0 0 0
0
t

=
A
=

(2.20)
This time series needs to be made by a separate routine (not part of XBeach). An example of
the required input format is given in the Appendix B
Second-order longcrested, irregular wave groups (instat = 3)
In this case a bound wave is added to the wave groups using Longuet-Higgins and Stewarts
(1964) theory. The format is prescribed in Appendix B
Standard J ONSWAP spectrum, based on user-input spectrum coefficients (instat = 4)
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 8
With this option alongshore varying timeseries of the wave energy E and bound long wave
z
s
are generated on the basis of a specified analytical 2D J onswap-type spectrum. With this
option realistic second-order bound, directionally-spread seas can be created.
The spectrum is determined by the peak period, wave height, spectral peakedness, mean
angle and directional spreading (see section 5.7 for details on the input values). The routine
follows the procedure as outlined by Van Dongeren et al. (2003), see next paragraph.
Unmodified SWAN 2D spectrum output file (instat = 5)
This option uses a SWAN 2D output file (sp2 file) in unmodified form. The procedure to
calculate the boundary conditions is analoguous to the instat=4 option.
Formatted variance-density spectrum file (instat = 6)
This option uses a formatted variance-density spectrum file, which needs to adhere to
certain criteria (see section 5.5). On the basis of this formatted spectrum, boundary
conditions are calculated using the procedure as outlined under instat = 4.
Reuse boundary condition files from an earlier XBeach simulation (instat = 7)
If the user does not wish to recalculate spectrum-based boundary condition files or
specifically wants to reuse the spectrum-based boundary condition files of another XBeach
simulation, it is possible to do so. In this case the user should select instat = 7 in
params.txt. No further wave boundary condition data need be given in params.txt.
Obviously, the calculation grid should remain the same between runs, as the angles and
number of grid points are embedded in the boundary condition files. In order to use instat 7,
the user should copy ebcflist.bcf and qbcflist.bcf to the current directory. Additionally, the
user should also copy all files listed in ebcflist.bcf and qbcflist.bcf. Generally, these files
have E_ and q_ prefixes.
With instat=4, 5 and 6 time-varying (on the scale of wave groups) boundary conditions are
computed on the basis of stationary input spectra. It is also possible to use a sequence of
varying spectra to compute boundary conditions which not only vary on the time scale of
the wave groups but also have a variation on the longer timescale. The procedure is similar
to the one described above, only the implementation is through the specification of a list of
spectrum files, see section 5.7.
Boundary conditions for non-hydrostatic model (instat = 8)
XBeach can be ran as a non-hydrostatic model, which is essentially the nonlinear shallow
water equations with dispersion terms and without a wave-action driver. The boundary
conditions are described in a separate chapter and are activated using instat=8
No boundary condition (instat = 9)
This is a simple no wave action boundary condition. It still allows for a tidal record to be
specified, however through the zs0file parameter.
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 9
Sequence of stationary sea states (instat = 40)
While instat=0 specifies one stationary sea state (stationary in the sense that there are no
wave groups), it is also possible to specify a series of seastates, each with a duration. This is
done through a file as
instat =40
bcfile =jonswap1.txt
where the name of the bcfile is free, but the structure of the contents is not. It should contain
lines with
Hm0, Tp, angle, gamma (3.3), spreading, duration (s), timestep (=0.05)
Ie
0.7 8 90 3.3 5 1000 0.05
0.8 7 110 3.3 5 1000 0.05
Sequence of sea states to make time-varying wave groups (instat = 41).
This is an extension of instat=4. With instat =41 it is possible to specify a sea state on the
basis of which wave groups are imposed on the model for a certain duration, then specify
another sea state and run wave groups again without having to stop the model.
This condition is specified as
instat =41
bcfile =jonswap1.txt
where the name of the file is free and the structure of the contents is as in instat =40.
Procedure for converting spectra to wave energy and bound long wave
boundary conditions
Using instat=4,5,6,7 40 or 41 measured or parametric spectra are used as input to create
time-varying wave amplitudes, the envelopes of wave groups. This is done using a
procedure as outlined by Van Dongeren et al. (2003).
In order to create a time series of wave energy along the offshore boundary, the input
spectrum is assumed to be composed of K single summation wave components [Miles and
Funke, 1989; in van Dongerenet al., 2003] in the range around the spectral peak where the
energy density is greater than a certain fraction of the peak energy density (proscribed by the
keyword sprdthr). Each wave component has a specific frequency, phase, amplitude and
direction. Summed together the wave components create a time series of the sea surface at
the offshore boundary:
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 1 0
, , ,
1
0, , cos sin 2
K
i i i i i
i
y t B k y f t q u t
=
= +

(2.21)
whereB
i
represents the amplitude of each wave component.
In order to determine the specific properties of the wave components, the frequencies of all
K components are distributed uniformly in the range around the spectral peak. This choice
leads to a frequency resolution which is dependent on K. Each wave component is given a
wave phase using the random phase model. The direction of each wave component is
determined randomly using the Cumulative Distribution Function of the wave direction of
the input spectrum, see Figure 2.7. At this stage the directional CDF is based on integration
across all frequencies. In the case of strong frequency-directional correlation, it may be
advisable to use frequency dependent CDFs instead.
Figure 2.3 Random wave angles are chosen using a random number generator and the cumulative
distribution function of the directional spreading of the input spectrum
Once the frequency and direction of each wave component has been selected, the amplitude
Bi can be calculated by two dimensional interpolations across the 2D-input wave variance
spectrum. A linear correction is made to the amplitudes of the wave components to ensure
the integrated wave variance of the K components is the same as that of the input spectrum.
The wave number k
i
is determined using the dispersion relation, given the mean still water
depth at the offshore boundary.
Equation (2.21) can now be used to generate a time series of the sea surface elevation at the
offshore boundary. The envelope of the sea surface time series can be calculated using a
Hilbert transform, the amplitude of which being a measure for the wave energy:
, ,
1
, ,
2
i i
E y t gA y t
u u
= (2.22)
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 1 1
The particle velocity due to bound infragravity waves at the offshore boundary is calculated
according to the expressions developed by Herbers et al. [1994]. It is stated that bound
infragravity waves are generated by the interaction of two wave components with different
frequencies. The frequency of the bound infragravity wave is given as:
3 2 1
f f f = (2.23)
In the equation above, the subscripts on the right hand side refer to the indices of interacting
short wave pairs. In order to ensure positive interaction frequencies, indices should be
ordered according to increasing frequency of the short wave components. It should be noted
that two interacting wave components contribute to only one infragravity wave frequency.
However, one infragravity wave frequency may be forced by many different wave
component interactions.
Similarly, other properties of the bound infragravity wave can be deduced from the
associated properties of the short wave components. The bound wave number, wave group
velocity and wave phase are given as:
,
2 2
3 1 2 1 2 1 2 3
2 cos k k k k k k k u = +

(2.24)
3
3
3
2
g
f
c
k
t
= (2.25)
3 2 1
t = + (2.26)
Note that equation (2.26) is based on the assumption that the short wave groups and bound
long waves are in equilibrium, and therefore are 180 out of phase. The angle of the bound
long wave can be found using the following relation:
2 2 1 1
3
2 2 1 1
sin sin
arctan
cos cos
k k
k k
u u
u
u u
| |
=
|

\ .
(2.27)
The energy related to a bound infragravity wave with a specific frequency is given as [van
Dongeren et al., 2003]:
, ,
, ,
2 2
2
3 3 3 1 2
0 0
3 1 2 2 1
2 , ,
, ,
f
E f D f f f
E f f E f d d df
t t
u u t
u u u u

A
= + +
+
} } }
(2.28)
Where in equation (2.28) the first term behind the triple integral is the interaction coefficient
as defined by Herbers et al.[1994]. This interaction coefficient is determined through a
perturbation expansion of the Bernouilli equation, details are in original publication. In the
wave boundary condition module a modification of the interaction coefficient is
implemented to convert the output to surface level elevation instead of bed level pressure:
,
, ,
3
1 2
cosh
cosh cosh
surface bed
k h
D D
k h k h
= (2.29)
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 1 2
The amplitude of each bound wave can be found from the bound wave energy [van
Dongerenet al., 2003]:
3 3
2 A E df = (2.30)
Wheredf refers to the frequency resolution of the short waves, i.e. the frequency step used
to generate all K frequency components around the peak of the short wave spectrum.
A time series for the cross shore water flux across the offshore boundary is generated by
means of an Inverse Fourier Transform:
3,
3,
3, 3,
1
(0, ) cos
2
i
K
i i
x g i i
i
A
q t IFFT e c

=
(
=
(

(2.31)
This cross-shore flux is phase-shifted along the offshore boundary as:
3, 3,
3,
3, 3,
1
( , ) cos
2
y i i
K
ik y i i
x g i i
i
A
q y t IFFT e c e

u

=
(
=
(

(2.32)
The cross-shore flux (2.32) and the wave energy (2.22) are specified along the boundary.
Lateral wave boundary conditions
For the lateral boundary conditions we make the following reasonable assumptions for the
incoming wave energy:
- In the stationary case, we assume that the alongshore gradient of the wave energy is
zero; this means we copy the value of one row inside the domain to the boundary, for
the directional bins where the direction is into the model domain;
- In the instationary case, we assume that the gradient along the crest of the wave group is
zero. The direction of the crest is derived from the local mean wave direction and the
values at the boundary are determined by interpolation between the two points on the
row inside around a virtual point taken along the crest direction; in the figure, for
example, the value at point (3,1) is interpolated from points (2,2) and (3,2).
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 1 3
Figure 2.4 Example of interpolation at the lateral boundary.
2.4 Roller energy equation solver
The roller energy balance is coupled to the wave action/energy balance where dissipation of
wave energy serves as a source term for the roller energy balance. Similar to the wave action
the directional distribution of the roller energy is taken into account whereas the frequency
spectrum is represented by a single mean frequency.
The roller energy balance is then given by:
y r
r x r r
r w
c S
S c S c S
D D
t x y
u
u
c
c c c
+ + + = +
c c c c
(2.33)
with the roller energy in each directional bin represented by ( , , , )
r
S x y t u .The roller energy
propagation speeds in x- and y-direction are given by:
( , , , ) cos( )
( , , , ) sin( )
L
x
L
y
c x y t c u
c x y t c v
u u
u u
= +
= +
(2.34)
1,1 2,1 3,1 nx-1,1 nx,1 nx+1,1
1,2 2,2 3,2 nx-1,2 nx,2 nx+1,2
1,3 2,3 3,3 nx-1,3 nx,3 nx+1,3
1,ny+1 2,ny+1 3,ny+1 nx-1,ny+1 nx,ny+1 nx+1,ny+1
1,ny 2,ny 3,ny nx-1,ny nx,ny nx+1,ny
1,ny-1 2,ny-1 3,ny-1 nx-1,ny-1 nx,ny-1 nx+1,ny-1
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 1 4
where u represents the angle of incidence with respect to the x-axis. If wave-current
interaction is turned off (wci=0) then the last terms in either equation are not taken into
account.
The propagation speed in u -space is identical to the expression used for the wave energy
density propagation, (eq. (2.4), thus assuming that waves and rollers propagate in the same
direction. The phase velocity is obtained from linear wave theory:
c
k
o
= (2.35)
The total roller energy dissipation is given by (Reniers et al., 2004a):
2
r r
r
g E
D
c
|
= (2.36)
which combines concepts by Deigaard (1993) and Svendsen (1984)
Next the total roller dissipation,
r
D , is distributed proportionally over the wave directions:
( , , , )
( , , , ) ( , , )
( , , )
r
r r
r
S x y t
D x y t D x y t
E x y t
u
u = (2.37)
This closes the set of equations for the roller energy balance.
The roller contribution to radiation stress is given by:
2
,
, ,
2
,
( , , ) cos
( , , ) ( , , ) sin cos
( , , ) sin
xx r r
xy r yx r r
yy r r
S x y t S d
S x y t S x y t S d
S x y t S d
u u
u u u
u u
=
= =
=
}
}
}
(2.38)
These roller radiation stress contributions are added to the wave-induced radiation stresses
(eq. (2.17) to calculate the wave forcing utilizing the radiation stress tensor:
, , , ,
, , , ,
( , , )
( , , )
xy w xy r xx w xx r
x
xy w xy r yy w yy r
y
S S S S
F x y t
x y
S S S S
F x y t
x y
c + c + | |
= +
|
c c
\ .
c + c + | |
= +
|
c c
\ .
(2.39)
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 1 5
2.5 Shallow water equations solver
2.5.1 Depth-averaged equations
For the low-frequency and mean flows we use the shallow water equations. To account for
the wave induced mass-flux and the subsequent (return) flow these are cast into a depth-
averaged Generalized Lagrangian Mean (GLM) formulation (Andrews and McIntyre, 1978,
Walstra et al, 2000). In such a framework, the momentum and continuity equations are
formulated in terms of the Lagrangian velocity, u
L
, which is defined as the distance a water
particle travels in one wave period, divided by that period. This velocity is related to the
Eulerian velocity (the short-wave-averaged velocity observed at a fixed point) by:
L E S L E S
u u u and v v v = + = + (2.40)
Hereu
S
, v
S
represents the Stokes drift in x- and y-direction respectively (Phillips, 1977):
cos sin
S S w w
E E
u and v
hc hc
u u

= = (2.41)
where the wave-group varying short wave energy and direction are obtained from the wave-
action balance (eq.(2.1). The resulting GLM-momentum equations are given by:
2 2
2 2
E L L L L L
L L L sx bx x
h
F u u u u u
u v f v g
t x y x y h h x h
t t q
v

| | c c c c c c
+ + + = +
|
c c c c c c
\ .
(2.42)
2 2
2 2
E
L L L L L
sy by y L L L
h
F
v v v v v
u v f u g
t x y x y h h y h
t t
q
v

| | c c c c c c
+ + + + = + +
|
c c c c c c
\ .
(2.43)
0
L L
hu hv
t x y
q c c c
+ + =
c c c
(2.44)
Here ,
bx by
t t are the bed shear stresses, q is the water level,
x
F ,
y
F are the wave-induced
stresses,
t
v is the horizontal viscosity and f is the Coriolis coefficient. The bottom shear
stress terms are calculated with the Eulerian velocities as experienced by the bed:
E L S E L S
u u u and v v v = = (2.45)
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 1 6
and not with the GLM velocities. Also, the boundary condition for the flow computations
are expressed in functions of (u
L
, v
L
) and not (u
E
, v
E
).
Smagorinsky viscosity
TEXTO BE INSERTED
White-Colebrook bottom roughness
TEXT TO BE INSERTED
2.5.2 Quasi-3D equations (advanced option)
NOT OPERATIONAL YET TO BE INSERTED
2.5.3 Flow boundary conditions
Offshore flow boundary conditions
Usually, the offshore boundary is an artificial boundary which has no physical meaning. On
the offshore boundary wave and flow conditions are imposed. In the domain waves and
currents will be generated which need to pass through the offshore boundary to the deep sea
with minimal reflection. One way to do this is to impose a weakly reflective-type boundary
condition.
The options are:
front abbreviated
name
description
0 abs1d absorbing-generating (weakly-reflective) boundary in 1D
1 abs2d same, in 2D (default setting)
2 wall no flux wall
3 wlevel water level specification (from file)
4 nonh_1d boundary condition for nonhydrostatic option
1D absorbing-generating boundary condition
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 1 7
In XBeach, there are two options with regard to the offshore absorbing-generating boundary
condition. With the parameter setting front = 0 a simple one-dimensional radiating
boundary condition is activitated.
It reads:
i s s0
g
gh g
u= 1+ u +u- (z -z )
c h
| |
|
|
\ .
(2.46)
Where u
i
is the incoming particle velocity and z
s
is the surface elevation of the incoming
bound long wave, and z
s0
is the mean water level (averaged over many wave groups), u is
the mean velocity (current). This boundary condition assumes all incoming and outgoing
waves propagate normal to the boundary. It is therefore only useful for 1D (flume like)
simulations.
2D absorbing-generating boundary condition
With option front =1 (default value) the formulation by Van Dongeren and Svendsen
(1997) is activated which in turn is based on Verboom et al. (1981) and is based on the
Method of Characteristics. This boundary condition allows for obliquely-incident and
reflected waves, and is therefore suited for 1D and 2D computations.
The boundary conditions satisfy the following two necessary conditions:
1. the region outside the computation domain can influence the motion within the domain
only through the incident (long) waves and through the currents along the boundaries;
and
2. the (long) waves propagating out of the computational domain must be allowed to freely
propagate through the open-ocean offshore boundary with minimal reflection.
By placing the open boundaries carefully, one can achieve weak local forcing near these
boundaries. In practice this means that the offshore boundary is placed in sufficiently deep
water, i.e. outside the shoaling zone. Then the dominant terms in the continuity and
momentum equations near these boundaries are the nonlinear shallow water equation.s
For the general case of an arbitrary angle between the boundary at a point and the
coordinate axes, one can follow the work of Abbott (1979) and Verboom et al. (1981) to
derive the governing equations, which are valid for an arbitrary angle between the
coordinate axes and the model boundary (Figure 2.5a).
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 1 8
Figure 2.5 Coordinate system (a) for arbitrary angle between domain boundary and x-axis; (b) for =0
The derivation becomes simplified if the coordinate system is defined in a way the the x-
axis is normally inward to the seaward boundary of the rectangular domain, which sets =0
(Figure 2.5b). The governing equations derived following Abbott (1979) and Verboomet al.
(1981) then simplify to:
0
( )
v h
u c v c g F
t x y y x
|
| | |

c c c c c
= + + +
c c c c c
(2.47)
0
( )
h v
u c v c g F
t x y y x
|
| | |
+
+ + +
c c c c c
= + + +
c c c c c
(2.48)
u v g F
t x y y

q c c c c
= +
c c c c
(2.49)
Where, F includes all local forcing and friction terms for the motion, c is the wave celerity,
andh
0
is the still water depth. The Riemann variable
-
is defined as:
0
2 2 ( ) u c u g h | q

= = + (2.50)
Here u is the depth-averaged velocity. The Riemann variable + is similarly defined
as 2 u c |
+
= + . The -equation is the y-momentum equation, which has the Riemann
variable:
v = (2.51)
The definition sketch in Figure 2.6 shows that - propagates in the negativex-direction, +
propogates in the positivex-direction, and in they-direction.
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 1 9
The forcing terms, F, in equations (2.47)-(2.49) originate from the right-hand side of the
nonlinear shallow water equations, which imply that - , +, and are variables rather than
constants.
Figure 2.6 Definition sketch of the characteristics.
The offshore boundary conditions uses the outgoing - variant which contains information
about the waves leaving the domain and the variant which propagates along the boundary.
The latter is extra information which we will use to estimate the direction of the outgoing
wave which is the innovation in Van Dongeren and Svendsen (1997).
The procedure is as follows: during the computation, at time step n, we know the values
of
n
q and the total velocities( , )
n n
u v at all points in the domain. The incoming wave is
specified along the open boundaries through the x and y components of the particle
velocities of the incident wave( , )
in in
u v . The numerical integration of nonlinear shallow
water equations will provide the values of the totalq , u, andv for the interior points in the
domain at time stepn+1, and then the equivalent total (incoming plus outgoing) values need
to be determined along the boundaries at n+1. In other words, given the incoming wave, the
outgoing wave needs to be determined.
In XBeach the lowest-order derived equations are implemented for the weakly reflective
boundary conditions, with x=0 at the boundary. The outgoing wave angle (
r
) and velocity
in thex-direction (u
r
) are solved iteratively. For specifics on this derivation we refer to Van
Dongeren and Svendsen (1997), the shorter outline is given below
The - is updated along the boundary only through (2.47) which discretized in XBeach
(similarly as thex-momentum equation) reads
,
1 n
, , i+1,j i,j , ,
,
, , i+1,j i,j , ,
n n n n n n
i j i j x i j
i j
i j i j u i j
h h F
v
u c v c g
t x y y x x h
| |
| |

+

c c c
= + + +
A c c c
(2.52)
And is thus known at the time level n+1. We can then solve for the outgoing velocity u
r
by
expanding the Riemann variant (2.50) to lowest order as
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 2 0
0 0
0
1
2 ( ) 2 1
2
u g h u gh
h
q
| q

| |
= + = +
|
\ .
(2.53)
We further have the identities
0
cos
cos
i r
i r
i g i i
r r r
u u u u
u c
u gh
q q q
q u
q u
= + +
= +
=
=
(2.54)
where the last two identities assume waves propagating with constant form. The incoming
bound wave propagates at the group speed and the reflected wave at the free shallow water
celerity, where
i
and
r
are the angles of the incoming (known) wave and the outgoing (yet
unknown) wave, relative to the x=0 boundary. u is the mean current. Inserting these
identities into (2.53) and re-arranging gives:
0
0
cos
cos
2
cos 1 cos
g i
r
r i
r g i
c gh
u u gh u
c
u
u
|
u u

( | |
| |
( = + |
|
|
+
(
\ .
\ .
(2.55)
All the terms on the right hand side are known except for r which can be solved from the
=v variant as:
arctan arctan
r r
r
r i
u u
v v v v
u
| | | |
= =
| |

\ . \ .
(2.56)
Eqs. (2.55) and (2.56) are then solved iteratively to yield bothu
r
and
r
. The final boundary
condition is then the total velocity u u = u
i
+ u
r
+u at the boundary at the time level n+1.
Wall boundary condition
This boundary condition type is activated using front =2 and is a simple no flux boundary
condition.
Water level boundary condition
This boundary sets the water level at a prescribed value. This can be constant or time-
varying, see description of tide varying surge in Chapter 5. With this option the outgoing
long waves are not absorbed.
Non-hydrostatic boundary condition
This option is to be used with the non-hydrostatic mode, see Appendix
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 2 1
Bay-side flow boundary condition
In many cases, the onshore boundary will be a land boundary. In some case the onshore
will be a bay behind a barrier island. In that case it is possible to specify a number of
boundary conditions.
The options are:
back abbreviated
name
description
0 wall no flux wall
1 abs1d absorbing-generating (weakly-reflective) boundary in 1D
2 abs2d same, in 2D (default setting)
3 wlevel water level specification (from file)
1D absorbing boundary condition
This boundary condition type is activated using back =0 and is analoguous to (2.46)
without the possibility of specifying an incoming particle velocity.
Wall boundary condition
This boundary condition type is activated using back =1 and is a simple no flux boundary
condition
2D absorbing boundary condition (default)
This boundary condition type is activated using back =2 and is analoguous to the
absorbing-generating boundary condition by Van Dongeren and Svendsen (1997), but
without the incoming particle velocities (so 2D absorbing only).
Water level boundary condition
See above.
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 2 2
Lateral flow boundary conditions
1,1 2,1 3,1 nx-1,1 nx,1 nx+1,1
1,2 2,2 3,2 nx-1,2 nx,2 nx+1,2
1,3 2,3 3,3 nx-1,3 nx,3 nx+1,3
1,ny+1 2,ny+1 3,ny+1 nx-1,ny+1 nx,ny+1 nx+1,ny+1
1,ny 2,ny 3,ny nx-1,ny nx,ny nx+1,ny
1,ny-1 2,ny-1 3,ny-1 nx-1,ny-1 nx,ny-1 nx+1,ny-1
uu,vu
vv,uv
zs,zb,u,v
Neumann boundary
Neumann boundary
Riemann
boundary
Figure 2.7 Stencil for Neumann-type boundary conditions.
Neumann boundaries (default)
Lateral boundaries are the boundaries perpendicular to the coastline. Usually these are
artificial, because the model domain is limited but the physical coast will continue. At these
boundaries we need to prescribe information about the area beyond the numerical model
domain in such a way that the boundary condition does not influence the results in an
adverse way. The best way to do this is to prescribe so-called no-gradient or Neumann
boundaries, which state that there is locally no change in surface elevation and velocities.
These boundary conditions are activitate using left =0 or equivalently left=neumann
and/or right =0 or equivalently right=neumann), where the longshore water level
gradient is prescribed. The alongshore gradient is prescribed by the difference in specified
water levels at the offshore corner points, divided by the alongshore length of the domain.
This type of Neumann boundary condition has been shown to work quite well with (quasi-)
stationary situations, where the coast can be assumed to be uniform alongshore outside the
model domain. So far we have found that also in case of obliquely incident wave groups this
kind of boundary conditions appears to give reasonable results, though rigorous testing still
has to be done. The implementation consists of copying water levels from row 2 to row 1
and from row ny to row ny+1, and doing the same for the cross-shore (along-boundary)
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 2 3
velocities. The alongshore velocities can now be computed from row 1 through rowny; no
additional boundary conditions are required for the alongshore velocity.
Simple no-flux boundary conditions (walls) can be set using left =1 (or left =wall) and/or
right =1 or right = wall. Wall boundary conditions are preferred over Neumann
boundary conditions in 1D (cross-shore) models.
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 2 4
2.6 Sediment transport
Advection-diffusion scheme
The sediment transport is modeled with a depth-averaged advection diffusion equation
[Galappatti and Vreugdenhil, 1985]:
E E
eq
h h
s
hC hC
hC hCu hCv C C
D h D h
t x y x x y y T

( c c c c c c c
(
+ + + + =
(
(
c c c c c c c


(2.57)
whereC represents the depth-averaged sediment concentration which varies on the wave-
group time scale, and D
h
is the sediment diffusion coefficient. The entrainment of the
sediment is represented by an adaptation timeT
s
, given by a simple approximation based on
the local water depth, h, and sediment fall velocityw
s
:
max 0.05 ,0.2
s
s
h
T s
w
| |
=
|
\ .
(2.58)
where a small value of T
s
corresponds to nearly instantaneous sediment response. The
entrainment or deposition of sediment is determined by the mismatch between the actual
sediment concentration, C, and the equilibrium concentration, C
eq
, thus representing the
source term in the sediment transport equation.
The bed-updating is discussed next. Based on the gradients in the sediment transport the bed
level changes according to:
,
0
1
y
b mor x
q
z f q
t p x y
c | | c c
+ + =
|
c c c
\ .
(2.59)
wherep is the porosity,
mor
f is a morphological acceleration factor of O(1-10) (e.g. Reniers
et al., 2004a) and q
x
and q
y
represent the sediment transport rates in x- and y-direction
respectively, given by:
( , , )
E
x h
hCu C
q x y t D h
x x x
( c c c ( (
= +
( ( (
c c c

(2.60)
and
( , , )
E
y h
hCv C
q x y t D h
y y y
( ( ( c c c
= +
( ( (
c c c

(2.61)
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 2 5
To account for bed-slope effects on sediment transport a bed-slope correction factor f
slope
is
introduced.
Transport formulations
The equilibrium sediment concentration can be calculated with various sediment transport
formulae. At the moment the sediment transport formulation of Soulsby-van Rijn (Soulsby,
1997) has been implemented. TheC
eq
is then given by:
2.4
0.5
2
2
| | 0.018 (1 )
E sb ss rms
eq cr b
d
A A u
C u u m
h C
o
| |
| | +
| = +
|
|
\ .
\ .
(2.62)
where sediment is stirred by the Eulerian mean and infragravity velocity in combination
with the near bed short wave orbital velocity, u
rms
. In the default mode, the sediment is
stirred due to mean and infragravity velocities. By settinglws = 0 the mean component can
be excluded. The shortwave stirring can be turned off by settingsws = 0. By default sws =
1. Theu
rms
obtained from the wave-group varying wave energy using linear wave theory as
rms
u
2sinh( )
rms
rep rms
H
T kh H
t
o
=
+
(2.63)
The combined mean/infragravity and orbital velocity have to exceed a threshold value, u
cr
,
before sediment is set in motion. The drag coefficient, C
d
, is due to flow velocity only
(ignoring short wave effects). To account for bed-slope effects on the equilibrium sediment
concentration a bed-slope correction factor is introduced, where the bed-slope is denoted by
m and
b
o represents a calibration factor. The bed load coefficients A
sb
and the suspended
load coefficient A
ss
are functions of the sediment grain size, relative density of the sediment
and the local water depth (see Soulsby (1997) for details). Note that the transport model
does not contain transport contributions related to wave skewness.
The Soulsby-Van Rijn formulation is not strictly valid for sheet flow conditions. If applied
in high velocity situations, the formulation as used in XBeach leads to unrealistically high
sediment transport rates. In order to compensate this, steady flow velocities used to mobilize
sediment are limited by an upper-bound Shields parameter for the start of sheet flow

sf
=0.8 1.0):
, ,
2 2
2 50
,
min ,
E E
flow stirring sf
f
gD
u u v
c
u
| |
A
= +
|
|
\ .
(2.64)
This approach assumes that in sheet flow conditions higher velocities lead to higher
sediment transport rates, but not to higher equilibrium sediment concentrations, which is not
necessarily correct. However, the assumption does cause sediment discharge under sheet
flow conditions to become a linear function of flow discharge, which is in line with
Kobayashi et al. (1996).
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 2 6
2.6.1 Wave asymmetry
The wave asymmetry enters the advection-diffusion equation, repeated here:
eq
AV AV
h h
s
hC hC
hCu hCv hC C C
D h D h
t x y x x y y T

( c c c c c c c (
+ + + + =
(
(
c c c c c c c


(2.65)
where
cos
sin
E
AV W m
E
AV W m
u V u
v V v
u
u
= +
= +
(2.66)
with m is the mean wave angle and
,
W ua rms k s
V u S A = (2.67)
is the velocity amplitude, with ua the free parameter, keyword facua. and
,
,
0.60
0.60
0.79
cos tanh 0.64/
0.61 log
2 2
1 exp
0.35
0.79
sin tanh 0.64/
0.61 log
2 2
1 exp
0.35
k r
r
s r
r
S U
U
A U
U
t t
t t
| |
= +
|

\ .
+

| |
= +
|

\ .
+

(2.68)
the skewness and asymmetry as parameterized as a function of the Ursell number by
Ruessink and Van Rijn.
2.7 Morphological updating
2.7.1 Avalanching
To account for the slumping of sandy material during storm-induced dune erosion
avalanching is introduced to update the bed-evolution. Avalanching is introduced when a
user-defined critical bed-slope (keywords: wetslp and dryslp) is exceeded:
b
cr
z
m
x
c
>
c
(2.69)
Where the estimated bed slope is given by:
, 1, , , b i j b i j
b
z z
z
x x
+

c
=
c A
(2.70)
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 2 7
The bed-change within one time step is then given by:
min , 0.05 , 0
max , 0.05 , 0
b b
b cr
b b
b cr
z z
z m x t
x x
z z
z m x t
x x
| | | c | c
A = A A >
| |
c c
\ . \ .
| | | c | c
A = A A <
| |
c c
\ . \ .
(2.71)
Where a threshold of 0.05 m/s has been introduced to prevent the generation of large
shockwaves. The corresponding bed update is given by:
1
, , , , , ,
1
, 1, , 1, , ,
n n
b i j b i j b i j
n n
b i j b i j b i j
z z z
z z z
+
+
+ +
= + A
= A
(2.72)
To account for continuity, e.g. when sand is deposited within the wet part of the domain, the
water level is also updated:
1
, , , , , ,
1
, 1, , 1, , ,
n n
s i j s i j b i j
n n
s i j s i j b i j
z z z
z z z
+
+
+ +
= + A
= A
(2.73)
Similar expressions are used for the subsequent avalanching in the y-direction. Here we
consider that inundated areas are much more prone to slumping and therefore we apply
separate critical slopes for dry and wet points; default values are 1 and 0.3, respectively. The
former value is consistent with the equilibrium profile according to Vellinga (1986); it is
higher than the angle of natural repose and must be seen as an average slope observed after
dune erosion, where some stretches may exhibit vertical slopes and other, drier parts may
have slumped further. The underwater critical slope is much lower, and our estimate is based
on the maximum underwater slopes we have observed in experiments, e.g. the Zwin test
(see below) and tests carried out at Oregon State University with initially rather steep
profiles.
When the critical slope between two adjacent grid cells is exceeded, sediment is exchanged
between these cells to the amount needed to bring the slope back to the critical slope. This
exchange rate is limited by a user-specified maximum avalanching transport rate, which for
sandy environments is usually set so high as to have no influence on the outcome, while
ensuring numerical stability.
In our model simulations, the avalanching mechanism is typically triggered when a high
infragravity wave reaches the dune front and partly inundates it. The critical underwater
slope is suddenly exceeded and the two grid cells at the dune foot are adjusted during the
first timestep when this happens. In subsequent timesteps a chain reaction may take place
both in points landward, where now the critical dry slope may be exceeded because of the
lowering of the last wet point, and in points seaward, where now the critical wet slope may
be exceeded. As a result, sediment is brought from the dry dune into the wet profile, where it
is transported further seaward by undertow and infragravity backwash. An essential
difference with similar procedures in other dune erosion models is the fact that avalanching
is only applied between adjacent grid cells, rather than extrapolating profile behaviour well
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 2 8
beyond the wet domain. This is made possible by explicitly resolving the long-wave swash
motions. Another big advantage with respect to existing procedures is that the simple
avalanching algorithm is readily applied in two dimensions.
2.7.2 Morfac options
The morfac is the morphological acceleration factor that speeds up the morphological time
scale relative to the hydrodynamic timescale. It means that if you have a simulation of 10
minutes with a morfac of 6 you effectively simulate the morphological evolution over one
hour. There are now two ways in which you can input the time-varying parameters in
combination with morfac; this is governed by the input option morfacopt.
morfacopt=1
All times are prescribed on input in morphological time. If you apply a morfac all input
timeseries and other time parameters are divided internally by morfac. This way, you can
specify all timeseries as real times, and vary the morfac without changing the rest of the
input files. A typical application is that you run over a storm of some days and specify time-
varying water level and wave conditions. If you now specify a morfac of 6, the model just
runs for 10 (hydrodynamic) minutes each hour, during which the bottom changes per step
are multiplied by a factor 6. This of course saves a factor of 6 in computation time.
An important thing to note is that the can only be done as long as the water level changes
that are now accelerated by morfac do not modify the hydrodynamics too much. This is the
case if the tide is perpendicular to the coast and the vertical variations do not lead to
significant currents. If you have an alongshore tidal current, as is the case in shallow seas,
you cannot apply this method because you would affect the inertia terms and thus modify
the tidal currents.
morfacopt=0
In this new option the philosophy is different: you run the model over, say, a tidal cycle and
apply the morfac without modifying the time parameters. This means you leave all the
hydrodynamic parameters unchanged and just exaggerate what happens within a tidal cycle.
As long as the evolution over a single tidal cycle is limited, the mean evolution over a tidal
cycle using a morfac is very similar to running morfac tidal cycles without morfac. See
Roelvink, 2006 for a more detailed description of this approach. This method is more
appropriate for longer-term simulations with not too extreme events.
2.8 Multiple sediment fractions (advanced option)
To model the overwash deposits at barrier islands during extreme conditions XBEACH has
been extended with a multiple sediment class formulation. This allows for the tracking of
sediment but also for assigning different sediment characteristics such as grain size
diameter, fall velocity, mobility, etc. For each sediment class, i, the equilibrium sediment
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 2 9
concentration, ) (
*
i c
eq
, is calculated according to the Soulsby-van Rijn formulation, see
equation (2.59) .
The actual concentration then depends on the mismatch with the equilibrium concentration
in combination with the available fraction at that location. It is assumed that a top-layer of
10 cm depth is readily available for sediment pick-up. So based on the fractions of the
various sediment classes present in the top-layer the equilibrium concentration per sediment
class can be expressed as:
*
( ) ( ,1) ( )
eq eq
c i frc i c i = (2.74)
Where the index 1 refers to the top layer and frc the fraction of a specific sediment class.
Next the advection-diffusion equation (see eq. 2.54) is solved independently for the different
sediment classes leading to class dependent sediment transport rates,
i
S , from which the
bottom changes per sediment class,
i
z A , can be derived:
, ,
1
i y i x
i
p
S S
t
z
n x y
c c ( A
A = +
(
c c

(2.75)
Changes in fractional composition of the sediment classes in the top-layer due to sediment
deposition are then calculated by:
1
( ,1) ( ,1)
n n i z
z z
z D z
frc i frc i
D D
+
A A
= + (2.76)
Given dz <=Dz, else:
1
( ,1)
n i
z
frc i
z
+
A
=
A
(2.77)
And similarly for erosion:
1
( ,1) ( ,1) ( ,2)
n n n i z
z z
z D z
frc i frc i frc i
D D
+
A + A
= (2.78)
Where the number 2 refers to the layer immediately below the top layer. D
z
is the constant
layer thickness of 10 cm and z A is the total change in bed elevation (all classes combined
and positive upward) at time stepn+1 where n represent the time index. Next the underlying
layers are updated according to:
1
( , ) ( , ) ( , 1)
n n n z
z z
D z z
frc i j frc i j frc i j
D D
+
+ A A
= + (2.79)
For erosion and:
1
( , ) ( , ) ( , 1)
n n n z
z z
D z z
frc i j frc i j frc i j
D D
+
A A
= + (2.80)
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 3 0
During sedimentation where the subscript j refers to the individual layers. In case of erosion,
sediment is thus moving from the bottom layers towards the top layer and vice versa.
.
2.9 Hard Layers (advanced option)
TEXT TO BE INSERTED
2.10 Groundwater flow (advanced option)
2.10.1 Physical and numerical principles
The groundwater module in XBeach utilizes the principle of Darcy flow and is therefore
limited to laminar flow conditions. In situations in which the groundwater flow may become
turbulent, the full momentum equations (e.g. van Gent, 1995) should be applied. The
module includes a vertical interaction flow between the surface water and groundwater. This
flow is assumed to be a magnitude smaller than horizontal flow and is not incorporated in
the momentum balance.
2.10.2 Determining groundwater head
The driving force behind groundwater flow according to Darcy is the groundwater head
gradient. In the XBeach module, the groundwater head
gw
p has the unit [m]. Where there is
no surface water, the groundwater head is equal to the groundwater surface level
gw
q :
1
1
,
, ,
if 0
n n
n
gw gw i j
i j i j
p wetz q


( ( = =

(2.81)
Where there is surface water and the groundwater surface level is just below the surface of
the bed
b
z , the groundwater head is affected by the surface water head
s
z . If the
groundwater surface level is equal to the bed level, the groundwater head is equal to the
surface water head. If the groundwater surface level is more than
wetlayer
d below the surface
of the bed, the groundwater head is unaffected by the surface water head and is equal to the
groundwater surface level. At intermediate depths a linear interpolation takes place, using
the relative groundwater level fac :
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 3 1
, j
,
j
1 1
1
1
, ,
,
, , ,
1
1
,
,
,
1 if 1
0 1
n n n
n
n n
gw gw i j s gw i j
i j
i j i j i j
n
n
b gw
i j
i j n
i j
wetlayer
p fac z wetz
z
fac fac
d
q q
q

( ( ( = + =

(

= s s
(2.82)
2.10.3 Momentum balance
Darcy flow is described by the following relationship between the groundwater head
gradient, the permeability k , and the horizontal velocity:
gw
gw x
gw
gw y
dp
u k
dx
dp
v k
dy
=
=
(2.83)
In the module, the head gradient is found numerically using:
1, ,
, 1, , ,
,
, 1 ,
, , 1 , ,
,
n
n n
gw i j i j
z i j z i j
i j
n
n n
gw i j i j
z i j z i j
i j
dp p p
dx x x
dp p p
dy y y
+
+
+
+
(
=
(


(
=
(


(2.84)
And horizontal flow is calculated by:
,
,
,
,
n
n
gw
gw x
i j
i j
n
n
gw
gw y
i j
i j
dp
u k
dx
dp
v k
dy
(
( =
(


(
( =
(


(2.85)
2.10.4 Determining vertical flow
In order to simulate the interaction between the surface water and groundwater, a vertical
flow between the surface water layer and groundwater layer ( w) is introduced. This flow
has the unit [ms
-1
] and is defined positive from surface water to ground water and is given in
terms of surface water for the continuity equation (i.e. 100% porosity).
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 3 2
Exfiltration
Exfiltration, or flow from the groundwater layer to the surface water layer, takes place if the
groundwater surface level exceeds the bed level. The volume of groundwater (including
porosity por ) exceeding the bed level is joins the surface water within the same numerical
time step. The vertical velocity can therefore be calculated by:
j
j
1
1
1
, 1
,
,
,
,
if
n
n
gw b n
i j n
i j n
i j gw b
i j
i j
z
w por z
t
q
q

| |
(

|
( = >

|
A
|
\ .
(2.86)
Vertical infiltration model
Surface water running up and down a dry slope will infiltrate into the ground. In order to
model this fully, a 3D model must be used. In the XBeach groundwater module, the option
is made to model infiltration using a quasi-3D model.
In areas where there is surface water and the groundwater level is not greater than the bed
level, infiltration can take place. To a certain degree of truth, infiltration can be calculated
using Darcy flow:
1
z
dp
w k
dz
| |
= +
|
\ .
(2.87)
In an area that is covered by surface water, the head on the top of the bed can be said to be
equal to the surface water head. In the absence of groundwater at the bed level, the head
under the bed level is zero. As the distance between the top and bottom of the bed level is
zero, the head gradient is infinite. The resulting vertical velocity becomes infinite and the
method becomes numerically unstable. In order to circumvent this problem the vertical
infiltration is divided into an instantaneous, but finite reaction in the upper ground layer and
Darcy flow across a non-zero depth. The proportion of the instantaneous part to the Darcy
flow part is governed by the relative groundwater level fac , as in section 2.10.2. The
instantaneous part is handled in the same way as exfiltration. The head gradient for the
Darcy flow is found by assuming the head at the bottom of the infiltration layer is zero, and
the head on the top of the infiltration layer is equal to the height of water standing on the
bed (
s b
z z ).
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 3 3
j j
,
j
j
1
-1 1
, ,
,
1
1
1
,
,
, ,
,
-1
,
,
if 1 and :
1
if 0 and
n
n n
gw b
i j i j
i j
n
n
n
gw b
i j
i j n n s b
i j i j z
infiltration
i j
n
n
gw
i j
i j
wetz z
z
z z
w fac fac k por
t d
wetz
q
q
q

( = <

| |
(
(

|
= +
(
|
A
( |

\ .
( =

j
1
1
,
,
:
0
n
b
i j
n
i j
z
w

<
=
(2.88)
The infiltration velocity is limited by the amount of surface water available in the cell:
1
,
,
n
i j n
i j
h
w
t

s
A
(2.89)
The thickness of the infiltration layer (
infiltration
d ) is increased at the end of every time step
by the infiltrating water. The infiltration speed in the next time step will therefore be less
than that in the current time step. Infiltrating water is assumed to immediately become part
of the groundwater for the purpose of groundwater level and groundwater head calculations.
This approach is therefore not fully 3D and only uses a quasi-3D approximation to limit the
infiltration speed.
1
,
, ,
n
n n
i j
infiltration infiltration
i j i j
w t
d d
por
A
( ( = +

(2.90)
For numerical stability, the infiltration layer thickness is restricted to a minimum of one
third of
wetlayer
d , corresponding with the centroid of the instantaneous infiltration part. The
maximum thickness of the infiltration layer is equal to the depth of the groundwater level
below the bed level. Once an area has no surface water, the thickness of the infiltration layer
is reset to the minimum value, representing the fact that the infiltrated water has sunk out of
the way of subsequent infiltrations.
1
-1
,
,
1 1
-1
,
, ,
1
if 0
3
1
if 1
3
n
n
infiltration wetlayer i j
i j
n n
n
wetlayer infiltration b gw i j
i j i j
d d wetz
d d z wetz q

( = =

( ( s s =

(2.91)
2.10.5 Mass balance
The continuity equation for the groundwater system can be written as:
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 3 4
gw gw ugw gw vgw
d du h dv h
w
dt dx dy por
q
+ + = (2.92)
The effective depths through which horizontal ground water flow takes place ( ,
ugw vgw
h h ),
are found by taking the mean difference between the groundwater level and bed of the
aquifer (
, b aquifer
z ) in the two surroundingq -points:
, ,
, 1,
,
, ,
, , 1
,
2
2
n n
gw b aquifer gw b aquifer n
i j i j
ugw
i j
n n
gw b aquifer gw b aquifer n
i j i j
vgw
i j
z z
h
z z
h
q q
q q
+
+
( ( +

( =

( ( +

( =

(2.93)
This method is faster, but less momentum conservative than the method used in the surface
water flow routine. Since large gradients in the groundwater level are not expected, the
scheme is assumed sufficient.
Groundwater flux is limited in cells that are empty of groundwater. For such cells,
groundwater may enter the cell, but no groundwater may leave until the amount of
groundwater exceeds a minimum value (eps ).
,
,
,
,
, ,
1, 1, 1
,
,
, ,
, 1 , 1
min ,0
max ,0
if
min ,0
max ,0
n n
gw ugw gw ugw
i j i j
n n
gw ugw gw ugw
i j i j n
gw b aquife
i j n n
gw vgw gw vgw
i j i j
n n
gw vgw gw vgw
i j i j
u h u h
u h u h
z
v h v h
v h v h
q

( ( =

( ( =


( s
`

( ( =

( ( =

)
1
,
n
r
i j
eps

( +

(2.94)
The continuity equation for the groundwater level is solved by the following:
1
, , , 1, , , 1 ,
, , , 1, , , , , 1
n n n n n n
n
gw gw gw ugw gw ugw gw vgw gw vgw
i j i j i j i j i j i j i j
u i j u i j v i j v i j
u h u h v h v h
w
t x x y y por
q q
+


( ( ( ( ( (

= +
A
(2.95)
To account for infiltrating and exfiltrating groundwater, an additional term is added to the
continuity equation of the surface water, but none to the momentum balance.
1 1 1 1 1
, , , , 1, 1, , i,j , 1 , 1
,
u,i,j u,i-1,j v,i,j v,i,j-1
n n n n n n n n n
i j i j i j i j i j i j i j i j i j n
i j
u h u h v h v h
w
t x x y y
q q
+ + + + +


=
A
(2.96)
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 3 5
2.10.6 Groundwater boundary conditions
Vertical boundary conditions
The groundwater level is bounded by the bottom of the aquifer. In the central domain the
groundwater level is adjusted naturally by infiltration and exfiltration. The groundwater
level has no bounding maximum in the vertical, except on the offshore, bay side and lateral
boundaries. Here the groundwater level is bounded vertically by the bed level on the
boundaries:
j
,
j
,
j
,
j
,
1,
1, 1,
1,
1, 1,
,1
,1 ,1
, 1
, 1 , 1
min ,
min ,
min ,
min ,
n n
n
gw gw b
j
j j
n n
n
gw gw b
nx j
nx j nx j
n n
n
gw gw b
i
i i
n n
n
gw gw b
i ny
i ny i ny
z
z
z
z
q q
q q
q q
q q
+
+ +
+
+ +
( ( =

( ( =

( ( =

( ( =

(2.97)
The bed of the aquifer is set equal to or less than the regular bed level:
,
, ,
min ,
b aquifer b aquifer b
z z z eps = (2.98)
Offshore boundary condition
At the offshore boundary, the groundwater head is set equal to the offshore surface water
head:
j
,1
,1
n
n
gw s
i
i
p z ( =

(2.99)
Bay side boundary condition
For cases in which a bay side water level is given explicitly with a tidal level record (tideloc
= 4, tideloc = 2 andpaulrevere = 0), the groundwater head on the bay side boundary is set
equal to the bay side surface water head:
j
, 1
, 1
n
n
gw s
i ny
i ny
p z
+
+
( =

(2.100)
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 3 6
In all other cases, the bay side groundwater head is kept at the initial value:
1
, 1 , 1
n
gw gw
i ny i ny
p p
+ +
( ( =

(2.101)
Lateral boundary conditions
Neumann boundary conditions are applied to the groundwater head on the lateral
boundaries:
,1 ,2
, 1 ,
n n
gw gw
i i
n n
gw gw
i ny i ny
p p
p p
+
( ( =

( ( =

(2.102)
Initial conditions
The bed of the aquifer and the initial groundwater head must be specified, see section 5.12
for a description. The initial groundwater level is calculated from the initial groundwater
head.
2.11 Drifters (advanced option)
Drifters are objects that move with the lagrangean mean velocity. They are defined by their
deployment location, deployment time and retrieval time. The positions of the drifters are
evaluated every timestep; every tintp seconds the results are written to files drifternnn.dat,
with nnn the drifter number padded with zeros. The files are in the usual binary format, with
each record containing x,y and time of the drifter; invalid timepoints are characterized by -
999 values.
The drifter module is activated by including the following statements in params.txt.
ndrifter=number of drifters
drifterfile=name of drifterfile
The drifter input file should contain for each drifter a record with 4 numbers indicating x
and y (in world coordinates) of release location, release time and retrieval time (in seconds
from start of simulation)
Example input in params.txt
ndrifter=5
drifterfile=drifters.txt
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 3 7
Content of drifters.txt
550. 100. 100. 1000.
550. 200. 100. 1000.
550. 250. 100. 1000.
550. 300. 100. 1000.
550. 400. 100. 1000.
Example function to read and plot drifter paths
f unct i on r ead_dr i f t er ( )
ch={' r . ' ; ' b. ' ; ' k. ' ; ' g. ' ; ' m. ' ; ' c. ' }
f or i d=1: 6
dep=mod( ( i d- 1) , 6) +1;
i f i d<10
f name=[ ' dr i f t er 00' num2st r ( i d) ' . dat ' ]
el se
f name=[ ' dr i f t er 0' num2st r ( i d) ' . dat ' ]
end
t r y
f i dr =f open( f name, ' r ' )
f or i =1: 10000
xyt =f r ead( f i dr , [ 3] , ' doubl e' ) ;
i f i sempt y( xyt )
br eak
end
xyt ( xyt ==- 999) =nan;
xd( i ) =xyt ( 1) ; yd( i ) =xyt ( 2) ; t d( i ) =xyt ( 3) ;
end
f cl ose( f i dr )
pl ot ( xd, yd, ch{dep}) ; axi s equal
cl ear xd yd t d
cat ch
end
end
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 3 8
Figure 2.8 Example of drifter path output.
2.12 Discharge boundaries (river input) (advanced option)
It is possible to define river discharges to simulate the inflow from a river (i.e. uni-
directional, no tides) and no sediment transport.
The user can define a number of sections along the boundaries of the grid through which
discharges can be defined as time series. For each boundary section, the program checks
how many grid cell centers lie within the section, given in world coordinates. The total
length of the section is computed as the sum of the lengths of the grid cells along the
boundary, and each cell length divided by this total length is used to divide the discharge
over the cells within the boundary section.
For each boundary section a separate time series of total discharge is defined; actual
discharges are determined by interpolation in time. Positive discharges mean discharges into
the model.
The discharges may be defined on initially dry boundaries. They must, however be located
on otherwise closed (wall) boundaries, see parameter (back_
The discharge boundaries are activated with the following keywords inparams.txt:
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 3 9
disch_loc_file=<discharge locations file>
disch_timeseries_file=<discharge timeseries file>
The contents of the discharge locations file are a number of records equal to the number of
discharge boundary sections, with for each record four numbers:
X_begin Y_begin X_end Y_end
These world coordinates must be chosen such that they are close to the desired boundary
and enclose the cell centers of the cells that must be part of the boundary section.
In the discharge time series file, the first column is time in seconds; the next columns give
the total discharge time series per section.
Example: (River Outflow)
disch_loc_file=discharge_locations.txt
disch_timeseries_file=discharge_timeseries.txt
Contents of discharge_locations.txt:
840 820 840 890
Contents of discharge_timeseries.txt:
0 150
1000000 150
-1000 -500 0 500 1000 1500
0
200
400
600
800
1000
1200
1400
1600
102 hrs
discharge boundary section
Figure 2.9 Example of coast with river inflow.
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
2 4 0
2.13 Dam break (advanced option)
Dam break problems require the same set of hydrodynamic equations which XBeach
contains. The only difference between these types of problems and convential coastal model
is the specification of the innitial condition file for the water level.
An non-uniform initial condition for the water level, for instance when modeling a dam
break problem, can be specified by a file with the same format as the depth file. It is
activated by the following keyword inparams.txt:
zsinitfile = <initial condition file>
Example (Dambreak)
Zsinitfile=ini.dep
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
0
5
10
15
bed level, water level
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
0
5
10
15
water depth
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
-10
0
10
20
velocity
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
-20
0
20
40
discharge
Figure 2.10 Dambreak problem: snapshots of the water level, water depth, velocity and discharge.
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
3 1
3 Descripton of program structure
3.1 Single domain setup
The program XBeach consists of a mainFortran 90 program, xbeach.f90, and a number of
subroutines that operate on twoderived types (structures):
- par this contains general input parameters
- s this contains all the arrays for a given computational domain
For a single-domain run, one structure s is passed between flow, wave, sediment and bed
update solvers, which extract the arrays they need from the structure elements to local
variables, do their thing and pass the results back to the relevant structure elements. This
makes the overall program clear, prevents long parameter lists and makes it easy to add
input variables or arrays where needed. The various subroutines and their purposes are
outlined in Table 3.1.
Table 3.1. Overview of Fortran 90 subroutine calls by xbeach.f90
Function call Purpose
wave_input(par) Creates elements of structure par containing wave
input parameters
flow_input(par) Adds elements of structure par containing flow
input parameters
sed_input(par) Adds elements of structure par containing
sediment input parameters
grid_bathy(s) Creates grid and bathymetry and stores them in
structure s
distribute_par(par) MPI
space_alloc_scalars(sglobal) allocates space
grid_bathy(s,par) sets up grid and bathymetry
xmpi_determine_processor_grid(s%nx s%ny)
readtide(s,par) Read tide levels
readwind (s, par) Read wind field
init_output Read output requests and initialize output files
wave_init (s,par) Initialises arrays (elements of s) for wave
computations
flow_init (s,par) Initialises arrays (elements of s) for flow
computations
gwinit(s,par) Initialises arrays (elements of s) for groundwater
module
sed_init (s,par) Initialises arrays (elements of s) for sediment
computations
init_output(sglobal,slocal,par,it)
call readkey
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
3 2
Start time loop
timestep (s,par,it) Calculate automatic timestep
wave_bc (s,par) Wave boundary conditions update, each timestep
gwbc(s,par) Groundwater boundary conditions update, each
timestep
flow_bc (s,par) Flow boundary conditions update, each timestep
wave_timestep(s,par) Carries out one wave timestep, OR
wave_stationary(s,par) Carries out stationary wave computation
gwflow(par,s) carries out flow timestep of groundwater flow
flow_timestep (s,par) Carries out one flow timestep
drifter (s,par)
transus(s,par) Carries out one suspended plus bedload transport
timestep
bed_update(s,par) Carries out one bed level update timestep
var_output(it,s,par) Performs output
End time loop
3.2 Implementation of parallel computing using MPI
We refer to the Parallelization Report.
XBeach Model Description and Manual 1002266 March 2010
Unesco-IHE, Deltares and Delft University of Technology
4 1
4 Compiling the model
To compile the source code you can download the source code from the subversion
repository, located at: https://repos.deltares.nl/repos/XBeach/trunk
The source code contains three different build environments:
GNU autotools, tested with gfortran >=4.2, tested with ubuntu, OSX and cygwin with
windows XP.
Visual studio 2008, tested with intel fortran 11, tested with Windows XP
Visual studio 6, tested with Compaq fortran 6.6, tested with Windows XP
We expect it to compile with only minor problems under other compilers and/or under
Linux, since only standard Fortran 90/95 is used. Once an executable has been created, it
will be called xbeach or xbeach.exe under windows.
Building with MPI support
XBeach can be built with support for mpi. To build with mpi under windows, make sure you
have mpi installed in c:\program files\mpich2. Select the appropriate configuration in visual
studio (mpi debug or mpi release) and use rebuild all to make a mpi enabled executable.
To build a mpi executable use ./configure with-mpi && make to build an executable with
mpi support.
XBeach is tested with openmpi on linux and OS X and with mpich2 on all platforms.
Building with NetCDF support
XBeach can be built with support for netcdf. To build with netcdf under windows, select the
appropriate configuration in visual studio (netcdf) and use rebuild all to make a netcdf
enabled executable.
To build a netcdf executable under linux make sure you have netcdf 4.1 or higher installed.
It should be automatically found by the configure script and compiled.
The combination of netcdf and mpi has not yet been tested.
For more details see the README that come with the source code.
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 1
5 Running the model
5.1 Input file structure
On execution thexbeach.exe executable will read the fileparams;txt in the current directory
and interpret the keyword=value combinations in it. The keywords refer to parameter values
or filenames and can be listed in any order. Lines not containing an = sign are ignored and
may be used for comments.
The params.txt file contains grid and bathymetry info, wave input, flow input and
morphological input. The tables below contain a description of the keywords, the default
values and recommended minimum and maximum values.
5.2 Physical processes
Physical processes can be turned on or off by choice. The commonly used ones are turned
on by default, so for common cases the user does not need to specify anything explicitly
keyword description default
value
minimum
value
maximum
value
unit remarks
swave include/exclude
shortwaves
1 0 1 - turned on
by default
lwave include/exclude
generation of
long waves
1 0 1 - turned on
by default
flow include or
exclude NLSW
equations
1 0 1 - turned on
by default
sedtrans include/exclude
sediment
transport
1 0 1 - turned on
by default
morphology include/exclude
bed updating
1 0 1 - turned on
by default
nonh turn on
nonhydrostatic
flow
0 0 1 - turnedoff
by default
gwflow turn on
groundwater
flow
0 0 1 - turnedoff
by default
q3d turn on quasi 3d 0 0 1 - turnedoff
by default
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 2
5.3 Specifying grid and depth
keyword description default
value
minimum
value
maximum
value
unit remarks
nx number of grid
points in x
50 2 10000 -
ny number of grid
points in y
2 0 10000 -
dx grid size in x -1 0 1e9 m required
value
dy grid size in y -1 0 1e9 m required
value
xori x - origin in
world
coordinates
0 -1e9 1e9 m
yori y origin in
world
coordinates
0 -1e9 1e9 m
alfa angle of grid 0 0 360 deg
posdwn depth defined
positive down
1 -1 1 1 =positive
down
-1 =positive
up
vardx option of
variable grid size
0 0 1 - 1 =varying
grid size
0 =non-
varying grid
size
depfile bathymetry file no values;
specify name
xfile variable gridsize
file x
no values;
specify
name;only
with vardx=1
yfile variable gridsize
file y
no values;
specify name
only with
vardx=1
thetamin lower directional
limit
-90 -180 180 deg angle w.r.t
computational
x-axis
thetamax upper directional
limit
90 -180 180 deg angle w.r.t
computational
x-axis
dtheta directional
resolution
10 0.1 20 deg
thetanaut option to enter
thetamin
thetamax in
nautical
convention
0 0 1 - 0 =cartesian,
1 =nautical
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 3
Thedepfile keyword contains the reference to a bathymetry file, which should contain, for
each of ny+1 rows, nx+1 depth values, which may be defined positive downward or upward,
depending on a keyword posdwn that may be 1 (depth positive downward) or -1 (positive
upward).
If a non-equidistant grid is chosen, the keywordvardx=1 should be selected and an xfile and
yfile have to be specified.
The directional grid for short waves and rollers can either be specified as Cartesian (angle
w.r.t. the computational x-axis) or as nautical directions (direction waves come from in deg.
N, so from W is 270 deg. N); this depends onthetanaut (1 means nautical, default Cartesian
is 0) .
5.4 Physical constants
keyword description default
value
minimum
value
maximum
value
unit remarks
rho density of water 1025 1000 140 kg/m3
g gravitational
acceleration
9.81 9.7 9.9 m/s2
5.5 Time management
The hydrodynamic simulation starts at time 0 and starts outputting at timetstart; it stops at
tstop. The time intervals are described in detail in section 1.1. The actual time step of the
hydrodynamic simulation is determined based on a given maximum Courant number CFL.
In future versions a nonhydrostatic pressure correction term will be implemented in the flow
solver, allowing the modelling of individual short waves. This is not available in the present
code yet.
keyword description default
value
minimum
value
maximum
value
unit remarks
tstart start time of
simulation
1 0 1000000 s
tint time interval
output global
values
1 0.01 100000 s
tintg time interval
output global
values
tint 0.01 100000 supercedes tint if
specified
tintm time interval
output mean
ting 0.01 tstop
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 4
global values
tintp time interval
output point
values
ting 0.01 100000
tsglobal file with list of
output times for
global values
no values;
specify name
tsmean file with list of
output times for
meanglobal
values
no values;
specify name
tspoints file with list of
output times for
point values
no values;
specify name
tscross file with list of
output times at
cross sections
NOT
IMPLEMENTED
YET
tstop stop time
simulation
2000 1 1000000 s
CFL maximum
courant number
0.7 0.1 0.9 - actual CFL
number varies
during
calculation
5.6 Wave input
5.6.1 Action balance
With the switchwci one can turn off or on the wave-current interaction, viz. the feedback of
currents on the wave propagation, see section 2.3. With scheme the numerical scheme can
be chosen, by default use the higher-order Lax Wendroff scheme to minimize numerical
dissipation.
keyword description default
value
minimum
value
maximum
value
unit remarks
wavint interval between
stationary wave
module calls
1 1 3600 s instat = 0
only
scheme Switch numerical
schemes for wave
action balance
2 1 2 - 1 =Upwind
2 =Lax
Wendroff
wci wave current
interaction option
0 0 1 - instat = 0
only
hwci Minimum depth above
which wave-current
interaction is used
0.1 0.0001 1 m
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 5
cats Current averaging time
scale for wci, in terms
of mean wave periods
20 1 50 Trep
5.6.2 Wave dissipation model
For instationary model runs, use the Roelvink (1993) model with either break=1or break=3.
Note that the standard valuegamma=0.55 andn=10 was calibrated for optionbreak=1. for
break=3 the wave dissipation is proportional to H
3
/h instead of H
2
; this affects the
calibration. For stationary runs the Baldock et al, 1998 the break=2 option, is suitable. The
break=4 option is a model in which waves start and stop breaking
keyword description default
value
minimum
value
maximum
value
unit remarks
break option breaker
model
3 1 3 - 1 =roelvink1
2 =baldock
3 =roelvink2
4 =
roelvink_daly
gamma breaker parameter
in Baldock or
Roelvink
formulation
0.55 0.4 0.9 -
gamma2 end of breaking 0.3 0.0 0.5 break 4 only
gammax maximum allowed
waveheight over
waterdepth
2 0.4 5 this cuts off the
waveheight
numerically.
alpha wave dissipation
coefficient
1 0.5 2 -
n power in roelvink
dissipation model
10 5 20 -
delta Fraction of wave
height to add to
water depth
0 0 1 -
fw Bed friction factor 0 0 1 - on wave action
equation only!
unrelated to
bed friction in
flow equation
breakerdelay 1 0 1 -
5.6.3 Roller model
Using the roller model will give a shoreward shift in wave-induced setup, return flow and
longshore current. This shift becomes greater for lower beta values.
keyword description default minimum maximum unit remarks
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 6
value value value
roller option roller model 1 0 1 - keyword not
implemented
yet. Roller
model is
turned on by
default
beta breaker slope
coefficient in roller
model
0.15 0.05 0.3 -
5.7 Wave boundary conditions
5.7.1 User input
The user input options include various options for wave boundary conditions in params.txt.
The meaning of each is summarised below:
instat abbreviated
name
description
0 stat stationary wave boundary condition (sea state)
1 bichrom bichromatic (two wave component) waves
2 ts_1 first-order timeseries of waves (generated outside XBeach)
3 ts_2 second-order timeseries of waves (generated outside
XBeach)
4 jons wave groups generated using a parametric (J onswap)
spectrum
5 swan wave groups generated using a SWAN 2D output file
6 vardens wave groups generated using a formatted file
7 reuse reuse of wave conditions
8 nonh boundary conditions for nonhydrostatic option
9 off no wave boundary condition
40 stat_table a sequence of stationary conditions (sea states)
41 jons_table a sequence of time-varying wave groups
.
The steps that should be taken for an XBeach simulation using the new wave boundary
conditions are explained in the following sections. The reader is advised to consult Figure
5.1 to determine which sections are relevant to the simulation.
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 7
Figure 5.1 Reading guide
5.7.2 Instat = 0 to 3. wave boundary condition parameters for non-
spectral input
keyword description default
value
minimum
value
maximum
value
unit remarks
dir0 mean wave direction 270 180 360 deg instat=0-3
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 8
(Nautical convention) only
Hrms rms wave height 1 0 10 m instat=0-3
only
m power in cos^m
directional distribution
10 2 128 - instat=0-3
only
Tm01 spectral period 10 1 20 s instat=0-3
only
Trep alternative keyword for
representative period
10 1 20 s instat=0-3
only;
overrules
value for
Tm01
Tlong long wave / wave group
period
80 20 300 s instat = 1
only
taper time to spin up wave
boundary conditions
100 1e-6 1000 s not for
stationary
waves
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 9
5.7.3 Instat = 4, spectral parameter input
The new XBeach module allows the user to provide J ONSWAP parameters from which
XBeach computes a spectrum. To make use of this option, the user must specify instat = 4
inparams.txt. XBeach will then attempt to read J ONSWAP parameters from a separate file
specified by bcfile = in params.txt. The user must also state in params.txt the required
record length for the boundary condition file and the boundary condition file time step
(keywords rt = and dtbc = respectively). If the record length (rt) is less than the total
simulation time, XBeach will reuse the boundary condition file until the simulation is
completed. The boundary condition file time step should be small enough to accurately
represent the bound long wave, but need not be as small as the time step used in XBeach,
seeExplanation of input/output.
keywor
d
description defaul
t value
minimu
m value
maximu
m value
uni
t
remarks
bcfile input file for
spectral
computations
- - - file, with contents
specified below
rt record length 3600 1200 7200 s
dtbc file time step 0.5 0.1 2 s
fcutoff low freq
cutoff
frequency for
boundary
conditions
0 0 40 Hz for instat=4,5,6
sprdthr threshold
variance
density above
which spec
dens are read
in
0.08 0 1 - this is to reduce the
spectral width
nspr set directional
spreading
long waves
0 0 1 - 1 =bin all incoming
long wave directions
(instat 4+) in the
centres of the short
wave directional grid
cells
trepfac Compute
mean wave
period over
energy band:
0.01 0 1 - par%trepfac*maxval(Sf
) for instat 4,5,6;
converges to Tm01 for
trepfac =0.0
random random
generator
1 0 1 Random seed on (1) or
off (0)
oldwbc old
waveboundar
y condition
0 0 1 - this value should stay at
oldwbc=0.
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 1 0
The contents of the file specified by bcfile = in params.txt is a list of keyword = value
combinations which determine the J ONSWAP spectrum. These keywords are:
Keyword Type Description Default Minimum Maximum
Hm0 = real H
m0
of the wave spectrum,
significant wave height [m]
0.0 0.0 5.0
fp = real Peak frequency of the wave
spectrum [s
-1
]
0.08 0.0625 0.4
gammajsp =

real Peak enhancement factor in

the J ONSWAP expression
[-]
3.3 1.0 5.0
s = real Directional spreading
coefficient, cosine law [-]
10. 1.0 1000.
mainang = real Main wave angle (in
nautical terms) []
270. 180. 360.
fnyq = real Highest frequency used to
create J ONSWAP spectrum
[s
-1
]
0.3 0.2 1.0
dfj = real Step size frequency used to
create J ONSWAP spectrum
[s
-1
]
fnyq/200 fnyq/1000 fnyq/20
All variables are optional. If no value is given, the default value is used. It is advised not to
specifydfj and allow XBeach to calculate the default value.
A typical input file contains the following:
Hm0 = 0. 8
f p = 0. 125
mai nang = 285.
gammaj sp = 3. 3
s = 10.
f nyq = 0. 3
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 1 1
5.7.4 Instat = 5, SWAN spectrum input
XBeach has been programmed to read standard SWAN 2D variance density or energy
density output files (.sp2 files), as specified in the SWAN v40.51 manual.
To make use of this option, the user must specify instat = 5 in params.txt. XBeach will
then read the SWAN output spectrum from a separate file specified by bcfile = in
params.txt. The user must also state in params.txt the required record length for the
boundary condition file and the boundary condition file time step (keywords rt = and
dtbc = respectively). If the record length (rt) is less than the total simulation time, XBeach
will reuse the boundary condition file until the simulation is completed. The boundary
condition file time step should be small enough to accurately represent the bound long
wave, but need not be as small as the time step used in XBeach, see Explanation of
input/output. For definitions, see above.
XBeach assumes the output of the SWAN file is in nautical terms. If the file is in Cartesian
angles, the user must specify the angle in degrees to rotate the x-axis in SWAN to the x-axis
in XBeach (in Cartesian terms). This value is specified in params.txt after the keyword
dthetaS_XB =.
keyword description default
value
minimum
value
maximum
value
unit remarks
dthetaS_XB conversion
of angles
between
SWAN and
XBeach
0 -360 360 deg If SWAN input is
not in nautical
degrees,
dthetaS_XB is the
angle from SWAN
x-axis to XBeach
x-axis in cathesian
degrees
An example of a SWAN 2D output file is given below:
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 1 2
SWAN 1 Swan st andar d spect r al f i l e, ver si on
$ Dat a pr oduced by SWAN ver si on 40. 51
$ Pr oj ect : ' pr oj name' ; r un number : ' r unnum'
LOCATI ONS l ocat i ons i n x- y- space
1 number of l ocat i ons
22222. 22 0. 00
RFREQ r el at i ve f r equenci es i n Hz
23 number of f r equenci es
0. 0545
0. 0622
0. 0710
0. 0810
0. 0924
0. 1055
0. 1204
0. 1375
0. 1569
0. 1791
0. 2045
0. 2334
0. 2664
0. 3040
0. 3470
0. 3961
0. 4522
0. 5161
0. 5891
0. 6724
0. 7675
0. 8761
1. 0000
CDI R spect r al Car t esi an di r ect i ons i n degr
12 number of di r ect i ons
30. 0000
60. 0000
90. 0000
120. 0000
150. 0000
180. 0000
210. 0000
240. 0000
270. 0000
300. 0000
330. 0000
360. 0000
QUANT
1 number of quant i t i es i n t abl e
VaDens var i ance densi t i es i n m2/ Hz/ degr
m2/ Hz/ degr uni t
- 0. 9900E+02 except i on val ue
FACTOR
0. 675611E- 06
51 242 574 956 1288 1482 1481 1286 957 579 244 51
129 610 1443 2402 3238 3725 3724 3234 2406 1454 613 128
273 1287 3054 5084 6846 7872 7869 6837 5091 3076 1295 271
665 3152 7463 12402 16712 19229 19221 16690 12419 7518 3172 662
1302 6159 14608 24275 32688 37618 37603 32644 24309 14716 6198 1296
2328 10989 26020 43341 58358 67109 67080 58281 43401 26213 11058 2317
3365 15922 37712 62733 84492 97150 97110 84380 62820 37991 16021 3349
3426 16230 38440 63939 86109 99010 98969 85995 64027 38724 16331 3410
2027 9612 22730 37790 50909 58529 58505 50841 37843 22898 9672 2018
672 3178 7538 12535 16892 19440 19432 16870 12552 7594 3198 669
101 479 1135 1890 2542 2924 2923 2539 1892 1144 482 101
2 11 26 43 57 66 66 57 43 26 11 2
0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 1 3
5.7.5 Instat = 6, Use of formatted variance density spectrum
If the user has 2D spectrum information, but not in SWAN or J ONSWAP form, the user can
create a formatted spectrum file which can be read by XBeach.
To make use of this option, the user must specify instat = 6 in params.txt. XBeach will
then attempt to read the formatted output spectrum from a separate file specified by bcfile
= in params.txt. The user must also state inparams.txt the required record length for the
boundary condition file and the boundary condition file time step (keywords rt = and
dtbc = respectively). If the record length (rt) is less than the total simulation time, XBeach
will reuse the boundary condition file until the simulation is completed. The boundary
condition file time step should be small enough to accurately represent the bound long
wave, but need not be as small as the time step used in XBeach, see Explanation of
input/output.
The contents of the file specified by bcfile = inparams.txt must follow a specified format.
The information should be as follows:
- First line: number of frequencies (integer)
- Second line onwards: column with frequencies in Hz (each on a new line)
- Then: number of angles (integer)
- Then onwards: column with angles in degrees (each on a new line)
- Then per line: row of variance density in each direction, per frequency (i.e. first row
corresponds to variance density per direction in the first frequency band, second row of
the second frequency band, etc.).
Note that the angles in the input file must be in the calculation coordinate system of
XBeach, i.e. 0 is in the direction of the x-axis, 90 is in the direction of the y-axis. Also, the
angles must be increasing.
An example of a formatted variance density file is given below:
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 1 4
15
0. 0418
0. 0477
0. 0545
0. 0622
0. 0710
0. 0810
0. 0924
0. 1055
0. 1204
0. 1375
0. 1569
0. 1791
0. 2045
0. 2334
0. 2664
13
- 180. 0000
- 150. 0000
- 120. 0000
- 90. 0000
- 60. 0000
- 30. 0000
0. 0000
30. 0000
60. 0000
90. 0000
120. 0000
150. 0000
180. 0000
0 0 0 0 0 0 0 0 0 0 0 0
51 242 574 956 1288 1482 1481 1286 957 579 244 51
129 610 1443 2402 3238 3725 3724 3234 2406 1454 613 128
273 1287 3054 5084 6846 7872 7869 6837 5091 3076 1295 271
665 3152 7463 12402 16712 19229 19221 16690 12419 7518 3172 662
1302 6159 14608 24275 32688 37618 37603 32644 24309 14716 6198 1296
2328 10989 26020 43341 58358 67109 67080 58281 43401 26213 11058 2317
3365 15922 37712 62733 84492 97150 97110 84380 62820 37991 16021 3349
3426 16230 38440 63939 86109 99010 98969 85995 64027 38724 16331 3410
2027 9612 22730 37790 50909 58529 58505 50841 37843 22898 9672 2018
672 3178 7538 12535 16892 19440 19432 16870 12552 7594 3198 669
101 479 1135 1890 2542 2924 2923 2539 1892 1144 482 101
2 11 26 43 57 66 66 57 43 26 11 2
0 0 0 1 1 1 1 1 1 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 1 5
5.7.6 Instat =7, reusing existing boundary condition files
If the user does not wish to recalculate boundary condition files or specifically wants to
reuse the boundary condition files of another XBeach simulation, it is possible to do so. In
this case the user should select instat = 7 in params.txt. No further wave boundary
condition data need be given in params.txt. Obviously, the calculation grid should remain
the same between runs, as the angles and number of grid points are embedded in the
boundary condition files.
In order to use instat 7, the user should copy ebcflist.bcf and qbcflist.bcf to the current
directory. Additionally, the user should also copy all files listed in ebcflist.bcf and
qbcflist.bcf. Generally, these files haveE_ andq_ prefixes.
5.7.7 Instat = 8, boundary conditions for non-hydrostatic model
XBeach can be ran as a non-hydrostatic model, which is essentially the nonlinear shallow
water equations with dispersion terms and without a wave-action driver. The boundary
conditions are described in a separate chapter and are activated using instat=8
5.7.8 Instat = 9, no boundary condition
This is a simple no wave action boundary condition. It still allows for a tidal record to be
specified, however through the zs0file parameter, see Chapter 5.9 below.
5.7.9 Instat = 40, sequence of stationary sea states
While instat=0 specifies one stationary sea state (stationary in the sense that there are no
wave groups), it is also possible to specify a series of seastates, each with a duration. This is
done through a file as
instat =40
bcfile =jonswap1.txt
where the name of the bcfile is free, but the structure of the contents is not. It should contain
lines with
Hm0, Tp, angle, gamma (3.3), spreading, duration (s), timestep (=0.05)
Ie
0.7 8 90 3.3 5 1000 0.05
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 1 6
0.8 7 110 3.3 5 1000 0.05
5.7.10 Instat = 41, sequence of sea states to make time-varying wave
groups
This is an extension of instat=4. With instat =41 it is possible to specify a sea state on the
basis of which wave groups are imposed on the model for a certain duration, then specify
another sea state and run wave groups again without having to stop the model.
This condition is specified as
instat =41
bcfile =jonswap1.txt
where the name of the file is free and the structure of the contents is as in instat =40.
5.7.11 Notes on generation of boundary conditions
At the start of the XBeach simulation, XBeach checks whether non-stationary varying wave
boundary conditions are to be used. If this is the case, it next checks whether the wave
spectrum of the wave boundary conditions is to change over time, or remain constant. If the
wave spectrum is to remain constant, XBeach will only read from one input file to generate
wave boundary conditions. If the wave spectrum is to vary in time, XBeach reads from
multiple files.
Whether or not the wave spectrum of the boundary conditions changes over time, the
XBeach module requires a record length during which the current wave spectral parameters
are to apply. For the duration of the record length, boundary conditions are calculated at
every boundary condition file time step. These time steps are not required to be the same as
the time steps in the XBeach main program; XBeach will interpolate where necessary. The
boundary condition time steps should therefore only be small enough to accurately describe
the incoming bound long waves. The statistical data for the generation of the wave boundary
conditions is read from user-specified files. At this stage the XBeach module can interpret
SWAN 2D variance density output files and J ONSWAP parameters.
The beginning and end of the boundary condition file is tapered by the XBeach module.
This is done to ensure smooth transitions from one boundary condition file to the next.
The combination of a large record length and a small time step lead to large demands on the
system memory. If the memory requirement is too large, the user must choose to either
enlarge the boundary condition time step, or to reduce the record length. In case of the latter,
several boundary condition files can be generated and read sequentially. It is unwise
however to reduce the record length too much, as then the transitions between the boundary
condition files start to play an important role.
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 1 7
Every time the XBeach wave boundary condition module is run, it outputs data to the local
directory. Metadata about the wave boundary conditions are stored in list files: ebcflist.bcf
andqbcflist.bcf. The main XBeach program uses the list files to know how and when to read
and generate boundary condition files. The actual incoming short-wave energy and long-
wave mass flux data is stored in other files. These files have E_ andq_ prefixes. The main
XBeach program uses these files for the actual forcing along the offshore edge.
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 1 8
5.8 Flow input
The bed friction is influenced by the dimensionless friction coefficient cf or the dimensional
C value. These values are uniform and stationary.
The horizontal viscosity is first composed by adding an overall background viscosity nuh
and a viscosity depending on the roller dissipation, tuned by nuhfac. In the alongshore
direction the viscosity may be multiplied by a factor nuhv to account for additional
advective mixing.
keyword description default
value
minimum
value
maximum
value
unit remarks
cf flow friction
coefficient
0.003 0 0.1 -
C Chezy
coefficient
sqrt(g/cf) 20 100 m
1/2
/s alternative
for cf,
overrules cf
if set
nuh horizontal
background
viscosity
0.15 0 1 m2/s
nuhfac viscosity
coefficient for
roller induced
turbulent
horizontal
viscosity
0 0 1 -
nuhv additional shear
dispersion factor
1 1 20 - svendsen
and putrevu
(1994)
smag Smagorinski 1 0 1
lat latitude 0 0 90 deg N
omega angular velocity
of earth
1/24 0 1 1/hour
5.9 Flow boundary conditions
Flow boundary conditions need to be specified on all sides of the domain. We will
differentiate between the offshore, lateral and landward boundaries.
The carspan keyword is useful when you want to specify timeseries of free long waves
incident on the sea boundary. In that case use the file bc\gen.ezs, specify the time in the first
column, the long wave elevation in the second column and zero wave energy in the third
column.
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 1 9
keyword description default
value
minimum
value
maximum
value
unit remarks
front seaward boundary
condition
1 0 4 - see table
below
ARC active reflection
compensation at
seaward boundary
1 0 1 - 0 =off, 1 =
on
order order of wave
steering at seaward
boundary
2 1 2 - 1 =first
order , 2 =
second order
carspan free long wave
input
0 0 1 - 0 =use cg
(default); 1 =
use sqrt(gh)
ininstat = 3
for c&g tests
back bayside boundary
condition
2 0 3 - see table
below
left left lateral
boundary
condition
0 0 1 - 0 =
neumann,
1 =wall
right right lateral
boundary
condition
0 0 1 - 0 =
neumann, 1
=wall
epsi weighting factor
between steady
flow and particle
velocity
0 0 1 - obsolete
front abbreviated
name
description
0 abs1d absorbing-generating (weakly-reflective) boundary in 1D
1 abs2d same, in 2D (default setting)
2 wall no flux wall
3 wlevel water level specification (from file)
4 nonh_1d boundary condition for nonhydrostatic option
back abbreviated
name
description
0 wall no flux wall
1 abs1d absorbing-generating (weakly-reflective) boundary in 1D
2 abs2d same, in 2D (default setting)
3 wlevel water level specification (from file)
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 2 0
5.9.1 Time-varying tide/surge
XBeach can take in up to four time-vary tidal signals to be applied to the four boundaries. A
time-varying water level signal is read into XBeach by readtide.f90, which is called from
XBeach.f90. readtide opens the specified file in par%zs0file, and uses par%tideloc (number
of water level signal locations to be read in) and par%tidelen (length of input signals) to
format the file reading statements. The par%zs0file must follow the format with the first
column being time and each subsequent column containing the water level signals. As a
note, the input signal will be interpolated to the local time step of the simulation; therefore
the signals only need to be long enough and temporally-fine enough to resolve the water
level phenomenon of interest (i.e. tide variations, surge event). The above parameters and
file name are specified inparams.txt. If par%tideloc is equal to zero, readtide is not utilized
and a uniform water level is applied according to par%zs0.
There are now four options for handling the tidal and/or surge contribution to the
boundaries:
- Uniform water level
- One time-varying water level signal
- Two time-varying water level signals, which requires point of application indication.
- Four time-varying water level signals
The uniform water level is applied according to par%zs0, as stated above. For par%tideloc
equal to 1, one water level signal is read and applied to the offshore boundary and par%zs0
is applied to the land boundary. If par%tideloc is equal to 2, two water level signals are read
by readtide. par%paulrevere is utilized to indicate the locations of application for the two
signals. For par%paulrevere equal to 0, one tidal record is applied to both sea corners and
one tidal record to both land corners, whereas if par%paulrevere is equal to 1, the first tidal
record is applied to the (x=1,y=1) sea corner and the second tidal record (third column) to
the (x=1,y=N) sea corner. For par%tideloc equal to 4, four tidal records are read. These four
time-varying signals need to be in ordered columns going clockwise around the domain,
with first column signal being applied to (x=1,y=1). Therefore, the columns must follow the
order of: 1. (x=1,y=1), 2. (x=1,y=N), 3. (x=N,y=N), 4. (x=N,y=1).
The input signal(s) is(are) interpolated spatially along the boundaries for par%tideloc is
equal to 2 (if par%paulrevere=2, and if par%paulrevere=1 and no land mass blocks the flow
between the land and sea boundaries) and higher. The spatial interpolation occurs initially in
wave_init, and subsequently at every time step in flow_bc. The input signal(s) is(are) also
interpolated to the local time step inflow_bc.
keyword description default
value
minimum
value
maximum
value
unit remarks
zs01 initial water level 0 -5 5 m
tideloc number of input tidal
time series
0 0 4 -
paulrevere option of sea/sea
corner or sea/land
corner specification
0 0 1 -
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 2 1
5.10 Wind
keyword description default
value
minimum
value
maximum
value
unit remarks
rhoa air density 1.25 1 2 kg/m3
Cd wind drag
coefficient
0.002 0.0001 0.01 -
windv wind velocity 0 0 200 m/s
windth wind direction
(nautical
convention)
270 -360 360 deg .
The implementation of the wind stress in the momentum balance in XBeach is based upon
the work of Ruessink et al. (2001). The wind field is specified by the following parameters
inparams.txt:
- par%windth, the wind direction (Nautical coordinates)
- par%windv, the wind velocity specified in m/s
- par%rhoa, the air density (default 1.25 kg/m
3
)
- par%Cd, the air drag coefficient (default 0.002, but should be looked into for
simulations with wind velocities exceeding 50 m/s)
The input wind direction is converted to radians and Cartesian coordinates prior to use in the
momentum balance. The wind contribution is computed as an additional term in the x and y
momentum balances.
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 2 1
5.11 Limiters
Especially in very shallow water some processes need to be limited to avoid unrealistic
behaviour. The parameters below can be adjusted by the user. Reducinggammax will reduce
wave heights in very shallow water, probably 2 is a reasonable value. hmin prevents very
strong return flows or high concentrations. eps determines whether points are dry or wet and
can be taken quite small. hwci limits the computation of wave-current interaction in very
shallow water where the procedure may not converge.
keyword description default
value
minimum
value
maximum
value
unit remarks
gammax maximumratio
Hrms/hh
2 0.4 5 -
hmin threshold water
depth for
concentration and
return flow
0.05 0.001 1 m
eps threshold depth for
drying and flooding
0.005 0.001 0.1 m
umin threshold velocity
upwind scheme
0.0 0.0 5 m/s
hwci depth below which
wci is not applied
0.1 0.0001 1 m
5.12 Sediment transport
keyword description default
value
minimum
value
maximum
value
unit remarks
waveform option for
waveshape model
1 1 1 - 1=ruessin
k_vanrijn
2=vanthiel

form equilibrium sed.

conc. formulation
1 1 2 - 1 =
soulsby_
vanrijn,
2 =
vanthiel_v
anrijn
thetanum option to switch
between central
and upwind
scheme
1 0.5 1 [-] 1=upwind
0.5=central
z0 zero flow velocity
level in Soulsby
van Rijn (1997)
sed.conc.
0.006 0.0001 0.05 m
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 2 2
expression
facsl bed slope factor 0 0 1.6
BRfac Calibration factor
surface slope
1 0 1
turb Equilibrium
sediment
concentration
computation
option
2 - - - 0=none,
1=wave_a
veraged
2=bore_av
eraged
rhos density of
sediment (no
pores)
2650 2400 2800 kg/
m3
por porosity 0.4 0.3 0.5 -
smax max shields value
for overwash
-1 -1 3 [-]
lws longwave stirring 1 0 1 1 =on; 0 =
off
sws shortwave stirring 1 0 1 1 =on; 0 =
off
tsfac max value for fall
velocity
0.1 0.01 1 [-]
facua asymmetry
transport
0 0 1 -
rfb 1 0 1 - ONLY IF
FORM=2
If rfb =1
then
maximum
wave
surface
slope is fed
back in
roller
energy
balance;
else rfb =
par%Beta
5.13 Multiple sediment fractions and hard layers
Keywords related to hard layers and sediment classes:
struct option for hard
structures
0 0 1 0 =no
revetment, 1
=multiple
sediment
classes
ne_layer
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 2 3
ngd number of
sediment classes
1 1 20 -
nd number of
sediment class
layers
3 3 1000 -
dzg1
dzg2
dzg3
thickness of
sediment class
layers
0.1 0.01 1 -
D50 uniform D50
sediment diameter
0.0002 - - m
D90 uniform D90
sediment diameter
0.0003 - - m
sedcal factor on sediment
transport
1 - - - advanced
option, leave
at sedcal=1
ucrcal factor on critical
velocity for
erosion
1 - - - advanced
option, leave
at ucrcal=1
NEEDS MORE EXPLANATION FOR MULTIPLE SEDIMENT FRACTIONS
5.14 Morphological updating and avalanching
keyword description default
value
minimum
value
maximum
value
unit remarks
morfac morphological
factor
0 0 1000 -
morstart start time of
morphological
updates
120 0 10000 s
morfacopt type of morfac
updating
1 0 1 1 =on; 0
=off
sourcesink morphological
updating based on
sediment transport
gradients or source
sink
0 0 1 - 0 =
gradients
1=source
sink
wetslp critical avalanching
slope under water
0.3 0.1 1 -
dryslp critical avalanching
slope above water
1 0.1 2 -
hswitch water depth at
interface from
wetslp to dryslp
0.1 0.01 1
dzmax 0.05 0.00 1
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 2 4
5.15 Groundwater
In order to use the groundwater module in XBeach, a number of parameters have been
added to params.txt. The keywords, their type, default value and description are given
below.
keyword description default
value
minimum
value
maximum
value
unit remarks
gwflow switch
parameter
0 0 1 - Turns the
groundwater
module on
(1) or off
(2)
kx The Darcy
permeability
coefficient of
the aquifer in
x-direction
1E-4 1E-5 1E-2 ms
-
1
ky The Darcy
permeability
coefficient of
the aquifer in
y-direction
kx 1E-5 1E-2 ms
-
1
kz The Darcy
permeability
coefficient of
the aquifer in
z-direction
kx 1E-5 1E-2 ms
-
1
dwetlayer Thickness of
the
interaction
layer
wetlayer
d
0.2d0 0.01 1 m
aquiferbot The bed level
of the aquifer
is set to the
value of
aquiferbot
across the
entire domain
minval(zb)-
3.d0
-100 100 m
aquiferbotfile The initial
bed level of
the aquifer
from file
- - This
keyword
overrides
the
aquiferbot
keyword
gw0 The initial 0.d0 -5.0 5.0 m uniform
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 2 5
groundwater
head
over domain
gw0file The initial
groundwater
head from
file
- - This
keyword
overrides
thegw0
keyword
Example
An example is given in which the groundwater module is used and the vertical permeability
coefficient in the vertical is set differently to that in the horizontal. The initial bed level of
the aquifer is read from a file and the initial groundwater head is set to a uniform value.
Other parameters are kept at their default values.
5.16 Drifters
keyword description default
value
minimum
value
maximum
value
unit remarks
ndrifter number of drifters 0 0 50 -
drifterfile name of drifter file - - - -
5.17 Discharge boundaries (river input)
keyword description default
value
minimum
value
maximum
value
unit remarks
disch_loc_file discharge
locations
- - - - see
description
- - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Gr oundwat er opt i ons
gwf l ow = 1
kx = 0. 00015
kz = 0. 0005
aqui f er bot f i l e = zbaqui f er . t xt
gw0 = 0. 5
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 2 6
file below
disch_timeseries_file discharge
time series
fiel
- - - see
description
below
The contents of the discharge locations file are a number of records equal to the number of
discharge boundary sections, with for each record four numbers:
X_begin Y_begin X_end Y_end
These world coordinates must be chosen such that they are close to the desired boundary
and enclose the cell centers of the cells that must be part of the boundary section.
In the discharge time series file, the first column is time in seconds; the next columns give
the total discharge time series per section.
5.18 Nonhydrostatic
keyword description default
value
minimum
value
maximum
value
unit remarks
secorder second order
corrections to
advection/non-
linear terms based
on mcCormack
scheme
0 0 1 -
solver_maxit Maximum number
of iterations in the
linear SIP solver
30 1 1000
solver_acc accuracy with
respect to the right-
hand side used
0.005 0.00001 0.1
solver_urelax Underrelaxation
parameter
0.92 0.5 0.99
solver Solver used to
solve the linear
system, 1=SIP,
2=TRIDIAG (only
for 1d)
1 0 2
kdmin Minimum value of
kd (pi/dx >minkd )
0.0 0.0 0.05
dispc Coefficient in front
of the vertical
pressure gradient
1 0.1 2
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 2 7
Topt Absolute period to
optimize
coefficient
10 1 20
5.19 Data types
XBeach allows the user to choose between three types of output data: regular spatial output,
time-averaged spatial output and point output. Of these three types, only the regular spatial
output existed in previous versions of XBeach.
keyword description default
value
minimum
value
maximum
value
unit remarks
nglobalvar number of global
output variables
-1 -1 20 - see below
nmeanvar number of mean
variables
0 0 15 -
npoints number of point
output locations
0 0 50 -
nrugauge number of runup
gauges
0 0 50 -
If nglovalvar is set to -1 then the standard output variables consist of 'H ', 'zs ', 'zs0 ',
'zb ', 'hh ', 'u ', 'v ', 'ue ', 've ', 'urms ', 'Fx ', 'Fy ', 'ccg ', 'ceqsg', 'ceqbg', 'Susg ',
'Svsg ', 'E ', 'R ', 'D ', 'DR '
5.19.1 Regular spatial output
Regular spatial output describes the instantaneous state of variables across the entire model
domain at various points in time. In previous versions of XBeach, the program
automatically generated regular spatial output of 20 variables. In the current version of
XBeach, the user may determine how many and which variables are output.
To make use of this option the user must specify the number of output variables required
using thenglobalvar= keyword inparams.txt. The value of nglobalvar is always integer.
This line is immediately followed by the keywords of the required variables, one per line.
Refer to theVariable keywords section for a list of keywords. An example of the declaration
of regular spatial output is given below.
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 2 8
Example of the declaration of regular spatial output variables
Output files generated using this option are given the namekeyword.dat, for examplezs.dat.
The only exception is that files containing information about the wave height of the short
waves are calledhrms.dat instead of H.dat to maintain backward compatibility.
5.19.2 Time-averaged spatial output
This output describes the time-averaged state of variables across the entire model domain at
various points in time. The period used to average the variables is selected by the user and is
described in the section Choosing output times. The user may determine how many and
which variables are output. To make use of this option the user must specify the number of
output variables required using the nmeanvar= keyword in params.txt. The value of
nmeanvar is always integer. This line is immediately followed by the keywords of the
required variables, one on each subsequent line. Refer to theVariable keywords section for a
list of keywords. This statement block may be made before or after the regular spatial output
block. An example of the declaration of time-averaged spatial output is given below.
5.20 Selecting output data types
XBeach v12 allows the user to choose between three types of output data: regular spatial
output, time-averaged spatial output and point output. Of these three types, only the regular
spatial output existed in previous versions of XBeach.
5.20.1 Regular spatial output
Regular spatial output describes the instantaneous state of variables across the entire model
domain at various points in time. In previous versions of XBeach, the program
automatically generated regular spatial output of 20 variables. In the current version of
XBeach, the user may determine how many and which variables are output.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Out put opt i ons
nmeanvar = 2
u
v
- - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Out put opt i ons
ngl obal var = 5
zb
zs
u
v
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 2 9
To make use of this option the user must specify the number of output variables required
using thenglobalvar= keyword inparams.txt. The value of nglobalvar is always integer.
This line is immediately followed by the keywords of the required variables, one per line.
Refer to theVariable keywords section for a list of keywords. An example of the declaration
of regular spatial output is given below.
Example of the declaration of regular spatial output variables
Output files generated using this option are given the namekeyword.dat, for examplezs.dat.
The only exception is that files containing information about the wave height of the short
waves are calledhrms.dat instead of H.dat to maintain backward compatibility.
5.20.2 Time-averaged spatial output
This output describes the time-averaged state of variables across the entire model domain at
various points in time. The period used to average the variables is selected by the user and is
described in the section Choosing output times. The user may determine how many and
which variables are output. To make use of this option the user must specify the number of
output variables required using the nmeanvar= keyword in params.txt. The value of
nmeanvar is always integer. This line is immediately followed by the keywords of the
required variables, one on each subsequent line. Refer to theVariable keywords section for a
list of keywords. This statement block may be made before or after the regular spatial output
block. An example of the declaration of time-averaged spatial output is given below.
Example of the declaration of time-averaged spatial output variables
Output files generated using this option are given the namekeyword_mean.dat, for example
zs_mean.dat.
5.20.3 Point output
Point output allows the user to select one or more locations for which a time series of data is
provided. This output type is divided into two classes: fixed point output and runup gauge
output.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Out put opt i ons
nmeanvar = 2
u
v
- - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Out put opt i ons
ngl obal var = 5
H
zb
zs
u
v
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 3 0
Fixed point output
This output describes a time-series of a number of variables at one point in the model
domain. To make use of this option, the user must specify the number of output locations
using the npoints= keyword in params.txt. The value of npoints is always integer. This
statement is always immediately followed by one line per output location describing the
location coordinates, the number of output variables required at that location and the output
variable keywords. The location coordinates are given separately as x-coordinate and y-
coordinate, in world coordinates. XBeach will link the output location to the nearest
computational point. The number of required variables is integer. Variables are stated using
their respective variable keywords, see the Variable keywords section. Each keyword,
including the last keyword, must be followed by a number symbol (#).This statement block
may be made before or after any other output option block. All variables will be outputted
for all points.
An example is given below in which two output locations are defined. The first point is
located on the offshore boundary (x-coordinate =0.0) and somewhere in the middle of the
model domain in y-direction (y-coordinate =800.0). This location has two output variables,
u and v. The second point is located on the lateral boundary (y-coordinate =1600.0) and
somewhere in the middle of the domain in x-direction (x-coordinate =2000.0). This location
has four output variables, H, zs, zb andD.
Runup gauge output
This output describes a time-series of a number of variables at the waterline. In this case
XBeach scans in a x-directional transect defined by the user to find the last wet point before
shore (defined as the first dry point in a transect). Output information is recorded for this
point. This is particularly useful to keep track of runup levels in cross shore transects.
The definition of runup gauges is similar to the definition of fixed point output. To specify
runup gauges, the user should specify the number of runup gauges using thenrugauge=
keyword in params.txt. The value of nrugauge is always integer. This statement is always
immediately followed by one line per output location describing the location coordinates,
the number of output variables required at that location and the output variable keywords.
The location coordinates are given separately as x-coordinate and y-coordinate, in world
coordinates. XBeach will link the output location to the nearest computational grid row. The
number of required variables is integer. Variables are stated using their respective variable
keywords, see the Variable keywords section. Each keyword, including the last keyword,
must be followed by a number symbol (#).This statement block may be made before or after
any other output option block. All variables will be used for each point.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Out put opt i ons
npoi nt s = 2
0. 800. 6 H#zs#zb#D#u#v#
2000. 1600. 6 H#zs#zb#D#u#v#
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 3 1
Point output file structure
All data corresponding to point locations will be stored in files called pointxxx.dat in the
case of fixed point data andrugauxxx.dat in the case of runup gauge data. In the filenames
xxx represents a number between 001 and 999, corresponding to the order in which the
points are declared inparams.txt. The data files contain one row per output time step. The
first position on each row is the time at which the output is given. The subsequent positions
in the row are the values of the variables at the given point at that given moment. The order
of the variables is the same as the order in which they are defined for that point in
params.txt.
5.20.4 Cross-sections
NOT YET IMPLEMENTED
5.20.5 Default values
If no output information is given with respect to the numbers and types of output, XBeach
assumes the user requires only the original 20 regular spatial output types. The default
number of time-averaged output variables and point output locations is zero.
5.21 Choosing output times
The user may determine the output times for regular spatial output variables, time averaged
spatial variables and point location variables individually. Runup gauge output and fixed
point output are given at the same points in time. For all three types of output the user may
choose to either state a fixed interval time at which output is given, or supply an additional
file containing times at which output should be given.
5.21.1 Output at fixed intervals
In this case the user should define a point in time (in seconds) after the start of the
simulation at which the first output is generated. The user can do this by using thetstart=
keyword in params.txt. All output that is being generated at fixed intervals uses tstart as
their base. The interval for regular global output is given by the tintg= keyword in
params.txt. The keywords for the interval of time-averaged global output and point output
aretintm= and tintp= respectively. Note that tintg, tintm and tintp supersede the older
tint parameter. An example of the definition of fixed intervals is given below.
Example of the declaration of fixed interval output times
- - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Out put opt i ons
t st ar t = 100.
t i nt g = 100.
t i nt p = 2.
t i nt m= 3600.
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 3 2
In the case of regular global output and point output, the first output is given at tstart. In the
case of time-averaged global variables, the first output is given at tstart+tintm. This output
represents the average condition over the interval betweentstart andtstart+tintm.
5.21.2 Output times defined by external file
The user is given the option to have output at a set of points in time that are not separated by
regular intervals. In this case the user must supply an additional file for each output type
(regular, time-averaged or point). The user specifies the name of the output time series file
for regular spatial output using the tsglobal= keyword in params.txt. The keywords for
time series files for time-averaged output and point output are tsmean= and tspoint=
respectively. All time series files must contain on the first line the number of output times
(integer) followed by every output time (in seconds after the start of the simulation) on a
new line. An example of the definition of varying output time intervals is given below. An
example of the contents of one of the time series definition files is given in the second
example below.
Example of the declaration of varying output time intervals
Example of an output time series definition file (in the previous example this could be timeseries1.txt)
In the case of regular global output and point output, the first output is given at the first
stated point in time. In the case of time-averaged global variables, the first output is given at
the second stated point in time. This output represents the average condition over the
interval between first and second stated point in time. Subsequent averaging is made over
every interval.
18
0. 05
0. 15
0. 2
0. 8
12. 0
12. 5
19. 124
30.
60.
90.
120.
150.
160.
170.
177.
178.
179.
180.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Out put opt i ons
t sgl obal = t i meser i es1. t xt
t spoi nt s = t i meser i es2. t xt
t smean= t i meser i es3. t xt
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 3 3
5.21.3 Combinations of fixed interval and external files
The user is allowed to define certain types of output using fixed intervals and others using
external files. The use of an external file supersedes the use of fixed intervals. Note that
tstart will only apply to output of fixed interval type. An example of mixing fixed and
varying output time intervals is given below.
Example of the declaration of fixed and varying output time intervals
5.21.4 Default values
The default value of tintg is one second. If tintp or tintm is not stated, but output is declared
(npoints, nrugauge or nmeanvar is stated larger than zero), XBeach assumes the same
output interval as tintg. XBeach v12 has been designed to be backward compatible with
older versions. Therefore if tint is defined, this replaces the default value of tintg.
5.22 NetCDF CF output
The netcdf output will be stored in the file xboutput.nc. Netcdf output is generated using the
keywordoutputformat.
An example of the layout of the netcdf file is given below. One thing to note is the different
time dimensions. This is because of the different resolutions of point output and global
output. The temporal unit can be specified in the params.txt file using the keyword tunits.
This unit does not affect calculations and is only used for output.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Out put opt i ons
out put f or mat = net cdf
t uni t s = seconds si nce 1970- 01- 01 +1
- - - - - - - - - - - - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Out put opt i ons
t st ar t = 100.
t i nt g = 100.
t spoi nt s = t i meser i es2. t xt
t i nt m= 3600.
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 3 4
5.23 Reading output files in Matlab
Various functions have been developed to help users read XBeach output in Matlab. The
following script can be used to gather information about the XBeach output before reading
the output.
net cdf xbout put {
di mensi ons:
x = 565 ;
y = 101 ;
wave_angl e = 9 ;
bed_l ayer s = 3 ;
sedi ment _cl asses = 1 ;
i nout = 2 ;
gl obal t i me = 2 ;
t i det i me = 435 ;
t i decor ner s = 2 ;
wi ndt i me = 2 ;
var i abl es:
doubl e x( x) ;
x: uni t s = " m" ;
x: l ong_name = " l ocal x coor di nat e" ;
doubl e y( y) ;
y: uni t s = " m" ;
y: l ong_name = " l ocal y coor di nat e" ;
doubl e gl obal t i me( gl obal t i me) ;
gl obal t i me: uni t s = " s" ;
doubl e H( gl obal t i me, y, x) ;
H: uni t s = " m" ;
H: l ong_name = " wave hei ght " ;
doubl e zs( gl obal t i me, y, x) ;
zs: uni t s = " m" ;
zs: l ong_name = " wat er l evel " ;
doubl e zb( gl obal t i me, y, x) ;
zb: uni t s = " m" ;
zb: l ong_name = " bed l evel " ;
doubl e ue( gl obal t i me, y, x) ;
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 3 5
Matlab function to extract output metadata
f unct i on XBdi ms=get di mensi ons
%XBdi ms=get di mensi ons
%
%Must st ar t i n XBeach out put di r ect or y
%Out put st r uct ur e XBdi ms
%XBdi ms. nt = number of r egul ar spat i al out put t i mest eps
%XBdi ms. nx = number of gr i d cel l s i n x- di r eci on
%XBdi ms. ny = number of gr i d cel l s i n y- di r eci on
%XBdi ms. ngd = number of sedi ment cl asses
%XBdi ms. nd = number of sedi ment cl ass l ayer s
%XBdi ms. nt p = number of poi nt out put t i mest eps
%XBdi ms. nt m = number of t i me- aver age out put t i mest eps
%XBdi ms. t sgl obal = t i mes at whi ch r egul ar out put i s gi ven
%XBdi ms. t spoi nt s = t i mes at whi ch poi nt out put i s gi ven
%XBdi ms. t smean = t i mes at whi ch t i me- aver aged out put i s gi ven
%XBdi ms. x = x- coor di nat es gr i d
%XBdi ms. y = y- coor di nat es gr i d
%
%Cr eat ed 19- 06- 2008 : XBeach- gr oup Del f t
f i d=f open( ' di ms. dat ' , ' r ' ) ;
XBdi ms. nt =f r ead( f i d, [ 1] , ' doubl e' ) ;
XBdi ms. nx=f r ead( f i d, [ 1] , ' doubl e' ) ;
XBdi ms. ny=f r ead( f i d, [ 1] , ' doubl e' ) ;
XBdi ms. ngd=f r ead( f i d, [ 1] , ' doubl e' ) ;
XBdi ms. nd=f r ead( f i d, [ 1] , ' doubl e' ) ;
XBdi ms. nt p=f r ead( f i d, [ 1] , ' doubl e' ) ;
XBdi ms. nt m=f r ead( f i d, [ 1] , ' doubl e' ) ;
XBdi ms. t sgl obal =f r ead( f i d, [ XBdi ms. nt ] , ' doubl e' ) ;
XBdi ms. t spoi nt s=f r ead( f i d, [ XBdi ms. nt p] , ' doubl e' ) ;
XBdi ms. t smean=f r ead( f i d, [ XBdi ms. nt m] , ' doubl e' ) ;
f cl ose( f i d) ;
f i dxy=f open( ' xy. dat ' , ' r ' ) ;
XBdi ms. x=f r ead( f i dxy, [ XBdi ms. nx+1, XBdi ms. ny+1] , ' doubl e' ) ;
XBdi ms. y=f r ead( f i dxy, [ XBdi ms. nx+1, XBdi ms. ny+1] , ' doubl e' ) ;
f cl ose( f i dxy) ;
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 3 6
Regular spatial output and time-averaged output can then be read using the following
function.
Matlab function to read spatial output variables
To read point data the following function can be used.
Matlab function to read point output variables
f unct i on Poi nt dat a=r eadpoi nt ( f name, XBdi ms, nvar )
%Poi nt dat a=r eadpoi nt ( f name, XBdi ms, nvar )
%
%Out put Poi nt i s [ nt p, nvar +1] ar r ay, wher e nt p i s XBdi ms. nt p
% Fi r st col umn of Poi nt dat a i s t i me
% Second and f ur t her col umns of Poi nt dat a ar e val ues of
% var i abl es
%I nput - f name : name of dat a f i l e t o open, e. g. ' poi nt 001. dat ' or ' r ugau001. dat '
% - XBdi ms: di mensi on dat a pr ovi ded by get di mensi ons f unct i on
% - nvar : number of var i abl es out put at t hi s poi nt l ocat i on
%
%Cr eat ed 19- 06- 2008 : XBeach- gr oup Del f t
Poi nt dat a=zer os( XBdi ms. nt p, nvar +1) ;
f i d=f open( f name, ' r ' ) ;
f or i =1: XBdi ms. nt p
Poi nt dat a( i , : ) =f r ead( f i d, nvar +1, ' doubl e' ) ;
end
f cl ose( f i d) ;
f unct i on Var dat a=r eadvar ( f name, XBdi ms)
%Var =r eadvar ( f name, XBdi ms)
%
%Out put Var i s [ nx+1, ny+1, nt ] ar r ay, wher e nt i s XBdi ms. nt or XBdi ms. nt m
%I nput - f name : name of dat a f i l e t o open, e. g. ' zb. dat ' or ' u_mean. dat '
% - XBdi ms: di mensi on dat a pr ovi ded by get di mensi ons f unct i on
%
%Cr eat ed 19- 06- 2008 : XBeach- gr oup Del f t
i f ( l engt h( f name) >9 && st r cmp( f name( end- 8: end) , ' _mean. dat ' ) )
nt =XBdi ms. nt m;
el se
nt =XBdi ms. nt ;
end
f i d=f open( f name, ' r ' ) ;
Var dat a=zer os( XBdi ms. nx+1, XBdi ms. ny+1, nt ) ;
f or i =1: nt
Var dat a( : , : , i ) =f r ead( f i d, si ze( XBdi ms. x) , ' doubl e' ) ;
end
f cl ose( f i d)
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 3 7
5.24 Output variable keywords
All variables in spaceparams.tmpl can be defined as output.
real*8 2 x s%nx+1 s%ny+1 d [m] x-coord. comp. grid (positive shoreward, perp. to coastline)
real*8 2 y s%nx+1 s%ny+1 d [m] y-coord. comp. grid
real*8 1 xz s%nx+1 d [m] x-coord. comp. grid (positive shoreward, perp. to coastline)
real*8 1 yz s%ny+1 d [m] y-coord. comp. grid
real*8 1 xu s%nx+1 d [m] x-coord. in u points
real*8 1 yv s%ny+1 d [m] y-coord. in v points
real*8 2 xw s%nx+1 s%ny+1 d [m] world x-coordinates
real*8 2 yw s%nx+1 s%ny+1 d [m] world y-coordinates
real*8 0 dx b [m] grid size x-direction
real*8 0 dy b [m] grid size y-direction
real*8 0 xori b [m] x-origin of grid in world coordinates
real*8 0 yori b [m] y-origin of grid in world coordinates
real*8 0 alfa b [rad] (deg on input) angle of grid w.r.t. East
real*8 0 posdwn b [-] depths defined positive downwards (1) or upwards(-1)
real*8 0 thetamin b [rad] minimumangle of computational wave grid (carth. in rad)
real*8 0 thetamax b [rad] minimumangle of computational wave grid (carth. in rad)
integer 0 nx b [-] local number of grid cells x-direction
integer 0 ny b [-] local number of grid cells y-direction
real*8 2 zb s%nx+1 s%ny+1 d [m] bed level
real*8 2 zb0 s%nx+1 s%ny+1 d [m] initial bed level
real*8 1 theta s%ntheta b [rad] wave angles directional distribution w.r.t. comp. x-axis
integer 0 ntheta b [-] number of wave direction bins
real*8 0 dtheta b [rad] wave direction bin size
real*8 0 theta0 b [rad] mean incident wave angle
real*8 1 cxsth s%ntheta b [-] cos(theta)
real*8 1 sxnth s%ntheta b [-] sin(theta)
real*8 2 thetamean s%nx+1 s%ny+1 d [rad] mean wave angle
real*8 2 Fx s%nx+1 s%ny+1 d [N/m2] wave force x-direction
real*8 2 Fy s%nx+1 s%ny+1 d [N/m2] wave force y-direction
real*8 2 Sxy s%nx+1 s%ny+1 d [N/m] radiation stress
real*8 2 Syy s%nx+1 s%ny+1 d [N/m] radiation stress
real*8 2 Sxx s%nx+1 s%ny+1 d [N/m] radiation stress
real*8 2 n s%nx+1 s%ny+1 d [-] ratio group velocity/wave celerity
real*8 2 H s%nx+1 s%ny+1 d [m] wave height
real*8 3 cgx s%nx+1 s%ny+1 s%ntheta d [m/s] group velocity x-direction
real*8 3 cgy s%nx+1 s%ny+1 s%ntheta d [m/s] group velocity y-direction
real*8 3 cx s%nx+1 s%ny+1 s%ntheta d [m/s] wave celerity x-direction
real*8 3 cy s%nx+1 s%ny+1 s%ntheta d [m/s] wave celerity y-direction
real*8 3 ctheta s%nx+1 s%ny+1 s%ntheta d [rad/s] wave celerity theta-direction (refraction)
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 3 8
real*8 3 ee s%nx+1 s%ny+1 s%ntheta d [J /m2/rad] directionally distributed wave energy
real*8 3 thet s%nx+1 s%ny+1 s%ntheta d [rad] wave angles
real*8 3 costhet s%nx+1 s%ny+1 s%ntheta d [-] cos of wave angles
real*8 3 sinthet s%nx+1 s%ny+1 s%ntheta d [-] sin of wave angles
real*8 3 sigt s%nx+1 s%ny+1 s%ntheta d [rad/s] relative frequency
real*8 3 rr s%nx+1 s%ny+1 s%ntheta d [J /m2/rad] directionally distributed roller energy
real*8 2 k s%nx+1 s%ny+1 d [rad/m] wave number
real*8 2 c s%nx+1 s%ny+1 d [m/s] wave celerity
real*8 2 cg s%nx+1 s%ny+1 d [m/s] group velocity
real*8 2 sigm s%nx+1 s%ny+1 d [rad/s] mean frequency
real*8 2 wm s%nx+1 s%ny+1 d [rad/s] mean abs frequency
real*8 2 hh s%nx+1 s%ny+1 d [m] water depth
real*8 2 zs s%nx+1 s%ny+1 d [m] water level
real*8 2 zs0 s%nx+1 s%ny+1 d [m] water level due to tide alone
real*8 1 tideinpt par%tidelen b [s] input time of input tidal signal
real*8 2 tideinpz par%tidelen par%tideloc b [m] input tidal signal
real*8 1 windinpt par%windlen b [s] input time of input wind signal
real*8 1 windvel par%windlen b [m/s] input wind velocity
real*8 1 winddir par%windlen b [deg nautical] input wind direction
real*8 0 windvnow b [m/s] uniformwind velocity current time
real*8 0 winddirnow b [m/s] uniformwind direction current time
real*8 2 dzsdt s%nx+1 s%ny+1 d [m/s] rate of change water level
real*8 2 dzsdx s%nx+1 s%ny+1 d [m/s] water surface gradient in x-direction
real*8 2 dzsdy s%nx+1 s%ny+1 d [m/s] water surface gradient in y-direction
real*8 2 dzbdx s%nx+1 s%ny+1 d [-] bed level gradient in x-direction
real*8 2 dzbdy s%nx+1 s%ny+1 d [-] bed level gradient in y-direction
real*8 2 dzbdt s%nx+1 s%ny+1 d [m/s] rate of change bed level
real*8 2 uu s%nx+1 s%ny+1 d [m/s] (GLM) x-velocity in u-points
real*8 2 vv s%nx+1 s%ny+1 d [m/s] (GLM) y-velocity in v-points
real*8 2 qx s%nx+1 s%ny+1 d [m2/s] x-discharge in u-points
real*8 2 qy s%nx+1 s%ny+1 d [m2/s] y-discharge in u-points
real*8 2 sedero s%nx+1 s%ny+1 d [m] cum. sedimentation/erosion
real*8 2 dcbdx s%nx+1 s%ny+1 d [kg/m3/m] bed concentration gradient x-dir.
real*8 2 dcbdy s%nx+1 s%ny+1 d [kg/m3/m] bed concentration gradient y-dir.
real*8 2 dcsdx s%nx+1 s%ny+1 d [kg/m3/m] suspended concentration gradient x-dir.
real*8 2 dcsdy s%nx+1 s%ny+1 d [kg/m3/m] suspended concentration gradient y-dir.
real*8 2 ui s%nx+1 s%ny+1 d [m/s] incident bound wave velocity
real*8 2 E s%nx+1 s%ny+1 d [Nm/m2] wave energy
real*8 2 R s%nx+1 s%ny+1 d [Nm/m2] roller energy
real*8 2 urms s%nx+1 s%ny+1 d [m/s] orbital velocity
real*8 2 D s%nx+1 s%ny+1 d [W/m2] dissipation
real*8 2 Qb s%nx+1 s%ny+1 d [-] fraction breaking waves
real*8 2 ust s%nx+1 s%ny+1 d [m/s] Stokes drift
real*8 2 tm s%nx+1 s%ny+1 d [rad] mean wave direction
real*8 2 ueu s%nx+1 s%ny+1 d [m/s] Eulerian mean velocity x-dir.
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 3 9
real*8 2 vev s%nx+1 s%ny+1 d [m/s] Eulerian mean velocity y-dir.
real*8 2 vmagu s%nx+1 s%ny+1 d [m/s] (GLM) velocity magnitude u-points
real*8 2 vmageu s%nx+1 s%ny+1 d [m/s] (GLM) velocity magnitude u-points
real*8 2 vmagv s%nx+1 s%ny+1 d [m/s] (GLM) velocity magnitude v-points
real*8 2 vmagev s%nx+1 s%ny+1 d [m/s] (GLM) velocity magnitude v-points
real*8 2 u s%nx+1 s%ny+1 d [m/s] (GLM) x-velocity cell centre (for output)
real*8 2 v s%nx+1 s%ny+1 d [m/s] (GLM) y-velocity cell centre (for output)
real*8 2 ue s%nx+1 s%ny+1 d [m/s] Eulerian mean x-velocity cell centre (for output)
real*8 2 ve s%nx+1 s%ny+1 d [m/s] Eulerian mean y-velocity cell centre (for output)
real*8 2 hold s%nx+1 s%ny+1 d [m] water depth previous time step
integer 2 wetu s%nx+1 s%ny+1 d [-] mask wet/dry u-points
integer 2 wetv s%nx+1 s%ny+1 d [-] mask wet/dry v-points
integer 2 wetz s%nx+1 s%ny+1 d [-] mask wet/dry eta-points
real*8 2 hu s%nx+1 s%ny+1 d [m] water depth in u-points
real*8 2 hv s%nx+1 s%ny+1 d [m] water depth in v-points
real*8 2 hum s%nx+1 s%ny+1 d [m] water depth in u-points
real*8 2 hvm s%nx+1 s%ny+1 d [m] water depth in v-points
real*8 2 vmag s%nx+1 s%ny+1 d [m/s] velocity magnitude in cell centre
real*8 3 ccg s%nx+1 s%ny+1 par%ngd d [m3/m3] depth-averaged suspended concentration for each sediment
fraction
real*8 3 ccbg s%nx+1 s%ny+1 par%ngd d [m3/m3] depth-averaged bed concentration for each sediment fraction
real*8 2 uwf s%nx+1 s%ny+1 d [m/s] x-comp. Stokes drift
real*8 2 vwf s%nx+1 s%ny+1 d [m/s] y-comp. Stokes drift
real*8 2 ustr s%nx+1 s%ny+1 d [m/s] return flow due to roller
real*8 2 usd s%nx+1 s%ny+1 d [m/s] return flow due to roller after breaker delay
real*8 1 bi s%ny+1 d [m] incoming bound long wave
real*8 2 DR s%nx+1 s%ny+1 d [W/m2] roller energy dissipation
real*8 2 umean 2 s%ny+1 2 [m/s] longtermmean velocity at bnds in x-direction
real*8 2 vmean 2 s%ny+1 2 [m/s] longtermmean velocity at bnds in y-direction
integer 0 vardx b [-] 0 =uniformgrid size, 1 =variable grid size
real*8 2 vu s%nx+1 s%ny+1 d [m/s] y velocity in u points
real*8 2 uv s%nx+1 s%ny+1 d [m/s] x velocity in v points
real*8 1 D50 par%ngd b [m] D50 grain diameters for all sediment classses
real*8 1 D90 par%ngd b [m] D90 grain diameters for all sediment classses
real*8 1 sedcal par%ngd b [-] equilibrium sediment concentartion factor for each sediment class
real*8 1 ucrcal par%ngd b [-] calibration factor for u critical for each sediment class
real*8 3 Tsg s%nx+1 s%ny+1 par%ngd d [s] sediment response time for each sediment class
real*8 3 Susg s%nx+1 s%ny+1 par%ngd d [m2/s] suspended sediment transport x-dir. for each sediment class
(excluding pores)
real*8 3 Svsg s%nx+1 s%ny+1 par%ngd d [m2/s] suspended sediment transport y-dir. for each sediment class
(excluding pores)
real*8 3 Subg s%nx+1 s%ny+1 par%ngd d [m2/s] bed sediment transport x-dir. for each sediment class (excluding
pores)
real*8 3 Svbg s%nx+1 s%ny+1 par%ngd d [m2/s] bed sediment transport y-dir. for each sediment class (excluding
pores)
real*8 3 ceqbg s%nx+1 s%ny+1 par%ngd d [m3/m3] depth-averaged bed equilibrium concentration for each sediment
class
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 4 0
real*8 3 ceqsg s%nx+1 s%ny+1 par%ngd d [m3/m3] depth-averaged suspended equilibrium concentration for each
sediment class
real*8 2 ua s%nx+1 s%ny+1 d [m/s] time averaged flow velocity due to wave assymetry
real*8 2 BR s%nx+1 s%ny+1 d [-] maximumwave surface slope used in roller dissipation formulation
real*8 2 kb s%nx+1 s%ny+1 d [m^2/s^2] near bed turbulence intensity due to depth induces breaking
real*8 2 Tbore s%nx+1 s%ny+1 d [s] wave period interval associated with breaking induced turbulence
real*8 2 dzav s%nx+1 s%ny+1 d [m] total bed level change due to avalanching
real*8 2 maxzs s%nx+1 s%ny+1 d [m] maximumelevation in simulation
real*8 2 minzs s%nx+1 s%ny+1 d [m] minimumelevation in simulation
real*8 2 L1 s%nx+1 s%ny+1 d for subroutine dispersion
real*8 2 Sk s%nx+1 s%ny+1 d [-] skewness of short waves
real*8 2 As s%nx+1 s%ny+1 d [-] asymmetry of short waves
! real*8 2 Fimpact s%nx+1 s%ny+1 d ! [N/m^2] Force on dune face
real*8 2 gwhead s%nx+1 s%ny+1 d [m] groundwater head (differs from gwlevel)
real*8 2 gwlevel s%nx+1 s%ny+1 d [m] groundwater table (min(zb,gwhead))
real*8 2 gwheight s%nx+1 s%ny+1 d [m] vertical size of aquifer through which groundwater can flow
real*8 2 gwbottom s%nx+1 s%ny+1 d [m] level of the bottom of the aquifer
real*8 2 gwu s%nx+1 s%ny+1 d [m/s] groundwater flow in x-direction
real*8 2 gwv s%nx+1 s%ny+1 d [m/s] groundwater flow in y-direction
real*8 2 gww s%nx+1 s%ny+1 d [m/s] groundwater flow in z-direction (interaction between surface and ground
water)
real*8 2 dinfil s%nx+1 s%ny+1 d [m] Infiltration layer depth used in quasi-vertical flow model for groundwater
real*8 2 gw0back 2 s%ny+1 2 [m] boundary condition back boundary for groundwater head
real*8 2 kturb s%nx+1 s%ny+1 d [m^2/s^2] depth averaged turbulence intensity due to long wave breaking
real*8 3 ero s%nx+1 s%ny+1 par%ngd d [m/s] bed erosion rate per fraction
real*8 3 depo_im s%nx+1 s%ny+1 par%ngd d [m/s] implicit bed deposition rate per fraction
real*8 3 depo_ex s%nx+1 s%ny+1 par%ngd d [m/s] explicit bed deposition rate per fraction
integer 2 nd s%nx+1 s%ny+1 d [-] number of bed layers (can be different for each computational cell)
real*8 4 pbbed s%nx+1 s%ny+1 max(par%nd,2) par%ngd d [-] NO DESCRIPTION
real*8 3 dzbed s%nx+1 s%ny+1 max(par%nd,2) d [-] NO DESCRIPTION
real*8 2 z0bed s%nx+1 s%ny+1 d [-] NO DESCRIPTION
real*8 2 ureps s%nx+1 s%ny+1 d [m/s] representative flow velocitie for sediment advection and diffusion in x-
direction
real*8 2 vreps s%nx+1 s%ny+1 d [m/s] representative flow velocitie for sediment advection and diffusion in y-
direction
real*8 2 urepb s%nx+1 s%ny+1 d [m/s] representative flow velocitie for sediment advection and diffusion in x-
direction
real*8 2 vrepb s%nx+1 s%ny+1 d [m/s] representative flow velocitie for sediment advection and diffusion in y-
direction
real*8 2 umwci s%nx+1 s%ny+1 d [m/s] u-velocity (time-averaged) for wci
real*8 2 rolthick s%nx+1 s%ny+1 d [m] long wave roller thickness
real*8 2 vmwci s%nx+1 s%ny+1 d [m/s] v-velocity (time-averaged) for wci
real*8 2 zswci s%nx+1 s%ny+1 d [m] waterlevel (time-averaged) for wci
real*8 2 pres s%nx+1 s%ny+1 d [m^2/s^2] normalized dynamic pressure
real*8 2 wb s%nx+1 s%ny+1 d [m/s] vertical velocity at the bottom
real*8 2 ws s%nx+1 s%ny+1 d [m/s] vertical velocity at the free surface
real*8 2 taubx s%nx+1 s%ny+1 d [N/m^2] x-component of bed shear stress
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
5 4 1
real*8 2 tauby s%nx+1 s%ny+1 d [N/m^2] y-component of bed shear stress
real*8 2 Df s%nx+1 s%ny+1 d [W/m^2] dissipation rate due to bed friction
real*8 2 Dp s%nx+1 s%ny+1 d [W/m^2] dissipation rate in the swash due to transformation of kinetic wave energy
to potential wave energy
real*8 2 Sutot s%nx+1 s%ny+1 d [m2/s] Sediment transport integrated over bed load and suspended and for all
sediment grains
real*8 2 Svtot s%nx+1 s%ny+1 d [m2/s] Sediment transport integrated over bed load and suspended and for all
sediment grains
real*8 2 cctot s%nx+1 s%ny+1 d [m3/m3] Sediment concentration integrated over bed load and suspended and for all
sediment grains
real*8 2 wi 2 s%ny+1 2 [m/s] Vertical velocity at boundary due to (short) waves2
real*8 2 zi 2 s%ny+1 2 [m] Surface elevation at boundary due to (short) waves
real*8 2 nuh s%nx+1 s%ny+1 d [m2/s] horizontal viscosity coefficient
real*8 2 cf s%nx+1 s%ny+1 d [-] Friction coefficient flow
real*8 2 D50top s%nx+1 s%ny+1 d [-] Friction coefficient flow
real*8 2 D90top s%nx+1 s%ny+1 d [-] Friction coefficient flow
real*8 2 structdepth s%nx+1 s%ny+1 d [m] Depth of structure in relation to instantaneous bed level
5.25 Example params.txt file
The following example shows a partial params.txt file in which all three types of output are
specified.
Example of a shortened params.txt file with various output options
Out put opt i ons
t st ar t = 100.
t i nt g = 60.
t i nt m= 720.
t spoi nt s = t i mespoi nt s. t xt
ngl obal var = 4
zs
zb
H
ue
nmeanvar = 2
ue
H
npoi nt s = 3
124 12 2 H#zs#
174 12 2 H#zs#
199 12 4 H#zs#ue#ve#
nr ugauge = 1
12 3 zs#ue#ve#
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
6 1
6 References
Abbott, M.B., 1979. Computational Hydraulics. Elements of the theory of free surface flow. Pitman Publishing,
London, 324 pp.
Arcilla, A.S., Roelvink, J.A., O'Connor, B.A. Reniers, A., and J imenez. J .A. (1994). The Delta Flume '93
Experiment. Coastal Dynamics '94. Arcilla, Stive and Kraus (eds), ASCE, New York, pp. 488-502.
Carrier, G.F. and H.P. Greenspan (1958). Water waves of finite amplitude on a sloping beach. J. Fluid Mech.,
Vol. 4, pp. 97-109.
Galapatti, R., 1983. A depth integrated model for suspended transport. Report 83-7, Communications on
Hydraulics, Department of Civil Engineering, Delft University of Technology.
Reniers, A.J .H.M., J .A. Roelvink and E.B. Thornton. (2004). Morphodynamic modelling of an embayed beach
under wave group forcing. J. of Geophysical Res. , VOL. 109, C01030, doi:10.1029/2002J C001586,
2004
Reniers, A.J .H.M., E.B. Thornton, T. Stanton and J.A. Roelvink. (2004b) Vertical flow structure during Sandy
Duck: Observations and Modeling. Coastal Engineering, Volume 51, Issue 3, May 2004, Pages 237-
260
Roelvink, J.A. (1993a) Dissipation in random wave groups incident on a beach. Coastal Eng., 19 (1993) pp. 127-
150.
Roelvink, J.A. (1993b) Surf beat and its effect on cross-shore profiles. Ph.D. thesis, Delft University of
Technology.
Ruessink, B.G., J .R. Miles, F. Feddersen, R.T. Guza and S. Elgar, 2001. Modeling the alongshore current on
barred beaches. Journal of Geophysical Research, 106, 22451-22464.
Soulsby, R. (1997). Dynamics of marine sands. Thomas Telford Publications, London, ISBN 0 7277 2584 X.
Stelling GS, Duinmeijer SPA. A staggered conservative scheme for every Froude number in rapidly varied
shallow water flows. International J ournal for Numerical Methods in Fluids. 2003; 43:1329-1354.
Van Dongeren, A.R. and I.A. Svendsen (1997). An Absorbing-Generating Boundary condition for Shallow
Water Models. J. of Waterways, Ports, Coastal and Ocean Engineering, vol. 123, no. 6, pp. 303-313.
Van Dongeren, A.R., A.J.H.M. Reniers, J.A. Battjes and I.A. Svendsen (2003). Numerical modeling of
infragravity wave response during Delilah. J. Geoph. Res, 108, (C9) , 3288,
doi:10.1029/2002J C001332
van Gent, M.R.A., 1995. Wave Interaction with Permeable Coastal Structures, Delft University of Technology,
Delft.
Verboom, G.K., G.S. Stelling and M.J. Officier (1981). Boundary conditions for the shallow water equations. In:
Abbott, M.B. and J .A. Cung, eds. Engineering Applications of Computational Hydraulics, vol. 1., pp.
230-262.
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
A 1
A Description of gen.ezs file (instat+2,3)
This file should be in subdirectory bc and has a fixed name of gen.ezs.
The contents are as follows:
- An arbitrary header line
- Number of time points, number of columns
- Values for time, long wave elevation and wave energy (3 columns)
Example:
BL01
10 3
0 0.10 0
2 0.25 200
4 0.20 300
6 0.15 250
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
A 1
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
B 1
B Numerical implementation
B.1 Wave action equation
We use an up-wind schematisation to solve the wave action balance (keyword scheme =1
and activated by default). The wave action is given at the same points at the water level. The
advection of wave action is then discretized as follows:
, , , , , , 1, , 1, ,
, , ,
i,j i-1,j
, 1, , 1, , , , , , ,
, , ,
i+1,j i,j
(i,j,k) , 0
(i,j,k) , 0
n n n n
n n
x i j k i j k x i j k i j k n x
x i j k
n n n n
n n
x i j k i j k x i j k i j k n x
x i j k
c A c A
c A
c
x x x
c A c A
c A
c
x x x

+ +

c
= >
c

c
= <
c
(B.1)
The discretization using a Lax-Wendroff scheme (scheme =2) is:
, ,
, 1, , 1, , , 1, , 1, ,
i+1,j i-1,j
2 2 2
, 1, , 1, , , , , , , , 1, , 1, ,
i+1,j i,j i,j i-1,j
(i,j,k)
2
2
n n n n
n n
x i j k i j k x i j k i j k
x
n n n n n n
x i j k i j k x i j k i j k x i j k i j k
c A c A
c A
x x x
c A c A c A
t
x x x x
+ +
+ +

c
=
c
( ( ( +
A


(B.2)
In y-direction, the upwind discretization is
, , , , , , , 1, , 1,
, , ,
i,j i,j-1
, , 1, , 1, , , , , ,
, , ,
i,j+1 i,j
(i,j,k) , 0
(i,j,k) , 0
n n n n n n
y y i j k i j k y i j k i j k n
y i j k
n n n n n n
y y i j k i j k y i j k i j k n
y i j k
c A c A c A
c
y y y
c A c A c A
c
y y y

+ +
c
= >
c
c
= <
c
(B.3)
In the Lax-Wendroff discretization:
, ,
2 2 2
,, 1, , 1, , , , , , ,, 1, , 1, , , 1, , 1, ,, 1, , 1,
i,j+1 i,j-1 i,j+1 i,j i,j i,j-1
2
(i,j,k)
2
n n n n n n n n n n n n
yi j k i j k yi j k i j k yi j k i j k y yi j k i j k yi j k i j k
c A c A c A c A c A c A
t
y y y y y y y
+ + + +
( ( ( + c
A

=
c
(B.4)
In theta-space an upwind scheme is used only:
, , , , , , , , 1 , , 1
, , ,
i,j,k i,j,k-1
, , , 1 , , 1 , , , , ,
, , ,
i,j,k+1 i,j,k
(i,j,k) , 0
(i,j,k) , 0
n n n n
n n
i j k i j k i j k i j k n
i j k
n n n n
n n
i j k i j k i j k i j k n
i j k
c A c A
c A
c
c A c A
c A
c
u u
u
u
u u
u
u
u u u
u u u

+ +

c
= >
c

c
= <
c
(B.5)
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
B 2
Similar for the wave action balance:
1
, , , ,
, , , , , ,
, ,
n
n n n n n
n n
i j k i j k y
x
i j k i j k i j k
i j k
A A c A
c A c A D
t x y
u
u o
+
c
c c
=
A c c c
(B.6)
We apply an upwind schematisation, since the horizontal scale of the problem is limited and
such a scheme deals with shocks in a natural way.
B.2 Shallow water equations
We apply a staggered grid, where bed levels and water levels are defined in the centre of
cells, and velocity components at the cell interfaces.
If nx,ny are the number of cells in both directions, the water level points are numbered from
1 tonx+1 and from 1 tony+1.
The water level gradients are computed at the cell interfaces and are given by:
i+1,j i,j
i+1,j i,j
(i,j)
x x x
q q
q

c
=
c
(B.7)
i,j+1 i,j
i,j+1 i,j
(i,j)
y x x
q q
q

c
=
c
(B.8)
For computing the shear stresses at the cell interfaces we need the velocity magnitudes at
these interfaces. These are composed by combining the normal velocity component at the
interface and the average of the 4 adjacent tangential components:
, , , 1 , 1, 1 1,
, , 1, , 1, 1 , 1
1
( )
4
1
( )
4
u i j i j i j i j i j
v i j i j i j i j i j
v v v v v
u u u u u
+ +
+ +
= + + +
= + + +
(B.9)
The water depth in each cell is computed as:
, , , , i j i j b i j
h z q = (B.10)
For the depth at cell interfaces, following Stelling and Duinmeijer (2003) we distinguish
between the depth used in the continuity equation and that used in the momentum equation.
The depth at the interfaces for the continuity equation is taken as the upwind depth in case
the velocity is greater than a minimum velocity, or the maximum water level minus the
maximum bed level in case the velocity is less than this minimum velocity:
, , , , min
, , 1, , min
, , , , , 1, , , , 1, , min
,
,
max( , ) max( , ) ,
u i j i j i j
u i j i j i j
u i j s i j s i j b i j b i j i j
h h u u
h h u u
h z z z z u u
+
+ +
= >
= <
= <
(B.11)
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
B 3
, , , , min
, , , 1 , min
, , , , , , 1 , , , , 1 , min
,
,
max( , ) max( , ) ,
v i j i j i j
v i j i j i j
v i j s i j s i j b i j b i j i j
h h v v
h h v v
h z z z z v v
+
+ +
= >
= <
= <
(B.12)
For the depthin the momentum balance we take the average depth between the cell centers:
, , , 1,
1
( )
2
mu i j i j i j
h h h
+
= + (2.13)
, , , , 1
1
( ),
2
mv i j i j i j
h h h
+
= + (2.14)
The advection terms in x-direction are approximated as follows:
, , , , 1, 1, , 1,
,
,
, , , 1,
, , , , 1, 1, 1, ,
,
,
, , 1, ,
1
, 0
2
1
, 0
2
n n n
u i j i j u i j i j i j i j n
i j n n
i j
mu i j i j i j
n n n
u i j i j u i j i j i j i j n
i j n n
i j
mu i j i j i j
h u h u u u
u
u u
x h x x
h u h u u u
u
u u
x h x x

+ + +
+
+
c
= >
c
+
c
= <
c
(B.15)
, , , , , , 1 , , 1 , , 1
, ,
, , , , 1 ,
, , , , , , 1 , , 1 , 1 ,
, ,
, , , 1 , ,
1
, 0
2
1
, 0
2
n n n n n
u i j u i j u i j u i j i j i j n
u i j n n
mu i j i j i j i j
n n n n n
u i j u i j u i j u i j i j i j n
u i j n n
mu i j i j i j i j
h v h v u u
u
v v
y h y y
h v h v u u
u
v v
y h y y

+ + +
+
+
c
= >
c
+
c
= >
c
(B.16)
The advection terms in y-direction are approximated as follows:
, , , , , 1 , 1 , , 1
,
, , , , , 1
, , , , , 1 , 1 , 1 ,
,
, , , , 1 ,
1
, 0
2
1
, 0
2
n n n n n n n
v i j i j v i j i j i j i j n
i j n n n
i j mv i j i j i j
n n n n n n n
v i j i j v i j i j i j i j n
i j n n n
i j mv i j i j i j
h v h v v v
v
v v
y h y y
h v h v v v
v
v v
y h y y

+ + +
+
+
c
= >
c
+
c
= <
c
(B.17)
, , , , , 1, , 1, , 1,
, ,
,
, , , 1,
, , , , , 1, , 1, 1, ,
, ,
,
, , 1, ,
1
, 0
2
1
, 0
2
n n n n n
v i j v i j v i j v i j i j i j n
v i j n n
i j
mv i j i j i j
n n n n n
v i j v i j v i j v i j i j i j n
v i j n n
i j
mv i j i j i j
h u h u v v
v
u u
x h x x
h u h u v v
v
u u
x h x x

+ + +
+
+
c
= >
c
+
c
= >
c
(B.18)
The momentum equation is discretized as follows:
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
B 4
1 n
, , , i+1,j i,j , , 2 2 2
, , ,
,
, , i+1,j i,j , , ,
1.16
n n n n n n
i j i j i j x i j n n
f rms i j u i j n
i j
u i j u i j i j
u u u F
u u
u v c u u v g
t x y h x x h
q q

+

c c
= + + +
A c c
(B.19)
1
, , , i,j+1 i,j , , 2 2 2
, , ,
,
, , i,j+1 i,j , , ,
1.16
n n n n n
i j i j i j y i j n n
f rms v i j i j n
i j
v i j v i j i j
v v v F
v v
v u c u u v g
t y x h y y h
q q

+

c c
= + + +
A c c
(B.20)
From this, the velocities at the new time step level are computed. The water level is then
updated by:
1 1 1 1 1
, , , , 1, 1, , i,j , 1 , 1
u,i,j u,i-1,j v,i,j v,i,j-1
n n n n n n n n n
i j i j i j i j i j i j i j i j i j
u h u h v h v h
t x x y y
q q
+ + + + +


=
A
(B.21)
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
B 5
which yields the wave energy at the new time level.
B.3 Advection-diffusion equations
The differential equations for the advection diffusion of sediment is solved with finite
differences using the first order up-wind scheme discussed earlier with the water depths at
the old time level and the corresponding velocities at the new time level. The horizontal x-
advection is then given by:
, ,
, ,
, 1 , 1
, 1, , 1
,
, 1,
,
, 1 , 1
1, , , 1
,
1, ,
,
, 0
, 0
n n E n n n E n
E
i j i j E n
i j
i j i j
i j
n n E n n n E n
E
i j i j E n
i j
i j i j
i j
h C u h C u
hCu
u
x x x
h C u h C u
hCu
u
x x x
+ +
+

+ +
+ +
+

| | c
= >
|
c
\ .

| | c
= <
|
c
\ .
(2.22)
A similar expression for the horizontal advection in the y-direction:
, ,
, ,
, 1 , 1
, 1, , 1
,
, , 1
,
, 1 , 1
, 1 , , 1
,
, 1 ,
,
, 0
, 0
n n E n n n E n
E
i j i j E n
i j
i j i j
i j
n n E n n n E n
E
i j i j E n
i j
i j i j
i j
h C v h C v
hCv
v
y y y
h C v h C v
hCv
v
y y y
+ +
+

+ +
+ +
+

| | c
= >
|
c
\ .

| | c
= <
|
c
\ .
(2.23)
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
B 6
The horizontal diffusion is evaluated at the old time level n and approximated by:
, ,
1, ,
1, , ,
H x H x
i j i j
H
i j i j i j
D hC D hC
C
D h
x x x x
c c
+
+

c c | | | |
=
| |
c c
\ . \ .
(2.24)
Where the cross-shore gradient in the sediment concentration is given by:
1, ,
, 1, ,
i j i j
x
i j i j i j
C C
C
C
x x x
+
c
+

c
| |
= =
|
c
\ .
(2.25)
And similarly for the y-direction:
, ,
, 1 ,
,
, 1 ,
,
, 0
H y H y
i j i j E
H i j
i j i j
i j
D hC D hC
C
D h v
y y y y
c c
+
+

| | | | c c
= <
| |
c c
\ . \ .
(2.26)
Where the along-shore gradient in the sediment concentration, C
y
, is given by:
, 1 ,
, 1 , ,
i j i j
i j i j i j
C C
C
y y y
+
+

| | c
=
|
c
\ .
(2.27)
The time up-date of the sediment concentration is then given by:
1 1
, , , ,
, ,
, , ,
n n
n n n n
E E
i j i j i j i j
i j i j
n n n
eq
h h
s i j i j i j
h C h C
hCu hCv
t x y
hC hC
C C
D h D h
x x y y T
+ +
( ( c c
+ + +
( (
A c c

( ( ( c c c c ( (
+ + =
( ( ( ( (
c c c c

(2.28)
B.4 Bed update
The bed-update is then approximated by:
1
, , , , , , , 1, , , , , 1
0
(1 )
n n n n n n
b i j b i j x i j x i j y i j y i j
mor
z z S S S S
f
t p x y
+

(
+ + =
(
A A A

(2.29)
Where f
mor
represents a morphological factor to speed up the bed evolution (see e.g.
Roelvink, 2006).
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
C 1
C Multiple sediment fractions proof of concept
To test the implementation of the sediment class formulation a comparison is made with
observations of pre-and post hurricane Ivan cross-barrier island profiles (see Figure 6.1) at
Beasly Park, Florida, USA (Wang and Horwitz, 2007).
The hurricane Ivan impact is simulated with a constant surge level of 1.8 m present for 10
hours at which time the offshore incident significant wave height is kept at 10 m with a
mean wave period of 12 s. The sediment class distribution used in the calculations
discriminates between sand located within the frontal dune (class 1) and sand located on and
behind the barrier island (class 2) (see upper panel in ). The sand on the barrier island is
mostly vegetated which mitigates the erosion. Hence this sand has been given a mobility
restriction that makes it more difficult to pick-up by means of a reduction factor of 0.25 on
the equilibrium concentration. Grain sizes for both sand composites are the same with a D
50
of 0.0035 m and a D
90
of 0.005 mm. The initial sediment class distribution is presented in
the top panel of Figure 1, where an intensity of 1 corresponds to sediment class one only and
-1 to the presence of sediment class 2 only.
The bed-elevation and sediment class distribution after 10 hours are shown in the lower
panel of . The calculated bed-level is similar to the observations although differences are
apparent. These differences can be related to the fact that the hurricane impact is simply
modeled (i.e. constant conditions) and the fact that the post-survey was performed
approximately 10 months after the hurricane had past. Still the overall evolution is
consistent with the observations. The calculated changes in the sediment classes are also
consistent with the observations of Wang and Horwitz (2007) based on a number of cores
showing that the intersection of the new washover with the pre-hurricane sediment occurs
approximately at the original bed level.
Figure 6.1 Top panel: Initial pre-hurricane bed elevation (green line) and sediment class
distribution. A value of 1 corresponds to 100% of sediment class 1, a value of 0 to 50% of
XBeach Model Description and Manual Z4175 June 21, 20
Unesco-IHE, Deltares and Delft University of Technology
C 2
class 1 and 50 % of class 2, and a value of -1 corresponds to 100 % of sediment class two.
Post-hurricane bed elevation (dashed red line) given as a reference. Bottom panel:
Calculated bed-evolution (corresponding to the position of the top layer) and corresponding
sediment class distribution showing the thickness of the wash-over layer located behind the
initial dune. Pre- (green line) and post-hurricane (red dashed line) bed elevation given as a
reference