Coastal Engineering 172 (2022) 104068

Contents lists available at ScienceDirect

Coastal Engineering
journal homepage: www.elsevier.com/locate/coastaleng

SWAN SurfBeat-1D
Ad Reniers ∗, Marcel Zijlema
Delft University of Technology, Faculty of Civil Engineering and GeoSciences, Department of Hydraulic Engineering, Stevinweg 1, 2628CN, Delft, The Netherlands

ARTICLE INFO ABSTRACT

Keywords: The Simulating WAves Nearshore (SWAN) model has been extended with an infragravity module to predict
Bound and free infragravity waves the Wave-Group-Forced (WGF) infragravity response to a frequency-directional sea-swell spectrum at a mildly
Spectral modeling sloping alongshore uniform beach. To that end the SWAN model has been extended with an WGF-infragravity
Field and laboratory validation
source term denoted 𝑆𝑠𝑏 where the subscript denotes surfbeat. The corresponding WGF infragravity energy
model has been verified with a set of benchmark tests using the infragravity amplitude model of Reniers et al.
(2002). Next the implementation of the energy balance in SWAN has been validated with both prototype-scale
laboratory experiments and field observations, showing a good comparison with observations not affected by
the nodal structure of the (partially) standing infragravity waves. This suggests that the model is capable
of providing improved infragravity boundary conditions in relatively shallow water compared to the typical
assumption of equilibrium forcing conditions using for instance Hasselmann’s equilibrium theory (Hasselmann,
1962). These infragravity boundary conditions can subsequently can be used by other more sophisticated
models to compute runup, overtopping and dune erosion.

1. Introduction approach. First, both phase-resolving modeling and observations have

shown that the local equilibrium approach at times may significantly
The frequency-directional spectrum of the incident sea-swell waves overestimate the sea-swell forced infragravity waves (Schäffer, 1993;
controls the wave interference patterns forming wave groups with peri- Herbers and Burton, 1997; Battjes et al., 2004). This is related to
ods between approximately 25 s to 250 s. This results in a modulation the fact that the transfer of energy from the sea-swell waves to the
of the radiation stress, forcing bound infragravity waves that propagate underlying infragravity waves on a sloping bed needs time and thus
with the wave groups (Biesel, 1952; Longuet-Higgins and Stewart,
distance to occur. In case the changes in water depth and accompanying
1962; Hasselmann, 1962). As the wave groups approach the beach
sea-swell conditions are relatively fast with respect to the infragravity
the sea-swell waves become gradually steeper and ultimately break
wave length the equilibrium condition is not reached. Secondly, by
dissipating their energy in the surfzone. As a result the wave-group
modulation reduces and the bound infragravity waves get released and using the local equilibrium it is implicitly assumed that offshore only
(partially) reflect at the shoreline (Herbers et al., 1995; Van Dongeren bound infragravity waves are incident on the coast. Ardhuin et al.
et al., 2007; Thompson et al., 2006). On mildly sloping beaches this (2014) and Rawat et al. (2014) have shown that infragravity waves
can result in a dominance of the infragravity wave energy at the wa- generated at one coast can radiate away and propagate across oceanic
terline (Holman, 1981; Guza and Thornton, 1982; Ruessink, 1998a,b; scales where they arrive as free infragravity waves at another coast.
De Bakker et al., 2014), controlling the run-up and potential overtop- This implies that the local equilibrium approximation may in fact
ping at the beach (e.g. Stockdon et al. (2006), Ruggiero et al. (2004) underestimate the incident infragravity energy. This contribution of the
and Guza and Feddersen (2012)) as well as dune erosion (Van Thiel de free infragravity waves to the total infragravity spectrum will depend,
Vries et al., 2008; Roelvink et al., 2007) and overwash [e.g. McCall among other things, on the coastal configuration, where a larger part
et al. (2010)]. For a comprehensive review on infragravity waves refer of the infragravity energy can be trapped on the sloping bathymetry
to Bertin et al. (2018). with increasing offshore water depth (Herbers et al., 1995; Smit et al.,
Modeling the impact of incident infragravity waves requires an off-
2018). As a result the free infragravity energy arriving from distant
shore boundary condition. This boundary condition is usually obtained
storms is relatively unimportant along open coasts when the local sea-
by assuming a local equilibrium between the incident sea-swell wave
swell forcing is significant (Smit et al., 2018). However, in regional
forcing and the bound infragravity wave (Hasselmann, 1962; Longuet-
Higgins and Stewart, 1962). There are two potential problems with this seas where the offshore water depth is limited, a significant part of the

∗ Corresponding author.
E-mail address: [email protected] (A. Reniers).

https://doi.org/10.1016/j.coastaleng.2021.104068
Received 15 September 2020; Received in revised form 4 October 2021; Accepted 27 November 2021
Available online 17 December 2021
0378-3839/© 2021 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
A. Reniers and M. Zijlema Coastal Engineering 172 (2022) 104068

infragravity energy may travel from coast to coast affecting the local 2. Methods
infragravity conditions (Rijnsdorp et al., 2021) and subsequent runup
and associated coastal safety during severe storms. Hence, depending 2.1. Infragravity Amplitude Model (IAM)
on the coastal configuration, the local equilibrium approximation may
result in an overestimation or underestimation of the incident infra- The wave-group-forced (WGF in the following) infragravity model
gravity waves. Refer to Fiedler et al. (2019) for an assessment of the developed by Reniers et al. (2002) is based on the linearized shallow
infragravity wave boundary effects on runup and wave shape. These water equations (Mei and Benmoussa, 1984). This model has been
boundary problems can be overcome by using a much larger regional- compared with both analytical solutions and field observations showing
scale infragravity model domain with the offshore boundary condition
a good correspondence (Reniers et al., 2002, 2010) making it a suitable
located in deeper water.
benchmark. A brief description of the model is presented here for ease
However, computing infragravity waves at a regional scales (O(100)
of reference. The spectral WGF long wave equation is given by:
km) prevents the use of deterministic models as they need a fine
numerical grid to resolve the infragravity waves leading to very long 𝑑 2 𝜂̂ 𝑑ℎ 𝑑 𝜂̂
𝑔ℎ +𝑔 + 𝛥𝜔2 𝜂̂ − 𝑔ℎ𝛥𝑘2𝑦 𝜂̂ = (1)
computational times. This can be overcome by considering the wave 𝑑𝑥2 ( 𝑑𝑥 𝑑𝑥
energy balance of the infragravity waves using spatial propagation, 𝑥 𝑥
1 𝑑 𝑆̂𝑥𝑥 exp 𝑖 ∫0 𝛥𝑘𝑥 𝑑𝑥
2 𝑑 𝑆̂𝑥𝑦 exp 𝑖 ∫0 𝛥𝑘𝑥 𝑑𝑥
source and sink terms to describe the evolution over long time and − + 2𝑖𝛥𝑘𝑦
𝜌 𝑑𝑥2 𝑑𝑥
spatial scales. This approach is used to describe the evolution of sea- )
𝑥
swell waves at global scales with WaveWatch 3 (Tolman, 1991) and −𝛥𝑘2𝑦 𝑆̂𝑦𝑦 exp 𝑖 𝛥𝑘𝑥 𝑑𝑥
WAM (Hasselmann et al., 1988) as well as regional scales using SWAN ∫0
(Booij et al., 1999) and Tomawac (Benoit et al., 1996). To be able to where 𝑥 and 𝑦 are directed onshore and alongshore respectively, ℎ is
compute the infragravity energy balance, the non-linear source term the local water depth, 𝑔 the gravitational acceleration, 𝜂̂ is the complex
describing the transfer of energy from the incident waves to the infra-
valued infragravity surface elevation amplitude.
gravity waves needs to be included. This requires an evolution equation
Considering the short-wave group to be made up of two incident
of the triad interaction between two incident waves making up a wave
sea-swell components with amplitudes 𝑎1 and 𝑎2 , frequencies 𝜔1 and
group and the accompanying forced infragravity wave (Phillips, 1977;
𝜔2 and directions 𝜃1 and 𝜃2 respectively, the corresponding radiation
Madsen and Sorensen, 1993; Herbers and Burton, 1997). Eldeberky and
Battjes (1995) formulated a source term in SWAN, 𝑆𝑛𝑙3 , to describe stress amplitudes at the right hand side are given by:
( )
the generation and evolution of forced super harmonics within the 𝑐𝑔 1
𝑆̂𝑥𝑥 = (1 + cos2 𝛼) − 𝜌𝑔𝑎1 𝑎2 (2)
spectrum. To be able to describe the evolution of the infragravity 𝑐 2
energy balance this term needs to be complemented with a source term
and
for the forced subharmonics that we will denote 𝑆𝑠𝑏 where the subscript
𝑐𝑔
𝑠𝑏 stands for surfbeat.1 𝑆̂𝑥𝑦 = (cos 𝛼 sin 𝛼)𝜌𝑔𝑎1 𝑎2 (3)
To expedite computations SWAN uses a Lumped Triad Approxima- 𝑐
tion (LTA) (Eldeberky and Battjes, 1995), considering self interactions and
( )
of co-linear waves only to represent 𝑆𝑛𝑙3 . The LTA makes use of a non- 𝑐𝑔 1
linear coupling function based on the work by Madsen and Sorensen 𝑆̂𝑦𝑦 = (1 + sin2 𝛼) − 𝜌𝑔𝑎1 𝑎2 (4)
𝑐 2
(1993) and an estimate of the bi-phase with a parametrization based on
the local Ursell number to compute 𝑆𝑛𝑙3 . The expression for the bi-phase representing the slow modulation of the radiation stresses related to
predicts the phase coupling between the two primary waves and the mild changes in depth. The corresponding phase velocity, 𝑐 and group
forced super harmonic thus controlling the skewness and asymmetry velocity, 𝑐𝑔 , are computed with linear wave theory using the mean
of the incident waves (Elgar and Guza, 1985). This superharmonic bi- frequency of the two components. The bi-chromatic wave group creates
phase is different from the phase coupling of the sea-swell waves with a modulation in the radiation stress with difference frequency 𝛥𝜔 =
the forced sub-harmonics and the local Ursell parametrization of Elde- 𝜔2 − 𝜔1 , difference cross-shore wave number 𝛥𝑘𝑥 = 𝑘2,𝑥 − 𝑘1,𝑥 and
berky and Battjes (1995) can therefore not be used. Furthermore, the alongshore wave number 𝛥𝑘𝑦 = 𝑘2,𝑦 − 𝑘1,𝑦 forcing an infragravity
LTA considers self interactions only, whereas the infragravity waves are response. In case of an alongshore uniform beach the alongshore wave
forced by difference interactions, requiring a different description for number is constant, whereas the cross-shore wave number changes
the non-linear coupling. And finally, the LTA uses co-linear waves only, with depth, hence the integral in the phase function to represent the
whereas the forced infragravity waves typically have a much broader modulation at the wave group scale in the cross-shore direction in
directional distribution resulting from the directional spreading of the Eq. (1) . Given the fact that the wave direction of the individual sea-
sea-swell waves forcing them (Herbers and Burton, 1997). Instead, as a swell waves, 𝜃1 and 𝜃2 , can be computed with Snel’s law in case of
first step, the non-linear subharmonic coupling and accompanying bi- an alongshore uniform beach, the difference wave numbers are readily
phase in the presence of directional sea-swell waves at an alongshore retrieved and the mean sea-swell wave direction 𝛼 is subsequently
uniform coast are derived making use of the radiation stress concept obtained from:
(Longuet-Higgins and Stewart, 1964) following Reniers et al. (2002), ( )
𝑘1,𝑦 + 𝑘2,𝑦
Janssen et al. (2003) and Battjes et al. (2004). This results in a direc- 𝛼 = 𝑎𝑡𝑎𝑛 (5)
tional infragravity energy source term, 𝑆𝑠𝑏 , that is used to predict the 𝑘1,𝑥 + 𝑘2,𝑥
evolution of the wave-group-forced infragravity wave energy described Given the offshore boundary conditions for the two incident waves
in Section 2. Predictions with this infragravity energy model are first and a bottom profile the slow cross-shore variation in the radiation
compared with the infragravity amplitude model developed by Reniers stress forcing, 𝑆𝑖𝑗 (𝑥), can be computed with a mean wave energy bal-
et al. (2002) for simplified conditions in Section 3.1. Next the infragrav- ance taking into account shoaling, refraction and wave breaking. Next,
ity energy balance is implemented in SWAN and compared with both
with the appropriate boundary conditions, the complex infragravity
large scale laboratory experiments in Section 3.2 and field observations
wave amplitudes can be computed with Eq. (1). The corresponding
in Section 3.3. The general applicability and model limitations are
infragravity surface elevation can be re-constructed from the complex
discussed in Section 4 in view of potential applications at the regional
amplitudes through:
scale followed by conclusions in Section 5.
( )
𝜂(𝑥, 𝑦, 𝑡) = ℜ 𝜂(𝑥)
̂ exp(𝑖(𝛥𝜔𝑡 − 𝛥𝑘𝑦 𝑦)) (6)
1
The infragravity waves associated with breaking wave groups was coined where ℜ corresponds to the real part and similar expressions are
surfbeat by Munk (1949). used for the wave-group modulation and infragravity velocities (see

2
A. Reniers and M. Zijlema Coastal Engineering 172 (2022) 104068

Reniers et al., 2002 for details). The fully reflective shoreline boundary (14)
condition of Reniers et al. (2002) has been replaced with an absorbing
with the radiation stress amplitudes, 𝑆̂𝑖𝑗 , given by Eqs. (2)–(4) and
boundary condition to retain the incident infragravity waves only to
the bi-phase, 𝛥𝜓, represents the phase coupling between the radiation
facilitate the comparison with the WGF infragravity energy balance
stress forcing and the accompanying WGF infragravity wave. The first
discussed next.
two terms on the right hand side are related to the harmonic modula-
tion of the wave group and the third term is related to the slow changes
2.2. Infragravity Energy Model (IEM)
in the radiation stress amplitude due to cross-shore depth variations.
Combining the left hand side and right hand side the WGF-infragravity
Instead of solving for the infragravity surface elevation amplitude
energy balance, Eq. (7) can be written as:
with Eq. (1), an infragravity energy balance can be used in combination
̂2
with a non-linear source term to account for the transfer of energy 𝜌 𝜕𝑐𝑔 cos 𝜃𝑓 ℎ𝑈𝑓 1( )
from the sea-swell wave group to the accompanying WGF infragrav- = 𝛥𝑘𝑥 𝑈̂ 𝑓 ,𝑥 𝑆̂𝑥𝑥 + 𝛥𝑘𝑦 𝑈̂ 𝑓 ,𝑦 𝑆̂𝑥𝑦 sin 𝛥𝜓
2 𝜕𝑥 2
ity wave. To that end we extend the normally incident approach of 𝑈̂ 𝑓 ,𝑥 𝜕 𝑆̂𝑥𝑥
Battjes et al. (2004) by defining the steady state wave energy balance + cos 𝛥𝜓 (15)
2 𝜕𝑥
for obliquely incident WGF infragravity waves on a mildly sloping,
To compute the non-linear transfer the bi-phase, 𝛥𝜓, needs to be
alongshore uniform beach (Phillips, 1977):
known as well. To that end the WKB-based expression derived by
𝜕𝐹𝑥 Janssen et al. (2003) is extended to provide the evolution equation for
= 𝑆𝑠𝑏 (7)
𝜕𝑥 the bi-phase for obliquely incident waves resulting in (see Appendix A):
where 𝐹𝑥 is the cross-shore infragravity energy flux and 𝑆𝑠𝑏 represents 𝛥𝑘 𝜇 𝑆̂𝑥𝑥
𝑑𝛥𝜓
the source term associated with the forcing by a bichromatic wave train =− 𝑥 +
𝑑𝑥 2 2 𝑔ℎ|𝜂̂𝑓 |
made up of two obliquely incident sea-swell components. The energy ( ( ) )
−1
flux 𝐹𝑥 in Eq. (7), is approximated using the velocity potential for the 2 𝑑 𝑆̂𝑥𝑥 cos 𝜃𝑓 𝑑𝑐𝑔 (cos 𝜃𝑓 )
× 𝛥𝑘𝑥 cos 𝛥𝜓 − − sin 𝛥𝜓
long waves (Battjes et al., 2004): 𝑆̂𝑥𝑥 𝑑𝑥 𝑐𝑔 𝑑𝑥
𝜕𝜙 𝜕𝜙 (16)
𝐹𝑥 = −𝜌ℎ (8) ( ( ) )
𝜕𝑡 𝜕𝑥 𝑆̂𝑥𝑦 𝛥𝑘𝑦 ̂
1 𝜕 𝑆𝑥𝑦
where the overline corresponds to wave-group averaging. The time + 𝛥𝑘𝑦 cos 𝛥𝜓 − sin 𝛥𝜓
𝑔ℎ|𝜂̂𝑓 | 𝛥𝑘𝑥 𝑆̂𝑥𝑦 𝜕𝑥
derivative of the velocity potential corresponds to the combined instan- ( 2
)
taneous kinetic and potential energy of the infragravity wave motion 𝑆̂𝑦𝑦 𝛥𝑘𝑦
− sin 𝛥𝜓
that is subsequently transported by the cross-shore component of the 2 𝑔ℎ|𝜂̂𝑓 | 𝛥𝑘𝑥
instantaneous velocity of the WGF long wave:
where 𝜇 represents the resonant mismatch:
𝜕𝜙 ( )
= 𝑈𝑓 ,𝑥 = 𝑈𝑓 cos 𝜃𝑓 (9) 𝛥𝜔2 − 𝑔ℎ𝛥𝑘2𝑦
𝜕𝑥 𝜇 = 1− (17)
where the subscript 𝑓 stands for the WGF motion. The direction of 𝑔ℎ𝛥𝑘2𝑥
the WGF infragravity wave is obtained from the refraction of the two Summarizing, the infragravity energy model consists of two coupled
individual sea-swell waves making up the wave group: equations, Eq. (15) for the infragravity energy and Eq. (16) for the
( ) bi-phase, that need to be solved in tandem. Note that the evolution
𝛥𝑘𝑦
𝜃𝑓 = 𝑎𝑡𝑎𝑛 (10) equation of the bi-phase depends on the WGF infragravity surface
𝛥𝑘𝑥
elevation, 𝜂̂𝑏 , that can be expressed as function of the WGF infragravity
and the corresponding alongshore infragravity velocity is given by
velocity 𝑈̂ 𝑓 using the continuity equation assuming a mild bed-slope:
𝑈𝑓 ,𝑦 = 𝑈𝑓 sin 𝜃𝑓 . For a mildly sloping cross-shore profile the time
𝑐𝑔
derivative in Eq. (8) can be approximated as: 𝑈̂ 𝑓 = 𝜂̂𝑓 (18)
−𝑐𝑔 𝜕 ℎ
𝜕
= (11) leaving two unknowns, i.e. 𝑈̂ 𝑓 and 𝛥𝜓, that can be solved for a given
𝜕𝑡 cos 𝜃𝑓 𝜕𝑥
offshore boundary condition. For 𝐹𝑥 this is given by the incoming
which upon substitution in Eq. (8) yields the WGF-infragravity energy energy flux of the WGF infragravity wave at 𝑥 = 0:
flux equation:
1
( )2 𝐹𝑥,0 = 𝜌𝑐 𝑐𝑜𝑠𝜃𝑓 ,0 ℎ0 𝑈̂ 𝑓2,0 (19)
𝜌𝑐𝑔 ℎ 𝜕𝜙 𝜌𝑐𝑔 ℎ 𝜌 2 𝑔,0
𝐹𝑥 = = 𝑈𝑓2,𝑥 = 𝑐𝑔 ℎ𝑈̂ 𝑓2 cos 𝜃𝑓 (12) where the subscript 0 refers to the conditions at 𝑥 = 0 m. Using Eq. (18)
cos 𝜃𝑓 𝜕𝑥 cos 𝜃𝑓 2
the offshore WGF velocity is given by:
where 𝑈̂ 𝑓 corresponds to the velocity amplitude of the 𝛥𝑘
WGF-infragravity wave. 𝑐𝑔,0 𝑆̂𝑥𝑥,0 + 2𝑆̂𝑥𝑦,0 𝛥𝑘 𝑦 𝑐
𝑥,0 𝑔,0
To solve the infragravity energy balance the right-hand side of 𝑈̂ 𝑓 ,0 = 𝜂̂𝑓 ,0 = (20)
ℎ0 𝜇0 𝑔ℎ0 ℎ0
Eq. (7) has to be evaluated next. The source term 𝑆𝑠𝑏 corresponds to
where the corresponding WGF surface elevation, 𝜂̂𝑓 ,0 , is based on a
the work done by the radiation stresses on the WGF-infragravity wave
horizontal bed with a bi-phase of 0◦ corresponding to the offshore
in the presence of obliquely incident waves (Phillips, 1977):
boundary condition of 𝛥𝜓 (see Appendix A). Casting the coupled dif-
𝜕𝑆𝑥𝑥 𝜕𝑆𝑥𝑦 ferential equations in finite difference form using an upwind scheme
𝑆𝑠𝑏 = −𝑈𝑓 ,𝑥 − 𝑈𝑓 ,𝑦 (13) the WGF infragravity energy, amplitude and bi-phase can be integrated
𝜕𝑥 𝜕𝑦
along the 𝑥-axis (see Appendix B) and subsequently be compared with
operating in both the cross-shore and alongshore direction in the
the IAM results.
presence of obliquely incident waves. In case the forcing is a wave
group made up by two sea-swell components with different frequencies
2.3. SWAN implementation
and directions, 𝑆𝑠𝑏 can be approximated by (following the approach by
Battjes et al. (2004)):
2.3.1. Sea-swell forced infragravity energy
𝜕 𝑆̂ SWAN (Booij et al., 1999) computes the evolution of the frequency-
1 1 1
𝑆𝑠𝑏 = 𝛥𝑘𝑥 𝑈̂ 𝑓 ,𝑥 𝑆̂𝑥𝑥 sin 𝛥𝜓 + 𝛥𝑘𝑦 𝑈̂ 𝑓 ,𝑦 𝑆̂𝑥𝑦 sin 𝛥𝜓 + 𝑈̂ 𝑓 ,𝑥 𝑥𝑥 cos 𝛥𝜓 directional sea-swell spectrum, 𝐸(𝜔, 𝜃) with the angular frequency 𝜔
2 2 2 𝜕𝑥

3
A. Reniers and M. Zijlema Coastal Engineering 172 (2022) 104068

larger than the infragravity frequency cut-off 𝜔𝑐 , subject to refraction, surface elevation 𝜂̂𝑓 (𝛥𝜔, 𝜃𝑓 ). The corresponding WGF surface elevation
shoaling and sink and source terms, 𝑆(𝜔, 𝜃), related to bottom friction, variance is given by:
wave breaking, white capping, non-linear interactions and wind forc- 1 2
ing. Assuming stationary conditions subject to wave breaking only the 𝑆𝑓 (𝛥𝜔, 𝜃𝑓 ) = 𝜂̂ (𝛥𝜔, 𝜃𝑓 ) (28)
2 𝑓
wave energy balance reduces to:
Solving Eq. (15) for all possible ((𝑝, 𝑚), (𝑞, 𝑛)) pairs with 𝛥𝜔 ≤ 𝜔𝑐 and
collecting terms with the same difference frequency and directional bin
𝜕𝐸(𝜔, 𝜃)𝑐𝑔,𝑥 𝜕𝐸(𝜔, 𝜃)𝑐𝑔,𝑦 𝜕𝐸(𝜔, 𝜃)𝑐𝜃 yields the total WGF forced surface elevation spectrum:
+ + = −𝑆𝑏𝑟 (𝜔, 𝜃) (21)
𝜕𝑥 𝜕𝑦 𝜕𝜃
𝛴𝑆𝑓 (𝛥𝜔, 𝜃𝑓 )
where 𝑐𝑔,𝑥 , 𝑐𝑔,𝑦 and 𝑐𝜃 are the energy density transport velocities in 𝐸𝑓 (𝛥𝜔, 𝜃) = (29)
𝛿𝛥𝜔𝛿𝜃
𝑥, 𝑦 and 𝜃 direction respectively. Wave breaking, 𝑆𝑏𝑟 , is computed
The corresponding incident sea-swell root-mean-square (RMS) wave
with the dissipation model of Battjes and Janssen (1978) with 𝛾𝑏 as a
height is defined as:
calibration coefficient controlling the maximum sea-swell wave height,
√
𝐻𝑚𝑎𝑥 = 𝛾𝑏 ℎ. Given a bathymetry and an offshore sea-swell boundary √ 𝜔𝑚𝑎𝑥 𝜋∕2
condition the frequency-directional spectrum can be solved for (Booij 𝐻𝑟𝑚𝑠,ℎ𝑖,𝑐 = 2 2 𝐸(𝜔, 𝜃)𝑑𝜃𝑑𝜔 (30)
∫𝜔𝑐 ∫−𝜋∕2
et al., 1999).
To include infragravity waves in SWAN with frequencies less or where 𝜃 is defined with respect to the shore normal. The concurrent
equal to 𝜔𝑐 Eq. (15) is used, which is representative for the forcing WGF RMS wave height is given by:
by a single bichromatic group. However, in the presence of a realistic
√
directionally spread sea-swell wave field all frequency-directional pairs √ 𝜔𝑐 𝜋∕2
with a difference frequency less or equal to the infragravity frequency 𝐻𝑟𝑚𝑠,𝑙𝑜,𝑓 =2 2 𝐸𝑓 (𝛥𝜔, 𝜃)𝑑𝜃𝑑𝛥𝜔 (31)
∫𝜔𝑚𝑖𝑛 ∫−𝜋∕2
cutoff, 𝜔𝑐 , create bichromatic wave groups contributing to the WGF
source term 𝑆𝑠𝑏 . For a specific combination within the frequency- with 𝜔𝑚𝑖𝑛 as the minimum angular infragravity frequency.
directional spectrum this contribution to the radiation stresses can be
expressed as: 2.3.2. Reflected infragravity energy
( )
𝑐𝑔 1 The sea-swell forced infragravity waves are assumed to reflect at the
𝑆̂𝑥𝑥 (𝛥𝜔, 𝛥𝑘𝑦 ) = 𝜌𝑔 (1 + cos2 𝛼) − 𝑎𝑝,𝑛 𝑎𝑞,𝑚 (22) shoreline. This is achieved by defining an outgoing infragravity energy
𝑐 2
flux at a minimal depth, ℎ𝑐 :
and
𝑐𝑔 1 √
𝑆̂𝑥𝑦 (𝛥𝜔, 𝛥𝑘𝑦 ) = 𝜌𝑔 (cos 𝛼 sin 𝛼)𝑎𝑝,𝑛 𝑎𝑞,𝑚 (23) 𝐹𝑟,ℎ𝑐 = 𝜌ℎ𝑐 𝑔ℎ𝑐 𝑟2𝑐 𝑈̂ 𝑓2,ℎ 𝑐𝑜𝑠(𝜃𝑟 ) (32)
𝑐 2 𝑐

where the difference frequency is given by 𝛥𝜔 = 𝜔𝑞 − 𝜔𝑝 and the where 𝑟𝑐 is a reflection coefficient. The outgoing angle of the reflected
difference alongshore wave number is given by 𝛥𝑘𝑦 = 𝑘𝑞 𝑠𝑖𝑛(𝜃𝑚 ) − infragravity wave is computed according to:
𝑘𝑝 𝑠𝑖𝑛(𝜃𝑛 ), with 𝑝 and 𝑞 referring to the frequency bins and 𝑛 and 𝑚
to the directional bins in the offshore frequency-directional sea-swell 𝜃𝑟 = 𝜋 − 𝜃𝑓 (33)
spectrum. The corresponding amplitudes are obtained from:
√ assuming specular reflection at a shoreline aligned with the 𝑦-direction.
𝑎𝑖,𝑗 = 2𝐸(𝜔𝑖 , 𝜃𝑗 )𝛿𝜔𝛿𝜃 (24) Given the fact that the outgoing infragravity waves are free they can
with 𝑖 and 𝑗 representing the frequency and directional bins and 𝛿𝜃 be modeled with the regular wave energy balance equation, Eq. (21),
and 𝛿𝜔 corresponding to the directional and frequency resolution of applied to the reflected infragravity waves:
the sea-swell spectrum computed with Eq. (21). Instead of tracing the 𝜕𝐸𝑟 (𝛥𝜔, 𝜃)𝑐𝑔,𝑥 𝜕𝐸𝑟 (𝛥𝜔, 𝜃)𝑐𝜃
slow amplitude evolution of the individual frequency-directional wave + =0 (34)
𝜕𝑥 𝜕𝜃
components, 𝑎𝑝,𝑛 and 𝑎𝑞,𝑚 , the cross-shore distribution of the radiation thus simulating inverse shoaling and refraction. Depth-limited wave
stress amplitudes is obtained by scaling the offshore modulation of each breaking is assumed not to affect the dissipation of reflected infragrav-
pair with changes in the total variance according to: ity waves and hence absent in the infragravity energy balance Eq. (34).
1
𝛴 𝛴 𝑎2 (𝑥)
2 𝑖 𝑗 𝑖,𝑗 ( ) The reflected infragravity rms wave height is given by:
𝑎𝑝,𝑚 𝑎𝑞,𝑚 (𝑥) = 𝑎𝑝,𝑚 𝑎𝑞,𝑚 |𝑥=0 1 − 𝑄𝑏 (𝑥) (25) √
𝛴𝑖 𝛴𝑗 12 𝑎2𝑖,𝑗 |𝑥=0 √ 𝜔𝑐 𝜋
𝐻𝑟𝑚𝑠,𝑙𝑜,𝑟 = 2 2 𝐸 (𝛥𝜔, 𝜃)𝑑𝜃𝑑𝛥𝜔 (35)
∫𝜔𝑚𝑖𝑛 ∫−𝜋 𝑟
Furthermore, to account for the fact that the modulation is partly
reduced in the breakerzone, e.g. Schäffer and Svendsen (1988) and which is integrated over the full directional width to account for po-
De Bakker et al. (2015), the modulation is reduced proportional to the tential refractive trapping. The predicted total RMS infragravity wave
fraction of breaking waves, 𝑄𝑏 , obtained with (Baldock et al., 1998): height is given by the square root of the combined variances:
( ( ) ) √
𝐻𝑚𝑎𝑥 2 2 2
𝑄𝑏 = exp − (26) 𝐻𝑟𝑚𝑠,𝑙𝑜,𝑐 = 𝐻𝑟𝑚𝑠,𝑙𝑜,𝑓 + 𝐻𝑟𝑚𝑠,𝑙𝑜,𝑟 (36)
𝐻𝑚0
thus ignoring phase locking between the incoming and outgoing infra-
with 𝐻𝑚0 as the significant sea-swell wave height. gravity waves associated with standing wave patterns.
Within SWAN the phase and group velocity in Eqs. (15), (22) and
(23), have been computed with the mean wave period, 𝑇𝑚−10 , using
2.4. Laboratory observations
linear wave theory where:
𝜔 2𝜋
2𝜋 ∫𝜔 𝑚𝑎𝑥 ∫0 𝜔−1 𝐸(𝜔, 𝜃)𝑑𝜃𝑑𝜔 The laboratory observations were obtained during the LIP11D pro-
𝑐
𝑇𝑚−10 = (27) totype scale morphodynamic experiments in the Delta Flume (Arcilla
𝜔 2𝜋
∫𝜔 𝑚𝑎𝑥 ∫0 𝐸(𝜔, 𝜃)𝑑𝜃𝑑𝜔
𝑐 et al., 1994; Roelvink and Reniers, 1994). These experiments consid-
with 𝜔𝑐 as the infragravity frequency cutoff and 𝜔𝑚𝑎𝑥 as the maximum ered a range in wave conditions corresponding to random unidirec-
sea-swell frequency. For a specific pair of incident waves Eq. (15) tional sea and swell waves represented by a Jonswap spectrum (see
yields the cross-shore distribution of the WGF velocity, 𝑈̂ 𝑓 (𝛥𝜔, 𝜃𝑓 ) and Table 1).

4
A. Reniers and M. Zijlema Coastal Engineering 172 (2022) 104068

Fig. 1. Three-hourly FRF-8 m array conditions in October 1990. Panel A: Root mean square sea-swell wave height. Panel B: 𝑇𝑚,−10 wave period. Panel C: Mean wave direction
with respect to the shore normal. Panel D: Tidal elevation. Right panel: FRF (black dot) and array (red dots) instrument locations. Depth contour lines (white) in meters below
mean sea level. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Table 1 from the Field Research Field (FRF) pressure array at 8 m water depth
Overview of LIP11D test wave conditions with 𝐻𝑚0,ℎ𝑖 and 𝑇𝑝 as the significant sea-swell
(right panel of Fig. 1) at a three-hourly interval. During this time
wave height and the peak period respectively.
the incident wave conditions ranged from mild to moderate with a
Test case 1C0102 2A0102 2B0506
maximum root mean square wave height of 1.75 m (Panel A in Fig. 1).
𝐻𝑚0,ℎ𝑖 (m) 0.6 1.0 1.4
Local wave conditions were predominantly due to a North-Easterly
𝑇𝑝 (s) 8 5 5
swell resulting in long period waves (panel B in Fig. 1) with an oblique
incidence angle of approximately 25◦ with the shore normal. The tidal
range was in the order of 1 m and predominantly vertical (panel D in
The wave paddle was operated for an hour for a specific wave Fig. 1). The wave transformation was measured with a cross-shore array
condition after which the bed elevation changes were measured to of pressure sensors starting at an approximate water depth of 4 m (right
monitor the profile evolution. The hourly wave transformation was panel of Fig. 1).
measured with ten fixed pressure sensors operating at 10 Hz. Using Linear wave theory is used to translate the corresponding pressure
linear wave theory the pressure spectrum was translated to the surface spectra to surface elevation spectra from which the root mean square
elevation, 𝐸𝜂𝜂 , to obtain the sea-swell and infragravity wave heights sea-swell and infragravity wave heights have been computed according
given by: to Eqs. (37) and (38) using the same frequency intervals. In addition a
√
√ 0.3 Hz bi-spectral analysis (Hasselmann et al., 1963; Herbers et al., 1994), is
𝐻𝑟𝑚𝑠,ℎ𝑖,𝑚 = 2 2 𝐸 (𝑓 )𝑑𝑓 (37) used to compute the bound root mean square infragravity wave height:
∫0.05 Hz 𝜂𝜂
with the subscript ℎ𝑖 referring to the sea-swell band and 𝑚 denotes √
√ 0.04 Hz
measurements. The corresponding root mean square low frequency or 𝐻𝑟𝑚𝑠,𝑏,𝑚 =2 2 𝐸𝜂𝑏 𝜂𝑏 (𝑓 )𝑑𝑓 (39)
infragravity wave height is given by: ∫0.01 Hz
√
√ 0.04 Hz where the bound variance density at the infragravity frequency 𝛥𝑓 is
𝐻𝑟𝑚𝑠,𝑙𝑜,𝑚 = 2 2 𝐸 (𝑓 )𝑑𝑓 (38) given by:
∫0.01 Hz 𝜂𝜂
0.3−𝛥𝑓
2 ∫𝛥𝑓 𝐵(𝑓 , 𝛥𝑓 )𝑑𝑓
At five locations colocated measurements of velocity have been used 𝐸𝜂𝑏 𝜂𝑏 (𝛥𝑓 ) = (40)
to separate the incoming and outgoing infragravity wave heights using [ ]1
0.3−𝛥𝑓 2
2 ∫𝛥𝑓 𝐸(𝑓 + 𝛥𝑓 )𝐸(𝑓 )𝐸(𝛥𝑓 )𝑑𝑓
the method of Guza et al. (1984), allowing a direct comparison with
the model predictions of the incoming and reflected infragravity wave with 𝐵 as the bispectrum and a high-frequency cutoff of 0.3 Hz to avoid
heights given by Eqs. (31) and (35) respectively. This method uses the excessive amplification due to depth attenuation at higher frequencies.
shallow water phase speed for both incoming and reflected IG waves,
whereas bound IG waves propagate with the group velocity. As the
deepest sensor is located in approximately 2 m water depth the dif- 2.6. Error metrics
ferences between the group velocity and the shallow water wave speed
at all sensor locations is less than 15% for the conditions considered To evaluate the model performance three error metrics are consid-
here. This results in minor differences in the estimated incoming and ered. The root mean square error is given by:
reflected IG wave heights if the group velocity is used for the incoming √
IG waves. 𝜖(𝐻) = (𝐻𝑝 − 𝐻𝑜 )2 (41)

where the subscript 𝑝 refers to the predictions and 𝑜 to the observa-

2.5. Field observations tions of variable 𝐻. The overbar indicates averaging over all relevant
observations. The bias between predictions and observations is given
The field data were collected during an eleven day period in October by:
of 1990 at Duck, North Carolina as part of the DELILAH field experi-
ment. Offshore sea-swell frequency-directional spectra were obtained 𝑏(𝐻) = 𝐻𝑝 − 𝐻𝑜 (42)

5
A. Reniers and M. Zijlema Coastal Engineering 172 (2022) 104068

Fig. 2. Left Panels: IAM predictions for a normally incident bi-chromatic wave train with frequency difference of 0.04 Hz. Panel A: Snapshot of wave-group varying wave height
(solid line) and corresponding envelope (dashed line). Panel B: Snapshot of concurrent WGF infragravity surface elevation (solid line) and corresponding envelope (dashed lines).
Panel C: Similar for the cross-shore infragravity velocity. Panel D: Corresponding bed elevation. Right panels: Panel A: IAM-predicted (thick black line) and IEM-predicted (thick
red line) WGF infragravity wave amplitude. Equilibrium solution of Longuet-Higgins and Stewart (1964) is given by the open circles as a reference. Panel B: IAM-predicted phase
of the wave group forcing (dashed) and the accompanying WGF infragravity wave (solid). Panel C: IAM-predicted bi-phase (black) compared with IEM-prediction (red).

The corresponding scatter index is given by: Table 2

Deep water conditions for the numerical experiments of Herbers and Burton (1997).
𝜖(𝐻) The frequency increment in subset 1 is 0.005 Hz whereas the directional increment in
𝑠𝑐𝑖(𝐻) = 1 − (43)
𝑚𝑎𝑥(𝜎(𝐻𝑜 ), 𝐻̄ 𝑜 ) subset 2 is 5◦ . Deep water wave amplitudes are set at 0.2 m for all components.
Subset 𝑓1 (Hz) 𝑓2 (Hz) 𝜃1 (◦ ) 𝜃2 (◦ )
with 𝜎 as the standard deviation.
1 From 0.095 to 0.075 From 0.105 to 0.1250 0 0
2 0.08 0.12 0 From 0 to 60
3. Verification and validation

For the verification of the Infragravity Energy Model (IEM) the

infragravity response to a simple wave-group forcing made up of two height modulation varies between 0 and 0.8 m offshore (left panel A
incident sea-swell components of different frequencies and directions of Fig. 2). Wave breaking starts around 𝑥 = 900 m after which the
is compared with the IAM benchmark results. Model validation is per- modulation amplitude quickly decreases. Both the IAM-predicted WGF
formed in two steps. First by comparing the predicted, unidirectional infragravity surface elevation and velocity increase non-linearly with
infragravity response to normally incident random waves with large decreasing depth (left panels B and C in Fig. 2). The IAM-predicted WGF
scale laboratory observations obtained during the LIP11D experiments infragravity amplitude is generally lower than the equilibrium solution
in the Delta Flume as described in Section 2.4. Secondly by comparing of Longuet-Higgins and Stewart (1964) (right panel A in Fig. 2). This
the predicted directional infragravity response to frequency-directional is consistent with the results of Schäffer (1993), the observations of
sea-swell forcing observed during the Delilah field experiment at Duck, Battjes et al. (2004) and the experiments by Herbers and Burton (1997).
North Carolina as described in Section 2.5. Computing the difference in phase between the forced infragravity
wave and the corresponding wave group forcing (right panel B of
3.1. Benchmark testing Fig. 2) yields the bi-phase at every location (right panel C in Fig. 2).
Next we solve for the coupled wave energy flux and bi-phase with
To compare the IEM with the IAM predictions of the WGF infra- the IEM showing a favorable comparison with both the IAM-computed
gravity waves the setup of the numerical experiments performed by WGF infragravity wave amplitude (right panel A of Fig. 2) and the
Herbers and Burton (1997) is used. These experiments reflect typical corresponding bi-phase (right panel C of Fig. 2).
conditions at Duck, NC, with a constant slope of 0.01 starting at a The next example considers a bi-chromatic wave made up of the
water depth op 10 m (panel D on the left in Fig. 2). Herbers and same two sea-swell frequencies of 0.08 Hz and 0.12 Hz where the
Burton (1997) considered intermediate depths only, ranging from 10 m latter is now obliquely incident at an angle of 20◦ (see Fig. 3). Given
offshore to 4 m onshore. Here the model domain is extended to the the oblique incidence angle the cross-shore difference wave number is
water line with a minimal depth of 0.1 m with a constant grid spacing smaller resulting in slightly longer wave groups and a weaker response
of 1 m. The wave group forcing is made up of two sea-swell components in both the infragravity surface elevation and cross-shore velocity
with a range of different frequencies and directions. The first subset response (compare left panels of Fig. 3 with Fig. 2). Although the
consists of normally incident wave groups with a frequency difference equilibrium solution for obliquely incident sea-swell waves given by
ranging between 0.01 and 0.05 Hz centered around a mean frequency Reniers et al. (2002) is significantly smaller compared with the nor-
of 0.1 Hz. The second subset is made up of obliquely incident wave mally incident wave groups (compare the right panels in Figs. 2 and 3)
groups with a fixed frequency difference of 0.04 Hz centered around a it generally still overestimates the WGF infragravity wave amplitude.
mean frequency of 0.1 Hz. In these latter experiments the longer sea- Comparing the IEM-predictions with the IAM results shows a good
swell component is kept at normal incidence whereas the shorter one comparison for both the bi-phase and amplitude evolution (right panels
varies between normally incident and 60◦ in deep water. The conditions A and C in Fig. 3). Excluding the obliqueness of the WGF infragravity
for both subsets are summarized in Table 2. wave results in an overprediction of the WGF infragravity wave height
Starting with a normally incident wave group made up of two sea- (compare with green line in right Panel A).
swell waves with frequencies of 0.08 Hz and 0.12 Hz respectively and Analogous to Herbers and Burton (1997) the WGF infragravity
corresponding deep water amplitudes of 0.2 m (see Table 2). The wave amplitude at 4 m water depth is compared for a range in difference

6
A. Reniers and M. Zijlema Coastal Engineering 172 (2022) 104068

Fig. 3. Left panels: IAM predictions for an obliquely incident bi-chromatic wave train with frequency difference of 0.04 Hz and a deep water difference angle of 20◦ . Panel A:
Snapshot of wave-group varying wave height (solid line) and corresponding envelope (dashed line). Panel B: Snapshot of concurrent WGF infragravity surface elevation (solid
line) and corresponding envelope (dashed lines). Panel C: Similar but for the cross-shore infragravity velocity. Panel D: Corresponding bed elevation. Right panels: Panel A:
IAM-predicted (thick black line) and IEM-predicted (thick red line) WGF infragravity wave amplitude. Equilibrium solution of Longuet-Higgins and Stewart (1964) given by the
squares. IEM-predictions without taking obliqueness of infragravity waves into account (green line). Panel B: IAM-predicted phase of the wave group forcing (dashed) and the
accompanying WGF infragravity wave (solid). Panel C: IAM-predicted bi-phase (black) compared with IEM-prediction (red).

Fig. 4. Left panel: Forced infragravity amplitude at 4 m water depth predicted by IAM (black dots), IEM (red dots) and the equilibrium solution of Longuet-Higgins and Stewart
(1964) for normally incident bi-chromatic waves as function of the difference frequency. Right panel: Forced infragravity amplitude at 4 m water depth predicted by IAM (black
dots), IEM (red dots) and the equilibrium solution for obliquely incident bi-chromatic waves as function of the deep water difference angle. IEM-predictions without taking
obliqueness of infragravity waves into account (green dots).

frequencies and directions. Computing the WGF infragravity ampli- 3.2. Laboratory validation
tude for normally incident bi-chromatic wave for increasing difference
The following describes the first part of the validation of the WGF
frequencies, corresponding to subset 1 in Table 2, shows a gradual
infragravity predictions with SWAN for unidirectional random waves.
increase in both the IAM and IEM results (left panel of Fig. 4). The Model simulations use a constant cross-shore grid spacing of 2.5 m, a
largest difference with the equilibrium solution occurs for the smallest 0.05 Hz–0.3 Hz frequency range for the sea-swell waves and a 0.01 Hz-
difference frequency, consistent with the results obtained by Battjes 0.04 Hz frequency range for the infragravity waves, both with a 0.01 Hz
frequency resolution and 2 degree directional resolution. The normally
et al. (2004). Keeping the difference frequency at 0.04 Hz but gradually
incident waves are represented by a very narrow banded directional
increasing the deep water spreading angles, corresponding to subset 2 spectrum with a directional spreading of O(3)◦ . Both the wave breaking
in Table 2, shows a significant decay in the predicted WGF infragravity parameter 𝛾𝑏 and the reflection coefficient 𝑟𝑐 are used as calibration
amplitude at 4 m water depth consistent with the numerical results coefficients.
Starting with test 1C0102 corresponding to swell conditions (see Ta-
of Herbers and Burton (1997) (compare right panel in Fig. 4 with
ble 1). The swell wave height prediction with 𝛾𝑏 = 0.5 closely matches
their Figs. 2a and 3a). The mismatch with the equilibrium solution the observations with the onset of breaking at the bar crest, subsequent
decreases with increasing directional spreading. Overall the difference cessation in the trough and additional breaking close to shore (see left
between the IAM and IEM predicted amplitudes at 4 m depth are small. panel A of Fig. 5) resulting a mean rms error of less than 0.02 m
(see Table 3). The computed infragravity response shows a gradual
Excluding the directional effects in estimating the WGF infragravity
increase in the total wave height consistent with the observations (left
response results in a clear overestimation (compare with green dots in Panel B of Fig. 5) with the largest mismatch close to shore. Both the
the right panel of Fig. 4). reflected and WGF infragravity wave height compare favorably with

7
A. Reniers and M. Zijlema Coastal Engineering 172 (2022) 104068

Fig. 5. Left panels: Wave transformation for test 1C0102. Panel A: Predicted (solid black line) and measured (black dots) root mean square swell wave height. Bottom profile
given as a reference. Panel B. Predicted total (black), WGF (red) and reflected (blue) root mean square infragravity wave height. Measurements are indicated by the co-colored
markers. Right Panels: Similar but for test 2B0506.

Table 3 Model simulations use a constant cross-shore grid spacing of 5 m,

Error statistics of the sea-swell and infragravity transformation for the LIP11D
a 0.05 Hz–0.3 Hz frequency range for the sea-swell waves and a
laboratory experiments given in Table 1.
0.01 Hz–0.04 Hz frequency range for the infragravity waves, both with
LIP11D Test 1C0102 2B0506 2A0102
a 0.01 Hz frequency resolution and 10 degree directional resolution.
𝛾𝑏 0.5 0.6 0.6 The wave breaking parameter 𝛾𝑏 is used as a calibration parameter and
𝑟𝑐 0.89 0.8 0.92
𝜖(𝐻𝑟𝑚𝑠,ℎ𝑖 ) (cm) 2.0 2.5 2.7
the reflection coefficient 𝑟𝑐 is set to one. The model is forced with the
b(𝐻𝑟𝑚𝑠,ℎ𝑖 ) (cm) 1.1 −0.5 2.2 measured 3-hourly frequency-directional spectra. Daily surveys of the
sci(𝐻𝑟𝑚𝑠,ℎ𝑖 ) 0.06 0.05 0.06 bathymetry are used to construct the profile at the instrument array.
𝜖(𝐻𝑟𝑚𝑠,𝑙𝑜,𝑐 ) (cm) 1.2 2.0 1.7 The conditions during Oct15 at 10 pm are taken as an example. The bed
b(𝐻𝑟𝑚𝑠,𝑙𝑜,𝑐 ) 0.4 0.1 −0.1 profile on this day shows a typical sand bar located around 𝑥 = 700 m
sci(𝐻𝑟𝑚𝑠,𝑙𝑜,𝑐 ) 0.13 0.14 0.13 and a trough around 𝑥 = 800 m (panel C in Fig. 6). The frequency-
𝜖(𝐻𝑟𝑚𝑠,𝑙𝑜,𝑓 ) (cm) 0.9 3.3 2.5 directional spectrum is relatively narrow banded in frequency, with
b(𝐻𝑟𝑚𝑠,𝑙𝑜,𝑓 ) −0.1 −2.4 −2.2 a peak frequency at 0.095 Hz, combined with a directionally broad
sci(𝐻𝑟𝑚𝑠,𝑙𝑜,𝑓 ) 0.23 0.25 0.27 distribution, with a peak direction of −20◦ with respect to the shore
𝜖(𝐻𝑟𝑚𝑠,𝑙𝑜,𝑟 ) (cm) 1.0 0.7 0.9
normal (right panel in Fig. 6). The corresponding offshore root mean
b(𝐻𝑟𝑚𝑠,𝑙𝑜,𝑟 ) 0.1 −0.5 −0.8
square wave height is 0.55 m with a tidal elevation of −0.3 m with
sci(𝐻𝑟𝑚𝑠,𝑙𝑜,𝑟 ) 0.16 0.07 0.10
respect to mean sea level.
The SWAN-predicted sea-swell wave height transformation, ob-
tained with Eq. (21), shows a good match with the observations with
the measurements showing a steeper increase for the WGF response wave breaking on the bar and near the shore line. The corresponding
compared with the reflected infragravity waves, consistent with the WGF infragravity spectrum, given by Eq. (29), is integrated over the
non-linear forcing by the wave groups and inverse shoaling respectively infragravity frequency band, (0.01 Hz ≤ 𝛥𝑓 ≤ 0.04 Hz), showing a
(Battjes et al., 2004). The accompanying rms error estimates are in the strong increase with decreasing depth (panel A) consistent with the
order of 1 cm (see Table 3). benchmark testing. The directional spreading of the WGF infragravity
The wave transformation for test 2B0506 corresponds to erosive sea is restricted to incident wave angles only, i.e. (−90◦ < 𝜃 < 90◦ ).
conditions with an increased wave height and reduced wave period (see Offshore, the mean WGF direction aligns with the sea-swell waves of
Table 1). The waves start breaking at the beginning of the sloping bed approximately −20◦ (right panel of Fig. 6) gradually moving towards
and continue breaking all the way to the shoreline. Model predictions the shore normal as the sea-swell waves refract with decreasing depth.
obtained with a 𝛾𝑏 of 0.6 are consistent with the measurements with a Full reflection of the WGF infragravity energy is imposed at a water
mean rms error of less than 3 cm (see Table 3). The predicted total depth of 0.1 m. The reflected infragravity wave variance, computed
infragravity response shows an underprediction offshore of the bar with Eq. (34), shows a much broader directional distribution with
crest and an overprediction onshore caused by the mismatch of the a weak depth-dependent decay consistent with inverse shoaling and
refraction. Refractive trapping is apparent at the shore line and over
WGF infragravity wave height resulting in a rms error of O(2) cm. The
the bar (Bryan et al., 1998; Rijnsdorp et al., 2015), with reflected
steep increase in the WGF infragravity wave height observed in test
infragravity energy present at onshore wave directions.
1C0102 is mostly absent consistent with the reduced modulation by
Further integrating the WGF-infragravity spectrum over the direc-
wave breaking represented by Eq. (25). Similar results are obtained
tions yields the SWAN-predicted WGF infragravity root mean square
for test 1A0102 (not shown) with the error statistics summarized in
wave height (red line in panel D in Fig. 6). At the deepest sensor
Table 3 showing a favorable model performance to normally incident
location, CM90 (𝑥 ∼ 540) m (c.f. Fig. 1), the WGF predictions are
wave forcing.
close to the bispectrally estimated bound infragravity wave height,
suggesting a near balance between the WGF forcing and the accom-
3.3. Field validation panying bound infragravity wave. Closer to the shore line the WGF
infragravity wave height increases further, whereas the observed bound
Next the infragravity response to directionally spread sea-swell infragravity wave height decays, consistent with the release of the
waves is examined comparing SWAN predictions with field observa- bound infragravity wave and subsequent regular shoaling as the wave
tions obtained at Duck, North Carolina. group forcing reduces due to breaking. Including the contribution of

8
A. Reniers and M. Zijlema Coastal Engineering 172 (2022) 104068

Fig. 6. Left panels: Wave transformation on October 15 at 10 pm. Panel A: Frequency-integrated WGF infragravity variance (logarithmic color scale in m2 ∕◦ ). Panel B: Reflected
infragravity variance. Panel C: 𝐻𝑟𝑚𝑠,ℎ𝑖,𝑐 (black line) compared with observations 𝐻𝑟𝑚𝑠,ℎ𝑖,𝑚 (black dots). Panel D: WGF 𝐻𝑟𝑚𝑠,𝑙𝑜,𝑓 (red line) compared with bound infragravity wave
height (orange dots). Total predicted infragravity height 𝐻𝑟𝑚𝑠,𝑙𝑜,𝑐 (black line) compared with observations of 𝐻𝑟𝑚𝑠,𝑙𝑜,𝑐 (black dots). Reflected 𝐻𝑟𝑚𝑠,𝑙𝑜,𝑟 given as a reference (blue line).
Right panel: Sea-swell frequency-directional spectrum at FRF 8-m array for October 15 at 10 pm. Direction 𝜃 is with respect to the shore normal. (For interpretation of the
references to color in this figure legend, the reader is referred to the web version of this article.)

Table 4
Error statistics for the predicted infragravity conditions at CM90 for the total 𝐻𝑟𝑚𝑠𝑙 𝑜𝑐
for a range in frequency bands corresponding to Fig. 7.
f (Hz) 𝜖(𝐻𝑟𝑚𝑠𝑙 𝑜𝑐 ) (cm) b(𝐻𝑟𝑚𝑠𝑙 𝑜𝑐 ) (cm) sci(𝐻𝑟𝑚𝑠𝑙 𝑜𝑐 )
0.01 Hz 1.6 −1.3 0.26
0.02 Hz 1.7 −1.4 0.21
0.03 Hz 1.1 −0.1 0.15
0.04 Hz 0.9 −0.2 0.14
0.01–0.04 Hz 2.2 −1.4 0.16

frequencies, 𝑓 ≥ 0.03 Hz, is well predicted by the model (panels C and

D of Fig. 7) with an 𝜖 of O(1) cm and a scatter index 𝑆𝐶𝑖 of O(0.15)
(Table 4). For the lower infragravity frequencies, 𝑓 ≤ 0.02 Hz, the total
infragravity wave height is typically underestimated during moderate
sea-swell conditions (panels A and B of Fig. 7) resulting in a bias of
approximately 1.5 cm and an increase in 𝑆𝐶𝑖 to O(0.25) (Table 4).
The predicted WGF-infragravity wave height compares well with the
observations for all frequency bands, provided the sea-swell forcing
is mild. During storm conditions the WGF-infragravity wave height
exceeds the observed bound infragravity wave height consistent with
the release and subsequent regular shoaling as observed for Oct 15 at 10
pm (Panel D in Fig. 6). Fully integrating the three-hourly infragravity
frequency spectra shows a good match with the observations for the
total infragravity wave height (lower panel in Fig. 7) with a rms-error
of O(2) cm, with an underprediction during moderate conditions due
Fig. 7. Infragravity wave conditions at CM90 (see Fig. 1 for the location). Predicted to the diminished low frequency response (panels A and B in Fig. 7)
total (black) and WGF (red) infragravity wave height compared with observations resulting in a bias of −1.4 cm with an overall 𝑆𝐶𝑖 of 0.16 (Table 4).
of total (black markers) and bound (orange markers) infragravity wave heights
The model performance at the instrument locations closer to shore is
respectively at frequencies 0.01 Hz (Panel A), 0.02 Hz (panel B), 0.03 Hz (panel C) and
0.04 Hz (panel D) with a 0.01 Hz frequency bandwidth. Bottom panel: Corresponding considered next (see Fig. 1 for the locations). The frequency dependent
frequency-integrated, 0.01 Hz–0.04 Hz, infragravity wave heights. results show significant scatter around the optimal fit corresponding
to the diagonals in the left panels of Fig. 8. The mismatch with the
the reflected infragravity waves shows a good match with the measured observations seems to alternate between over/underpredictions for a
total infragravity wave height, especially for the deeper sensor locations given frequency band depending on the location of the instrument,
(black line in panel D). consistent with the presence of nodes and anti-nodes in the surface
The frequency-dependent infragravity response for all three-hourly elevation. This results in a small bias but a significantly increased scat-
predictions at location CM90 is examined next. To that end the pre- ter compared to the results obtained at CM90 (compare Tables 4 and
dicted infragravity wave height for a 0.01 Hz frequency bandwidth 5). This mismatch is reduced by integrating over the full infragravity
ranging from 0.01 Hz to 0.04 Hz is compared with the observations. spectrum (right panel in Fig. 8) resulting in an overall root mean square
On average the total infragravity wave height in the higher infragravity error of O(4) cm and a scatter index of 0.22 (see Table 5).

9
A. Reniers and M. Zijlema Coastal Engineering 172 (2022) 104068

Fig. 8. Comparison of the predicted and measured total infragravity wave height at all sensor locations. Gray scale corresponds to the distance from the shore line (where the black
dots correspond to CM90). Results are shown for the individual frequency bands of 0.01, 0.02, 0.03 0.04 Hz with a band width of 0.01 Hz in panels A, B, C and D respectively
and the total infragravity band from 0.01–0.04 Hz in the larger panel on the right.

Table 5 (refer to Guza and Thornton (1985) for improved estimates of standing
Error statistics for the predicted infragravity conditions at all sensors combined for the
infragravity wave patterns). Note that applying the same method for
total 𝐻𝑟𝑚𝑠𝑙 𝑜𝑐 for a range in frequency bands corresponding to Fig. 8.
the DELILAH field experiment is not straightforward as the cross-shore
F (Hz) 𝜖(𝐻𝑟𝑚𝑠𝑙 𝑜𝑐 ) (cm) b(𝐻𝑟𝑚𝑠𝑙 𝑜𝑐 ) (cm) sci(𝐻𝑟𝑚𝑠𝑙 𝑜𝑐 )
wave number now not only depends on the infragravity incidence angle
0.01 Hz 2.7 −1.5 0.35 of both the incoming and outgoing infragravity waves but also the
0.02 Hz 2.7 −1.0 0.29
refractive trapping that is present (panel B in Fig. 6).
0.03 Hz 2.7 0.2 0.30
0.04 Hz 2.3 0.1 0.29
Model predictions for DELILAH typically demonstrate an underesti-
0.01–0.04 Hz 3.9 −1.1 0.22 mation of the energy in the lower infragravity frequency bins (panels
A and B in Fig. 7 and Tables 4 and 5), especially at times of moderate
sea-swell conditions. In the current approach it is assumed that the
incoming infragravity energy is in equilibrium with the directionally
4. Discussion spread sea-swell forcing at the FRF 8 m array. This ignores any free
incident infragravity waves that originate from elsewhere (Bromirski
In the computations phase locking between the WGF and reflected et al., 2010; Ardhuin et al., 2014; Neale et al., 2015; Vrećica et al.,
infragravity energy is absent (Eq. (36)). This is expected to hold further 2019; Rijnsdorp et al., 2021). Ardhuin et al. (2014) use a empirical
away from the shore line but not close to the reflection point (Guza and approach to obtain the outgoing infragravity waves at the coast as a
Thornton, 1985). This inhibits a proper comparison with the observa- boundary condition for their global infragravity modeling. By coupling
tions of the total infragravity wave height closer to the water line using with a regional model like SWAN, a potentially more accurate predic-
Eq. (36). For normally incident regular infragravity waves with radial tion can be made of the outgoing infragravity waves in both frequency
frequency 𝜔 the expected standing surface elevation amplitude, 𝐴, is distribution and direction. This two-way coupling is expected to lead
given by: to improved predictions of the incoming infragravity waves that can
( 𝑋 )
subsequently be used by other models to examine the coastal safety
𝐴(𝜔, 𝑋) = 𝐴(𝜔)|𝑋=0 cos 𝑘(𝑋)𝑑𝑋 related to for instance beach and dune erosion as well overwash and
∫𝑋=0
( ) breaching.
𝑋
𝜔
≃ 𝐴|𝑋=0 cos √ 𝑑𝑋 (44) Model predictions have been performed without taking bed friction
∫𝑋=0 𝑔ℎ(𝑋)
into account. This compensates partly for the omission in the radiation
with 𝑋 starting at the reflection point being positive offshore. This stress forcing related to the roller contribution as demonstrated by
approach assumes that the infragravity waves close to shore have been Reniers et al. (2002). Additional energy dissipation can occur by infra-
released and behave as free infragravity waves. The cross-shore struc- gravity wave breaking within the swash zone (Henderson et al., 2006;
ture within a certain frequency bandwidth can then be approximated Van Dongeren et al., 2007; Lin and Hwung, 2012; De Bakker et al.,
by integrating Eq. (44) over the appropriate frequencies and scaling 2014). In the LIP11D laboratory cases this is represented with a reduced
with the sum of the predicted incoming, 𝐻𝑟𝑚𝑠,𝑙𝑜,𝑓 and reflected, 𝐻𝑟𝑚𝑠,𝑙𝑜,𝑟 , reflection coefficient ranging between 0.8 and 0.93 (Table 3. For the
infragravity wave heights for that frequency bandwidth. Applying this Delilah field experiment the reflection coefficient has been set to unity,
to LIP11D test 1C0102 for a frequency bandwidth of 0.01 Hz shows that although based on the work of Van Dongeren et al. (2007) a reduced
the cross-shore variability in the observations can be partly explained reflection coefficient for the higher infragravity frequency bands is
by the underlying standing wave patterns (left panels of Fig. 9). expected similar to what is observed in the laboratory experiment.
Integrating over the full infragravity frequency range from 0.01– Additional effects associated with the presence of an alongshore wave-
0.04 Hz the predicted total infragravity wave height, 𝐻𝑟𝑚𝑠,𝑙𝑜,𝑐 , shows driven current (Howd et al., 1992) and set-up have also not been
an improved match with the observations close to shore. This holds taken into account, although the latter is expected to be less important
for all tests (right panels in Fig. 9). Around 𝑥 = 100 m the nodal provided the sea-swell wave height is predicted well (Reniers et al.,
structure is mostly absent and Eq. (36) is valid. This means that for the 2002).
offshore sensors in both the laboratory and field experiments Eq. (36) Additional model limitations are related to the underlying assump-
is expected to hold, but that close to shore the predicted results for tion of alongshore bathymetric uniformity. This limits the applicability
the total infragravity wave height should be interpreted with care to coasts that have only mild alongshore variations in the bathymetry

10
A. Reniers and M. Zijlema Coastal Engineering 172 (2022) 104068

Fig. 9. Left panels: Infragravity wave heights for frequencies 0.01 Hz (panel A), 0.02 Hz (panel B), 0.03 Hz (panel C) and 0.04 Hz (panel D) with a 0.01 Hz bandwidth. Predicted
(red line) and observed (red markers) WGF heights, predicted (blue line) and observed (blue markers) reflected wave heights and predicted total infragravity wave heights with
(black dashed line) and without (black solid line) interference patterns and observations (black markers). Right panels: Similar but now for the frequency-integrated infragravity
wave height for tests 1C0102 (panel A), 2B0506 (panel B) and 1A0102 (panel C).

and shore line position. To overcome this problem a more general under consideration. The model application is restricted to alongshore
description of the infragravity source term, 𝑆𝑠𝑏 and associated bi-phase uniform beaches with a mildly sloping beach profile. In these cases it
𝛥𝜓, is required taking into account 2D bathymetric variability (Janssen, can be used to provide the combined sea-swell and infragravity bound-
2006) but retaining the current computational efficiency allowing for ary conditions for more sophisticated, but computationally expensive,
realistic applications. models at relatively shallow depths while reducing the error associated
The modeled transfer of wave energy to the WGF infragravity wave with the assumption of a local equilibrium between the WGF and the
is one-way, i.e. there is no equivalent loss of energy for the sea-swell bound infragravity wave. This is deemed especially important during
waves forcing the infragravity response, nor is there a mechanism to storm conditions that are used for coastal safety assessments as well as
extract energy from the infragravity waves and return it to the sea-swell engineering studies for harbor designs to optimize access and minimize
waves (e.g. Thompson et al., 2006; Henderson et al., 2006; De Bakker downtime.
et al., 2015). Additionally the effect of bore merging is absent (Sénéchal
et al., 2001a; Tissier et al., 2015, 2017). As a result the representation CRediT authorship contribution statement
of the infragravity dynamics in the inner surf and swash zone is limited.
Finally, the release of the bound infragravity waves is represented by Ad Reniers: Conceptualization, Methodology, Validation, Writing –
a reduction in the WGF depending on the fraction of breaking waves original draft. Marcel Zijlema: Software, Verification, Writing – review
in Eq. (25). This approach takes into account the changes in the WGF & editing.
forcing but does not account for the √ fact that the released infragravity
waves subsequently propagate with 𝑔ℎ instead of the group velocity Declaration of competing interest
𝑐𝑔 . In relatively shallow water, like the surfzones considered here, this
is not a problem as both velocities are similar, but it does not work The authors declare that they have no known competing finan-
properly in case the infragravity waves enter a deeper tidal channel cial interests or personal relationships that could have appeared to
after passing over for instance a shallow shoal. influence the work reported in this paper.

5. Conclusions Acknowledgments

Wave Group Forcing of infragravity waves has been implemented Data used in this study are provided by the Field Research Facility
in SWAN as a first step to predict the generation, propagation and of the US Army Engineer Waterways Experiment Station’s Coastal Engi-
evolution of infragravity waves at regional scales. The WGF is based neering Research Centre. Accessibility and permission to use these data
on a description of the energy balance of infragravity waves forced is appreciated very much. We also appreciate the discussions with Dr.
by directionally spread random waves at an alongshore uniform coast. Marion Tissier and Ir. Gal Akrish on the content of the work presented
The subsequent extension of the infragravity energy balance has been in this article.
verified with a benchmark comparison with the Infragravity Amplitude
Model developed by Reniers et al. (2002) and shows a good match for Appendix A. Derivation of the bi-phase evolution for obliquely
both normally incident and obliquely incident wave group forcing. The incident waves
subsequent implementation of the WGF in SWAN has been validated
with both large-scale laboratory experiments and field observations. The following is an extension of the work by Janssen et al. (2003)
The laboratory experiments consist of normally incident random waves by including the alongshore component of the incident and infragravity
representing both accretive and erosive conditions. The field observa- waves at an alongshore uniform beach, consistent with the modeling
tions correspond to the WGF by a fully directional sea-swell spectrum. approach of Reniers et al. (2002). This means that the WGF infragravity
In both cases the SWAN infragravity predictions show a favorable surface elevation is written as:
match with observations provided the measurement location is not af- 1
fected by the interference patterns created by the phase coupling of the 𝜂𝑓 (𝑥, 𝑦, 𝑡) = 𝜂̂𝑓 (𝑥) exp(𝑖(−𝛥𝜔𝑡 + 𝛥𝑘𝑦 𝑦))+ ∗ (A.1)
2
incident and reflected infragravity waves. As a result the mismatch with
and the radiation stress forcing as:
the observed total infragravity wave height generally increases closer to
shore although the specifics depend on the infragravity frequency band 𝑆𝑖𝑗 (𝑥, 𝑦, 𝑡) = 𝑆̂𝑖𝑗 exp(𝑖𝜓(𝑥)) exp(𝑖(−𝛥𝜔𝑡 + 𝛥𝑘𝑦 𝑦))+ ∗ (A.2)

11
A. Reniers and M. Zijlema Coastal Engineering 172 (2022) 104068
( )
which upon substitution in the linearized long wave equation, Eq. (1), 𝑑 𝑆̂𝑥𝑦
= 2𝛽𝑖𝛥𝑘𝑦 + 𝑖𝛽 −1 𝛥𝑘𝑥 𝑆̂𝑥𝑦 exp(𝑖𝜓) (A.15)
yields: 𝑑𝑋
𝜕 2 𝜂̂𝑓 𝜕ℎ 𝜕 𝜂̂𝑓 and lastly the third term:
𝑔ℎ +𝑔 + 𝛥𝜔2 𝜂̂𝑓 − 𝑔ℎ𝛥𝑘2𝑦 𝜂̂𝑓 = (A.3)
𝜕𝑥2 ( 𝜕𝑥 𝜕𝑥 ) −𝛥𝑘2𝑦 𝑆̂𝑦𝑦 𝑒𝑥𝑝(𝑖𝜓) = −𝛥𝑘2𝑦 𝑆̂𝑦𝑦 𝑒𝑥𝑝(𝑖𝜓) (A.16)
1 𝜕 𝑆̂𝑥𝑥 exp 𝜓
2 𝜕 𝑆̂𝑥𝑦 exp 𝜓 2 ̂
− + 2𝑖𝛥𝑘𝑦 − 𝛥𝑘𝑦 𝑆𝑦𝑦 exp 𝜓 Again collecting all the terms of the rhs up to order 𝛽 gives:
𝜌 𝜕𝑥2 𝜕𝑥
(
Following Janssen et al. (2003) a slow cross-shore spatial scale is 2𝑖𝛽 𝑑 𝑆̂𝑥𝑥 𝑖𝛽 𝑆̂𝑥𝑥 𝑑𝛥𝑘𝑥 𝑆̂𝑥𝑥
𝑔ℎ𝛥𝑘2𝑥 + −
introduced: 𝑔ℎ𝛥𝑘𝑥 𝑑𝑋 𝑔ℎ𝛥𝑘2𝑥 𝑑𝑋 𝑔ℎ
)
𝑋 = 𝛽𝑥 (A.4) 2𝑖𝛽𝛥𝑘𝑦 𝑑 𝑆̂𝑥𝑦 2𝛥𝑘𝑦 𝑆̂𝑥𝑦 𝛥𝑘2𝑦 𝑆̂𝑦𝑦
+ − − exp(𝑖𝜓) (A.17)
assuming that the changes in the water depth are small compared with 𝑔ℎ𝛥𝑘2𝑥 𝑑𝑋 𝑔ℎ𝛥𝑘𝑥 𝑔ℎ𝛥𝑘2𝑥
the WGF infragravity wave length expressed by the scaling parameter Returning to physical spatial variable 𝑥 and equating to the lhs with
𝛽: the rhs dividing both with 𝑔ℎ𝛥𝑘2𝑥 :
( )
1 𝑑ℎ
𝛽=𝑂 (A.5) ( ( ))
𝛥𝑘𝑥 ℎ 𝑑𝑥 𝑖 𝑑𝑇 𝑇 𝑑𝛥𝑘𝑥 𝑇 𝑑ℎ
−𝜇𝑇 + + 2 + exp 𝑖𝜓 = (A.18)
Next the complex surface elevation amplitude is written as a combi- 𝛥𝑘𝑥 𝑑𝑥 𝛥𝑘𝑥 𝑑𝑥 ℎ 𝑑𝑥
(
nation of the slow spatial scale 𝑋 and the faster phase changes at the 1 𝑆̂ 2𝛥𝑘𝑦 𝑆̂𝑥𝑦 𝛥𝑘2𝑦 𝑆̂𝑦𝑦 2𝑖 𝑑 𝑆̂𝑥𝑥
group scale: − 𝑥𝑥 − − +
𝜌 𝑔ℎ 𝑔ℎ𝛥𝑘𝑥 𝑔ℎ𝛥𝑘2𝑥 𝑔ℎ𝛥𝑘𝑥 𝑑𝑥
𝑋 )
𝜂̂𝑓 (𝑋) = 𝑇 (𝑋) exp(𝑖 𝛥𝑘𝑥 𝛽 −1 𝑑𝑋) = 𝑇 (𝑋) exp(𝑖𝜓(𝑋)) (A.6) 𝑖𝑆̂𝑥𝑥 𝑑𝛥𝑘𝑥 2𝑖𝛥𝑘𝑦 𝑑 𝑆̂𝑥𝑦
∫0 + + exp(𝑖𝜓)
𝑔ℎ𝛥𝑘2𝑥 𝑑𝑥 𝑔ℎ𝛥𝑘2𝑥 𝑑𝑥
that is substituted in the long wave equation, Eq. (A.3) where the
different terms will be evaluated based on their relative contribution This can be simplified by making use of the following expressions:
expressed by the scaling parameter 𝛽 where the following identity is 𝛥𝜔
𝑐𝑔 = (A.19)
used: 𝛥𝑘
𝜕 𝜕 and:
=𝛽 (A.7)
𝜕𝑥 𝜕𝑋 1 𝑑𝛥𝑘𝑥 1 𝑑 𝛥𝜔 cos 𝜃𝑓 1
Starting with the first term on the lhs of Eq. (A.3) we obtain: = =
𝛥𝑘𝑥 𝑑𝑥 𝛥𝑘 cos 𝜃𝑓 𝑑𝑥 𝑐𝑔 𝛥𝑘 cos 𝜃𝑓
( ( )) ( )
𝜕 2 𝜂̂𝑓 𝜕 2 𝜂̂𝑓 𝜕 𝜕𝑇 −𝛥𝜔 cos 𝜃𝑓 𝑑𝑐𝑔 𝛥𝜔 𝑑 cos 𝜃𝑓
𝑔ℎ = 𝑔ℎ𝛽 2 = 𝑔ℎ𝛽 2 exp 𝑖𝜓 + 𝑖𝛽 −1 𝑇 𝛥𝑘𝑥 exp 𝑖𝜓 × + = (A.20)
𝜕𝑥2 𝜕𝑋 2 𝜕𝑋 𝜕𝑋 𝑐𝑔2 𝑑𝑥 𝑐𝑔 𝑑𝑥
(A.8) ( ) 𝑔 𝑐
( −1 𝑑𝑐𝑔 1 𝑑 cos 𝜃𝑓 cos 𝜃𝑓 𝑑 cos 𝜃𝑓
𝜕2 𝑇
2 𝜕𝑇 + =−
= 𝑔ℎ𝛽 + 𝑖𝛽 −1 𝛥𝑘𝑥 𝑐𝑔 𝑑𝑥 cos 𝜃𝑓 𝑑𝑥 𝑐𝑔 𝑑𝑥
𝜕𝑋 2 𝜕𝑋
)
𝜕𝛥𝑘𝑥 𝜕𝑇 and multiplying with 𝛥𝑘𝑥 ∕2∕𝑖 gives (dropping exp 𝑖𝜓):
+𝑖𝛽 −1 𝑇 + 𝑖𝛽 −1 𝛥𝑘𝑥 − 𝛽 −2 𝑇 𝛥𝑘2𝑥 exp 𝑖𝜓
𝜕𝑋 𝜕𝑋 𝑔 𝑐
𝑖𝜇𝛥𝑘𝑥 ⎛ cos 𝜃 𝑑 cos 𝜃 ⎞ 𝑆̂
𝑑𝑇 ⎜ 𝑓 𝑓 1 𝑑ℎ ⎟
followed by the second term gives: + 𝑇 = − 𝑇 + 𝑥𝑥
𝑑𝑥 2 ⎜ 2𝑐𝑔 𝑑𝑥 2ℎ 𝑑𝑥 ⎟ 2𝜌𝑔ℎ
( ) ⎝ ⎠
𝑑ℎ 𝜕 𝜂̂𝑓 𝑑ℎ 𝜂̂𝑓 𝑑ℎ 𝑑𝑇
𝑔 = 𝑔𝛽 2 = 𝛽2𝑔 + 𝑖𝛽 −1 𝛥𝑘𝑥 𝑇 exp 𝑖𝜓 (A.9) 𝑐
𝑑𝑥 𝜕𝑥 𝑑𝑋 𝜕𝑋 𝑑𝑋 𝑑𝑥 ⎛ 𝑑 𝑔 ⎞
2 𝑑 𝑆̂𝑥𝑥 cos 𝜃𝑓 cos 𝜃𝑓 ⎟
the third term: × ⎜𝑖𝛥𝑘𝑥 + − + (A.21)
⎜ 𝑆̂𝑥𝑥 𝑑𝑥 𝑐𝑔 𝑑𝑥 ⎟
⎝ ⎠
𝛥𝜔2 𝜂̂𝑓 = 𝛥𝜔2 𝑇 exp 𝑖𝜓 (A.10) ( ) ( )
𝑆̂𝑥𝑦 𝛥𝑘𝑦 1 𝑑 𝑆̂𝑥𝑦 𝑆̂𝑦𝑦 𝛥𝑘2𝑦
and finally the fourth term: 𝑖𝛥𝑘𝑦 + +
𝜌𝑔ℎ 𝛥𝑘𝑥 𝑆̂𝑥𝑦 𝑑𝑥 𝜌𝑔ℎ 𝛥𝑘𝑥
−𝑔ℎ𝛥𝑘2𝑦 𝜂̂𝑓 = −𝑔ℎ𝛥𝑘2𝑦 𝑇 exp 𝑖𝜓 (A.11) where 𝜇 is given by Eq. (17).
In the case of normally incident waves, 𝛥𝑘𝑦 and cos 𝜃𝑓 both equal
Collecting all the terms up to 𝛽, thus ignoring 𝛽2
contributions, yields:
( ( )) 0 and Eq. (23) of Janssen et al. (2003) is retrieved. Following Janssen
𝑖𝛽 𝑑𝑇 𝑇 𝑑𝛥𝑘 𝑥 𝑇 𝑑ℎ et al. (2003) the equation is written as a function of phase difference
𝑔ℎ𝛥𝑘2𝑥 −𝜇𝑇 + 2 + + exp 𝑖𝜓 (A.12)
𝛥𝑘𝑥 𝑑𝑋 𝛥𝑘𝑥 𝑑𝑋 ℎ 𝑑𝑋 between the WGF infragravity wave, |𝑇 | = |𝜂𝑓 |, and the (negative)
where 𝜇 is the resonance mismatch: radiation stress forcing corresponding to the bi-phase, 𝛥𝜓:
( ) 𝑐
𝛥𝜔2 − 𝑔ℎ𝛥𝑘2𝑦 𝑑|𝜂𝑓 |
𝑔
⎛ cos 𝜃 𝑑 cos 𝜃 ⎞
𝜇 = 1− (A.13) 𝑑𝛥𝜓 𝑖𝜇𝛥𝑘𝑥 ⎜ 𝑓 𝑓 1 𝑑ℎ ⎟
𝑔ℎ𝛥𝑘2𝑥 + 𝑖|𝜂𝑓 | + |𝜂𝑓 | = − |𝜂 |+
𝑑𝑥 𝑑𝑥 2 ⎜ 2𝑐𝑔 𝑑𝑥 2ℎ 𝑑𝑥 ⎟ 𝑓
⎝ ⎠
Next the right hand side of Eq. (A.3) is evaluated. Starting with the
(A.22)
first term on the r.h.s.: 𝑐𝑔
𝑆̂𝑥𝑥 ⎛ 𝑑 ⎞
𝜕 2 𝑆̂𝑥𝑥 exp(𝑖𝜓) 𝜕 2 𝑆̂𝑥𝑥 exp(𝑖𝜓) 2 𝑑 𝑆̂𝑥𝑥 cos 𝜃𝑓 cos 𝜃𝑓
= 𝛽2 = (A.14) (cos 𝛥𝜓 + 𝑖 sin 𝛥𝜓) ⎜𝑖𝛥𝑘𝑥 + − ⎟+
𝜕𝑥 2 𝜕𝑋 2 2𝜌𝑔ℎ ⎜ 𝑆̂𝑥𝑥 𝑑𝑥 𝑐𝑔 𝑑𝑥 ⎟
( ) ⎝ ⎠
𝑑 𝑆̂𝑥𝑥
2 𝑑 𝑆̂ 𝑑𝛥𝑘𝑥 ( )
𝛽2 + 2𝑖𝛽 −1 𝛥𝑘𝑥 𝑥𝑥 + 𝑖𝛽 −1 𝑆̂ − 𝛽 −2 𝛥𝑘2𝑥 𝑆̂𝑥𝑥 exp(𝑖𝛹 ) 𝑆̂𝑥𝑦 𝛥𝑘𝑦 1 𝑑 𝑆̂𝑥𝑦
𝑑𝑋 2 𝑑𝑋 𝑑𝑋 𝑥𝑥 (cos 𝛥𝜓 + 𝑖 sin 𝛥𝜓) 𝑖𝛥𝑘𝑦 +
𝜌𝑔ℎ 𝛥𝑘𝑥 𝑆̂𝑥𝑦 𝑑𝑥
followed by the second term: ( 2)
𝑆̂𝑦𝑦 𝛥𝑘𝑦
𝜕 𝑆̂𝑥𝑦 𝑒𝑥𝑝(𝜓) 𝜕 𝑆̂𝑥𝑦 exp(𝑖𝜓) + (cos 𝛥𝜓 + 𝑖 sin 𝛥𝜓)
𝜌𝑔ℎ 𝛥𝑘𝑥
2𝑖𝛥𝑘𝑦 = 2𝛽𝑖𝛥𝑘𝑦
𝜕𝑥 𝜕𝑋

12
A. Reniers and M. Zijlema Coastal Engineering 172 (2022) 104068

Collecting the imaginary contributions the evolution equation for Battjes, J.A., Janssen, J.P.F.M., 1978. Energy loss and set-up due to breaking of random
the bi-phase is obtained: waves. In: Edge, B. (Ed.), Proceedings of the 16th International Conference on
Coastal Engineering. Am. Soc. of Civ. Eng., Reston, Va, pp. 569–587.
𝑑𝛥𝜓 𝜇𝛥𝑘𝑥 𝑆̂𝑥𝑥 Benoit, M., Marcos, F., Becq, F., 1996. Development of a third generation shallow
=− + water wave model with unstructured spatial meshing. In: Proc. 25th International
𝑑𝑥 2 2𝜌𝑔ℎ|𝜂𝑓 |
Conference on Coastal Engineering (ICCE’1996), Orlando, USA, pp. 465–478.
𝑔 𝑐
⎛ ⎛ cos 𝜃𝑓 𝑑 cos 𝜃𝑓 ⎞⎞ Bertin, X., et al., 2018. Infragravity waves: From driving mechanisms to impacts.
2 𝑑 𝑆̂𝑥𝑥
× ⎜𝛥𝑘𝑥 cos 𝛥𝜓 − ⎜ − sin 𝛥𝜓 ⎟⎟ + (A.23) Earth-Sci. Rev. 177, 774–799. http://dx.doi.org/10.1016/j.earscirev.2018.01.002.
⎜ ⎜ 𝑆̂𝑥𝑥 𝑑𝑥 𝑐𝑔 𝑑𝑥 ⎟⎟ Biesel, F., 1952. Equations generales au second ordre de la houle irreguleiere. La Houille
⎝ ⎝ ⎠⎠
( ) Blanch 371–376.
𝑆̂𝑥𝑦 𝛥𝑘𝑦 ̂
1 𝑑 𝑆𝑥𝑦 Booij, N., Ris, R.C., Holthuijsen, L.H., 1999. A third-generation wave model for coastal
𝛥𝑘𝑦 cos 𝛥𝜓 + sin 𝛥𝜓
𝜌𝑔ℎ|𝜂𝑓 | 𝛥𝑘𝑥 ̂
𝑆𝑥𝑦 𝑑𝑥 regions, 1. Model description and validation. J. Geophys. Res. 104, 7, 649–7, 666.
( 2 ) Bromirski, P.D., Sergienko, O.V., MacAyeal, D.R., 2010. Transoceanic infragravity waves
𝑆̂𝑦𝑦 𝛥𝑘𝑦 impacting antarctic ice shelves. Geophys. Res. Lett. 37.
− sin 𝛥𝜓 Bryan, K.R., Howd, P.A., Bowen, A.J., 1998. Field observations of bar-trapped edge
𝜌𝑔ℎ|𝜂𝑓 | 𝛥𝑘𝑥 waves. J. Geophys. Res. 103, 1285–1305.
De Bakker, A.T.M., Herbers, T.H.C., Smit, P.B., Tissier, M.F.S., Ruessink, B.G., 2015.
which reduces to the solution given by Janssen et al. (2003) for
Nonlinear infragravity-wave interactions on a gently sloping laboratory beach. J.
normally incident waves, i.e. their Eq. (25). The equilibrium solution Phys. Oceanogr. 45, 589–605. http://dx.doi.org/10.1175/JPO-D-14-0186.1.
for a horizontal bed with 𝛥𝜓 = 0 yields: De Bakker, A.T.M., Tissier, M.F.S., Ruessink, B.G., 2014. Shoreline dissipation of
infragravity waves. Cont. Shelf Res. 72, 73–82. http://dx.doi.org/10.1016/j.csr.
𝛥𝑘𝑥,0 𝜇0 𝑆̂𝑥𝑥,0 𝑆̂𝑥𝑦,0 2013.11.013.
− + 𝛥𝑘𝑥,0 + 𝛥𝑘𝑦 = 0 (A.24)
2 2𝜌𝑔ℎ0 |𝜂̂𝑓 ,0 | 𝜌𝑔ℎ0 |𝜂̂𝑓 ,0 | Eldeberky, Y., Battjes, J.A., 1995. Parameterisation of triad interactions in wave energy
models. In: Proc. Coastal Dynamics Conf. ’95, Gdansk, Poland, pp. 140–148.
from which it follows that: Elgar, S., Guza, R.T., 1985. Observations of bispectra of shoaling surface gravity waves.
𝛥𝑘 J. Fluid Mech. 161, 425–448.
𝑆̂𝑥𝑥,0 + 2𝑆̂𝑥𝑦,0 𝛥𝑘 𝑦 Fiedler, J.W., Smit, Pieter B., Brodie, Katherine L., McNinch, Jesse, Guza, R.T., 2019.
𝑥,0
|𝜂̂𝑓 ,0 | = (A.25) The offshore boundary condition in surf zone modeling. Coast. Eng. 143, 12–20.
𝜇0 𝜌𝑔ℎ0 http://dx.doi.org/10.1016/j.coastaleng.2018.10.014.
where the subscript 0 refers to the offshore conditions. Guza, R.T., Feddersen, F., 2012. Effect of wave frequency and directional spread
on shoreline runup. Geophys. Res. Lett. 39, L11607. http://dx.doi.org/10.1029/
2012GL051959.
Appendix B. Finite difference equations Guza, R.T., Thonrnton, E.B., Holman, R.A., 1984. Swash on steep and shallow beaches.
In: Proc. of the 19th Int. Conf. Coastal Eng.. ASCE, pp. 708–723.
Guza, R.T., Thornton, E.B., 1982. Swash oscillations on a natural beach. J. Geophys.
Writing the coupled differential equations for the bound infragravity Res. 87 (C1), 483–491.
energy flux and the bi-phase yields for Eq. (15) : Guza, R.T., Thornton, E.B., 1985. Observations of surf beat. J. Geophys. Res. 87,
( 483–491.
1
𝐹𝑥,𝑗+1 = 𝐹𝑥,𝑗 + 𝛥𝑥 𝛥𝑘 𝑈̂ 𝑆̂ sin 𝛥𝜓𝑗 Hasselmann, K., 1962. On the non-linear energy transfer in a gravity-wave spectrum
2 𝑥,𝑗+1 𝑓 ,𝑥,𝑗 𝑥𝑥,𝑗+1
) part 1. General theory. J. Fluid Mech. 12, 481–500.
1
+ 𝛥𝑘𝑦,𝑗+1 𝑈̂ 𝑓 ,𝑦,𝑗 𝑆̂𝑥𝑦,𝑗+1 sin 𝛥𝜓𝑗 (B.1) Hasselmann, S., Hasselmann, K., Janssen, P.A.E.M., et al., 1988. The WAM model
2 - a third generation ocean wave prediction model. J. Phys. Oceanogr. 18 (12),
and for Eq. (16): 1775–1810.
( ) Hasselmann, K., Munk, W., MacDonald, G., 1963. Bispectra of ocean waves. In:
𝛥𝑘𝑥,𝑗+1 𝜇𝑗+1 Rosenblatt, M. (Ed.), Time Series Analysis. John Wiley, New York.
𝛥𝜓𝑗+1 = 𝛥𝜓𝑗 − 𝛥𝑥 (B.2)
2 Henderson, S.M., Guza, R.T., Elgar, S., Herbers, T.H.C., Bowen, A.J., 2006. Nonlinear
generation and loss of infragravity wave energy. J. Geophys. Res. 111, C12007.
𝑆̂𝑥𝑥,𝑗+1 ( http://dx.doi.org/10.1029/2006JC003539.
− 𝛥𝑥𝛥𝑘𝑥,𝑗+1 cos 𝛥𝜓𝑗
2𝜌𝑔ℎ𝑗+1 |𝜂̂𝑓 ,𝑗 | Herbers, T.H.C., Burton, M.C., 1997. Nonlinear shoaling of directionally spread waves
( ) ) on a beach. J. Geophys. Res. 102, 21101–21114.
𝑆̂𝑥𝑥,𝑗+1 − 𝑆̂𝑥𝑥,𝑗 𝑐𝑔,𝑗+1 − 𝑐𝑔,𝑗 Herbers, T.H.C., Elgar, S., Guza, R.T., 1994. Infragravity-frequency (0.005–0.05 hz)
− 2 − sin 𝛥𝜓𝑗
𝑆̂𝑥𝑥,𝑗+1 𝑐𝑔,𝑗+1 motions on the shelf, part I : Forced waves. J. Phys. Oceanogr. 24, 917–927.
( ( ) ) Herbers, T.H.C., Elgar, S., Guza, R.T., 1995. Generation and propagation of infragravity
𝑆̂𝑥𝑦,𝑗+1 𝛥𝑘𝑦 𝑆̂𝑥𝑦,𝑗+1 − 𝑆̂𝑥𝑦,𝑗 waves. J. Geophys. Res. 100, 24863–24872.
+ 𝛥𝑥𝛥𝑘𝑦 cos 𝛥𝜓𝑗 − sin 𝛥𝜓𝑗
𝜌𝑔ℎ𝑗+1 |𝜂̂𝑓 ,𝑗 | 𝛥𝑘𝑥,𝑗+1 𝑆̂𝑥𝑦,𝑗+1
Holman, R.A., 1981. Infragravity energy in the surf zone. J. Geophys. Res. 86,
6442–6450.
𝛥𝑥𝑆̂𝑦𝑦,𝑗+1 𝛥𝑘2𝑦 Howd, P.A., Bowen, A.J., Holman, R.A., 1992. Edge waves in the presence of strong
− 𝑠𝑖𝑛(𝛥𝜓𝑗 ) longshore currents. J. Geophys. Res. 97, 11357–11371.
2𝜌𝑔ℎ𝑗+1 |𝜂̂𝑓 ,𝑗 | 𝛥𝑘𝑥,𝑗+1 Janssen, T.T., 2006. Nonlinear Surface Waves over Topography (Ph.D. thesis). Technical
respectively, where 𝛥𝑥 is the cross-shore grid spacing. Here, a uniform University of Delft, The Netherlands.
Janssen, T.T., Battjes, J.A., van Dongeren, A.R., 2003. Long waves induced by short
grid is 𝑗𝛥𝑥, 𝑗 = 0, 1, ..𝑁 is employed where 𝑁 is the number of grid cells. wave groups over a sloping bottom. J. Geophys. Res. 108 (C8), 3252. http:
Point 𝑗 = 0 indicates the offshore boundary where boundary condition //dx.doi.org/10.1029/2002JC001515.
Eq. (19) and 𝛥𝜓 = 0 are imposed for the WGF-infragravity energy Lin, Y.-H., Hwung, H.-H., 2012. Infra-gravity wave generation by the shoaling wave
flux and bi-phase respectively. Cross-shore-integration yields the bound groups over beaches. China Ocean Eng. 26, 1–18.
Longuet-Higgins, M.S., Stewart, R.W., 1962. Radiation stress and mass transport in
long wave energy, amplitude and bi-phase that can be compared with
gravity waves with applications to ’surf-beats’. J. Fluid. Mech. 8, 565–583.
the IAM results for obliquely incident bi-chromatic waves. Longuet-Higgins, M.S., Stewart, R.W., 1964. Radiation stresses in water waves. a
physical discussion, with applications. Deep Sea Res. 11, 529–562.
Madsen, P.A., Sorensen, O.R., 1993. Bound waves and triad interactions in shallow
References
water. J. Ocean Eng. 20 (4), 359–388.
McCall, R., Van Thiel de Vries, J.S.M., Plant, N., Van Dongeren, A.R., Roelvink, J.A.,
Arcilla, A.S., Roelvink, J.A., OConnor, B.A., Reniers, A., Jimenez, J.A., 1994. The delta Thompson, D., Reniers, A., 2010. Two-dimensional time dependent hurricane
flume 93 experiment. In: Coastal Dynamics 94 (Bareclona, Spain), pp. 488-502. overwash and erosion modeling at santa rosa island. Coast. Eng. 57, 668–683.
Ardhuin, F., Rawat, A., Aucan, J., 2014. A numerical model for free infragravity waves: http://dx.doi.org/10.1016/j.coastaleng.2010.02.006.
definition and validation at regional and global scales. Ocean Model. 77, 20–32. Mei, C.C., Benmoussa, C., 1984. Long waves induced by short wave groups over an
Baldock, T.E., Holmes, P., Bunker, S., Van Weert, P., 1998. Cross-shore hydrodynamics uneven bottom. J. Fluid Mech. 139, 219–235.
within an unsaturated surf zone. Coast. Eng. 34, 173–196. Munk, W., 1949. Surf beat. Eos Trans. AGU 30, 849–854.
Battjes, J.A., Bakkenes, H.J., van Dongeren, A.R., 2004. Shoaling of subharmonic gravity Neale, J., Harmon, N., Srokosz, M., 2015. Source regions and reflection of infragravity
waves. J. Geophys. Res. 109, C02009. http://dx.doi.org/10.1029/2003JC001863. waves offshore of the U.S. Pacific northwest. J. Geophys. Res. 120, 1–18.

13
A. Reniers and M. Zijlema Coastal Engineering 172 (2022) 104068

Phillips, O.M., 1977. The Dynamics of the Upper Ocean, second ed. Cambridge Schäffer, H.A., Svendsen, I.A., 1988. Surf beat generation on a mild-slope beach.
University Press, Cambridge-London-New York-Melbourne, p. 336. In: Proc. of the 21st International Conf. Coastal Eng.. Torremolinos, Spain, http:
Rawat, A., Ardhuin, F., Ballu, V., Crawford, W., Corela, C., Aucan, J., 2014. Infragravity //dx.doi.org/10.9753/icce.v21.79.
waves across the oceans. Geophys. Res. Lett. 41, 7957–7963. http://dx.doi.org/10. Sénéchal, N., Bonneton, P., Dupuis, H., 2001a. Field observations of irregular wave
1002/2014GL061604.2014GL061604. transformation in the surf zone. In: Proc. Fourth Conf. Coast. Dyn. pp. 62–70.
Reniers, A.J.H.M., van Dongeren, A.R., Battjes, J.A., Thornton, E.B., 2002. Linear Smit, P.B., Janssen, T.T., Herbers, T.H.C., Taira, T., Romanowicz, B.A., 2018. Infragrav-
modelling of infragravity waves during delilah. J. Geophys. Res. 107, 3137. http: ity wave radiation across the shelf break. JGR Oceans http://dx.doi.org/10.1029/
//dx.doi.org/10.1029/2001JC001083. 2018JC013986.
Reniers, A.J.H.M., Groenewegen, M.J., Ewans, K.C., Masterton, S., Stelling, G.S., Stockdon, H.F., Holman, R. a., Howd, P. a., Sallenger, A.H., 2006. Empirical
Meek, J., 2010. Estimations of infragravity waves at intermediate depth. Coast. parameterization of setup, swash, and runup. Coast. Eng. 53, 573–588.
Eng. 57, 52–61. Thompson, J., Elgar, S., Raubenheimer, B., Herbers, T.H.C., Guza, R.T., 2006. Tidal
Rijnsdorp, D.P., Reniers, A.J.H.M., Zijlema, M., 2021. Free infragravity waves in the modulation of infragravity waves via nonlinear energy losses in the surfzone.
north sea. J. Geophys. Res. 126 (8), http://dx.doi.org/10.1029/2021JC017368. Geophys. Res. Lett. 33, L05601. http://dx.doi.org/10.1029/2005GL025514.
Rijnsdorp, D.P., Ruessink, B.G., Zijlema, M., 2015. Infragravity-wave dynamics in a Tissier, M., Bonneton, P., Michallet, H., Ruessink, B.G., 2015. Infragravity-wave
barred coastal region, a numerical study. J. Geophys. Res. 120 (6), 4068–4089. modulation of short-wave celerity in the surf zone. J. Geophys. Res. Oceans 120,
http://dx.doi.org/10.1002/2014JC010450. 6799–6814. http://dx.doi.org/10.1002/2015JC010708.
Roelvink, J.A., Reniers, A.J.H.M., 1994. LIP11D Delta Flume Experiments. Data report Tissier, M., Bonneton, P., Ruessink, B., 2017. Infragravity waves and bore merging. In:
H2130, WL/Delft Hydraulics, The Netherlands. Proceedings of the Conference Coastal Dynamics 2017. ASCE, Helsingør, Denmark,
Roelvink, D., Reniers, A., van Dongeren, A., van Thiel de Vries, J., McCall, R., pp. 451–460.
Lescinski, J., 2007. Modelling storm impacts on beaches, dunes and barrier islands. Tolman, H.L., 1991. A third-generation model for wind waves on slowly varying,
Coast. Eng. 56, 1133–1152. unsteady and inhomogeneous depths and currents. J. Phys. Oceanogr. 21, 782–797.
Ruessink, B.G., 1998a. Bound and free infragravity waves in the nearshore zone under Van Dongeren, A., Battjes, J., Janssen, T., van Noorloos, J., Steenhauer, K., Steen-
breaking and nonbreaking conditions. J. Geophys. Res. 103, 12.795–12.805. bergen, G., Reniers, A., 2007. Shoaling and shoreline dissipation of low-frequency
Ruessink, B.G., 1998b. The temporal and spatial variability of infragravity energy in a waves. J. Geophys. Res. 112, C02011. http://dx.doi.org/10.1029/2006JC003701.
barred nearshore zone. Cont. Shelf Res. 18, 585–605. Vrećica, T., Soffer, R., Toledo, Y., 2019. Infragravity wave generation by wind gusts.
Ruggiero, P., Holman, R.A., Beach, R.A., 2004. Wave run-up on a high-energy Geophys. Res. Lett. 46 (16), 9728–9738.
dissipative beach. J. Geophys. Res. 109 (C06025). Van Thiel de Vries, J., van Gent, R.A., Walstra, D.J., Reniers, A.J.H.M., 2008. Analysis
Schäffer, H.A., 1993. Infragravity waves induced by short wave groups. J. Fluid Mech. of dune erosion processes in large-scale flume experiments. Coast. Eng. 55 (12),
247, 551–588. 1028–1040. http://dx.doi.org/10.1016/j.coastaleng.2008.04.004.

14