Invited paper

Surface waves in deep OCEANOLOGIA, 52 (1), 2010.

pp. 5 – 52.
and shallow waters
C 2010, by Institute of
Oceanology PAS.

KEYWORDS
Surface waves
Non-linear interactions
Energy dissipation
Aerosol fluxes

Stanisław R. Massel
Institute of Oceanology
of the Polish Academy of Sciences,
Powstańców Warszawy 55, PL–81–712 Sopot, Poland

Received 8 January 2010, revised 24 February 2010, accepted 10 March 2010.

Abstract

The motion of water due to surface waves is the most dynamic factor observed in
the marine environment. In this review various aspects of the wave modelling of
non-linear, steep surface waves and their role in the atmosphere-ocean interaction
are discussed. Significant improvements in wave forecasting have been made in
the last ten years. This is to a large extent related to substantial progress in the
description of wind forcing and other processes, as well as to the more efficient
use of satellite observations and assimilation methods. One striking observation is
the increasing variety and complexity of models in which more physical processes
are implemented, greater precision and resolution achieved and extended ranges
of applicability demonstrated. However, in order to evaluate the applicability of
particular models, comparison with high quality experimental data, collected in
nature or under laboratory conditions, is necessary.

Notation list
a – wave amplitude, constant
a1 – function
The complete text of the paper is available at http://www.iopan.gda.pl/oceanologia/
6 S. R. Massel

ak – wave steepness
(E)
az – Eulerian downward acceleration
(L)
az – Lagrangian downward acceleration
aγ – coefficient
A – complex amplitude, constant
Abr – critical amplitude
Arms – root-mean-square amplitude
à – non-dimensional complex amplitude

b – constant

C – phase velocity
Cbr – phase speed of breaking waves
Cf r – bottom friction coefficient
Cg – group velocity
C+ , C− , ϕ – functions of interaction coefficients
Cθ – phase velocity in θ – space
Cσ – phase velocity in σ – space

Db – dissipation factor due to wave breaking

Df r – energy-dissipation rate due to bottom friction
D(θ) – directional spreading function

Edissrate – rate of energy loss

Ediss – energy dissipation for a very narrow spectrum

f – probability density function

Fbr – probability of wave breaking
Fcov – percentage of sea surface covered by breaking waves
F (k) – wave number spectrum

h – water depth
Hs – significant wave height
Hmax – max. wave height

Ic , Is – integrals
I0 – modified Bessel function
I(s) – spreading function
J0 ,J1 – Bessel functions of the first kind

kc – carrier wave number

k – wave number vector
KT – transmission coefficient

mn – spectral moment of the nth-order

N – wave action density

Nb – number of breaking waves per wave
 Surface waves in deep and shallow waters 7

Rmax – max. run-up height

s – directional spreading parameter

S – source function
Sdiss – source term due to energy loss
s – directional spreading parameter
S – source function
Sdiss – source term due to energy loss
Sin – source term due to input from wind
Snl – source term due to non-linear interaction
S(ω, θ) – two-dimensional frequency spectrum

T – wave period

u∗ – friction velocity
urms,bottom – root-mean-square orbital velocity at bottom

V10 – wind speed at 10 m altitude

w – vertical velocity component

X – wind fetch

Z0 – function of propagation mode

Zn – function of evanescent mode
Z−1 – function of sloping mode

α – constant

β – Phillips’ constant
β1 – bottom slope
βK – growth rate due to Krasitskii
βx , βBF – growth rate by Benjamin & Feir

γ – peak enhancement factor

γbr – energy dissipation due to wave breaking
γf r – energy dissipation due to bottom friction
γ1 – numerical constant

δ – small non-dimensional quantity

∆ – relative local energy growth rate

ǫ – small quantity
εd – dominant wave steepness
εl – local surface slope
εs – significant steepness
εth – threshold steepness
8 S. R. Massel

ζ – surface wave ordinate

ζmax – max. crest height

η – mean water level

θ – wave component direction, inclination of a breaking wave’s

forward face

ρa – density of air
ρw – density of water

σ – intrinsic frequency

ωc – carrier wave frequency

ωp – peak frequency of wave spectra
Ω, ω – observed (absolute) frequency

1. Introduction
The atmosphere and ocean form a coupled system which continuously
exchanges heat, momentum and mass at the air-sea interface. Owing to
energy flow from the atmosphere to the ocean, this interface presents an
aerodynamically rough surface that can consist of dynamic, unsteady, very
high and steep surface waves. The hydrodynamics of this process is still not
fully understood.
In the past, the study of surface wave mechanics concentrated predom-
inantly on two aspects. On the one hand, wave mechanics was treated
as a purely mathematical problem, important in numerical modelling
(Lamb 1932, Stoker 1957, Phillips 1977, Le Blond & Mysak 1978, Davidan
et al. 1978, 1985, Mei 1989, Komen et al. 1994, Massel 1989, 1996, Ochi
1998, Lavrenov 2003, Holthuijsen 2007). On the other, flow kinematics due
to surface waves was used to estimate the resultant loadings on offshore
and coastal structures under extreme and operating conditions. Textbooks
and scientific papers provide many examples of the direct applications of
wave mechanics to ocean and coastal engineering (Krylov et al. 1976, 1986,
Sarpkaya & Isaacson 1981, Dean & Dalrymple 1998, Goda 2000, Mei et al.
2006).
However, in recent years, the focus has been redirected to climate change
and the role played by the oceans. The interaction between the ocean and
atmosphere at the air-sea interface is critical to our understanding of the
Earth’s climate as the ocean surface forms a filter to the exchange of heat,
moisture, momentum and trace constituents (Massel 2007). In particular,
the energy transfer from the atmosphere to the ocean enhances the heat
flux and mixed layer during the circulation of the upper ocean, while the
 Surface waves in deep and shallow waters 9

energy fluxes from the ocean to the atmosphere affect gas and aerosol
concentrations, as well as the atmospheric circulation, weather and climate.
This review examines the physics and modelling of non-linear, steep
surface waves and their role in the atmosphere-ocean interaction. A detailed
description of particular wave models is not given here, but the governing
equations and references to the subject literature are used to illustrate the
structure of particular models.
For the presentation of the problems it is useful to distinguish between
deep ocean and shallow waters. In both cases, the dominant physical wave
processes are characterised first. To a large extent these processes determine
the selection of suitable mathematical and numerical models. Moreover,
special wave events, such as whitecapping, freak waves, tsunamis and wave-
induced groundwater circulation, are discussed. Knowledge of these events
is of special importance for modern oceanography, and oceanic and coastal
technology.
For the purposes of this paper, Table 1 classifies in a simple way the
various types of waves and sets out suggested methods for their solution.
Although such a classification is not unique, it will provide some useful
insight into the complexity of surface waves.
The paper is organised as follows. The energy balance equation for
deep water waves and its constituents are discussed in Section 2: particular

Table 1. Classification of surface waves and possible methods for their solution

Type of waves Methods of solution

linear sinusoidal wave small amplitude wave theory
L 2
H



(typical of deep seas) for H/h ≪ 1, L/h < 10 and U r = h h
< 75
Stokes short waves 5th order approximation (see Fenton 1985)
(typical of intermediate for H/h ≤ 1, L/h < 10(20) and U r > 75
wave depths)
long, cnoidal and a Boussinesq-type approximation (see Fenton 1986)
solitary waves or higher order cnoidal wave theory (see Fenton
(typical of shallow water) 1979) for H/h ≤ 1, L/h > 20(20) and U r > 75
freak waves simulation using a Schrödinger-type equation
when Hmax /Hs > 2 at any depth
steep, near breaking local solution by Sobey (1992) or Baldock & Swan
waves (1994) and the fully non-linear model by Clamond
& Grue (2001a, b)
tsunami waves numerical solution on the global scale (see Kowalik
et al. 2005); in the coastal zone-run-up, wave
models with breaking (see Pelinovsky 1996)
10 S. R. Massel

emphasis is placed on the kinematics and dynamics of extreme waves as well

as the link between sea surface state and aerosol fluxes. Section 3 deals with
the motion of waves in shallow waters, including the interaction of waves
with a porous seabed. Finally, Section 4 gives the main conclusions.

2. Waves in deep waters

2.1. Energy balance equation
Surface waves are mostly generated by winds. They are highly dispersive
and interact with each other; however, the time of interaction is finite
because each wave propagates with a different phase speed. Therefore, in
practice, there is no need for detailed information regarding the phases of
ocean waves. Knowledge about the distribution of wave energy over the
wave number space k (e.g. wave number spectrum F (k, t)) or over the wave
frequency ω and direction θ (e.g. the wave frequency-directional spectrum,
S(ω, θ, t)) is normally sufficient. In order to derive the governing equations
for functions F (k, t) or S(ω, θ, t), two methods are usually applied: the
Hamiltonian approach (Zakharov 1968) or the Lagrangian approach (Luke
1967). The resulting action balance equation takes the form (Janssen 2004)

∂N
+ ∇x · (∇x ΩN ) − ∇k · (∇x ΩN ) = S = Sin + Snl + Sdiss , (1)
∂t
in which the wave action density N is given by
gF (k)
Nk = (2)
σ
and σ is the so-called intrinsic frequency:
q
σ= gk tanh(kh), (3)
where k is the wave number modulus k = |k| and h is the water depth.
The observed or absolute frequency Ω obeys the following dispersion
relation:
Ω = k · U + σ, (4)
in which U is the water current vector. The right-hand side of the equation
describes the source function S, including the spectral input from wind
(Sin ), the net spectral flux of energy action through the wave number k
by non-linear wave-wave interactions (Snl ) and the energy loss by wave
breaking (Sdiss ).
Equation (1) gives us the rate of change in time of the action density
spectrum due to advection with group velocity Cg = ∇k Ω, to refraction
 Surface waves in deep and shallow waters 11

resulting from the temporal and spatial dependence of the local wavenum-
ber, and due to physical processes contained in the source term S. Therefore
the direction of wave energy transport is generally not normal to the wave
crest when an ambient current is present. Moreover, the wave energy is
not conserved as the wave propagates through a current. The conserved
quantity is the wave action N defined in eq. (1) – see Bretherton & Garrett
(1969). This equation is currently regarded as the starting point of
many modern wave forecasting models (Massel 1996, Holthuijsen 2007).
Observations of the two-dimensional spectrum F (k) are rare. It is much
easier to obtain the frequency-directional spectrum based on the analysis
of a time series at a given location. The relationship between both spectra
takes the form:
S(ω, θ)dωdθ = F (k)dk = F (k, θ)kdkdθ; (5)
thus:
k
S(ω, θ) = F (k, θ), (6)
Cg
in which Cg = ∂ω/∂k and θ is the direction of the particular wave
component.
In the following we will discuss the processes included in the source term
S in equation (1).

2.2. The generation of ocean waves by wind

This problem has led to much debate and much controversy. First of
all, we are dealing with the extremely difficult problem of turbulent airflow
over a surface varying in space and time. On the other hand, in order
to measure the growth rate of waves by wind we have to make certain
assumptions regarding the process that causes waves to grow as a result of
the work done by the wind pressure on the sea surface. However, because
of the small air-water density ratio, growth rates are small and a very
accurate determination of amplitude and phase of the wave-induced pressure
fluctuations is required.
Nevertheless, considerable progress in the prediction of wave growth
has been made over the past 40 years. Substantial contributions were
made in the mid-1950 s, when Phillips (1957) and Miles (1957) published
their theories of wave generation by wind. Both theories are based on the
assumption that wave growth is due to a resonance phenomenon. Phillips
considered the resonant forcing of waves by turbulent pressure fluctuations,
while Miles assumed resonant interaction between wave-induced pressure
fluctuations and free surface waves. A consequence of this assumption is
that Phillips’ mechanism gives rise to a linear growth of the wave spectrum
12 S. R. Massel

in time, although this growth is independent of the spectrum itself. In the

case of Miles’ mechanism, wave growth is exponentially proportional to the
spectrum.
Field measurements have shown an order of magnitude agreement with
Miles’ theory, although this still predicts energy-transfer rates that are
smaller than measured values. One of the reasons for these discrepancies is
that Miles’ theory assumes inviscid airflow and so neglects the role of air
turbulence. Moreover, non-linear effects such as wave-mean flow interaction
have been neglected.
There have been several attempts to overcome these shortcomings by
numerical modelling of the turbulent boundary layer flow over a moving
water surface (Janssen 2004). Results from this modelling that include the
effects of small-scale turbulence and effects of gustiness agree with Miles’
theory for the case when the phase speed is smaller than the wind speed
(young waves). However, when the phase speed is larger than the wind speed
(old waves), numerical models predict wave damping in contrast to Miles’
theory. These simulations clearly demonstrated that the generation of ocean
waves is in fact an example of two-way interaction: as soon as the waves
become sufficiently steep the associated wave-induced stress slows down the
airflow. This should result in dependence of the air-sea momentum transfer
on the sea state.

2.3. Non-linear wave-wave interaction

Ocean waves are usually regarded as weakly non-linear, dispersive

waves, and the effect of non-linearity on wave propagation is the result
of a perturbation expansion starting with linear, freely propagating waves.
Theoretical and experimental studies provide convincing evidence that
regular wave trains in deep water are liable to a number of instabilities
that lead to wave breaking without external forcing. In particular, a small
disturbance in the form of two modes with ‘sideband’ frequencies, adjacent
to the fundamental frequency, will be forced to increase exponentially owing
to non-linear interaction mechanisms, and primary wave motion becomes
unstable due to this form of disturbance. Such instability of periodic
wave trains is known as Benjamin-Feir instability. Benjamin & Feir (1967)
examined a primary, deep-water wave of finite amplitude a and frequency
ω superimposed by two ‘sidebands’ of infinitesimal amplitude ǫa and of
frequencies ω(1 ± δ) respectively, where ǫ and δ are small quantities. They
found that motion becomes unstable and the sidebands tend to grow in
amplitude at the expense of the primary wave, provided that
2(ak)2 > δ2 . (7)
 Surface waves in deep and shallow waters 13

Then the corresponding initial growth rate becomes

d(ln a)
βx = , (8)
d (kx)
which is presented by Benjamin & Feir (1967) in the form
βx = βBF = ε2 δ̂(2 − δ̂2 )1/2 , (9)
δω
where ε = ak, δ̂ = and ωc is the carrier frequency. The instability
εωc
√
requires that 0 < δ̂ ≤ 2, and the maximum growth appears when δ̂ = 1.0.
However, the theory of Benjamin & Feir (1967) is asymptotically valid
only for sufficiently small values of wave steepness ak and modulation
frequency δ. In particular, Longuet-Higgins (1978) found that subharmonic
instabilities of the Benjamin-Feir type are confined to waves whose ‘steep-
ness’ ak lay within a certain finite range with the upper limit at ak ≈ 0.37.
The growth rate is about 14% per period, until ak ≈ 0.32. As ak increases
beyond ≈ 0.346, the wave modes become stable again. Comparison of
the calculated growth rates showed good agreement with the observations
reported by Benjamin (1967) in the wave steepness range 0.07 < ak < 0.17.
The experimental and numerical study of Lake et al. (1977) showed that
the evolution of a nonlinear wave train, in the absence of dissipative effects,
exhibited the so-called ‘Fermi-Pasta-Ulam’ recurrence phenomenon, where
the modulation periodically increases and decreases, and the wave form
periodically returns to its previous form. Modulation was caused by the
growth of the two dominant sidebands of the Benjamin-Feir instability at
the expense of the carrier. For small steepness, the original near-three-wave
system was almost recovered. Figure 1 gives an example of such a system.
In Tulin & Waseda’s (1999) wave tank experiment, a wave 1 m in length
was considered whose initial wave carrier frequency ωc = 1.23 Hz, while the
sideband frequencies were 1.12 Hz and 1.34 Hz. Indicated fetches in the wave
tank were 3.6 m, 9.0 m, 14.4 m, 19.8 m, 25.2 m, 30.6 m, 36 m and 41.4 m. The
maximum modulation was observed in the middle of the channel and the
spectrum recovered at a fetch of 41.4 m.
When the initial steepness is large enough, wave trains experience strong
modulations followed by demodulations with the spectral peak downshifted
to a lower sideband. Recently, it was found that Krasitskii’s modification
of the Zakharov evolution equation (Zakharov 1968) for the four-wave
interaction correctly predicts the major features of the energy increase in
the lower sideband relative to the upper sideband (Krasitskii 1994, Tulin
& Waseda 1999). Downshifting to a lower sideband of the spectral peak
appears in the absence of breaking, which demonstrates the role of the
balance between momentum losses and energy dissipation in the exchange
14 S. R. Massel

Figure 1. Fermi-Pasta-Ulam recurrence phenomenon (Tulin & Waseda 1999)

of energy between sidebands. Krasitskii’s (1994) four-wave reduced equation

yields the growth rate in the form
βK = ε2 (C+ C− )1/2 sin ϕ, (10)
in which C+ , C− and ϕ are functions of interaction coefficients given in
Krasitskii (1994). The experimental data of Waseda & Tulin (1999) have
been compared with the theoretical predictions of Benjamin & Feir (1967)
and Krasitskii (1994) (see Figure 2). Clearly, Benjamin and Feir’s theory
 Surface waves in deep and shallow waters 15

Figure 2. Comparison of the theoretical growth rate with experiments (Tulin

& Waseda 1999)

overestimates the growth rate, but Krasitskii’s solution agrees very well with
measurements.
For a more complicated situation, analytical solutions of the evolution
equations fail and only numerical simulations are possible. In particular,
Dold & Peregrine (1986) established a connection between a weakly non-
linear four-wave interaction process and a truly non-linear wave-breaking
phenomenon (see also Song & Banner 2002, Banner & Song 2002). Dold
& Peregrine showed that initial carrier wave steepness (ak)c differentiates
between two modes of behaviour, i.e. recurrence of the initial state without
breaking or the rapid onset of breaking. These two modes of behaviour are
controlled by the non-dimensional relative local energy growth rate ∆:

1 Dχ2
∆= , (11)
ωc Dt

in which χ2 is the mean relative local energy growth rate.

The recurrence happens when the maximum value of ∆ is less than some
threshold level ∆th . Thus we have
(
< (1.30 − 1.50) × 10−3 − recurrence case
∆th = (12)
> (1.30 − 1.50) × 10−3 − breaking case
16 S. R. Massel

For wave breaking, the value of ∆ continues to increase beyond the threshold
level.
A common feature of the observed evolution is the fact that either for
breaking or for recurrence towards the original wave group, the evolution
induced by the non-linear group dynamics is accompanied by a systematic
mean convergence of the energy density towards the local maximum of the
evolving wave group.
The question now arises as to how the above discoveries may be used
to predict wave propagation in practice. In particular, we are interested
in a statistical description of the sea surface in terms of the evolution of
the energy of an ensemble of waves when non-linear transfer gives rise to
a downshift of the wave number. In order to conserve energy and wave
action, considerable amounts of energy are transferred from the region just
beyond the location of the spectral peak to the high wave number part of the
spectrum. Therefore, the rate of change of the spectrum due to non-linear
interactions demonstrates the typical three-lobe structure (Figure 3).

Figure 3. Three-lobe structure of the rate of change of the spectrum (Janssen

2004)

It should be noted that both resonant and non-resonant wave-wave

interactions are included in the evolution equation. For short times the
evolution of the wave action density N is due to both resonant and non-
resonant quadruplet wave-wave interactions. However, the time scale on
which non-resonant interactions operate is short, typically of the order of
10 to 20 wave periods. Thus, for wave prediction on the global scale there
is interest only in the slow time evolution of the wave spectrum when only
 Surface waves in deep and shallow waters 17

resonant interactions contribute to spectral change. Numerical experiments

using the exact non-linear transfer function suggest that the spectra tend
to adjust in such a way that the directionally averaged spectra are close to
k−5/2 (or to k−7/2 for the two-dimensional spectrum). This result follows
from Kolmogorov’s concept of an energy cascade when a constant non-linear
energy flux is maintained through the system. Then the Kitaigorodskii
(1983) scaling yields the frequency spectrum S(ω) in the form suggested by
Toba (1973):
1/2
ρa

S(ω) = 2A u∗ gω −4 , (13)
ρw
where ρa and ρw are the air and water density respectively, u∗ is the friction
velocity, g is the acceleration due to gravity and A is a constant. The ω −4
law is typically valid for the frequency range (1.3–3.0)ωp , where ωp is the
peak frequency of the waves.
Calculation of the Snl term in eq. (1) requires an enormous computation
effort. Thus, some form of parameterisation of Snl is needed. After several
past attempts there is now some consensus that the discrete interaction
approximation (DIA) developed by Hasselmann et al. (1985) offers the
best parameterisation of Snl . It was incorporated into the well-known
third-generation prediction model WAM4. For more information on the
numerical implementation techniques for the energy balance equation, see,
for example, WAMDI (1988), Janssen (2004) or Holthuijsen (2007).

2.4. Whitecap coverage and energy dissipation

2.4.1. Whitecap coverage of the sea surface
The energy flow from the atmosphere to the ocean generates an
aerodynamically rough ocean surface. If the energy flow is sufficiently
strong, at some points of the surface, waves lose their stability and
eventually break in the form of whitecaps of various scales (see colour
photo). The percentage of sea surface covered by whitecaps is usually
parameterised in terms of the wind speed, when this speed is greater than
about 4 m s−1 ; below this speed, whitecaps are not observed. However,
wind speed alone cannot fully parameterise the complex process of wave
breaking and whitecap formation.
The rate of energy supplied by the wind is closely related to the wind
stress and to the atmospheric stability conditions. Approximately, the
percentage of sea surface covered by breaking waves Fcov , can be written in
the form (Massel 2007)
b
Fcov = a · V10 , (14)
18 S. R. Massel

where Fcov is expressed as percentage, V10 is the wind speed in m s−1 at

10 m altitude, while a and b are empirical constants. The experimental data
(Monahan 1971, Stramska & Petelski 2003, Massel 2007) indicate that

10−7 < a < 10−5 and 3.0 < b < 3.75. (15)

Photo 1. Whitecapping on the sea surface

A very important conclusion resulting from Stramska & Petelski’s

(2003) observations is that whitecap coverage depends on the history of
wave field development. They distinguished three sea states, namely,
a developed sea, an undeveloped sea and a decreasing sea. In terms
of the mechanics of wave generation, all of these states are related
to the wind fetch X and wind duration t, or more precisely to non-
gX gt

dimensional quantities such as 2 or . As shown in Figure 4,
V10 V10
at a given wind velocity V10 , fully developed seas (denoted by crosses) are
generally characterised by a greater whitecap coverage than undeveloped
seas (triangles) and seas corresponding to decreasing winds (diamonds).
Fully developed seas occur in regions where the trade winds blow, because
there winds vary at smaller time scales. On the other hand, highly
variable atmospheric conditions, as in the northern polar regions, result
in lower whitecap coverage. The least squares fitting procedure provides
 Surface waves in deep and shallow waters 19

the following relationships (in %) for the whitecap coverage of developed

and undeveloped seas (see Figure 4):
(dev)
Fcov = 0.005 · (V10 − 4.47)3 (16)
and
(undev)
Fcov = 0.00875 · (V10 − 6.33)3 . (17)
There is no doubt that a correlation between wind speed and whitecap
coverage exists, but the premise that whitecap coverage is a function
of wind speed alone is unacceptable. Whitecaps are generated by wave
breaking; nevertheless, there are many processes that control breaking and
the resulting whitecap coverage. Theoretical models based on the various
breaking criteria show close links between whitecap coverage and wave
breaking probability.

10
6
4
2
1
0.6
whitecap coverage [%]

0.4
0.2
eq. (16)
0.1
0.06
0.04
0.02
eq. (17)
0.01
0.006
0.004
0.002
0.001
4 5 6 7 8 9 10 11 12 13 14
-1
wind speed V10 [m s ]

Figure 4. Oceanic whitecap coverage – a function of wind speed in the northern

polar waters: crossess – developed sea, triangles – underdeveloped sea, diamonds
– decreasing wind (Massel 2007)

2.4.2. Wave breaking criteria and the probability of breaking

Breaking waves are usually associated with steep waves that occur in
a given sea. Most descriptions of the breaking phenomenon are still based
20 S. R. Massel

on the linear wave theory. This theory provides some convenient measures of
incipient wave breaking such as kinematic, geometric and dynamic criteria
of wave breaking (Massel 2007). In particular, a wave starts to break when
the horizontal fluid velocity at the surface u exceeds the phase velocity C
(kinematic criterion). Downstream from this point, fluid particles tend to
escape from the water surface.
According to the geometric criteria for breaking to occur, the local
surface slope εl should exceed some threshold steepness εth , i.e.
∂ζ
εl = ≥ εth . (18)
∂x
The spatial fraction of sea surface covered by whitecaps Fcov , identified with
the probability of breaking and characterised by probability density f (εl ),
becomes
Z ∞
Fcov ≈ Fbr = f (εl ) dεl . (19)
εth

For wind-induced waves, the probability density of the wave slope f (εl ) is
given by (Massel 2007):
" # " #
εl ε2 ε2l (Ic − Is )
f (εl ) = m4 √ exp − m4 l I0 , (20)
g2 Ic Is 4 g2 Ic Is 4m g 2 Ic Is
4

in which I0 (x) is the modified zero-order Bessel function (Abramowitz

& Stegun 1975), m4 is the fourth spectral moment, and the integrals Ic
and Is are
Zπ
Ic = cos2 ΘD(Θ) dΘ, (21)
−π
and
Zπ
Is = sin2 ΘD(Θ) dΘ, (22)
−π

in which D(Θ) is the directionality function of the wave field.

Let us assume that a fully developed sea is characterised by the Pierson-
Moskowitz spectrum and spilling breakers predominate. Therefore, using
the relationship (19), we obtain (Massel 2007)
" 0.22 # " 0.22 #
gX gX
 
Fcov = exp −8.265ε2th 2 = exp −0.5587 2 , (23)
V10 V10
where εth = 0.26.
If the wind fetch in formula (23) is known, Fcov becomes a function
of wind speed only. This function is shown in Figure 5 for wind fetches
 Surface waves in deep and shallow waters 21

10

1
whitecap coverage [%]

10

0.1 25

50

100
0.01 200
300
500

1000

0.001
1 2 3 4 5 6 7 8 9 10 20
wind speed V10 [m s-1]

Figure 5. Whitecap coverage, based on the limiting steepness criterion, as

a function of wind speed for chosen wind fetches. Experimental data from
Monahan & O’Muircheartaigh (1981) and Stramska & Petelski (2003) are added
for comparison (Massel 2007)

X = 10, 25, 50, 100, 200, 300, 500 and 1000 km. The experimental data, re-
analysed by Monahan & O’Muircheartaigh (1981), as well as the data
reported by Stramska & Petelski (2003), are given in the same figure.
Despite scattering, the computed whitecap coverage provides realistic values
under the assumption of an adopted wind fetch range, as the true fetches are
unknown. Closer comparison shows that small whitecap coverage appears
only for the case of decreasing winds in northern polar waters.
In a simpler way Banner et al. (2000) showed that the probability of
breaking increases close to quadratically for so-callled dominant steepness
εd = 1/2kp Hd when

Fbr = a (εd − 0.055)b , (24)

where
(Z )1/2
1.3ωp
Hd = 4 S(ω)dω , (25)
0.7ωp

in which ωp and kp are the frequency and wave number of the spectral
peak, a is in the 13.0–37.2 range (mean = 22.0) and b is in the 1.78–2.30
22 S. R. Massel

range (mean = 2.01). Massel (2007) showed that when the JONSWAP
spectrum is used, the dominant steepness εd is related to the wind velocity
V10 and wind fetch X as follows:
−0.11
gX

εd = aγ 2 , (26)
V10
where

 0.1973
 for enhancement parameter γ = 1.0
aγ = 0.2661 for enhancement parameter γ = 3.3 (27)
 0.3367 for enhancement parameter γ = 7.0


Parameter εd reflects not only the mean steepness of dominant waves, but
also the fundamental role of non-linear group dynamics in determining the
onset of breaking.
It is known that near-breaking regular waves have rounded crests with
a small radius of curvature, while the vertical downward acceleration at
the crest is less than 0.5g (it is approximately equal to 0.39g) (Massel
2007). A distinction should be made between two accelerations of the water
particles (Longuet-Higgins 1985). The first acceleration is known as the
(E)
apparent (or Eulerian) acceleration, az = ∂ 2 ζ(t)/∂t2 = ∂w(t)/∂t, while
the second one, including the convective terms, is known as the real (or
(L)
Lagrangian) acceleration az :

∂w ∂w ∂w
a(L)
z (x, ζ, t) = +u +w , (28)
∂t ∂x ∂z
in which w is the vertical velocity component at the sea surface. In the
linearised theory both acceleration definitions are equal, but in steep waves
they are different. To extend this result to the breaking of irregular waves,
let us assume that for breaking to occur, the downward acceleration at the
crest of the wave should be greater than αg, i.e.

d2 ζ
> αg, (29)
dt2

in which α is a constant. According to Snyder et al. (1983), α varies from

0.4 to 0.52, while the laboratory experiments of Ochi & Tsai (1983) provide
the value α ≈ 0.4.
The starting point for calculating the probability of wave crests breaking
in a given wave train is the probability density function of maxima
(crests) with a downward acceleration greater than αg. Under the as-
sumption that breaking will occur at a crest point on the sea surface
 Surface waves in deep and shallow waters 23

satisfying condition (29), the probability of wave breaking becomes (Srokosz

1986)
!
(αg)2
Fbr = exp − , (30)
2m4
where m4 is the spectral moment of the fourth order. As m4 → ∞, Fbr → 1
and the whole surface is covered by whitecaps. It should be noted that
the probability Fbr is independent of any assumption regarding the spectral
width, assuming that moment m4 exists.
To illustrate the dependence of the probability of breaking Fbr on wind
conditions, let us substitute a value for the spectral moment m4 following
from the JONSWAP spectrum. We then obtain the probability of breaking
in the form (Massel 2007)
" 0.22 #
gX

2
Fbr = exp −3.858α 2 . (31)
V10
The fetch dependence of the probability  of breaking is substantial. For
gX

a given fetch X, a small value of 2 is associated with a higher wind
V10
velocity
 and a higher value of the Phillips constant β. The small value of
gX

2 also provides a high value of ωp (or small Tp value). This finally
V10
yields a higher value of moment m4 and a higher probability of breaking.
Comparison with the experimental data gathered by Xu et al. (2000) in
Bohai Bay shows close agreement for the constant α ≈ 0.35 between the
experimental data and the theoretical formula.

2.4.3. Energy dissipation due to wave breaking

Whitecaps are evidence that waves are breaking and energy is being
dissipated. However, despite a great research effort, our knowledge of energy
dissipation is still fragmentary. This is mostly due to the absence of good
quantitative measures of the distribution of breaking and the rates of energy
dissipation in breaking waves.
Potentially, energy dissipation could be estimated from the energy
budget expressed in wind-wave evolution models. At present, two such
approximate methods are considered, namely, the equilibrium range model
(Phillips 1985, Hanson & Phillips 1999), and the whitecap model for
a fully developed sea (Komen et al. 1984). Tables 2 and 3 list known
experimental data and theoretical estimates of energy loss due to breaking.
The following notation is used in both tables: C is the dominant phase
speed, Cbr is the phase speed of the breaking waves, θ is the inclination
24 S. R. Massel

Table 2. Summary of experimental data on wave energy dissipation due to wave

breaking (Massel 2007)

Quantity Mathematical expression Source

Number of breaking Nb = (4.0 ± 2.0)× Thorpe (1993)

3
waves per wave ×10−3 VC10
Rate of energy Edissrate = (3.0 ± 1.8)×

5 3
loss per unit ×10−5 ρw CCbr V10 Thorpe (1993)
surface [J/m2 /s]
5
Cbr
Rate of energy Edissrate ∼ 0.009 ρw g sin θ
Duncan (1981)
Cbr5
loss per unit crest Edissrate ∼ 0.0075 ρw g sin θ
Duncan (1983)
length [J/m2 /s] Edissrate ∼ Rapp & Melville
5

Cbr
3.2 × 10−3 – 1.6 × 10−2 ρw g
(1990)
Rate of total
3.74
energy dissipation Edissrate = 4.28 × 10−5 V10 Hanson & Phillips
in the equilibrium (1990)
range [J/m2 /s]

Table 3. Summary of theoretical formulae for wave energy dissipation due to wave
breaking (Massel 2007).

Quantity Mathematical expression Source

γ1 ρw I(3s)
Rate of total energy Edissrate = 16[I(s)] 3 g3 × Hanson & Phillips
∞
11 3
R
dissipation in the × ω S (ω)dω (1999)
ωp
 2
m4
equilibrium range Edissrate = 1.59ρw g mm
0 m2
1
1
g 2 m3
Komen et al.
0
2
range [J/m /s] (1984)
Rate of energy
 2
m4
dissipation for an Edissrate = 1.59ρw gm1 1
g 2 m3
Komen et al.
0
extremely narrow (1984)
spectrum [J/m2 /s]
Energy dissipation  
A2
for a very narrow Ediss = 12 ρw gA2rms exp − A2br Longuet-Higgins
rms
spectrum [J/m2 /s] (1969)
Energy dissipation
R∞
for a two-dimensional Ediss = 12 ρw g a1 IA (T )dT Massel (2007)
0
probability density
f (A, T ) [J/m2 /s]
 Surface waves in deep and shallow waters 25

of a breaking wave’s forward face, mn is the spectral moment of order

n, γ1 is the numerical constant, I(s) is the spreading function, s is the
directional spreading parameter, Arms and Abr are the root-mean-square
and critical amplitudes respectively, a1 and IA (T ) are the functions of the
period probability density.
For example, in the Phillips equilibrium range model, the whitecapping
and the breaking process is assumed to be essentially local in the wave
number space. The components with larger wave numbers approach the
statistical equilibrium determined by the balance among the input from
the wind, wave-wave interactions and energy loss by breaking. If the wave
frequency spectra take the form of eq. (13), we obtain the total energy
dissipation rate as given in the first integral in Table 3.
The second approach by Komen et al. (1984), also shown in this Table,
is in fact an extension of the original model by Hasselmann (1974). In this
model the whitecaps are treated as a random distribution of perturbation
forces (pressure pulses), of a spatial and temporal scale that is small
compared to the wavelength and period of waves. From the physical point
of view, it is assumed that the attenuation factor in Hasselmann’s model is
due to whitecaps situated on the forward faces of waves and to downward
pressure on the upward moving water, which results in negative work on the
waves. The pressure exerted by the whitecap on the surface of the waves
and induced energy decay combined with the attenuation of short waves by
the passage of large whitecaps yields the dissipation function in the form
(Komen et al. 1984):
m  n
α̂ ω

Sdiss = Cdiss ρg ωS(ω), (32)
α̂P M ω
in which Cdiss , m, and n are fitting parameters, ω is the mean radian
frequency and α̂/α̂P M is a measure of the overall steepness of the wave field.
It should be noted that for m = 0 and n = 1, the expression (32) agrees with
Hasselmann’s (1974) result. Representation of α̂ and ω in terms of spectral
moments yields the expressions given in Table 2.
Another estimation of energy loss due to breaking for narrow spectra,
given in Table 3, results from Longuet-Higgins’ (1969) assumption that
a given wave breaks in such a way that its energy decreases from some
present value to the critical saturation level.
Figure 6 presents an example of the total dissipation rate according to
the Phillips model. The set of two lines, corresponding to various fetches,
shows almost linear dependence of the total dissipation rate on the wind
speed, and very weak influence of the wind fetches. The model results are
in satisfactory agreement with experimental data collected in the Gulf of
Alaska.
26 S. R. Massel

Toba type spectrum

1 km
20
x=
total dissipation rate [J m-2 s-1]

0.1 km
100
x=

)
0.01 999
s1
hillip
&P
on
ans
ata (H
kad
las
0.001 fA
lf o
Gu

0.0001
2 3 4 5 6 7 8 9 10 20
-1
wind speed [m s ]

Figure 6. Total dissipation rate as a function of the wind speed according to

Phillips (1985) (Massel 2007)

Figure 7. Arrangement of wave gauges in the Ocean Basin (MARINTEK)

experiment

Another comparison of observed energy dissipation with results from

a narrow spectral model is available from the Ocean Basin Experiment in
Trondheim (Norway), where mechanically generated waves were recorded at
 Surface waves in deep and shallow waters 27

Figure 8. Comparison of the experimental energy attenuation with the theoretical

attenuation resulting from Longuet-Higgins’ (1969) approach

various points for different values of incident wave steepness (Massel et al.
2001) – see Figure 7.
Using the Longuet-Higgins (1969) solution for the energy dissipation rate
(see Table 3), we obtain the following governing equation for the spectral
energy density gradient along a distance x (Massel 2007):

dE(x) ω2 b3
 
+ E(x) exp − = 0, (33)
dx 2πg E(x)
where
2
1 αg

b3 = ρg 2 , (34)
2 ωrms

in which ω is the mean frequency and ωrms is the root-mean-square

frequency. Figure 8 compares the wave energy measured at the wave staffs
distributed along the main profile during the Ocean Basin Experiment
with theoretical results. In all the tests, mechanically generated waves
corresponding to the JONSWAP spectrum with enhancement parameter
γ = 7.0 and directional distribution D(θ) ≈ cos2 Θ were used. The initial
wave train characteristics were assumed at wave staff 1 located 3.45 m from
28 S. R. Massel

the wave generator (see Figure 7). The rate of energy dissipation is shown
for three tests of different significant values of wave steepness, where
Hs
εs = , (35)
gTp2
in which Hs is the significant wave height.

2.5. Development of extreme waves

In deep water the non-linearities result in focusing of wave energy and

therefore in a greater probability of freak wave events. Freak or giant
waves correspond to unusually large amplitude waves appearing on the sea
surface. Figure 9 shows the time-history of the famous ‘New Year wave’
about 26 m in height, observed at the Draupner platform in the North Sea
on 1 January 1995.
Although various definitions of freak waves have been proposed in the
literature, it is usual to characterise a freak wave by its height Hmax , when
Hmax /Hs > 2 (Hs is the significant wave height), or by the crest height ζmax ,
when ζmax /Hs > 1.2 (Bitner-Gregersen & Hagen 2004). Accordingly, from
the Rayleigh distribution it follows that the probability of such extreme
wave formation is less than 0.000336 or one wave among more than 2980
waves.

Figure 9. ‘The New Year wave’ recorded at the Draupner platform in the North
Sea on 1 January 1995 (Janssen 2004)
 Surface waves in deep and shallow waters 29

Modelling of special wave events such as extreme, very steep and rapidly
varying waves, as well as freak waves and tsunamis, requires a phase-
resolving approach. The generation of extreme waves is usually explained by
the presence of ocean currents or specific bottom topography, energy focus
by refraction or reflection and trapping. However, the question arises as
to why exceptionally large waves occur in the open ocean away from non-
uniform currents or a special type of bathymetry? During the last 30 years,
various mathematical models of freak wave phenomena have been developed
and many laboratory experiments conducted, so that great progress has
been achieved in the understanding of the physical mechanisms involved
(Dysthe 1979, Lo & Mei 1985, Trulsen & Dysthe 1997, Onorato et al.
2000, 2001, 2006, Kharif & Pelinovsky 2003, Kurkin & Pelinovsky 2004,
Bitner-Gregersen & Hagen 2004). In particular, the following mechanisms,
based on the linear or non-linear description of wave mechanics, are
the principal potential sources of extreme wave generation: dispersion
enhancement of transient wave groups, spatial focusing of waves, wave-
current interaction.
To a first approximation, ocean surface waves can be regarded as
narrow spectrum waves when the complex envelope of the sea elevation
is described by the non-linear Schrödinger equation (Kharif & Pelinovsky
2003):
∂A ∂A ωc ∂ 2 A ωc kc2 2
 
i + Cg = + |A| A, (36)
∂t ∂x 8kc2 ∂x2 2
in which kc and ωc are the wave number and frequency of the carrier wave
respectively. The complex amplitude A is a slowly varying function of x
and t. These assumptions are physically quite realistic. For example, the
above-mentioned ‘New Year Wave’ had a maximum height of 25.6 m but

Figure 10. Formation of a highly energetic wave group ζ(x) in a slowly modulated
wave train. The time (0 on the left and 320 on the right) is normalised by the
fundamental wave period and the coordinate is in radians of the fundamental mode
(Kharif & Pelinovsky 2003)
30 S. R. Massel

the envelope of the wave train as a whole was slowly modulated, weakly
non-linear and had a relatively small bandwidth. Figure 10 demonstrates
that modulational instability leads to the decomposition of an initially
homogeneous wave train into a system of envelope quasi-solutions, and that
the spatial distribution of wave energy displays significant intermittency.
Since freak waves have a large amplitude and are of short duration,
the assumption of weak non-linearity and a narrow-banded spectrum
does not correspond exactly to real data. Substantial improvement in the
reconstruction of high-amplitude freak waves has been achieved through
the extension of weakly non-linear models by the inclusion of higher-order
terms of wave steepness, wave-induced mean flow and higher-order linear
dispersive terms (Dysthe 1979, Lo & Mei 1985, Trulsen & Dysthe 1996,
Dysthe et al. 2003).
In some cases the initial conditions for a numerical simulation are given
in the form of frequency spectra, and the so-called time-like non-linear
Schrödinger equation is used for analysis. In particular, in terms of the non-
dimensional complex amplitude Ã, normalised to the carrier wave amplitude
Ac , we have
2
∂ Ã ∆ω ∂ Ã

+i + iε2 |Ã|2 Ã = 0, (37)
∂x ωc ∂2t
in which ε = kc Ac is the carrier wave steepness and 1/∆ω is the characteristic
time scale.
Onorato et al. (2000) reported numerical simulations of the non-linear
Schrödinger equation (36) and the Dysthe-Lo-Mei equation (Lo & Mei
1985), which takes the form
2
∂ Ã ∆ω ∂ Ã ∂ Ã 8ε2 ∆ω 2 ∂ Ã ∆ω 2 ∂φ
  
2 2
+i + iε |Ã| Ã + |Ã| + 4iε Ã = 0,(38)
∂x ωc ∂2t ∂2t ωc ∂t ωc ∂t
This equation is able to account for higher-order physical effects such as the
asymmetric evolution of wave packets and side-bands; it also controls the
size of the instability region by limiting energy leakage to higher modes. In
numerical simulations of the non-linear Schrödinger and Dysthe equations,
JONSWAP-type spectra have been used, i.e.
βg2 5
 
S (ω̂) = 4 ω̂ −5 exp − ω̂ −4 γ r , (39)
ωp 4
in which ω̂ = ω/ωp , γ is the peak enhancement factor of the standard value
of 3.3, and function r takes the form
1 (ω̂ − 1)2
" #
r = exp − , (40)
2 σ02
 Surface waves in deep and shallow waters 31

where σ0 = 0.07 for ω < ωp and σ0 = 0.09 for ω > ωp . Onorato et al.
(2000) showed that the probability density of wave height substantially
depends on the enhancement factor γ and on the Phillips constant β. When
the coefficients β and γ increase, the effects of non-linearity become more
important and freak waves are more likely to occur. For a JONSWAP
spectrum with β = 0.0081 and γ = 6, simulation using the Dysthe-Lo-
Mei equation gives the probability of recording a freak wave five times
greater than the one predicted by the Rayleigh distribution. When the
linear Schrödinger equation was used in the simulation, no freak waves were
found. Moreover, the focusing of frequency modulated wave groups and the
blocking effects of spectral components on opposing currents becomes very
sensitive to the spectrum width.
Another type of extreme waves are tsunami waves. Tsunami is a Japanese
word, in fact a combination of two words: ‘ami’, which means wave, and
‘tsu’, which denotes a particular point at the waterline. Thus, a tsunami
is ‘a wave that approaches the shoreline’. A tsunami is an impulsively
driven water wave, caused by the sudden displacement of a large mass of
water. The most common cause of such a displacement is a large earthquake,
volcanic eruption, landslide either above or below the water surface, or
a large meteor impact. Out in the open ocean tsunamis have very long
wavelengths and very small amplitudes. They carry enormous energy, and
running up onto a shore, their amplitudes can increase very substantially,
causing tremendous destruction to populations and structures on shore. On
26 December 2004 a strong tsunami was generated by a magnitude 8.3
earthquake along the Andaman-Sumatra fault. Tsunami waves travelled to
every location in the World Ocean. The model by Kowalik et al. (2005)
shows that in the Southern Ocean surrounding Antarctica, in the Pacific,
and especially in the Atlantic, waves propagate over large distances by
energy trapping over oceanic ridges, which causes the amplitude to increase
over shallower depths. As the wave speed over a ridge is slower and the wave
speed away from a ridge is faster, the joint tsunami wave front is curved in
such a way that the energy is directed towards the ridge.

2.6. Steep-wave kinematics

The wave kinematics and loadings due to steep and extreme waves are
very important for offshore engineering operations, offshore structures and
ship performance. An accurate assessment of the maximum water particle
velocities beneath a high wave crest is required for drag force calculations,
while the maximum accelerations beneath the steepest section of the wave
profile are critical for inertial force calculations. The wave models used to
determine the water particle kinematics associated with a measured time
32 S. R. Massel

history of the surface elevation ζ(t) are traditionally based either on a non-
linear regular wave theory (for example, a higher-order Stokes’ theory) or
on an unsteady linear wave theory (a Fourier transform approach).
Gudmestad (1993) provided a comprehensive review of the measured
and predicted deep-water kinematics of regular and irregular waves. The
main conclusion from his review is that the velocities of regular waves at
points below the mean water level are relatively accurately predicted by
the non-linear wave theory. However, in random seas and in unsteady sea
states, where very steep waves occur, this prediction is not in agreement
with experiments, especially near the tip of the waves. In particular, a fifth-
order solution suitable for accurately predicting regular waves (Tørum
& Gudmestad 1990, Gudmestad 1993) is no longer valid close to the breaking
limit for regular waves (kH/2 ≈ 0.44) and higher-order terms must be
included. In the case of random waves, the near-surface velocities beneath
a large wave crest are significantly overestimated because the linear theory
does not allow individual wave components to ride over one another; rather,
all the components oscillate about the still water level.
In recent years, much theoretical and experimental effort has been
expended in developing better prediction models for velocities and accel-
erations. We are particularly interested in the wave characteristics at the
wave crest, where the highest velocity and acceleration, as well as wave
breaking, are most likely to occur. A full solution of the Laplace equation
under periodic lateral boundary conditions is numerically possible but time-
consuming. However, we are not usually interested in wave parameters
at any time within the wave period. Therefore, it is more useful to
concentrate on methodologies that seek only to represent the local (close
to the wave crest) behaviour of waves. Moreover, this region coincides with
the region of maximum errors in theoretical predictions. To overcome this
difficulty, an empirical stretching technique (Wheeler 1970), the best fit of
the experimental profile to the fifth-order Stokes profile (Massel 2007) and
two alternative local methods for solving irregular wave problems (Sobey
1992, Baldock & Swan 1994) were propounded.
Sobey’s (1992) local methodology compromises applicability in a global
sense to achieve as exact a representation of wave motion as possible in
a local sense. In particular, in his local Fourier approximation methodology,
the field equation throughout the fluid domain, the bottom boundary
condition at the bed and the free surface boundary conditions at the water
surface are satisfied within a window of duration τ , which is small in
comparison with the local zero-crossing period. This solution gives an
excellent description of the crest kinematics, but is unable to model the
 Surface waves in deep and shallow waters 33

global non-linearities and results in a poor description of the kinematics in

the lower layers of the flow.
In contrast to the Sobey model, the method put forward by Baldock
& Swan (1994) provides a ‘global’ solution in the sense that it represents
a complete irregular wave record of several waves through the inclusion
of both time and space dependence. Thus it can model both local
and global non-linearities, but is limited in terms of the total number
of Fourier components that can practically be included, resulting in
some underestimation of the near-surface kinematics. Also, in certain
circumstances, Baldock & Swan’s method overestimates the kinematics
beneath the still water level. Overall, however, the double Fourier series
solution provides both the most accurate and the most reliable estimate of
the water particle kinematics for extreme 2D waves (Baldock et al. 1996,
Smith & Swan 2002).
This is not the case for freak waves and near-breaking waves when
the kinematics is not well predicted by any existing theory. A new,
fully non-linear unsteady wave model has recently been developed by
Clamond & Grue (2001a,b), which offers a substantial improvement in
the modelling of steep wave kinematics. The model is based on potential
theory and assumes the integration of the kinematic and dynamic boundary
conditions at the free surface of very steep (freak) waves. The theoretical
results compare favourably with various available sets of experimental data.

Figure 11. Comparison of experimental velocity profile with theoretical results:

solid line – nonlinear solution, broken line – exponential profile (Grue et al. 2003)
34 S. R. Massel

Figure 11 compares the experimental velocity profile with the theoretical

results of Grue et al. (2003). Measurements and computations show some
deep water waves with a fluid velocity up to 75% of the phase speed.
The PIV and LDA data demonstrate excellent agreement with the fully
non-linear theory. The experimental data correspond to the JONSWAP
spectrum with ωp2 Hs /2g = 0.15, and the range of the wave slope is 0.21 <
ε < 0.34. These data are compared to the non-linear computations with
steepness ε = 0.29. A surprising result is that the exponential profile eky
compares well with all measurements in deep water.
Accelerations at the sea surface are required to estimate the inertia
forces on offshore installations. Very few measurements exist of the real
acceleration below steep waves (see, for example, Bonmarin & Kjeldsen
2000, Grue & Jensen 2006). In particular, Grue and Jensen reported
a maximum negative vertical acceleration at the wave crest of about 1.1g,
while the upward vertical acceleration grows to about 1.5g in the front face of
the wave at the base below the overturning jet of the plunging breakers. The
fully non-linear theoretical model for accelerations in unsteady waves shows
very good agreement with experiments. It was shown that the convective
term is of the same order of magnitude and of opposite sign to the local
acceleration, and it cannot be neglected when estimating the acceleration
and forces (Jensen et al. 2007).

2.7. Dependence of aerosol fluxes on wave breaking

conditions
All non-gaseous particles suspended in the atmosphere are called
aerosols. Aerosols are composed mainly of droplets and crystals but also
of organic matter particles and large ions. These constituents can be both
natural and anthropogenic: liquid seawater drops, dry sea salt particles,
dust transport from the deserts by wind, as well as particles resulting from
human activities related to industrial processes and agriculture, volcanic
eruptions, and meteorite destruction in the atmosphere. Aerosol emission
from the global ocean is one of the major natural sources of aerosols in the
atmosphere. Because approximately 71% of the Earth’s surface is covered
by oceans, marine aerosols play an important role in various geochemical
and geophysical processes and in the Earth’s climate as a whole. It is
estimated that the annual production of marigenic aerosols varies from
0.3 × 1012 kg to 30 × 1012 kg, corresponding to a sea aerosol mass flux
over the oceans of 0.03 × 10−3 kg m−2 s−1 to 3 × 10−3 kg m−2 s−1 (Lewis
& Schwartz 2004).
The marine aerosol generation functions depend on the product of the
(wc)
so-called size-dependent production flux fprod (r) and whitecap coverage
 Surface waves in deep and shallow waters 35

Table 4. Monthly sea salt production (×106 kg) in the Baltic Sea in 1999 (Massel
2007)

Month Criterion for determination of whitecap coverage

limiting steepness limiting vertical acceleration
January 1.39 4.87
February 3.11 8.28
March 0.68 3.07
April 0.62 2.96
May 0.15 1.13
June 0.09 0.69
July 0.17 1.23
August 0.06 0.61
September 0.26 1.60
October 1.70 5.83
November 3.37 8.66
December 7.97 14.96

(wc)
Fcov . The production flux fprod (r) is only a function of radius r, while the
whitecap coverage Fcov depends on the rate of energy dissipation and sea
state. Therefore seasonal wind and wave data for a given sea basin provide
an opportunity to estimate the aerosol production there. For example,
Table 4 summarises the monthly sea salt production for 1999 in the Baltic
Sea. Sea salt production is highest during winter and lowest during summer.
Taking into account the fact that the surface area of the Baltic Sea is about
384 700 km2 , the monthly average sea salt production per km2 is 20.72–
38.90 kg during stormy weather.
This estimate of sea salt production is based on the assumption that
the whitecap coverage is parameterised in terms of the sea state parameters
through two wave breaking criteria, i.e. the limiting steepness and the
limiting vertical acceleration. The table indicates that sea salt production
based on the limiting vertical acceleration criterion is higher than that
calculated using the limiting steepness criterion. In particular, during the
winter months, the ratio of both production rates is about 2 and for smaller
waves in summer, the difference between the methods of calculation is even
greater.

3. Waves in shallow waters

3.1. Action balance equation components for shallow water
Present-day coastal engineering projects are becoming more complex,
and advanced near-shore wave models are required to provide results of
36 S. R. Massel

greater accuracy. In particular, design calculations are needed to assess the

applied water loading on structures, the stability and transport of bed
material, and the run-up on beaches and over-topping of breakwaters. The
great accuracy of models requires the formulation of predictive models in
which non-linearities and the unsteady character of the wave field are both
taken into account.
Predictive models in shallow waters are based on the action balance
equation, similar to eq. (1). In Cartesian co–ordinates we have (Holthuijsen
2007)
∂N ∂[Cg,x N ] ∂[Cg,y N ] ∂[Cθ N ] ∂[Cσ N ]
+ + + + = S, (41)
∂t ∂x ∂y ∂θ ∂σ
in which Cg,x and Cg,y are group velocities in x- and y-space respectively.
The fourth term, with phase velocity Cθ in θ-space, represents depth-
induced and current-induced refraction. The fifth term is the shifting of
the frequency due to variation in depth and currents with phase velocity Cσ
in σ-space. In shallow water, the source term S includes generation by wind,
non-linear wave-wave interaction (triad interaction), energy dissipation due
to wave breaking and bottom friction.
In shallow water a transfer of energy takes place from two primary waves
to a third wave, through the near-resonant triad interactions, in contrast
to the quadruplet wave-wave interactions in deep waters. The necessary
requirements are that the sum of the frequencies and the vector wave
numbers of the interacting waves must be (nearly) zero (Massel 1996), i.e.
ω1 ± ω2 ± ω3 = 0 (42)
and
k1 ± k2 ± k3 = kδ (43)
The mismatch in the wave number kδ approaches zero only for extremely
shallow water, where waves are non-dispersive.
In sea waters of limited depth with a sandy bottom, bottom friction
is the dominant mechanism for dissipation. This mechanism covers the
complicated processes in the relatively thin turbulent layer at the bottom.
Using the quadratic law to estimate the shear stress at the bottom, the
energy-dissipation rate for random waves becomes (Collins 1972)
Df r = −ρw Cf r u2rms,bottom · urms,bottom , (44)
in which urms,bottom is the root-mean-square orbital velocity at the bottom.
In general, if an accurate description of water particle kinematics and
dynamics is required, in particular close to the water surface, then the wave
models applied must incorporate both the non-linearity and unsteadiness
of events. Some recent theoretical developments in single frequency wave
 Surface waves in deep and shallow waters 37

mechanics for constant as well as for varying water depths are discussed in
some detail in the following Sections.

3.2. Waves in waters of constant depth

In waters of intermediate depths the water particle kinematics and
dynamics, corresponding to a measured or predicted time-history of surface
elevation ζ(t), are usually determined according to Stokes’ non-linear steady
wave theory (Fenton 1985), while in very shallow waters of constant depth,
a non-linear cnoidal theory is more appropriate (Fenton 1979). In linear
models, the unsteadiness and irregular nature of surface waves may be
included, but the non-linearity is not. The observed surface elevation is
assumed to be the sum of freely propagating wave components, each of
which satisfies the linear dispersion relation. However, this type of solution,
particularly for near-surface velocities, results in significant errors due to
high frequency contamination.
For non-linear and unsteady surface waves Longuet-Higgins & Cokelet
(1976) in their pioneer study developed an exact method for extreme 2D
waves. They showed that the non-linear free surface conditions can be re-
arranged in such a way that a spatial representation of surface elevation
and velocity potential Φ are defined at some initial time t = t0 , and the
entire solution can be time-matched to give values for ζ(x) and Φ(ζ, x) at
all subsequent times. This procedure was extended by Dold & Peregrine
(1984) for a spatially periodic wave field with no restriction on the water
depth. First, they considered six laboratory tests with different periods,
corresponding to deep, intermediate and shallow water conditions, and
two spectra – a broad-banded spectrum and a narrow-banded spectrum.
These tests were successfully reproduced by Dold and Peregrine’s numerical
model for adopted initial conditions. Subsequently, Smith and Swan (2002)
used these exact numerical calculations rather than laboratory data as
benchmark data to compare numerical results with other prediction models.
These comparisons indicate that the fifth-order Stokes’ solution or linear
random wave theory provides a poor description of water particle kinematics
as it is unable to model the unsteadiness and the non-linearities that are
common in extreme wave events. On the other hand, the local Fourier series
solution (Sobey 1992) and the double Fourier series (Baldock & Swan 1994)
mentioned above provide improved representations of the surface profile and
wave kinematics, also for waves propagating in finite water depths.
Fenton (1986) developed an alternative local approximation method for
unsteady waves in shallow water. In this method, the local solution is
represented by a truncated polynomial series for the complex potential
function. The unknown polynomial terms are determined numerically to
38 S. R. Massel

fit the non-linear free surface conditions using a measured time history of

the surface elevation. This procedure is best suited for long waves where the
vertical variation in the fluid velocities can be described by a polynomial
function.

3.3. Wave propagation over gradually varying water depths

Propagation of surface waves over a bottom of varying depth is
a mathematically difficult problem. Therefore, although the non-linear
effects become significant close to the shoreline, a linear solution is still very
useful, and approximate wave models retaining only the essential features
of the problem are used. If relative wave height and bottom slope are small,
non-linearities are weak, the mild-slope equation developed by Berkhoff
(1972) and many of its alternative derivations and numerous extensions
provide effective tools to predict wave variation in regions of moderate
size. In all these solutions, the vertical structure of the wave potential
is predetermined.
According to Booij (1983), the mild-slope equation gives accurate results
even with a plane bottom slope up to 1:3. However, steep bottom slopes,
such as underwater shoals and coral reefs (Lie & Tørum 1991, Massel 1993,
1996) require the inclusion of higher-order terms and possibly wave breaking
mechanisms in the mild-slope equation for the wave amplitude A variation:
d2 A d(CCg ) dA
2
+ (CCg )−1 + [k2 (1 + ψ) + iγk]A = 0, (45)
dx dx dx
in which
 2
dh g d2 h
ψ = E1 (kh) + E2 (kh) 2 , (46)
dx ω dx2
and γ = γbr + γf r . The ψ term describes the influence of bottom slope
dh/dx and bottom shape d2 h/dx2 , and the γ term represents the sum of
energy dissipated due to wave breaking γbr and bottom friction γf r , and E1
and E2 are functions of non-dimensional water depth (Massel 1996).
The modelling of energy dissipation due to breaking in the wave train,
is usually based on four main assumptions:
– dissipation is equivalent to dissipation in a bore connecting two regions
of uniform flow (Battjes & Janssen 1978),
– dissipation is proportional to the difference between the local energy
flux and the stable energy flux (Dally et al. 1985),
– the breaking wave height is saturated, i.e. the wave height is propor-
tional to the local water depth and the proportionality coefficient is
assumed to be constant across the surf zone,
– dissipation is controlled by the presence of a surface roller (Svendsen
1984).
 Surface waves in deep and shallow waters 39

For example, using the first approach, the quantity γbr becomes (Massel
1996)
√
3α0 ωp gh Hrms Hrms 4
 
γbr = √ , (47)
4 π CCg h Γh
in which Hrms is the root-mean-square wave height, ωp is the peak frequency,
α0 is a coefficient of O(1) and Γ ≈ 0.3–0.5.
The rate of energy dissipation due to bottom friction γf r is given by
(Massel 1996):
16Cf r u3bottom
γf r = 2
. (48)
3π gCg Hrms
where Cf r is the friction factor, and ubottom is the velocity amplitude at the
bed.
In recent years, many papers on the applicability of the mild-slope
equation to coastal problems have been published. For example, the
extended refraction-diffraction equation has been applied to predict wave
transformation and breaking as well as a wave-induced set-up on two-
dimensional reef profiles of various shapes (Massel & Gourlay 2000).
Comparison of predicted and observed wave heights and set-up values
showed good agreement.
However, equation (45) does not satisfy exactly the Neumann condition
on a sloping bottom. This means that the velocity field in the vicinity of the
bottom is poorly represented and wave energy is not generally conserved.
In order to improve the mild-slope representation in the bottom layer,
Athanassoulis & Belibassakis (1999) developed the consistent coupled-mode
theory, in which an additional term, called the sloping-bottom mode, was
introduced to satisfy the bottom condition exactly. The equation for wave
amplitude A then takes the form
∞
X
A(x, z) = A−1 (x) · Z−1 (z, x) + A0 (x) · Z0 (z, x) + An (x) · Zn (z, x),(49)
n=1

where A0 (x) Z0 (z, x) denotes the propagating mode, and the remaining
terms An (x) Zn (z, x) are the evanescent modes. The functions Z0 (z, x)
and Zn (z, x) are the classical functions representing the z-dependence of
wave motion for propagating and evanescent modes respectively. The addi-
tional sloping-bottom mode A−1 (x) Z−1 (z, x) provides a proper Neumann
condition over a non-horizontal bottom when the function Z−1 (z, x) takes
the form
" 3 2 #
z z

Z−1 (z, x) = h(x) + . (50)
h(x) h(x)
40 S. R. Massel

This idea was further extended to a 3D environment (Belibassakis et al.

2001) and to second-order Stokes waves over variable bathymetry (Belibas-
sakis & Athanassoulis 2002). By way of example, Figure 12 presents the
resulting equipotential lines (real and imaginary parts) when 6 evanescent
modes are included. Now we can see that the equipotential lines intersect
the bottom profile perpendicularly.

Figure 12. Equipotential lines for wave motion over topography: a) real part of
the potential, b) imaginary part of the potential (Athanassoulis & Belibassakis
1999)

3.4. Run-up of waves at a beach and wave-induced groundwater

circulation
3.4.1. Waves on beaches
Sandy beaches are highly exploited but very dynamic and fragile
environments. The beach system is driven largely by the physical energy
induced by waves and tides. In tideless seas, the flow is totally controlled by
the dynamics of surface waves on the beach. Wave motion on beaches is very
complex and the groundwater flow is different in different beach regions. In
Region 3, (see Figure 13), the wave run-up contributes mainly to the raising
of the coastal water table. On the other hand, the beach groundwater flow
in the set-up region induces a groundwater circulation that contributes to
the submarine groundwater discharge.
 Surface waves in deep and shallow waters 41

Figure 13. Reference scheme and relationships between wave run-up, infiltration
and coastal watertable (Massel & Pelinovsky 2001)

Propagating waves transport not only energy but also momentum, which
is a vector quantity. The transport of momentum is equivalent to a stress
and known as radiation stress (Longuet-Higgins & Stewart 1964). As waves
propagate towards the shore, they become steeper as the water depth
becomes shallower; at a certain depth they lose their stability and finally
start to break. When waves break, wave energy is dissipated and the
radiation stress is reduced, which gives rise to changes in the mean sea level
(MSL). The balance of the sea level gradient and the gradient of radiation
stress takes the form
dSxx dη
+ ρg (h + η) = 0, (51)
dx dx
in which η is the change of MSL due to wave action, and Sxx is the
radiation stress tensor component. The change of η due to wave action is
shown schematically in Figure 13. For water depth decreasing monotonically
towards the beach, the maximum set-down of η br = −1/16Hbr 2 /h appears
br
close to the breaking point, while the maximum set-up η max = 5/16Hbr 2 /h
br
occurs at the dynamic waterline. The set-up depends on the incoming wave
height. Thus, for a stationary situation, the set-up is stationary. However,
the waves usually arrive in a group, which causes the set-up to fluctuate.
The wave run-up height Rmax is defined here as the maximum vertical
height above still water level reached by the wave uprush. The run-up height
is always greater than the wave set-up. On the other hand, wave run-down
is defined as the lowest vertical height reached by the backwash of a wave
before the uprush of the next wave starts to run-up the beach face.
42 S. R. Massel

The wave run-up limit and induced water infiltration into a beach body
is a response to the instantaneous flow of the surface water. Therefore,
modelling the surface oscillation should be based on the phase-resolving
wave type model. Available run-up models usually assume that waves are
non-dispersive and that the phase velocity depends on the water depth only
(Carrier & Greenspan 1958, Pelinovsky 1996). This assumption is applicable
to tsunami and wind-induced waves very close to the shoreline.
However, in deeper water, waves are usually dispersive. Thus, we
need an approach in which the dispersive character of waves is maintained
seawards and the approximation of shallow water is used close to the
waterline. Massel & Pelinovsky (2001) attempted to develop a more
complex approach for the run-up of dispersive breaking and on-breaking
waves. Waves approaching the shallow water area were modelled by
the mild-slope equation. At very small water depths, the non-linear
and linear equations for shallow water waves are considered and the
dissipation due to wave breaking is included, providing a more realistic
estimation of run-up characteristics. For long, non-dispersive waves, the
governing equations are usually based on the Carrier & Greenspan (1958)
transformation and its various modifications (Pelinovsky 1996, Belibassakis
& Athanassoulis 2006).
In the simple case when a plane slope merges into a horizontal bottom,
the surface elevation ζ(x, t) over the sloping bottom takes the form
s 
Hi iDb 4ω 2 (−x) 
 
ζ(x, t) = ℜ KT J0  1+ exp(−iωt), (52)
2 ω gβ1
in which Hi is the incident wave height, KT is the transmission coefficient
of wave motion from the horizontal bottom region to the sloping bed region
β1 is the beach slope, Db is the dissipation factor due to wave breaking and
J0 (x) is a zero-order Bessel function of the first kind. The transmission
coefficient KT becomes
2
KT = q , (53)
J0 (ǫ) − i 1 + i Dωb J1 (ǫ)
in which
s
iDb 4ω 2 (−x)

ǫ= 1+ . (54)
ω gβ
The values of run-up, observed in the experiments, also include the set-up
mechanism. Therefore, the final maximum run-up height becomes
Hi
Rmax = | KT | +η max , (55)
2
where η max is the solution of eq. (51).
 Surface waves in deep and shallow waters 43

The experimental data on wave run-up are numerous and some compar-
isons between experiments and theory can be found Massel & Pelinovsky
(2001).
Another convenient method of treating the problem of wave motion
in a plane-beach (wedge) region was published recently by Belibassakis
& Athanassoulis (2006), who used cylindrical-polar coordinates (x = r cos θ,
z = r sin θ) in the vertical plane with the origin at the waterline. If we ignore
the evanescent modes, the governing equation for the wave amplitude A(r),
close to the waterline (r → 0) has the form
∂ 2 A 1 ∂A 1 µ µθ
 
+ + − A = 0, (56)
∂r 2 r ∂r r θ 3
where µ = ω 2 /g and h = −r sin β1 .
Belibassakis and Athanassoulis showed that the solution of the above
equation is in agreement with that of equation (51). This solution is
sometimes used as the initial value for solving the more complex problem
of wave propagation over a sloping sea bottom.

3.4.2. Wave-induced groundwater circulation

Water flow through the beach body is of great importance for introduc-
ing water, organic materials and oxygen to the ground environment, as well
as for sediment transport, coastal structure stability and modern beach
nourishment techniques. Water flow controls the vertical and horizontal,
chemical and biological gradients, and nutrient exchange in the beach,
which helps to maintain biological activity in the porous media. This
provides a basis for assessing the vulnerability of a beach’s biodiversity
and the functioning of the interaction between tourism, natural changes
and the physical properties of the sea in such an ecosystem (Węsławski
et al. 2000).
As it is diffcult to estimate the infiltration of water into beach sand
under real sea conditions, a controlled, almost full-scale experiment was
carried out in the Large Wave Channel in Hanover (Germany) (Massel et
al. 2004, 2005). During the experiment, a water depth of 4 m in front of
the beach was assumed. Natural beach sand with fine-grained sand of mean
diameter D50 = 0.24 mm was used and a uniform beach with a 1/20 slope was
created in the channel. To measure the pore water pressure, four systems
of pressure gauges were installed along the beach face. In each system, four
piezoelectric pressure sensors were fixed to a metal rod arranged in the form
of a cross. Such an arrangement enables not only the pore pressure to be
estimated, but also the horizontal and vertical water velocities in the beach
body through the simultaneous measurements of pore pressures.
44 S. R. Massel

The experiment revealed the dual nature of the recorded pore pressures.
Beyond the breaker zone, only the rapidly-varying phase-resolving pore
pressure component due to surface variation is observed, while within the
surf zone, the phase-resolving component as well as the slowly varying phase-
averaged pore pressure component due to wave set-up were present (Massel
2001, Massel et al. 2004, 2005).
To explain the nature of the rapidly varying pore-pressure component
and the resulting velocity circulation, an exact close-form solution based on
the Biot theory for multiphase flow has been developed. This solution takes
into account soil deformations, volume change and pore-water pressure.
When the stiffness ratio G/Ew ′ ≥ 100 (G is the shear modulus of the soil and,
′
Ew is the apparent bulk modulus of the pore water), the vertical distribution
of the pore pressure is very close to the Moshagen & Tørum (1975) solution
assuming that the soil is rigid and the fluid compressible.
The apparent bulk modulus depends on the degree of saturation by air,
which is very difficult to estimate under experimental conditions. Extensive
field measurements carried out by de Rouck & Troch (2002) showed that
there was approximately 3% gas in the soil pores. Tørum (2007) argued that
in laboratory conditions, the air/gas content can be in the range 3–10%.
Figure 14 compares the theoretical results with the Large Wave Channel
experimental tests (Massel et al. 2005). It shows the long-wave case

normalised pressure amplitude

-0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
-2.0

-2.5 zero gas content

Test 2 (long surface waves)

depth z [m]

approx. sol.
exact sol.
-3.0 T=8s
Hin = 0.5 m
h = 2.0 m
hn = 4.0 m
n = 0.26
-3.5 E'w = 2.8 × 105 N m-2
Es= 108 N m-2
Kf = 2.9 × 10-4 m s-1

-4.0
non-porous sea bottom
experimental data

Figure 14. Comparison between experimental and theoretical values of pore

pressure (Massel et al. 2005)
 Surface waves in deep and shallow waters 45

and three solutions, namely the exact closed-form solution, the Moshagen
& Tørum (1975) solution and the special case of a rigid soil and incompress-
ible water. The solution for a partly saturated soil, when G/Ew ′ → ∞, is

very close to the exact solution and compares well with the experimental
data.
The radiation stress tensor Sxx induces a change in the mean water
level that exhibits two different horizontal steady pressure gradients. These
steady pressure gradients induce two systems of pore water circulation,
related to the signs of different gradients (Massel 2001). For the offshore
gradient, the horizontal excess pressure carries the flow in the offshore
direction. However, closer to the shore, the pressure gradient is reversed
and the resulting flow moves shorewards. The final circulation pattern due
to wave run-up on a porous beach is the result of the combined impact of
the phase-resolving and phase-averaged pore pressure components.

4. Conclusions
This overview examines the role that ocean waves play in the interactions
of atmosphere and ocean. In particular, it is demonstrated that ocean waves
evolve in space and in time according to the well-known energy balance
equation. This equation is the basis for modern wave forecasting techniques.
Although these techniques have not been discussed in the review, it should
be stressed that significant improvements in wave forecasting have been
made in the last ten years. To a large extent this is related to substantial
progess in the description of wind forcing and other processes, as well as
to the more efficient use of satellite observations and assimilation methods.
An observation that strikes one when reviewing the modelling techniques
of surface wave propagation is the increasing variety and complexity of
models in which more physical processes are included, greater precision
and resolution are achieved and extended ranges of applicability are
demonstrated. In particular, substantial progress has been made in the
modelling of freak and tsunami waves. Estimated velocities and the forces
induced by these events can help improve the design of ships and offshore
structures.
In coastal waters, more processes have to be taken into account
than in oceanic waters. The modified mild-slope equation approach offers
a more accurate description of wave propagation over a sloping bed.
However, highly non-linear phenomena such as wave breaking and wave
run-up require new theoretical ideas and more precise experimental data.
In general, the selection of any model should be based on a proper
appreciation of the physical processes to be modelled. Finally, in order to
estimate the applicability of particular models, comparison with high quality
46 S. R. Massel

experimental data, collected in nature or under laboratory conditions, is

necessary.

Acknowledgements
The author would like to thank the anonymous reviewers for their very
helpful suggestions and improvements to this paper.

References
Abramowitz M., Stegun I. A., 1975, Handbook of mathematical functions, Dover
Publ., New York, 1045 pp.
Athanassoulis G. A., Belibassakis K. A., 1999, A consistent coupled-mode theory
for the propagation of small-amplitude water waves over variable bathymetry
regions, J. Fluid Mech., 389, 275–301.
Baldock T. E., Swan C., 1994, Numerical calculations of large transient water
waves, Appl. Ocean Res., 16 (2), 101–112.
Baldock T. E., Swan C., Taylor P. H., 1996, A laboratory study of non-linear surface
waves on water, Philos. T. Roy. Soc. A, 354 (1707), 1–28.
Banner M. L., Babanin A. V., Young I. R., 2000, Breaking probability for dominant
waves on the sea surface, J. Phys. Oceanogr., 30 (12), 3145–3160.
Banner M. L., Song J. B., 2002, On determining the onset and strength of breaking
for deep water waves. Part II: Influence of wind forcing and surface shear,
J. Phys. Oceanogr., 32 (9), 2559–2570.
Battjes J. A., Jansen J. P. F. M., 1978, Energy loss and set-up due to breaking of
random waves, Proc. 16th Coastal Eng. Conf., 1, 563–587.
Belibassakis K. A., Athanassoulis G. A., 2002, Extension of second-order Stokes
theory to variable bathymetry, J. Fluid Mech., 464, 35–80.
Belibassakis K. A., Athanassoulis G. A., 2006, A coupled-mode technique for the
run-up of non-breaking dispersive waves on plane beaches, Proc. 25th Int. Conf.
Offshore Mech. Arctic Eng., OMAE 2006 – 92162, 1–8.
Belibassakis K. A., Athanassoulis G. A., Gerostathis T. P., 2001, A coupled-mode
model for the refraction-diffraction of linear waves over steep three-dimensional
bathymetry, Appl. Ocean Res., 23 (6), 319–336.
Benjamin T. B., 1967, Instability of periodic wavetrains in nonlinear dispersive
systems, P. Roy. Soc. Lond. A Mat., 299 (1456), 59–76.
Benjamin T. B., Feir J. E., 1967, The disintegration of wave trains in deep water.
Part 1. Theory, J. Fluid Mech., 27 (3), 417–430.
Berkhoff J. C. W., 1972, Computation of combined refraction-diffraction, Proc. 13th
Coastal Eng. Conf., 1, 471–490.
Bitner-Gregersen E. M., Hagen Ø., 2004, Freak wave events within the second order
wave model, Proc. 23rd Int. Conf. Offshore Mech. Arctic Eng., OMAE 2004 –
51410.
 Surface waves in deep and shallow waters 47

Bonmarin P., Kjeldsen P., 2000, Some geometric and kinematic properties of
breaking waves, [in:] Rogue waves 2000, M. Olagnon & G. A. Athanassoulis
(eds.), Edn. Ifremer, Brest, 169–180.
Booij N., 1983, A note on the accuracy of the mild-slope equation, Coast. Eng.,
7 (3), 191–203.
Bretherton F. P., Garrett C. J. R., 1969, Wave trains in inhomogeneous moving
media, P. Roy. Soc. Lond. A Mat., 302 (1471), 529–554.
Carrier G. F., Greenspan H. P., 1958, Water waves of finite amplitude on a sloping
beach, J. Fluid Mech., 4 (1), 97–109.
Clamond D., Grue J., 2001a, A fast method for fully non-linear water-wave
computations, J. Fluid Mech., 447, 337–355.
Clamond D., Grue J., 2001b, On efficient numerical simulations of freak waves,
Proc. 11th Int. Offshore Polar Eng. Conf.
Collins J. I., 1972, Prediction of shallow-water spectra, J. Geophys. Res., 77 (15),
2693–2707.
Dally W. R., Dean R. G., Dalrymple R. A., 1985, Wave height variation across
beaches of arbitrary profile, J. Geophys. Res., 90 (C6), 11 917–11 927.
Davidan I. N., Lopatukhin L. I., Rozkhov W. A., 1978, Wind waves as random
hydrodynamic process, Gidrometeoizdat, Leningrad, 287 pp., (in Russian).
Davidan I. N., Lopatukhin L. I., Rozkhov W. A., 1985, Wind waves in World Ocean,
Gidrometeoizdat, Leningrad, 254 pp., (in Russian).
Dean R. G., Dalrymple R. A., 1998, Water wave mechanics for engineers and
scientists, Adv. Ser. Ocean Eng., Vol. 2, World Sci. Publ., Singapore, 353 pp.
de Rouck J., Troch P., 2002, Pore water pressure response due to tides and waves
based on prototype measurements, PIANC Bull., 110, 9–31.
Dold J. W., Peregrine D. H., 1984, Steep unsteady waves: an efficient computational
scheme, Proc. 19th Int. Conf. Coastal Eng., 1, 955–967.
Dold J. W., Peregrine D. H., 1986, Water-wave modulation, Proc. 20th Int. Conf.
Coastal Eng., 1, 163–175.
Duncan J. H., 1981, An investigation of breaking waves produced by a towed
hydrofoil, P. Roy. Soc. Lond. A Mat., 377 (1770), 331–348.
Duncan J. H., 1983, The breaking and non-breaking wave resistance of a two-
dimensional hydrofoil, J. Fluid Mech., 126, 507–520.
Dysthe K. B., 1979, Note on a modification to the non-linear Schrödinger equation
for application to deep water waves, P. Roy. Soc. Lond. A Mat., 369 (1736),
105–114.
Dysthe K. B., Trulsen K., Krogstad H. E., Socquet-Juglard H., 2003, Evolution of
a narrow-band spectrum of random surface gravity waves, J. Fluid Mech., 478,
1–10.
Fenton J. D., 1979, A high-order cnoidal wave theory, J. Fluid Mech., 94 (1), 129–
161.
48 S. R. Massel

Fenton J. D., 1985, A fifth order Stokes’ theory for steady waves, J. Waterw. Port C.
Div., 111 (2), 216–234.
Fenton J. D., 1986, Polynomial approximation and water waves, Proc. 20th Coastal
Eng. Conf., 1, 193–207.
Goda Y., 2000, Random seas and design of maritime structures, World Sci. Publ.,
Singapore, 443 pp.
Grue J., Clamond D., Huseby M., Jensen A., 2003, Kinematics of extreme waves
in deep water, Appl. Ocean Res., 25 (6), 355–366.
Grue J., Jensen A., 2006, Experimental velocities and accelerations in very steep
wave events in deep water, Eur. J. Mech. B-Fluid., 25 (5), 554–564.
Gudmestad O. T., 1993, Measured and predicted deep water wave kinematics in
regular and irregular seas, Mar. Struct., 6 (1), 1–73.
Hanson J. L., Phillips O. M., 1999, Wind sea growth and dissipation in the open
ocean, J. Phys. Oceanogr., 29 (8), 1633–1648.
Hasselmann K., 1974, On the spectral dissipation of ocean waves due to white
capping, Bound.-Lay. Meteorol., 6 (1–2), 107–127.
Hasselmann S., Hasselmann K., Allender J. H., Barnett T. P., 1985, Computations
and parameterizations of the non-linear energy transfer in a gravity wave
spectrum. Part II: Parameterizations of the non-linear energy transfer for
application in wave models, J. Phys. Oceanogr., 15 (11), 1378–1391.
Holthuijsen L. H., 2007, Waves in oceanic and coastal waters, Cambridge Univ.
Press, Cambridge, 387 pp.
Janssen P., 2004, The interaction of ocean waves and wind, Cambridge Univ. Press,
Cambridge, 300 pp.
Jensen A., Clamond D., Huseby M., Grue J., 2007, On local and convective
acceleration in steep wave events, Ocean Eng., 34 (3–4), 426–435.
Kharif C., Pelinovsky E., 2003, Physical mechanisms of the rogue wave
phenomenon, Eur. J. Mech. B-Fluids, 22 (6), 603–634.
Kitaigorodskii S. A., 1983, On the theory of the equilibrium range in the spectrum
of wind-generated gravity waves, J. Phys. Oceanogr., 13 (5), 816–827.
Komen G. J., Cavaleri L., Donelan M., Hasselmann K., Hasselmann S., Janssen
P. A. E. M., 1994, Dynamics and modelling of ocean waves, Cambridge Univ.
Press, Cambridge, 532 pp.
Komen G. J., Hasselmann S., Hasselmann K., 1984, On the existence of a fully
developed wind-sea spectrum, J. Phys. Oceanogr., 14 (8), 1271–1285.
Kowalik Z., Knight W., Logan T., Whitmore P., 2005, Numerical modeling of the
global tsunami: Indonesian tsunami of 26 December 2004, Sci. Tsunami Haz.,
23 (1), 40–56.
Krasitskii V. P., 1994, On reduced equations in the Hamiltonian theory of weakly
non-linear surface waves, J. Fluid Mech., 272, 1–20.
Krylov J. M. (ed.), 1986, Wind, waves and marine ports, Gidrometeoizdat,
Leningrad, 264 pp., (in Russian).
 Surface waves in deep and shallow waters 49

Krylov J. M., Strekalov S. S., Tsyplukhin W. F., 1976, Wind waves and their
interaction with structures, Gidrometeoizdat, Leningrad, 256 pp., (in Russian).
Kurkin A. A., Pelinovsky E. N., 2004, Freak waves: Facts, theory and modelling,
Tech. Univ. Nizhny Novgorod, 157 pp., (in Russian).
Lake B. M., Yuen H. C., Rungaldier H., Ferguson W. E., 1977, Nonlinear deep-water
waves: Theory and experiment. Part 2. Evolution of a continuous wave train,
J. Fluid Mech., 83 (1), 49–74.
Lamb H., 1932, Hydrodynamics, Dover Publ., Inc., New York, 738 pp.
Lavrenov I. V., 2003, Wind-waves in oceans. Dynamics and numerical simulation,
Springer-Verlag, Berlin, 376 pp.
LeBlond P. H., Mysak L. A., 1978, Waves in the ocean, Elsevier Oceanogr. Ser., 20,
Elsevier, Amsterdam, 602 pp.
Lewis E. R., Schwartz S. E., 2004, Sea salt aerosol production. Mechanisms,
methods, measurements, and models, Geophys. Monogr., Vol. 152, American
Geophys. Union, Washington, 412 pp.
Lie V., Tørum A., 1991, Ocean waves over shoals, Coast. Eng., 15 (5–6), 545–562.
Lo F., Mei C. C., 1985, A numerical study of water-wave modulation based on
a high-order non-linear Schrödinger equation, J. Fluid Mech., 150, 395–416.
Longuet-Higgins M. S., 1969, On wave breaking and the equilibrium spectrum of
wind-generated waves, P. Roy. Soc. Lond. A Mat., 310 (1501), 151–159.
Longuet-Higgins M. S., 1978, The instabilities of gravity waves of finite amplitude in
deep water. II. Subharmonics, P. Roy. Soc. Lond. A Mat., 360 (1703), 489–505.
Longuet-Higgins M. S., 1985, Acceleration in steep gravity waves, J. Phys.
Oceanogr., 15 (11), 1570–1579.
Longuet-Higgins M. S., Cokelet E. D., 1976, The deformation of steep surface waves
on water. I. A numerical method of computation, P. Roy. Soc. Lond. A Mat.,
350 (1660), 1–26.
Longuet-Higgins M. S., Stewart R. W., 1964, Radiation stress in water waves:
A physical discussion with applications, Deep-Sea Res., 11 (4), 529–562.
Luke J. C., 1967, A variational principle for a fluid with a free surface, J. Fluid
Mech., 27 (2), 395–397.
Massel S. R., 1989, Hydrodynamics of coastal zones, Elsevier, Amsterdam, 316 pp.
Massel S. R., 1993, Extended refraction-diffraction equations for surface waves,
Coastal Eng., 19 (1–2), 97–126.
Massel S. R., 1996, Ocean surface waves: Their physics and prediction, World Sci.
Publ., Singapore, 491 pp.
Massel S. R., 2001, Circulation of groundwater due to wave set-up on a permeable
beach, Oceanologia, 43 (3), 279–290.
Massel S. R., 2007, Ocean waves breaking and marine aerosol fluxes, Springer, New
York, 323 pp.
Massel S. R., Gourlay M. R., 2000, On the modelling of wave breaking and set-up
on coral reefs, Coast. Eng., 39 (1), 1–27.
50 S. R. Massel

Massel S. R., Pelinovsky E. N., 2001, Run-up of dispersive and breaking waves on
beaches, Oceanologia, 43 (1), 61–97.
Massel S. R., Przyborska A., Przyborski M., 2004, Attenuation of wave-induced
groundwater pressure in shallow water. Part 1, Oceanologia, 46 (3), 383–404.
Massel S. R., Przyborska A., Przyborski M., 2005, Attenuation of wave-induced
groundwater pressure in shallow water. Part 2. Theory, Oceanologia, 47 (3),
291–323.
Mei C. C., 1989, The applied dynamics of ocean surface waves, World Sci. Publ.,
Singapore, 740 pp.
Mei C. C., Stiassnie M., Yue D. K.-P., 2006, Theory and applications of ocean
surface waves. Part 1: Linear aspects, Part 2: Nonlinear aspects, Adv. Ser.
Ocean Eng., Vol. 23, World Sci. Publ., Singapore, Vol. 1 – 1071 pp., Vol. 2 –
569 pp.
Miles J. W., 1957, On the generation of surface waves by shear flows, J. Fluid
Mech., 3 (2), 185–204.
Monahan E. C., 1971, Oceanic whitecaps, J. Phys. Oceanogr., 1 (2), 139–144.
Monahan E. C., O’Muircheartaigh I., 1981, Optimal power-law description of
oceanic whitecap coverage dependence on wind speed, J. Phys. Oceanogr.,
10 (12), 2094–2099.
Moshagen H., Tørum A., 1975, Wave induced pressure in permeable sea beds,
J. Waterway. Div. – ASCE, 101 (1), 49–57.
Ochi M. K., 1998, Ocean waves. The stochastic approach, Cambridge Univ. Press,
Cambridge, 319 pp.
Ochi M. K., Tsai C.-H., 1983, Prediction of occurrence of breaking waves in deep
water, J. Phys. Oceanogr., 13 (11), 2008–2019.
Onorato M., Osborne A. R., Serio M., Bertone S., 2001, Freak waves in random
oceanic sea states, Phys. Rev. Lett., 86 (25), 5831–5834.
Onorato M., Osborne A. R., Serio M., Cavaleri L., Braudini C., Stansberg C. T.,
2006, Extreme waves, modulational instability and second order theory: Wave
flume experiments on irregular waves, Eur. J. Mech. B-Fluids, 25 (5), 586–601.
Onorato M., Osborne A. R., Serio M., Damiani T., 2000, Occurrence of freak waves
from envelope equations in random ocean wave simulations, [in:] Rogue waves
2000, M. Olagnon & G. A. Athanassoulis (eds.), Edn. Ifremer, Brest, 181–192.
Pelinovsky E. N., 1996, Tsunami waves hydrodynamics, Inst. Appl. Phys., Nizhny
Novgorod, 274 pp., (in Russian).
Phillips O. M., 1957, On the generation of waves by turbulent wind, J. Fluid Mech.,
2 (5), 417–445.
Phillips O. M., 1977, The dynamics of the upper ocean, Cambridge Univ. Press,
Cambridge, 336 pp.
Phillips O. M., 1985, Spectral and statistical properties of the equilibrium range in
wind-generated gravity waves, J. Fluid Mech., 156, 505–531.
Rapp R. J., Melville W. K., 1990, Laboratory measurements of deep-water breaking
waves, Philos. T. Roy. Soc. A, 331 (1622), 735–800.
 Surface waves in deep and shallow waters 51

Sarpkaya T., Isaacson M. St. Q., 1981, Mechanics of waves forces on offshore
structures, Van Nostrand Reinhold, New York, 651 pp.
Smith S. F., Swan C., 2002, Extreme two-dimensional water waves: An assessment
of potential design solutions, Ocean Eng., 29 (4), 387–416.
Snyder R. L., Smith L., Kennedy R. M., 1983, On the formation of whitecaps by
a threshold mechanism. Part III: Field experiment and comparison with theory,
J. Phys. Oceanogr., 13 (8), 1505–1518.
Sobey R. J., 1992, A local Fourier approximation method for irregular wave
kinematics, Appl. Ocean Res., 14 (2), 93–105.
Song J.-B., Banner M. L., 2002, On determining the onset and strength of breaking
for deep water waves. Part I: Unforced irrotational wave groups, J. Phys.
Oceanogr., 32 (9), 2541–2558.
Srokosz M. A., 1986, On the probability of wave breaking in deep water, J. Phys.
Oceanogr., 16 (2), 382–385.
Stoker J. J., 1957, Water waves, the mathematical theory with applications,
Interscience Publ., Inc., New York, 567 pp.
Stramska M., Petelski T., 2003, Observations of oceanic whitecaps in the
north polar waters of the Atlantic, J. Geophys. Res., 108 (C3), 3086,
doi:10.1029/2002JC001321.
Svendsen I. A., 1984, Wave heights and set-up in a surf zone, Coast. Eng., 8 (4),
303–329.
Thorpe S. A., 1993, Energy loss by breaking waves, J. Phys. Oceanogr., 23 (11),
2498–2502.
Toba Y., 1973, Local balance in the air-sea boundary process. III. On the spectrum
of wind waves, J. Oceanogr. Soc. Jpn., 29 (5), 209–220.
Tørum A., 2007, Wave induced pore pressure–air/gas content, J. Waterw. Port C.
Div., 133 (1), 83–86.
Tørum A., Gudmestad O. T. (eds.), 1990, Water wave kinematics, Kluwer Acad.
Publ., Dordrecht, 771 pp.
Trulsen K., Dysthe K. B., 1996, A modified non-linear Schrödinger equation for
broader bandwidth gravity waves on deep water, Wave Motion, 24 (3), 281–
289.
Trulsen K., Dysthe K., 1997, Freak waves – a three-dimensional wave simulation,
Proc. 21 Symp. Nav. Hydrodyn., Nat. Acad. Press, 550–560.
Tulin M. P., Waseda T., 1999, Laboratory observations of wave group evolution
including breaking effects, J. Fluid Mech., 378, 197–232.
WAMDI group: Hasselmann S., Hasselmann K., Bauer E., Janssen P. A. E. M.,
Komen G. J., Bertotti L., Lionello P., Guilaune A., Cordone V. C., Greenwood
M., Reistad L., Zambresky L., Ewing J. A., 1988, The WAM model:
A third generation ocean wave prediction model, J. Phys. Oceanogr., 18 (12),
1775–1810.
Waseda T., Tulin M. P., 1999, Experimental study of the stability of deep-water
wave trains including wind effects, J. Fluid Mech., 401, 55–84.
52 S. R. Massel

Węsławski J. M., Urban-Malinga B., Kotwicki L., Opaliński K., Szymelfenig M.,
2000, Sandy coastlines – are there conflicts between recreation and natural
values?, Oceanol. Stud., 29 (2), 5–18.
Wheeler J. D., 1970, Method for calculating forces produced by irregular waves,
Proc. Offshore Technol. Conf., 1, 71–82.
Xu D., Liu X., Yu D., 2000, Probability of wave breaking and whitecap coverage in
a fetch-limited sea, J. Geophys. Res., 105 (C6), 14 253–14 259.
Zakharov V. E., 1968, Stability of periodic waves of finite amplitude on the surface
of a deep fluid, J. Appl. Mech. Tech. Phy., 9 (2), 190–194.