Download as pdf or txt
Download as pdf or txt
You are on page 1of 19

Nonlinear Analysis: Real World Applications 57 (2021) 103176

Contents lists available at ScienceDirect

Nonlinear Analysis: Real World Applications


www.elsevier.com/locate/nonrwa

Resonances for water waves over flows with piecewise constant


vorticity
Calin Iulian Martin a ,∗, Biswajit Basu b
a
Faculty of Mathematics, University of Vienna, Oskar-Morgenstern-Platz 1, 1090 Vienna, Austria
b
School of Engineering, Trinity College Dublin, Dublin 2, Ireland

article info abstract

Article history: We investigate three and four-wave resonances of capillary–gravity water waves
Received 27 March 2020 arising as the free surface of water flows exhibiting piecewise constant vorticity.
Accepted 27 June 2020 More precisely, our type of flow has a jump in the vorticity distribution that
Available online xxxx separates a rotational layer at the top (commonly generated by wind-shear) from
Keywords: another rotational layer adjacent to the sea-bed, region that accommodates strong
Three-and four-wave resonances sheared currents. Instrumental in deriving our findings is the dispersion relation
Dispersion relation for such flows having a jump in the distribution of the vorticity. In general, for
Piecewise constant vorticity rotational flows of non-constant vorticity, the dispersion relation is very intricate.
However, we show that a disentanglement occurs in the case of capillary–gravity
water waves for which the ratio “thickness of the near-surface vortical layer/the
wavelength of the surface wave” is sufficiently large. More precisely, we find
explicitly two solutions, λ0 and λ1 that represent the (relative) surface wave
speed. We then confirm analytically that λ1 gives rise to three-wave resonances
for capillary–gravity water waves with wavelengths not exceeding 2 cm. However,
for significant wavelength range, we establish that λ1 does not lead to four-
wave resonances. In contrast to the previous conclusion λ0 does not bring about
three-wave resonances, but is able to generate four-wave resonances.
© 2020 The Author(s). Published by Elsevier Ltd. This is an open access article under
the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction

The confluence of two, or more, (mother) water waves giving rise to another (daughter) wave bears the
name resonance. From a historical perspective, the study of resonances was initiated by Phillips [1], who
showed that two (or three) trains of gravity waves may interact so as to produce a fourth (tertiary) wave
whose wave-number is different from any of the three (four) primary wave-numbers, and whose amplitude
grows in time. The issue was further developed by Longuet-Higgins [2], Longuet-Higgins and Phillips [3],
Phillips [4]; for an account of the early findings on resonances we refer the reader to the paper by Hammack
and Henderson [5].

∗ Corresponding author.
E-mail addresses: calin.martin@univie.ac.at (C.I. Martin), basub@tcd.ie (B. Basu).

https://doi.org/10.1016/j.nonrwa.2020.103176
1468-1218/© 2020 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
2 C.I. Martin and B. Basu / Nonlinear Analysis: Real World Applications 57 (2021) 103176

A nonlinear process by necessity – for it is known that wave fields governed by linear propagation
equations evolve independently – the resonance is an event that has a marked influence on the evolution
of waves. Indeed, significant energy transfer occurs during the resonance process, cf. [5–7] and the references
therein. The transfer of energy among wave components might give rise to the formation of large waves.
In line with these ideas, it was established recently [8] that three-wave resonant interaction (a prominent
subject of our study) gives rise to rogue waves, which are oceanic phenomena responsible for a number
of maritime disasters.1 Experimental quantification of the possibility to predict the emergence of rogue
waves were recently performed in [9,10]. The possibility of formation of rogue waves from resonance is also
confirmed in the numerical study [11] on “long wave-short wave resonance” in the case of a density stratified
fluid in the absence of vorticity.
Moreover, resonant interactions of water waves may give rise to the development of weak turbulence
[12,13]. Valuable insights into the onset and the evolution of turbulence in shear flows are given by
Eckhart [14–16], by Moehlis, Faisst and Eckhardt [17] and Schumacher and Eckhardt [18]. For insightful
discussions on the identification of coherent structures in turbulent flows we refer the reader to Haller [19,20].
Along with the previous ideas, such resonant interactions may generate a significant change in the
spectrum of ocean waves within a few hours. An average spectrum represented in the frequency domain
(satisfying stochastic criteria) provides the basis for generating waveforms, a process that usually follows
empirical approaches geared towards engineering applications, e.g. the Joint North Sea Wave Project
(JONSWAP). Such applications, like vibration control of offshore platforms and floating wind turbine
structures [21,22] or control of wave energy converters for maximizing power output [23], require the
development of time-histories derived from simulations of water wave profiles [24,25].
However, most of the applications mentioned earlier adopt an empirical approach, which bypasses the
difficulties associated with the generation of wave profiles based on a firm theoretical description that
demands the fulfillment of the nonlinear water wave equations together with their nonlinear kinematic free
surface boundary condition. A further aspect that is (to a great extent) overlooked (in both applied and
analytical studies) is the presence of currents in water flows and their interactions with the surface waves,
as documented by [26,27]. To accommodate the latter facets one has to work within the setting of rotational
water waves, perspective that is, by far, more demanding if compared to the scenario of irrotational water
waves, where the availability of a velocity potential significantly simplifies the arguments. With respect
to this aspect we refer the reader to the comprehensive book [28] for the treatment of various topics on
rotational water flows; for discussions relating the vorticity and swirling motions of fluid flows, and on recent
progress in the identification of vortices from features of the velocity field, see [29–31])
The integration of vorticity in the analytical investigation of nonlinear water waves commenced with the
study by Constantin and Strauss [32], where the case of a regular vorticity distribution was considered,
the occurrence of a rough vorticity being considered by the same authors in [33]. The results from [32,33],
delivering existence of (global) curves of solutions to the water wave problem, were enhanced by studies on
the symmetry of rotational water waves [34–36], on regularity [37–39], or on the stability [40], an issue whose
treatment relies much upon the existence of suitable variational formulations [41,42].
In the present paper we account for both nonlinearity and rotationality in the investigation of resonances
with the hope that the consideration of these effects will lead to relevant engineering applications. However,
this is a task that will be undertaken in future studies. A pioneering work on understanding the nonlinear
resonances for constant vorticity water waves was performed by Constantin and Kartashova [43] in the
context of capillary waves. Resonances for capillary–gravity water waves above linearly sheared flows were
investigated in [44,45]. As far as our paper is concerned, we derive in Section 2 the dispersion relation for
capillary–gravity water waves in the case of a sheared water flow with a jump in the vorticity distribution.

1
7 rogue wave disasters, from Columbus to cruise ships, New Scientist, 29 July 2014.
C.I. Martin and B. Basu / Nonlinear Analysis: Real World Applications 57 (2021) 103176 3

The dispersion relation is a formula that displays the (relative) speed of the surface wave in terms of the
physical parameters of the flow like the mean depth, the vorticity, the (constant) density, the wavelength and
the position at which the jump in vorticity occurs. While, for the type of flows we analyze, the dispersion
relation is concealed as the positive solution of a convoluted third order algebraic equation, we succeed
to explicitly derive it in the case of capillary–gravity water waves over flows for which the quantity dL0 is
sufficiently large–d0 represents the thickness of the near surface vortical layer and L is the wavelength. More
precisely, we find two positive explicit solutions of the third order algebraic equation, λ0 and λ1 , which give
the relative surface wave speed. Moreover, in the setting of pure gravity waves having moderate wavelength
L, we are still able to provide the explicit dispersion relation as long as L is not too big if compared with
the size d0 of the near surface vortical layer.
Based on the derived dispersion relation we establish in Section 3 that capillary–gravity water waves
(corresponding to λ1 with wavelengths not exceeding 2 cm) satisfy the three-wave resonance conditions (3.7)
for exactly one (negative) constant value of the vorticity γ1 . Moreover, the special value of the vorticity that
triggers resonances is given in terms of the wave numbers of the two mother waves and of the physical
parameters g (gravity), σ (surface tension) and ρ (the constant density). On the contrary, capillary–gravity
waves corresponding to λ0 do not exhibit three-wave resonances.
We would like to mention that resonant three-wave interactions for irrotational capillary–gravity waves
occur in the case of non-collinear wave trains as confirmed experimentally in a closed laboratory tank by
Haudin et al. [46]. In contrast, rotational collinear gravity water waves do not exhibit three-wave resonances,
as it is evidentiated at the end of Section 3.
We further in Section 4 our investigation to the analysis of four-wave resonances. While capillary–
gravity waves corresponding to λ0 give rise to four-wave resonances (irrespective of the wave number) those
pertaining to λ1 behave quite differently, as explained below. We derive first a necessary condition for
four-wave resonance to take place and then show that the obtained condition is not satisfied for capillary–
gravity water waves (corresponding to λ1 ) with wavelengths comprised between 1.7 cm and 2.7 cm. This
seems to constitute a marked difference when compared with the irrotational case. Indeed, employing the
Zakharov equation [47], and making use of the deep-water linear dispersion relation (for gravity waves),
Shemer and Stiassnie [48] have studied the evolution characteristics of an integrable system of four waves
above irrotational flow.
The notable difference between three and four-wave resonance found in Sections 3 and 4 is also
perceptible within the weak turbulence theory where energy transfer through the scales occurs by three-wave
interactions, instead of four-wave interactions [49].
However, trivial resonances occur not only for capillary–gravity water waves but also in the pure
gravity case. That is, denoting with k1 , k2 , k3 , k4 the involved wave numbers, we obtain that the resonance
conditions hold true for tuples of type (k1 , k2 , k1 , k2 ) and (k1 , k2 , k2 , k1 ). Four-wave resonances were observed
experimentally by Bonnefoy et al. [50] for oblique gravity waves obeying the linear dispersion relation

ω(k) = gk in the absence of currents. For an extensive description concerning the state of the art on
analytical, computational and more applied aspects concerning water wave resonances we refer the reader
to [51–57].

2. The water wave problem and the dispersion relation

2.1. Formulation of the water wave problem

Under consideration are steady periodic (of period L > 0) capillary–gravity water waves which arise as
the free surface – denoted y = η(x) – of a two-dimensional rotational, inviscid and incompressible water flow
that is bounded below by the flat bed y = −d, cf. Fig. 1.
4 C.I. Martin and B. Basu / Nonlinear Analysis: Real World Applications 57 (2021) 103176

Fig. 1. A cross-section of the fluid domain.

The water waves we consider propagate in the positive x-direction under the influence of gravity and
surface tension. In a reference frame moving in the same direction with the wave and with speed c > 0, the
equations of motion are Euler’s equations
{
(u − c)ux + vuy = −Px
(2.1)
(u − c)vx + vvy = −Py − g,
and the equation of mass conservation
ux + vy = 0, (2.2)
whereby (u, v) denotes the velocity field, P is the pressure and g is the gravitational constant. The shear in
the flow is captured by the vorticity function, defined as

ω(x, y) := uy − vx .

The specification of the water wave problem is completed by the kinematic and dynamic boundary
conditions. The kinematic boundary conditions require that the free surface, y = η(x), and the bed y = −d
(where d > 0 is a constant) consists of the same particles, throughout the evolution of the flow. As kinematic
boundary conditions we have {
v = (u − c)ηx on y = η(x)
(2.3)
v = 0 on y = −d
The dynamic boundary condition accounts for the idea that the pressure jump across the free surface is
proportional to its mean curvature. Therefore, denoting with Patm the constant atmospheric pressure and
with σ the coefficient of surface tension, the dynamic boundary condition is written in mathematical terms
as
ηxx
P = Patm − σ 3 , (2.4)
(1 + ηx2 ) 2
The considerations in this paper concern the situation of the absence of stagnation point, that is we will
assume that
u < c throughout the fluid. (2.5)
We aim to simplify the presentation of the problem (2.1)–(2.4) and notice that due to (2.2) we obtain the
existence of a stream function ψ, which is defined (up to a constant) by the relations

ψx = −v, ψy = u − c.
C.I. Martin and B. Basu / Nonlinear Analysis: Real World Applications 57 (2021) 103176 5

Moreover, the no stagnation condition (2.5) guarantees cf. [28,33] that the vorticity function, ω, is a single
valued function of ψ. More precisely, there is a function s → γ(s) such that

ω(x, y) = γ(−ψ(x, y)),

for all x, y with −d ≤ y ≤ η(x).


The previous considerations about the stream function allow us to write the governing equations and the
boundary conditions (2.1)–(2.4) only in terms of the stream function and of η as the system

∆ψ = γ(−ψ) in −d < y < η(x),
ηxx


⎨|∇ψ|2 + 2g(y + d) − 2σ

= Q on y = η(x),
(1 + ηx2 )3/2 (2.6)


⎪ ψ = 0 on y = η(x),
ψ = −p0 on y = −d,

where Q is a constant related to the total head, and p0 < 0 is a constant representing the relative mass flux,
given by ∫ η(x)
p0 = (u(x, y) − c) dy.
−d
We aim now to further simplify the system (2.6). To this end we use the Dubreil-Jacotin transform

q(x, y) = x, p(x, y) = −ψ(x, y) (2.7)

and obtain that (2.6) is equivalent to the quasilinear elliptic boundary value problem


⎪ (1 + h2q )hpp − 2hp hq hpq + h2p hqq − γh3p = 0 in Ω ,

h2p hqq

2 2 (2.8)
1 + hq + (2gh − Q)h p − 2σ = 0 on p = 0,


⎪ (1 + h2q )3/2
h = 0 on p = p0 ,

where the unknown function h defined on Ω := [0, L] × [p0 , 0] by h(q, p) := y + d represents the height above
the flat bed and is even and of period L in the q-variable. The condition of no stagnation points in the fluid
is equivalent to the elliptic non degeneracy condition

hp > 0 in Ω. (2.9)

We are interested in this paper in the situation when the range of the vorticity function consists of two
values. More precisely, we assume that there are γ1 , γ2 ∈ R with γ1 ̸= γ2 and p1 ∈ (p0 , 0) such that
{
γ2 , for p ∈ [p0 , p1 ]
γ(p) =
γ1 , for p ∈ (p1 , 0).

Indeed, we would like to remark that Fedorov and Brown [58] give the values γ1 ≈ −1.25 × 10−2 s−1 and
γ2 ≈ 2.5 × 10−4 s−1 for the Equatorial Undercurrent.
The latter vorticity distribution requires the utilization of a diffraction (or transmission) problem. Namely,
setting Ω1 = (0, L) × (p0 , p1 ), Ω2 = (0, L) × (p1 , 0) we look for (h, H) ∈ C 2+α (Ω 1 ) × C 2+α (Ω 2 ) satisfying
(1 + hq2 )hpp − 2hp hq hpq + hp2 hqq − γ1 hp3 = 0


⎪ in Ω1 ,
2 2 3




⎪ (1 + Hq )Hpp − 2Hp Hq Hpq + Hp Hqq − γ2 Hp = 0 in Ω2 ,
⎪ 2
⎪ Hp Hqq
1 + Hq2 + (2gH − Q)Hp2 − 2σ

= 0 on p = 0, (2.10)
⎪ (1 + Hq2 )3/2



⎪ h = H on p = p1 ,



⎪ hp = Hp on p = p1 ,

h = 0 on p = p0 ,
6 C.I. Martin and B. Basu / Nonlinear Analysis: Real World Applications 57 (2021) 103176

Using Crandall–Rabinowitz theory we proved in [59] the existence of local bifurcation curves consisting of
solutions of (2.10) (under the assumption (2.9)) and emerging from laminar (trivial) solutions of (2.10). The
laminar solutions represent water waves with a flat free surface and parallel streamlines. If (h, H) denotes
a laminar flow solution which depends only on the variable p then it solves (2.10) and (2.9) if and only if it
solves the system ⎧ ′′ ′3

⎪ h = γ1 h in p0 < p < p1 ,
⎪ ′′ ′3
H = γ2 H in p1 < p < 0,



′2


1 + (2gH(0) − Q)H (0) = 0, (2.11)

⎪ h(p1 ) = H(p1 ),
′ ′





⎩ h (p1 ) = H (p1 ),
h(p0 ) = 0.
The family of laminar solutions solving (2.11) is given by
∫ p
1
h(p) := h(p; λ) := √ ds, p ∈ [p0 , p1 ],
p0 λ − 2Γ (s)
∫ p (2.12)
1
H(p) := H(p; λ) := √ ds, p ∈ [p1 , 0],
p0 λ − 2Γ (s)
∫p
where the parameter λ satisfies λ > 2 max[p0 ,0] Γ , with Γ being the function p ∈ [p0 , 0] → Γ (p) = 0
γ(s) ds.
Moreover, the parameter λ is related to the speed at the surface of the laminar flows by
√ 1
λ= = (c − u)|free surface . (2.13)
Hp (0)

2.2. The derivation of the dispersion relation



Let us set a(λ, p) = λ − 2Γ (p), p ∈ [p0 , 0] and let k := 2π L be the wave number. The necessary and
sufficient conditions for the existence of waves of small amplitude obtained as perturbations of the laminar
flows (2.12) is (see [59]) that the problem

(a3 mp )p = k 2 mv in (p0 , p1 ), (2.14a)


3 2
(a Mp )p = k aM in (p1 , 0), (2.14b)
m=M on p = p1 , (2.14c)
mp = Mp on p = p1 , (2.14d)
σ
a3 Mp = (g + k 2 )M on p = 0, (2.14e)
ρ
m = 0 on p = p0 , (2.14f)

has a non-trivial solution (m, M ) ∈ C 2+α ([p0 , p1 ]) × C 2+α ([p1 , 0]).


Let p1 ∈ (p0 , 0) denote the height of the discontinuity of the vorticity such that the upper layer
[p1 , 0] has vorticity γ1 and the bottom layer [p0 , p1 ) has vorticity γ2 . We will focus on finding (m, M ) ∈
C 2+α ([p0 , p1 ]) × C 2+α ([p1 , 0]) that obey the system (2.14a)–(2.14f). To this end we compute first
{√
√ λ − 2γ1 p, if p ∈ [p1 , 0],
a(p, λ) = λ − 2Γ (p) = √ (2.15)
λ − 2(p − p1 )γ2 − 2p1 γ1 , if p ∈ [p0 , p1 ).

To find m that satisfies Eq. (2.14a), we make the Ansatz


( )
2γ2 a(p)
m(p) = m̃ k (2.16)
a(p) γ2
C.I. Martin and B. Basu / Nonlinear Analysis: Real World Applications 57 (2021) 103176 7

which together with the bottom boundary condition (2.14f) gives that
( )
2γ2 a(p) − a(p0 )
m(p) = c̃ sinh k ,
a(p) γ2

for some constant c̃. Employing a similar Ansatz as in (2.16) we find that
( ( ) ( ))
2γ1 ka(p) ka(p)
M (p) = c̃1 cosh + c̃2 sinh ,
a(p) γ1 γ1
for some real constants c̃1 , c̃2 .
Utilizing the matching conditions (2.14c) and (2.14d) we find that
( ( ))
sinh(k(θ − ρ)) sinh(kθ) sinh(kρ) γ2
c̃1 = c̃γ2 + −1 (2.17)
γ1 ka(p1 ) γ
( ( 1 ))
cosh(k(θ − ρ)) sinh(kθ) cosh(kρ) γ2
c̃2 = c̃γ2 + 1− (2.18)
γ1 ka(p1 ) γ1
where
a(p1 ) − a(p0 ) a(p1 )
θ := , ρ := .
γ2 γ1
Making use of the top boundary condition (2.14e) we obtain the relation
( ( )) ( ( √ ) ( √ ))
1 σ 2 k λ k λ
γ1 − √ g+ k c̃1 cosh + c̃2 sinh
λ ρ γ1 γ1
( ( √ ) ( √ )) (2.19)
√ k λ k λ
−k λ c̃1 sinh + c̃2 cosh = 0.
γ1 γ1

To make (2.19) more explicit we obtain, after a routine calculation, that


( √ ) ( √ )
k λ k λ
c̃1 cosh + c̃2 sinh (2.20a)
γ1 γ1
[ ( ( √ )) ( ) ( ( √ ))]
1 λ sinh(kθ) γ2 λ
= c̃γ2 sinh k θ − ρ + + − 1 sinh k ρ −
γ1 γ1 ka(p1 ) γ1 γ1

and
( √ ) ( √ )
k λ k λ
c̃1 sinh + c̃2 cosh (2.20b)
γ1 γ1
[ ( ( √ )) ( ) ( ( √ ))]
1 λ sinh(kθ) γ2 λ
= c̃γ2 cosh k θ − ρ + − − 1 cosh k ρ −
γ1 γ1 ka(p1 ) γ1 γ1

We denote now with d0 the average depth corresponding to the value p1 . That is, we have
∫ 0
p1 = (u(y) − c) dy, (2.21)
−d0

where u represents here the horizontal component of the velocity field corresponding to the laminar flow.

This means that uy = ω and, since (c − u)(0) = λ, we have that
{ √
−√λ + γ1 y, y ∈ [−d0 , 0],
u(y) − c = (2.22)
− λ − γ1 d0 + γ2 (y + d0 ), y ∈ [−d, −d0 ].
8 C.I. Martin and B. Basu / Nonlinear Analysis: Real World Applications 57 (2021) 103176

Returning now to (2.21) we obtain the equation in the unknown d0


γ1 2 √
d + λd0 + p1 = 0, (2.23)
2 0
whose only solution is √ √
− λ + λ − 2γ1 p1 a(p1 ) − a(0)
d0 = = , (2.24)
γ1 γ1
the last equality being a consequence of the formula (2.15). From the last equality we obtain immediately
that √
a(p1 ) λ
ρ= = + d0 . (2.25)
γ1 γ1
We seek now to find a formula for θ in terms of d, d0 and (possibly) λ. As before, we start with the mass
flux p0 . That is, we write
∫ −d0 ∫ 0
p0 = (u(y) − c)dy + (u(y) − c)dy
−d −d0
∫ −d0 √ (2.26)
∫ 0 √
= [− λ − γ1 d0 + γ2 (y + d0 )]dy + [− λ + γ1 y]dy
−d −d0

and get the second order equation in d


γ2 2 √ γ2 − γ1 2
d + [ λ + (γ1 − γ2 )d0 ]d + d0 + p0 = 0. (2.27)
2 2
Making use of the formula (2.23) we obtain that the discriminant, ∆, of the previous equation equals a(p0 ).
Routine calculations give that
γ1 a(p0 ) − a(0)
d − d0 = − d0 + . (2.28)
γ2 γ2
From the last equality we infer that γ1 d0 + γ2 (d − d0 ) = a(p0 ) − a(0) which, together with (2.24), gives
a(p1 ) − a(p0 )
θ= = d0 − d. (2.29)
γ2
Thus, we infer from equalities (2.25) and (2.29) that
√ √
λ λ
ρ− = d0 and θ − ρ + = −d,
γ1 γ1
which substituted in (2.19) and (2.20) lead to following dispersion relation
3 1
cosh(kd)λ 2 + [(γ2 − γ1 ) sinh(k(d − d0 )) cosh(kd0 ) + γ1 kd0 cosh(kd) + γ1 sinh(kd)] λ
k
g̃(k) − γ12 d0 √
[ ]
γ1 (γ1 − γ2 )
− 2
sinh(k(d − d0 )) sinh(kd 0 ) + sinh(kd) λ (2.30)
k k
g̃(k)
− 2 [γ1 kd0 sinh(kd) + (γ2 − γ1 ) sinh(k(d − d0 )) sinh(kd0 )] = 0,
k
where g̃(k) = g + σρ k 2 .
We note that setting d0 := d in (2.30) (that is, considering the constant vorticity case) and dividing by
cosh(kd) we get the equation
3 γ1 tanh(kd) g̃(k) − γ12 d √ g̃(k)
λ 2 + [γ1 d + ]λ − tanh(kd) λ − γ1 d tanh(kd) = 0,
k k k
which can be rewritten as
√ γ1 tanh(kd) √ √ g̃(k) √
λ( λ + γ1 d) + ( λ + γ1 d) λ − tanh(kd)( λ + γ1 d) = 0.
k k
C.I. Martin and B. Basu / Nonlinear Analysis: Real World Applications 57 (2021) 103176 9


Appealing
√ to (2.22), and recalling that u < c, we see that λ + γ1 d > 0. Therefore, we can divide above by
λ + γ1 d and obtain that
γ1 tanh(kd) √ g̃(k)
λ+ λ− tanh(kd) = 0,
k k
that is, we recover the dispersion relation for water flows with constant vorticity [28], thus modeling sheared
currents [60–62].

2.3. The formula for λ pertaining to flows with piecewise constant vorticity

We proceed to study the resonance problem for water waves over flows displaying a piecewise constant
distribution in vorticity. It is to note that, in the capillary–gravity regime, kd and kd0 are significantly large,
so that tanh(kd) = 1. Thus, Eq. (2.30) becomes
[ ( ) ]
3 1 (γ2 − γ1 ) sinh(kd − 2kd0 )
λ2 + · 1+ + γ1 kd0 + γ1 λ
k 2 cosh(kd)
g̃(k) − γ12 d0 √
[ ( ) ]
γ1 (γ1 − γ2 ) cosh(kd − 2kd0 )
− · 1− + λ (2.31)
2k 2 cosh(kd) k
[ ( )]
g̃(k) γ2 − γ1 cosh(kd − 2kd0 )
− 2 γ1 kd0 sinh(kd) + · 1− = 0.
k 2 cosh(kd)
For the capillary–gravity waves regime (for which tanh(kd) = 1) we have that
sinh(kd − 2kd0 )
= tanh(kd) cosh(2kd0 ) − sinh(2kd0 ) = e−2kd0 ,
cosh(kd)
cosh(kd − 2kd0 )
= cosh(2kd0 ) − sinh(2kd0 ) tanh(kd) = e−2kd0 .
cosh(kd)
We recall that d0 represents the thickness of the vortical layer adjacent to the free surface. It is reasonable
to assume that the thickness of this layer is (in many real world situations) about 1 m, cf. [63]. The latter
and the fact that the wavelength of capillary–gravity waves is of the order of centimeters imply that one can
effectively consider that e−2kd0 vanishes in (2.31). The previous considerations lead then to the following
dispersion relation
γ1 − γ2 √
( ) ( )
3 1 γ1 + γ2 γ1
λ2 + + γ1 kd0 λ + γ1 d0 − λ
k 2 k 2k
(2.32)
g̃(k) √
( )
γ1 − γ2
− λ + γ1 d0 − =0
k 2k
Note that we can rewrite the first line of (2.32) as
γ1 − γ2 √
( ) ( )
3 γ1 − γ2 γ1 γ1
λ + γ1 d0 −
2 + λ+ γ1 d0 − λ
2k k k 2k
(√ γ1 ) (
γ1 − γ2 √ √
) ( γ1 )
=λ λ+ + γ1 d0 − λ λ+
k 2k k
√ (√ γ1 ) √
( )
γ1 − γ2
= λ λ+ λ + γ1 d0 − .
k 2k
Thus, the dispersion relation becomes
γ1 √ √
( )( )
g̃(k) γ1 − γ2
λ+ λ− λ + γ1 d0 − = 0. (2.33)
k k 2k

We denote with λ1 the positive solution of the equation obtained by setting the first bracket equal to zero,

and with λ0 the solution that emerges from the second bracket above, that is,
√ γ1 − γ2
λ0 = −γ1 d0 + .
2k
10 C.I. Martin and B. Basu / Nonlinear Analysis: Real World Applications 57 (2021) 103176

3. Three-wave resonances

3.1. Investigation of resonances pertaining to λ0

We argue now that the solution


√ γ1 − γ2
λ = −γ1 d0 +
2k
gives rise to non-resonant interactions. Indeed, denoting with u0 the strength of the current at the surface
of laminar flows and using relation (2.13) we find that
γ1 − γ2
c = u0 − γ1 d0 + . (3.1)
2k
Therefore, the frequency corresponding to the wave number k is
γ1 − γ2
ω(k) = (u0 − γ1 d0 )k + (3.2)
2
If the resonant condition ω(k1 ) + ω(k2 ) = ω(k3 ) (for wave numbers k1 , k2 and k3 with k1 + k2 = k3 ) were to
hold, then we would obtain from (3.2) that γ1 − γ2 = 0. We recall now that the object of our investigations
concerns the scenario when γ1 ̸= γ2 . We would also like to note that neither gravity nor surface tension
enters (3.2), situation suggesting that such waves are triggered by the vorticity jump between the layers.


3.2. Resonances pertaining to λ1

We have from (2.33) that λ1 satisfies the equation


γ1 √ g̃(k)
λ+ λ− = 0, (3.3)
k k
whose unique positive solution is

( √ )
1 2 σ 3
λ= −γ1 + γ1 + 4 kg + 4 k . (3.4)
2k ρ
Taking into account relation (2.13) we find that for the wave speed c it holds
( √ )
1 2 σ 3
c = u0 + −γ1 + γ1 + 4 kg + 4 k , (3.5)
2k ρ
where u0 denotes the strength of the current at the surface of laminar flows.
We remark that in (3.5) only the vorticity of the top layer appears in the formula. This was also
a realization of the paper [64] by Ivanov, where it was shown that (in certain situations) wave–current
interactions are influenced only by the current profile in the strip adjacent to the surface.
The frequency corresponding to the wave number k is
( √ )
1 σ
ω(k) = ku0 + −γ1 + γ12 + 4 kg + 4 k 3 . (3.6)
2 ρ
To shed light on the rather convoluted nature of (3.6) we present a few figures delivering the behavior of ω as
a function of k for given u0 and γ1 and then as a function of k and γ1 for a fixed value of u0 (see Figs. 2–6).
Our attention will be geared towards showing that the frequencies corresponding to two wave number k1
and k2 add up properly. More precisely, we will find out for which values of γ1 is possible to solve for k1 , k2
and k3 in the system
ω(k1 ) + ω(k2 ) = ω(k3 ),
(3.7)
k1 + k2 = k3 .
C.I. Martin and B. Basu / Nonlinear Analysis: Real World Applications 57 (2021) 103176 11

Fig. 2. The frequency ω(k) for several values of the surface current and for the fixed value γ1 = −1.25 · 10−2 s−1 of the vorticity in
the top layer. A “straightening” of the graph of ω is visible for large values of the wave number k. This fact alludes to the possibility
of existence of three-wave resonances, result that we confirm further in the paper.

Fig. 3. The frequency ω(k) for several values of the surface current and for the fixed value γ1 = −1 s−1 of the vorticity in the top
layer. We note again the tendency to smooth out of the graph of ω for large values of the wave number.

Fig. 4. The frequency ω(k) as a function of the wave number k and of the vorticity γ1 for a significant range of γ1 .
12 C.I. Martin and B. Basu / Nonlinear Analysis: Real World Applications 57 (2021) 103176

Fig. 5. The frequency ω(k) as a function of the wave number k and of the vorticity γ1 assuming the fixed value of the surface current
u0 = 2 m s−1 .

Fig. 6. A closer look at the behavior of ω(k) whose graph is plotted for a smaller range of the vorticity γ1 . A tilt of the graph is
noticeable as γ1 varies.

With the substitutions Σ := γ21 σρ and g := gρ



σ the first equation in the previous system transforms to
√ √ √
Σ 2 + k3 g + k33 + Σ = Σ 2 + k1 g + k13 + Σ 2 + k2 g + k23 . (3.8)

By squaring in (3.8) (after observing that its left hand side is positive) we obtain

k 3 + k23 − k33 + 2 (Σ 2 + k1 g + k13 )(Σ 2 + k2 g + k23 )
Σ= 1 √ (3.9)
2 Σ 2 + k3 g + k33
A further squaring of (3.8) delivers
F2
Σ2 = , (3.10)
8(k13 + k23 − k33 ) [F + 2k3 (g + k32 )(k13 + k23 − k33 )]
where
F = (k13 + k23 − k33 )2 − 4k13 k23 − 4k1 k2 g 2 + g(k12 + k22 ) .
[ ]
(3.11)

3.3. The sign of the vorticity triggering resonances

While formulas (3.9), (3.10), (3.11) clearly supply the (unique) vorticity that is needed for resonance (3.7)
to occur, we aim now at gaining a better insight on the sign of the vorticity that triggers resonances of the
C.I. Martin and B. Basu / Nonlinear Analysis: Real World Applications 57 (2021) 103176 13

mentioned kind. To this end we introduce the function Σ → f (Σ ) defined by


√ √ √
f (Σ ) = Σ 2 + k1 g + k13 + Σ 2 + k2 g + k23 − Σ 2 + k3 g + k33 − Σ . (3.12)
Clearly,
lim f (Σ ) = ∞ (3.13)
Σ →−∞
and ( √ √ )
′ 1 Σ 2 + k3 g + k33 − Σ 2 + k2 g + k23
f (Σ ) = Σ √ + √ √ − 1 < 0, (3.14)
Σ 2 + k1 g + k13 Σ 2 + k2 g + k23 Σ 2 + k3 g + k33
for all Σ < 0, since k3 > k2 . Thus, the function f is strictly decreasing on (−∞, 0]. Therefore, the necessary
and sufficient condition for f (Γ ) = 0 to have a solution in (−∞, 0] is that f (0) < 0. The latter condition is
equivalent to √ √ √
k1 g + k13 + k2 g + k23 < k3 g + k33 . (3.15)
Squaring in the above inequality we obtain (taking also into account the requirement k1 + k2 = k3 ) that it
is equivalent to
9k1 k2 (k12 + k22 ) + 14k12 k22 > 4g(k12 + k22 ) + 4g 2 . (3.16)
We will show now that for waves with wavelengths of up to 2 cm the previous inequality holds. This
assumption is legitimate for capillary–gravity water waves. To see this, note first that for L ≤ 0.02 m we
−1
have that the wave number k = 2π 2
L ≥ 10 π m . Therefore,
9k1 k2 ≥ 9π 2 · 104 m−2 .
Working with the values ρ = 1000 kg m−3 for the constant density, σ = 72 · 10−3 N for the surface tension,
we obtain that g = gρ 4
σ = 13.61 · 10 m
−2
. Obviously, 9π 2 > 4 · 13.61, which implies that
9k1 k2 > 4g. (3.17)
The choice that me made for the wavelength L also implies that
14k12 k22 ≥ 14π 4 · 108 m−4 > 4 · (13.61)2 · 108 m−4 = 4g 2 . (3.18)
From (3.17) and (3.18) we conclude now that (3.16) holds true for all k1 , k2 ≥ 102 ·π m−1 . Thus, for capillary–
gravity waves with wavelength L ≤ 2 cm we have that f (0) < 0 which, corroborated with (3.13) and the
discussion thereafter, implies that the equation f (Σ ) = 0 has a unique negative solution in (−∞, 0].
From the previous discussion it emerges that two capillary–gravity water waves with given wavelengths
L1 , L2 , not exceeding 2 cm,
√exhibit the resonance conditions (3.7) for exactly one (negative) value of the
constant vorticity γ1 = 2Σ σρ with Σ being given by (3.9)–(3.10).

3.4. Absence of three-wave resonances for gravity waves

We consider gravity water waves for which kd0 is sufficiently large. We note that even in the scenario
when the wavelength L is of similar size as the thickness of the near surface vortical layer, we have that
e−2kd0 is about 3 · 10−6 m. Therefore, it is appropriate to work again with the decomposition (2.33) of the
dispersion relation (2.30).
Trying to find three-wave resonances for gravity waves we set σ = 0 N in formula (3.6). The system (3.7)
becomes then equivalent to
√ √ √
γ12 + 4k1 g + γ12 + 4k2 g = γ12 + 4k3 g,
(3.19)
k1 + k2 = k3 .
Squaring in the first equation above implies that 16k1 k2 g 2 vanishes, which is clearly impossible. Thus, no
three-wave resonances exist in the case of pure gravity waves.
14 C.I. Martin and B. Basu / Nonlinear Analysis: Real World Applications 57 (2021) 103176

4. Four-wave resonances

4.1. Resonances associated with λ0

The solution √ γ1 − γ2
λ = −γ1 d0 +
2k
gives rise to resonant four-wave interactions irrespective of the values of the wave numbers k1 , k2 , k3 , k4 .
Indeed, the condition ω(k1 ) + ω(k2 ) = ω(k3 ) + ω(k4 ) for k1 + k2 = k3 + k4 is equivalent to
−γ1 d0 (k1 + k2 ) + γ1 − γ2 = −γ1 d0 (k3 + k4 ) + γ1 − γ2 ,
which is trivially satisfied.

4.2. The derivation of the resonance condition for resonances associated with λ1

Again, we seek the values of the constant vorticity γ1 for which, given wave numbers k1 , k2 , k3 , k4 with
k1 + k2 = k3 + k4 , we have the resonance condition satisfied
ω(k1 ) + ω(k2 ) = ω(k3 ) + ω(k4 ), (4.1)
where, as before, ( √ )
1 σ
ω(k) = ku0 + −γ1 + γ12 + 4 kg + 4 k 3 . (4.2)
2 ρ
The substitution Σ := γ21 σρ transforms (4.1) in

√ √ √ √
Σ 2 + k1 g + k13 + Σ 2 + k2 g + k23 = Σ 2 + k3 g + k33 + Σ 2 + k4 g + k43 , (4.3)
with g = 13.61 · 104 m−2 . We will split the analysis of (4.3) in two cases.

4.2.1. The case k1 k2 > k3 k4


The previous assumption implies, using also k1 + k2 = k3 + k4 , that
k13 + k23 < k33 + k43 . (4.4)
Assuming that there are values of Σ that enable four-wave resonances to occur, we square in (4.3) and obtain
the equation

k33 + k43 − k13 − k23 + 2 (Σ 2 + k3 g + k33 ) (Σ 2 + k4 g + k43 )

= 2 (Σ 2 + k1 g + k13 ) (Σ 2 + k2 g + k23 ). (4.5)
Noticing that, due to (4.4), the left hand side of (4.5) is positive, we find (by squaring) that (4.5) is equivalent
to

(Σ 2 + k3 g + k33 ) (Σ 2 + k4 g + k43 ) + Σ 2
k1 k2 (g + k12 )(g + k22 ) − k3 k4 (g + k32 )(g + k42 ) k33 + k43 − k13 − k23
= − (4.6)
k33 + k43 − k13 − k23 4
Before continuing with the search for Σ , we notice that (4.6) places a necessary condition for four-wave
resonance to eventuate. Namely, the positivity of the left hand side above requires that, in addition to
k1 + k2 = k3 + k4 , the wave numbers k1 , k2 , k3 and k4 must satisfy also the inequality
k1 k2 (g + k12 )(g + k22 ) − k3 k4 (g + k32 )(g + k42 ) k33 + k43 − k13 − k23

k33 + k43 − k13 − k23 4

≥ k3 k4 (g + k32 )(g + k42 ), (4.7)
C.I. Martin and B. Basu / Nonlinear Analysis: Real World Applications 57 (2021) 103176 15

which is equivalent to
k1 k2 (g + k12 )(g + k22 ) − k3 k4 (g + k32 )(g + k42 ) 3(k1 + k2 )(k1 k2 − k3 k4 )

3(k1 + k2 )(k1 k2 − k3 k4 ) 4

2
≥ k3 k4 (g + k3 )(g + k4 ), 2 (4.8)

To ease the notation we set

A := k1 k2 (g + k12 )(g + k22 ) − k3 k4 (g + k32 )(g + k42 )

and
E := 3(k1 + k2 )(k1 k2 − k3 k4 ).
The necessary condition (4.8) for the existence of four-wave resonances for capillary–gravity waves can be
written as
A E

− ≥ k3 k4 (g + k32 )(g + k42 ). (4.9)
E 4
We pay attention in what follows to capillary–gravity waves with wavelengths L with 1.7 cm ≤ L ≤ 2.7 cm.
The wave number k = 2πL of such waves satisfies

k
0.74 m−1 ≤ ≤ 1.17 m−1 .
102 π
Setting k := k
102 π
and using that g = 1.37 · π 2 · 104 m−2 we see that (4.8) is equivalent to

k1 k2 (1.37 + k12 )(1.37 + k22 ) − k3 k4 (1.37 + k32 )(1.37 + k42 ) 3(k1 + k2 )(k1 k2 − k3 k4 )

3(k1 + k2 )(k1 k2 − k3 k4 ) 4

≥ k3 k4 (1.37 + k32 )(1.37 + k42 ). (4.10)

Using that k32 + k42 = k12 + k22 + 2(k1 k2 − k3 k4 ), we obtain (after some calculations) that

k1 k2 (1.37 + k12 )(1.37 + k22 ) − k3 k4 (1.37 + k32 )(1.37 + k42 )


= (k1 k2 − k3 k4 ) 1.372 + 1.37 · (k12 + k22 − 2k3 k4 ) + k12 k22 + k1 k2 k3 k4 + k32 k42
[ ]

≤ (k1 k2 − k3 k4 ) 1.372 + 1.37 · (k3 − k4 )2 + 3k12 k22 ,


[ ]
(4.11)

where the last inequality is obtained from utilizing k3 k4 < k1 k2 and k12 + k22 < k32 + k42 . Consequently, for
the first term on the left hand side of (4.10) holds

k1 k2 (1.37 + k12 )(1.37 + k22 ) − k3 k4 (1.37 + k32 )(1.37 + k42 )


3(k1 + k2 )(k1 k2 − k3 k4 )
1.372 + 1.37 · (k3 − k4 )2 + 3k12 k22
≤ (4.12)
3(k1 + k2 )

The verification of (4.10) will be split according to the position of k1 , k2 in the interval [0.74 m−1 ,
1.17 m−1 ].
Case 1: Both k1 and k2 belong to the interval [0.74 m−1 , 1 m−1 ].
We use that k1 + k2 ≥ 2 · 0.74 m−1 and, since |k3 − k4 | ≤ 0.43 m−1 , we obtain that

1.372 + 1.37 · (k3 − k4 )2 + 3k12 k22


≤ 1.15 m−1 ,
3(k1 + k2 )
while √
k3 k4 (1.37 + k32 )(1.37 + k42 ) ≥ 1.41 m−1 .
16 C.I. Martin and B. Basu / Nonlinear Analysis: Real World Applications 57 (2021) 103176

From (4.12) we see that (4.10) fails for k1 , k2 ∈ [0.74 m−1 , 1 m−1 ] and k3 , k4 ∈ [0.74 m−1 , 1.17 m−1 ].
Case 2: At least one of k1 , k2 , say k1 , belongs to (1 m−1 , 1.17 m−1 ].
Since k3 + k4 = k1 + k2 , it follows that max{k3 , k4 } > 12 (1 + 0.74) m−1 = 0.87 m−1 which implies that

k3 k4 (1.37 + k32 )(1.37 + k42 ) ≥ 1.62 m−1

and
1.372 + 1.37 · (k3 − k4 )2 + 3k12 k22
≤ 1.14 m−1 ,
3(k1 + k2 )
where, to obtain the last estimate, we used that k1 + k2 > 1.74 m−1 . Therefore, the latter two estimates
together with (4.12) show that (4.10) is violated also in this case. Since the case k2 ∈ (1 m−1 , 1.17 m−1 ] is
similar to Case 2, we have proved that the necessary condition (4.8) for occurrence of four wave resonances
(4.1) is hindered.

4.2.2. The case k1 k2 = k3 k4


Assuming that the wave numbers k1 , k2 , k3 and k4 with k1 + k2 = k3 + k4 also satisfy k1 k2 = k3 k4 . The
two equalities say that k1 , k2 and k3 , k4 , respectively, are solutions of the same quadratic equation. This
implies that k1 = k3 , k2 = k4 or k1 = k4 , k2 = k3 , thus the only resonant tuples are (k1 , k2 , k1 , k2 ) and
(k1 , k2 , k2 , k1 ). In these cases any constant vorticity γ1 triggers the resonance.

4.3. Gravity waves

We set σ = 0 in (3.6) and see that the resonance condition (4.1) becomes equivalent to
√ √ √ √
γ12 + 4k1 g + γ12 + 4k2 g = γ12 + 4k3 g + γ12 + 4k4 g,
(4.13)
k1 + k2 = k3 + k4 .

By squaring the first equation in the above system it becomes equivalent to the system

k1 k2 = k3 k4 ,
(4.14)
k1 + k2 = k3 + k4 .

Arguing as in Section 3.4, we infer that the only tuples that satisfy (4.14) are (k1 , k2 , k1 , k2 ) and (k1 , k2 , k2 ,
k1 ). Moreover, they arise as resonant tuples for all (constant) values of the vorticity γ1 .

5. Conclusion

We have presented in this paper an analytical investigation of three- and four-wave wave resonances in
the case of rotational capillary–gravity water waves. Instrumental in deriving our results was the dispersion
relation for capillary–gravity water waves in rotational flows with piecewise constant vorticity. This type
of vorticity describes a scenario of existence of non-uniform currents with a rapid change in the current
strength, cf. [65]. The distribution of vorticity in our setting is as follows: we consider a layer of constant
non-vanishing vorticity γ1 adjacent to the free surface above a rotational flow of vorticity γ2 . On physical
grounds, this situation is justified by the fact that rotational wind generated waves possess a layer of high
vorticity adjacent to the wave surface [66,67], while in the near bed region there may exist currents resulting
from sediment transport along the ocean bed [68].
While the dispersion relation is, in general, encapsulated as the positive root of an intricate third-degree
algebraic equation (involving physical quantities of the flow like the mean depth, the wave number, the
surface tension coefficient, the two vorticities γ1 , γ2 ), we are able to find it explicitly for scenarios for which
C.I. Martin and B. Basu / Nonlinear Analysis: Real World Applications 57 (2021) 103176 17

d0
e−2kd0 = e−4π L is sufficiently small. The latter occurs for capillary–gravity water waves, since in this case
the wavelength is of the order of centimeters. To be more precise, we found two positive roots of the third-
√ √
degree equation, denoted λ1 and λ0 . The latter two values deliver the relative surface wave speed. We also
note that only the vorticity γ1 , pertaining to the top layer, enters that analysis concerning the resonances

associated to λ1 .
Based on the explicit dispersion relation (3.5) we passed to the investigation of existence of three-wave
resonances. We found that two capillary–gravity water waves with given wavelengths L1 , L2 , not exceeding
2 cm, exhibit
√ the resonance conditions (3.7) for exactly one (negative) value of the constant vorticity
γ1 = 2Σ σρ with Σ being given (in terms of the wave numbers k1 , k2 ) by (3.9)–(3.10). These three-wave
√ √
resonances are brought about by frequencies associated to λ1 . Notably, frequencies corresponding to λ0
do not give rise to three-wave resonances. We also establish that pure gravity water waves do not exhibit
three-wave resonances. We point out that, as demonstrated in [8], three-wave resonances may give rise to
rogue waves (also known as monster waves) which can be extremely dangerous, even to large ships.2
Rather different is the case of four-wave resonances. More precisely, for capillary–gravity waves with
wavelengths comprised between 1.7 cm and 2.7 cm we establish that the four-wave resonance conditions do

not hold for frequencies arosen from λ1 . Nevertheless, four-wave resonances are triggered by frequencies

corresponding to λ0 .
Concerning four-wave resonances, we would like to note the study of Stuhlmeier and Stiassnie [69] where
resonances of the type k1 + k2 = 2k3 were considered.
Lastly, we would also like to point out that an important problem for future research is the resonant
interaction between wave packets leading to a system of ordinary differential equations and partial differential
equations like in [69,70].

Acknowledgments

The authors are indebted to Prof. Bruno Eckhardt whose suggestions significantly improved the results
of this paper. Moreover, the authors also acknowledge the suggestions of two referees whose remarks have
enhanced the presentation. C. I. Martin would like to acknowledge the support of the Austrian Science Fund
(FWF) under research grant P 30878-N32. The comments and suggestions of two anonymous referees are
gratefully acknowledged.

References

[1] O.M. Phillips, On the dynamics of unsteady gravity waves of infinite amplitude, J. Fluid Mech. 9 (1960) 193–217.
[2] M.S. Longuet-Higgins, Resonant interactions between two trains of gravity waves, J. Fluid Mech. 12 (3) (1962) 321–332.
[3] M.S. Longuet-Higgins, O.M. Phillips, Phase velocity effects in tertiary wave interactions, J. Fluid Mech. 12 (3) (1962)
333–336.
[4] O.M. Phillips, Theoretical and experimental studies of gravity wave interactions, Proc. R. Soc. Lond. Ser. A Math.
Phys. Eng. Sci. 299 (1967) 104–119.
[5] J.L. Hammack, D.M. Henderson, Resonant interactions among surface water waves, Ann. Rev. Fluid Mech. 25 (1993)
55–97.
[6] A.D.D. Craik, Wave Interactions and Fluid Flows, Cambridge University Press, Cambridge, 1986.
[7] M.D. Bustamante, B. Quinn, D. Lucas, Robust energy transfer mechanism via precession resonance in nonlinear turbulent
wave systems, Phys. Rev. Lett. 113 (8) (2014) 084502.
[8] F. Baronio, M. Conforti, A. Degasperis, S. Lombardo, Rogue waves emerging from the resonant interaction of three
waves, Phys. Rev. Lett. 111 (11) (2013) 114101.
[9] W. Cousins, M. Onorato, A. Chabchoub, T. Sapsis, Predicting ocean rogue waves from point measurements: An
experimental study for unidirectional waves, Phys. Rev. E 99 (3) (2019) 032201.

2
Monsters of the deep, The Economist, September 17, 2009.
18 C.I. Martin and B. Basu / Nonlinear Analysis: Real World Applications 57 (2021) 103176

[10] R. El Koussaifi, A. Tikan, A. Toffoli, S. Randoux, P. Suret, M. Onorato, Spontaneous emergence of rogue waves in
partially coherent waves: A quantitative experimental comparison between hydrodynamics and optics, Phys. Rev. E 97
(1) (2018) 012208.
[11] H.N. Chan, R.H.J. Grimshaw, K.W. Chow, Modeling internal rogue waves in a long wave-short wave resonance
framework, Phys. Rev. Fluids 3 (12) (2018) 124801.
[12] A.N. Pushkarev, V.E. Zakharov, Turbulence of capillary waves, Phys. Rev. Lett. 76 (1996) 3320–3323.
[13] A.N. Pushkarev, V.E. Zakharov, Turbulence of capillary waves-theory and numerical simulations, Physica D 135 (1–2)
(2000) 98–116.
[14] B. Eckhardt, Transition to turbulence in a shear flow, Phys. Rev. E. 60 (1) (1999) 509–517.
[15] B. Eckhardt, A critical point for turbulence, Science 333 (6039) (2011) 165–166.
[16] B. Eckhardt, Transition to turbulence in shear flows, Physica A 504 (2018) 121–129.
[17] J. Moehlis, H. Faisst, B. Eckhardt, A low-dimensional model for turbulent shear flows, New J. Phys. 6 (2004) 56.
[18] J. Schumacher, B. Eckhardt, Evolution of turbulent spots in a parallel shear flow, Phys. Rev. E 63 (2001) 046307.
[19] G. Haller, Lagrangian Structures and the rate of strain in a partition of two-dimensional turbulence, Phys. Fluids 13
(11) (2001) 3365–3385.
[20] M. Mathur, G. Haller, T. Peacock, J.E. Ruppert-Felsot, H.L. Swinney, Uncovering the Lagrangian skeleton of turbulence,
Phys. Rev. Lett. 98 (14) (2007) 144502.
[21] S. Colwell, B. Basu, Tuned liquid column dampers in offshore wind turbines for structural control, Eng. Struct. 31 (2)
(2009) 358–368.
[22] V.N. Dinh, B. Basu, Passive control of floating offshore wind turbine nacelle and spar vibrations by multiple tuned mass
dampers, Struct. Control Health Monit. 22 (1) (2015) 152–176.
[23] S.R.K. Nielsen, Q. Zhou, M.M. Kramer, B. Basu, Z. Zhang, Optimal control of nonlinear wave energy point converters,
Ocean Eng. 72 (2013) 176–187.
[24] M. Shinozuka, G. Deodatis, Simulation of stochastic processes by spectral representation, Appl. Mech. Rev. 44 (4) (1991)
191–204.
[25] J.N. Yang, Simulation of random envelope processes, J. Sound Vib. 21 (1) (1972) 73–85.
[26] C. Swan, I.P. Cummins, R.L. James, An experimental study of two-dimensional surface water waves propagating on
depth-varying currents, J. Fluid Mech. 428 (2001) 273–304.
[27] G.P. Thomas, Wave–current interactions: an experimental and numerical study, J. Fluid Mech. 216 (1990) 505–536.
[28] A. Constantin, Nonlinear Water Waves with Applications to Wave–Current Interactions and Tsunamis, in: CBMS-NFS
Regional Conference Series in Applied Mathematics, vol. 81, Society for Industrial and Applied Mathematics (SIAM),
Philadelphia, PA, 2011.
[29] G. Haller, An objective definition of a vortex, J. Fluid Mech. 525 (2005) 1–26.
[30] G. Haller, A. Hadjighasem, M. Farazm, F. Huhn, Defining coherent vortices objectively from the vorticity, J. Fluid
Mech. 795 (2016) 136–173.
[31] A. Hadjighasem, D. Karrasch, H. Teramoto, G. Haller, Spectral-clustering approach to Lagrangian vortex detection,
Phys. Rev. E 93 (6) (2016) 063107.
[32] A. Constantin, W.A. Strauss, Exact steady periodic water waves with vorticity, Comm. Pure Appl. Math. 57 (4) (2004)
481–527.
[33] A. Constantin, W. Strauss, Periodic traveling gravity water waves with discontinuous vorticity, Arch. Ration. Mech.
Anal. 202 (1) (2011) 133–175.
[34] A. Constantin, J. Escher, Symmetry of steady deep-water waves with vorticity, European J. Appl. Math. 15 (6) (2004)
755–768.
[35] A. Constantin, J. Escher, Symmetry of steady periodic surface water waves with vorticity, J. Fluid Mech. 498 (2004)
171–181.
[36] A. Constantin, M. Ehrnstrom, E. Wahlen, Symmetry of steady periodic gravity water waves with vorticity, Duke Math.
J. 140 (3) (2007) 591–603.
[37] J. Escher, A.-V. Matioc, B.-V. Matioc, On stratified steady periodic water waves with linear density distribution and
stagnation points, J. Differential Equations 251 (10) (2011) 2932–2949.
[38] J. Escher, Regularity of rotational travelling water waves, Philos. Trans. R. Soc. Lond. Ser. A Math. Phys. Eng. Sci.
370 (1964) (2012) 1602–1615.
[39] J. Escher, B.-V. Matioc, On the analyticity of periodic gravity water waves with integrable vorticity function, Differential
Integral Equations 27 (3–4) (2014) 217–232.
[40] A. Constantin, W.A. Strauss, Stability properties of steady water waves with vorticity, Comm. Pure Appl. Math. 60 (6)
(2007) 911–950.
[41] A. Constantin, D. Sattinger, W. Strauss, Variational formulations for steady water waves with vorticity, J. Fluid Mech.
548 (2006) 151–163.
[42] A. Constantin, R.I. Ivanov, C.I. Martin, Hamiltonian formulation for wave–current interactions in stratified rotational
flows, Arch. Ration. Mech. Anal. 221 (3) (2016) 1417–1447.
[43] A. Constantin, E. Kartashova, Effect of non-zero constant vorticity on the nonlinear resonances of capillary water waves,
Europhys. Lett. 86 (2009) 29001.
[44] B. Basu, C.I. Martin, Resonant interactions of rotational water waves in the equatorial f-plane approximation, J. Math.
Phys. 59 (10) (2018) 103101, 9 pp..
[45] C.I. Martin, Resonant interactions of capillary–gravity water waves, J. Math. Fluid Mech. 19 (4) (2017) 807–817.
[46] F. Haudin, A. Cazaubiel, L. Deike, T. Jamin, E. Falcon, M. Berhanu, Experimental study of three-wave interactions
among capillary–gravity surface waves, Phys. Rev. E 93 (2016) 043110.
C.I. Martin and B. Basu / Nonlinear Analysis: Real World Applications 57 (2021) 103176 19

[47] V.E. Zakharov, Stability of periodic wave of finite amplitude on the surface of a deep fluid, J. Appl. Mech. Tech. Phys.
9 (1968) 190–194.
[48] M. Stiassnie, L. Shemer, On the interaction of four-waves, Wave Motion 41 (2005) 307–328.
[49] M. Berhanu, E. Falcon, Space–time-resolved capillary wave turbulence, Phys. Rev. E 87 (2013) 033003.
[50] F. Bonnefoy, F. Haudin, G. Michel, B. Semin, T. Humbert, S. Aumaı̂tre, M. Berhanu, E. Falcon, Observation of resonant
interactions among surface gravity waves, J. Fluid Mech. 805 (2016).
[51] E. Kartashova, S. Nazarenko, O. Rudenko, Resonant interactions of nonlinear water waves in a finite basin, Phys. Rev.
E 78 (1) (2008) 016304.
[52] M.D. Bustamante, E. Kartashova, Dynamics of nonlinear resonances in hamiltonian systems, Europhys. Lett. 85 (1)
(2009) 14004.
[53] M.D. Bustamante, E. Kartashova, Effect of the dynamical phases on the nonlinear amplitudes’ evolution, Europhys.
Lett. 85 (3) (2009) 34002.
[54] M.D. Bustamante, E. Kartashova, Resonance clustering in wave turbulent regimes: Integrable dynamics, Commun.
Comput. Phys. 10 (5) (2011) 1211–1240.
[55] E. Kartashova, Nonlinear resonances of water waves, Discrete Contin. Dyn. Syst. Ser. B 12 (3) (2009) 607–621.
[56] E. Kartashova, Nonlinear resonance analysis, in: Theory, Computation, Applications, Cambridge University Press,
Cambridge, 2011.
[57] J. Harris, M.D. Bustamante, C. Connaughton, Externally forced triads of resonantly interacting waves: Boundedness
and integrability properties, Commun. Nonlinear Sci. Numer. Simul. 17 (12) (2012) 4988–5006.
[58] A.V. Fedorov, J.N. Brown, Equatorial waves, in: J. Steele (Ed.), Encyclopedia of Ocean Sciences, Academic Press, 2009,
pp. 3679–3695.
[59] C.I. Martin, B.-V. Matioc, Existence of capillary–gravity water waves with piecewise constant vorticity, J. Differential
Equations 256 (8) (2014) 3086–3114.
[60] D.H. Peregrine, Interaction of water waves and currents, Adv. Appl. Mech. 16 (1976) 9–117.
[61] V. Shrira, Surface waves on shear currents: Solution of the boundary-value problem, J. Fluid Mech. 252 (1993) 565–584.
[62] H.-C. Hsu, M. Francius, P. Montalvo, C. Kharif, Gravity-capillary waves in finite depth on flows of constant vorticity,
Proc. R. Soc. A: Math. Phys. Eng. Sci. 472 (2195) (2016) 20160363.
[63] S. Monismith, Personal communication.
[64] R.I. Ivanov, Hamiltonian model for coupled surface and internal waves in the presence of currents, Nonlinear Anal.:
Real World Appl. 34 (2017) 316–334.
[65] I.G. Jonsson, Wave–current interactions, in: The Sea, Wiley, New-York, 1990.
[66] K. Okuda, Internal flow structure of short wind waves, J. Oceanogr. Soc. Jpn. 38 (1982) 28–42.
[67] O.M. Phillips, M.L. Banner, Wave breaking in the presence of wind drift and swell, J. Fluid Mech. 66 (1974) 625–640.
[68] C. Pattiaratchi, M. Collins, Sediment transport under waves and tidal currents: A case study from the northern bristol
channel, U. K. Mar. Geol. 56 (1984) 27–40.
[69] R. Stuhlmeier, M. Stiassnie, Evolution of statistically inhomogeneous degenerate water wave quartets, Phil. Trans. R.
Soc. A 376 (2017) 20170101.
[70] R. Ivanov, C.I. Martin, On the time-evolution of resonant triads in rotational capillary-gravity water waves, Phys. Fluids
31 (2019) 117103.

You might also like