Linear Wave Theory

The most elementary wave theory is the small-

amplitude or linear wave theory. This theory,
developed by Airy (1845), is easy to apply, and gives
a reasonable approximation of wave characteristics
for a wide range of wave parameters. A more
complete theoretical description of waves may be
obtained as the sum of many successive
approximations, where each additional term in the
series is a correction to preceding terms. For some
situations, waves are better described by these higher-
order theories, which are usually referred to as finite-
amplitude wave theories.
Although there are limitations to its applicability,
linear theory can still be useful provided the
assumptions made in developing this simple theory
are not grossly violated.
The assumptions made in developing the linear wave
theory are:
 The fluid is homogeneous and incompressible.
 Surface tension is negligible.
 The fluid is ideal or inviscid.
 Coriolis effect due to the earth's rotation is
negligible.
  Pressure at the free surface is uniform and
constant.
 The particular wave being considered does not
interact with any other water motions. The flow is
irrotational so that water particles do not rotate
(only normal forces are important and shearing
forces are negligible).
 The bed is a horizontal, fixed, impermeable
boundary, which implies that the vertical velocity
at the bed is zero.
 The wave amplitude is small and the waveform is
invariant in time and space.
 Waves are plane or long-crested (two-
dimensional).
The theory (also known as Airy wave theory) is based
on only two equations: a mass balance equation and a
momentum balance equation.
u  v  w  0 (1)
x y z
Note: We take x as the horizontal component in the
direction of motion and z as the vertical coordinate.

Write down the differential momentum equations for

Newtonian fluid for the case of linear gravity waves
(incompressible and inviscid flow and gravity acting
only in the z-direction).

2
  (u  u u  v u  w u ) =  p
t x y z x
 (v  u v  v v  w v ) =  p (2)
t x y z y
 (w  u w  v w  w w) =  g z  p
t x y z z
In this case gz = -g since gravity acts in the –ve z-
direction.
Boundary Conditions1:
The kinematic surface boundary condition relates the
velocity of a particle on the surface to the motion of
the surface. Consider a two-dimensional situation, as
shown in Fig. 1.

If a particle is to remain on the surface, then the

velocity component of that particle in the direction
normal to the surface wcosα – usinα should be equal
to the velocity of the surface in that direction
( /  t)cosα

3
 wcos  u sin   cos at z = 
t
where  is the surface elevation above some reference
level and α is the slope of the surface. This can be re-
written as
w    u tan    u 
t t x
For a 3-dimensional situation, this kinematic
boundary condition at the surface can be written
similarly as
w    u   v  at z =  (3)
t x y
The kinematic bottom boundary condition is simply
that the water may not penetrate the fixed, horizontal
bottom, so that
w=0 at z = -d
The dynamic surface boundary condition (pressure is
zero) is
p=0 at z = 

The Velocity Potential Function:

The velocity potential function  is defined for

irrotational flows as

4
u   , v   and w  
x y z
Or, in vector form V = 

Conservation Equations(2):

We can express the continuity equation in terms of the

velocity potential as
2  2  2  0 or in vector form 2 = 0
x2 y2 z2
We can also express the equations of motion in terms
of the velocity potential function. We first invoke the
condition of irrotationality:
w  v  0  v  w
y z z y
u  w  0  u  w
z x z x
v  u  0  u  v
x y y x
The equation of motion in the x-direction can then be
written as
u  u u  v v  w w =  1 p
t x x x  x or

Or, moving u, v and w behind the differentiations,

u   (1 u2  1 v2  1 w2)  1 p
t x 2 2 2  x
This equation of motion can be expressed in terms of
the velocity potential
5
   2 2 2 
       1             1 p
 
   
t  x  x  2  x   y   z     x
  
This can be re-written as
  2 2 
    1           p   0
2
 
 
x  t 2  x   y   z    
   
The corresponding equations for the momentum in the
y- and z-directions may similarly be written as
  2 2 
    1         p   0
2
     
 
y  t 2  x   y   z    
   
  2 2 

2

   1         
       gz   0
p
  
z  t 2  x   y   z    
   
We can add the term gz to the x- and y-direction
momentum equations (it will disappear on
differentiation).

6
   2 2 
    1           p  gz   0
2
 
 
x  t 2  x   y   z    

   
  2 2 
    1         p  gz   0
2
     
 
y  t 2  x   y   z    

   
  2 2 

2

   1         
       gz   0
p
  
z  t 2  x   y   z    
   
The sum of the terms between the curly brackets
appears in all three equations, in each case expressing
the fact that this sum of terms is not a function of x, y
and z. This sum can therefore be only an arbitrary
function of time t, f(t), for which we take the simplest
possible: f(t) = 0. We therefore find from the
equations of motion
 2 2 2 

  1             p  gz  0

t 2 x   y   z   


 
 

Or, in vector notation,

2
  1   p  gz  0 (4)
t 2 
This is the Bernoulli’s equation for unsteady motion.
Removing the quadratic terms makes this equation the
linearized Bernoulli’s equation for unsteady motion.

7
   p  gz 0 (5)
t 
Boundary Conditions2:
The kinematic boundary condition at the surface in
terms of the velocity potential, from equation (3)
         at z = 
z t x x y y
Removing the non-linear terms (the last two terms on
the right-hand side) makes this boundary condition
linear
   at z = 0
z t
The dynamic boundary condition at the surface,
expressed in terms of the velocity potential, is
obtained by taking the Bernoulli’s equation (eqn 4) at
the surface z = , with p = 0.
2
  1   g  0
t 2
Removing the quadratic term makes this equation
linear:
  g  0 at z = 0 (6)
t
The kinematic boundary condition at the bottom, in
terms of the velocity potential function is
  0 at z = -d
z

8
 The linearized basic equations and boundary
conditions for the linear wave theory in terms of the
velocity potential
The Laplace equation and the kinematic boundary
conditions are used to obtain the solution for the
velocity potential and hence all kinematic aspects of
the waves. The linearized Bernoulli’s equation and the
linearlized dynamic boundary condition will
subsequently be used, in combination for the velocity
potential, to obtain the expressions for some dynamic
aspects of the waves.

Solution to Linearized Water Wave Boundary

Value Problem for a Horizontal Bottom:

A convenient method for solving some linear partial

differential equations is the method of separation of
variables. The solution can be expressed as:

9
   x, z,t   X (x).Z (z). (t)
where X(x) is some function that depends only on x,
Z(z) depends only on z and (t) varies only with time.
Since we know that  must be periodic in time, we
can specify (t) = sint, where  (symbol  is also
used) is the angular frequency of the wave (=2/T).
  x, z,t   X (x).Z (z).sin t
Substituting into the Laplace equation, we get
d 2 X ( x) .Z ( z).sin t  X ( x). d 2 Z ( z) .sin t  0
dx 2 dz 2

Dividing throughout by  gives

1 d 2 X  1 d 2Z  0
X dx 2 Z dz 2
The first term of this equation depends on x alone,
while the second term depends only on z. Thus, the
only way this equation will hold is if each term is
equal to the same constant with opposite signs, i.e.
1 d 2 X  k 2
X dx 2
1 d 2Z  k 2
Z dz 2
For real values of k2, the possible solution of the
above equation is
  x, z,t   ( Acos kx  Bsin kx).(Cekz  Dekz ).sin t

10
The periodic lateral boundary conditions apply in both
time and space
(x,t)= (x+L, t)
(x,t) = (x, t+T)
where L is the wavelength and T is the wave period.
The periodicity condition is satisfied for coskL = 1
and sinkL = 0, which gives kL = 2
or k(wave number) = 2 /L
The solution of the boundary value problem for small-
amplitude waves is given by
  H g cosh k (d  z) cos kx sin t
2  cosh kd
1 
From eqn (6),  ( x,t )     H cos kx.cos t
g t z 0 2
where 2 = gk tanh(kd), known as the dispersion
relation.

The wave form is shown in the figure below. At t =

/2, the wave form is zero for all x, at t = 0, it has a
cosine shape and at other times, the same cosine shape
with different magnitudes. This wave-form is a
standing wave, as it does not propagate in any
direction. Standing waves often occur when incoming
waves are completely reflected by vertical walls.

11
 Standing water wave

Consider another standing wave,

  H g cosh k (d  z) sin kx cos t
2  cosh kd
which is also a solution of the Laplace equation and
all the boundary conditions. It differs from the
previous solution in that the x and t terms are 90o out
of phase. The associated water surface position is
 ( x,t)   1   H sin kx.sin t
g t z 0 2
Subtracting the present velocity potential from the
previous one
  H g cosh k (d  z) (cos kx sin t  sin kx cos t)
2  cosh kd
   H g cosh k (d  z) sin(kx  t) (7)
2  cosh kd

12
This new velocity potential has a water surface
profile, given as
 ( x,t)   1   H cos(kx  t ) (8)
g t z 0 2
The same result is obtained by subtracting the two
(x,t).

Examining equation (8) for the water surface profile,

it is clear that this wave form moves with time. It is
the profile of a progressive wave. Combining equation
(7) with the dispersion relation,
   H C cosh k (d  z) sin(kx  t) (9)
2 sinh kd

where C is the wave celerity and is equal to L/T. The

wave form travels with this speed C.
The horizontal and vertical components of velocity
under the wave are obtained by
u    gHk cosh k (d  z) cos( t  kx)
x 2 cosh kd
w    H sinh k (d  z) sin( t  kx)
z 2 sinh kd

u  uˆ cos( t  kx)
w  wˆ sin( t  kx)
where û and ŵ are the amplitudes of the velocity
components.

13
The u and w components are clearly 90o out of phase;
the extreme values of the u-component occur at the
phase position t – kx = 0, , …(under the crest and
the trough), while the extreme values of the w-
component occur at the phase position /2, 3/2, …
(location of zero water displacement), as shown in the
figure below.

Water particle velocities in a progressive wave.

The above velocities are called ‘orbital velocities’

because they correspond to the motion of the particles
in close, circular or elliptical orbits. The velocities in
the crest of the wave are oriented in the down-wave
direction and those in the trough of the wave are
oriented in the up-wave direction.

14
In deep water when kd  , uˆ  aekz and wˆ  aekz
and the magnitude of the (resultant) total velocity
V  aekz

Thus, the wave-induced velocities decrease

exponentially with the distance to the surface (z < 0).
At the surface, where z = 0, the orbital velocity is
V = a which means that the radius of the orbit at the
surface is equal to the amplitude (a) of the wave.

In very shallow water, when kd  0,

uˆ   a and wˆ  a 1 z 
kd  d

Orbital motion in deep water (d/L > 1/2),

intermediate-depth water (1/25 < d/L < 1/2) and very
shallow water (d/L < 1/25)

The particle displacements are obtained by integrating

the velocities. Similarly, the accelerations in the two
directions are obtained by differentiating the
equations for u and w.
15
Each particle goes through an ellipse with horizontal
and vertical semi-main axes given by
A  a cosh k (d  z )
sinh kd

B  a sinh k (d  z )
sinh kd

where z is an arbitrarily chosen location.

In deep water, when kd  , A = B, so that the

particles move through circles with the radius
decreasing exponentially with the distance to the
surface r  aekz

In very shallow water when kd  0, the lengths of the

axes are A = a/kd and B = a(1+z/d). Closer to the
bottom, B  0, as z  -d; thus at the bottom the
ellipse degenerates into a straight, horizontal line.

16
Dynamics:

For a harmonic wave to be a free wave (subjected

only to gravitation, travelling at a precise speed), we
need to invoke the free-wave condition: the
atmospheric pressure at the water surface, where z =
 and p = constant = 0.

Re-writing the equation (5)

  p  gz 0 (5)
t 
This equation reduces to
  g  0 at z = 0 (6)
t

Re-writing the velocity potential function from eqn

(7):

   H g cosh k (d  z) sin(kx  t) (7)

2  cosh kd

and the water surface profile

 ( x,t)   1   H cos(kx  t ) (8)
g t z 0 2

17
Applying the surface kinematic condition   
z t
Differentiating eqn. (7) w.r.t. z and eqn. (8) w.r.t. t
and substituting z = 0 gives

2 = gk tanh kd (9)

gT 2  2 d 
L tanh   (10)
2  L 

For deep water, tanh(kd)  1 when kd  . We get

  gk0 or L0 = gT2/(2)

where k0 and L0 are the deep-water wave number and

wave length respectively.

From the above two equations,

 2 d 
L
 tanh 
L0  L 
 
Multiplying both sides with d/L,

d  d tanh  2 d 
L0 L  L 
 

18
For shallow water, tanh(kd)  kd when kd  0. We
get

  k gd or L  T gd

Phase velocity and group velocity

The propagation of the surface wave profile i.e. the

phase speed can be obtained from the dispersion
relation using C = L/T = /k

g
C  g tanh(kd )  tanh(kd ) (11)
k
This equation shows that, in general, the phase speed
depends on wave number and therefore on frequency;
long waves travel faster than short waves. Such
waves, the propagation speed of which depends on
wave length or frequency, are called dispersive
waves. In deep water tanh(kd)  1, this expression
reduces to
g
C0  or C0 = gT/(2) (12)
k0

In very shallow water, tanh(kd)  kd, so eqn (11)

reduces to
Cshallow  gd (13)

19
Eqn (13) shows that, in very shallow water, the phase
speed does not depend on wave length or frequency.
Under these conditions, the waves are said to be non-
dispersive.

The group velocity - defined as the velocity with

which the overall shape of the wave’s amplitudes
propagates through space - is given by

Cgroup  Cg    nC
k
where C is the phase velocity and n is given by
n  1 1 2kd 
2  sinh(2kd ) 
The value of kd lies between 0 and infinity. The value
on n varies between 1/2 (for deep water) and 1 (very
shallow water). This implies that the speed of an
individual wave is always greater than or equal to the
speed of the group i.e. C  Cg. One consequence of
this is that each wave travels forwards through the
group, until it reaches the front of the group, where it
disappears. The group is kept alive by the new waves
that are continuously formed at the tail of the group.
Since waves propagate across the ocean as groups,
travel times of ocean waves should be calculated with
the group velocity not with the phase speed.

In deep water, n = 1/2 and Cg = C/2

20
In very shallow water, n = 1 and Cg = C

This means that in very shallow water, each wave

maintains its position in the group; no waves
disappear at the front and no waves are generated at
the tail.

One important effect of the dependence of the group

velocity on frequency is that a field of waves with
various frequencies, as is normal for ocean waves,
disintegrates slowly into a sequence of wave fields
with the longer waves travelling ahead of the shorter
waves – the wave energy disperses across the ocean.
This phenomenon is therefore called frequency-
dispersion and hence the name dispersion relation for
the equation 2 = gk tanh(kd) and the term dispersive
waves. When storm-generated waves travel across the
ocean, it is this frequency-dispersion that transforms
the irregular storm waves into regular swell.

Wave-induced pressure:

The motions of the water particles in circles or

ellipses imply accelerations that can be caused only
by forces acting on these particles. These forces are
provided in this case by gradients in the wave-induced
pressure in the water. Substituting (differentiating) the
21
equation for  from eqn (7) in the linearized
Bernoulli’s eqn (5) yields
p  gz   ga cosh k (d  z) cos( t  kx)
cosh(kd )

The first term on the right-hand side is the hydrostatic

pressure; the second term represents the wave-induced
pressure, pwave.
pwave  pˆ wave cos( t  kx) where pˆ wave   ga cosh k (d  z)
cosh(kd )
This is a propagating pressure wave in the water body,
in phase with the surface elevation and with vertically
decreasing amplitude. This pressure distribution does
indeed provide vertical accelerations beneath the crest
and in the trough of the wave and horizontal
accelerations beneath the zero-crossings of the wave
surface

The wave-induced pressure superimposed on the

hydrodynamic pressure along the vertical, beneath a
wave crest and a wave trough.

22
Wave energy:

The presence of a wave at the water surface implies

that water particles were moved from their position at
rest to some other position. This change in position
requires work done against gravitation and this
represents potential energy. In addition, the moving
particles represent kinetic energy.

Considering a slice of water with thickness z in a

column and integrating to get the average potential
energy:
 0 
E potential    gz dz -   gz dz =   gz dz
d d 0

E potential  1  g 2 = 1  ga2
2 4

where the overbar represents time-averaging.

The time-averaged kinetic energy for a harmonic

wave with amplitude a is obtained as:


Ekinetic   1  u 2dz = 1  ga2
d 2 4

23
Hence, with the approximations of the linear wave
theory, Epotential = Ekinetic . The total time-averaged
wave-induced energy density is the sum of the
potential and kinetic energy densities.

E  1  ga2
2

Rate of Energy Transport:

The rate of energy transport by waves is

P  ECg

24