Feb 15: Tides and tidal
currents (3)
|

![]()
A widely used bottom drag parameterization is
where
is the magnitude of the near bottom
velocity.
This is in analogy to the quadratic drag
parameterization, that stems from a dimensional analysis, for a body with drag
coefficient CD in a fluid stream of velocity UB
and density r.
A typical value for the drag coefficient is CD
= 2 x 10-3 (dimensionless)
For a bottom current of 0.5 m s-1 this
gives a stress of 0.5 N m-2 (or Pa).
The occurrence of non-linear friction is going to make
simple analysis difficult.
An alterative parameterization is a linear drag
law:
where r
has dimensions of velocity.
Bowden explains (rather glibly) the rationale behind
introducing the depth average velocity and the relationship between r and CD.
The parameterization relies on the velocity profile
maintaining a constant shape. This is
generally a reasonable assumption.
Observations of the tidally reversing flow in the
To use this parameterization, r may need to be tuned to reflect the typical magnitude of the near
bottom velocity w.r.t. the depth-average.
By removing the dependence on the actual time-varying
velocity we have removed the nonlinearity of the friction from the equations.
Before we can proceed with a linear analysis of the
2-D depth averaged equations, we need to consider the non-linear momentum
advection terms.
These are of order U2/L, while the Coriolis
term is of order Uf (adopting the inertial period as the characteristic
time scale).
The ratio of non-linear to Coriolis terms is U/fL = the Rossby number.
For tidal currents on the shelf of order 0.5 m/s and
an along-shelf length scale of L = 100 km
U/fL
= 0.5/(10-4 x 10 x 103) = 5 x 10-2
so we can be reasonably justified in ignoring the
nonlinear terms (for now).
A third non-linearity arises from the occurrence of
terms involving
.
We will limit the situations we consider to cases
where the tidal height variation is modest compared with the water depth, i.e. ![]()
Then we get the linearized 2-D depth-averaged
equations for tidal flow:


Since these equations are linear, they could be solved
for each harmonic constituent independently, and the various solutions
superimposed to give the response.
To examine the dynamical processes that these
equations describe, it is instructive to consider the simple unforced
case:
How do the sea surface and currents respond if the
ocean is distorted in some way, but the disturbing force is removed?
The response is in the form of free waves.
Omit friction at first:

A solution can be obtained for the case of a wave
traveling in the x direction with no
movement in the y direction, i.e. v=0.
Differentiate (1) w.r.t. x, and (3) w.r.t. t, and difference them to get
This is a wave equation with wave speed given by c2= gh.
The general solution is:
![]()
from which it follows that
from equation (1).
Equation (2) shows that

The amplitude of the wave decreases exponentially in
the y direction.
The length scale of the exponential decay is

which we see can be interpreted as the distance the
Kelvin wave (or any long wave in shallow water) travels in one inertial period.
[Sketch schematic of
Kelvin wave - Bowden fig 2.7]
Amplitude decreases to the left of the direction the
Kelvin wave propagates.
Exponential term means the Kelvin wave must be limited
by a coast (where the pressure gradient can be balanced by the coastal
boundary)
Simple harmonic wave form: ![]()
where k = wavenumber
, and ![]()

In the deep ocean
·
h
= 4000 m, g = 9.81
gives c = 200 m/s
·
c
= wavelength/period for M2
= 12.4 hours
·
=> wavelength = 8850 km
·
R = 2720 km (at latitude 30o where f
= 7.29 x 10-5 s-1)
·
For
e.g. A = ~0.5 m we would get
U = ~2.5 cm/s

Eliminating
from (1) and (3) gives

which is a wave equation
with a damping term.
Anticipating that the same
form of solution will work provided we accommodate the possibility the wave
amplitude decays as the wave propagates, we try:

Differentiating and
substituting into the damped wave equation we find (by independently satisfying
the real and imaginary parts of the equation) that:
or 
showing the wave speed is
reduced.
We also so have that:

so after dividing through by
2k we estimate that frictional decay scale

provided we assume c/co
is approximately equal to 1 (i.e. friction is weak).
The solution for velocity
can be derived from continuity

after assuming a solution of
the form
![]()
which incorporates the
likelihood of a phase shift with respect to the sea level displacement:

The solution is:

Notice that the current:
·
is no longer in phase with the amplitude
o the current
peaks a time a/w before the sea level crest
·
the magnitude is reduced compared to the same
frictionless wave