Yamamoto 1982
Yamamoto 1982
Yamamoto 1982
submarine earthquakes
TOKUO YAMAMOTO
Ocean Engineering Division, Rosenstiel School of Marine and A tmospheric Science,
University of MiamL Miami, Florida 33149
The time harmonic problem of the propagation of gravity waves and acoustic waves in the ocean by
vertical oscillation of a block of ocean floor is treated analytically. It is found that the evanescent
modes of gravity waves change into the propagating modes of acoustic waves at frequencies above the
cut-off frequencies. Therefore, the common assumption of an incompressible ocean water invites
serious errors in the pressures and the tsunami waves near the source. From the propagating acoustic
wave spectra, it is shown that the size and the amplitude of the oscillating source can be determined
far in advance of tsunami arrival.
INTRODUCTION
The problem of the generation and the propagation of
tsunamis is important and thus has been investigated by
many engineers and geophysicists. Usually a large section of
ocean floor moves and vibrates for several minutes and
generates tsunamis, a Because the high frequency components of the ground motion is not important for the tsunamis
far from the source, only the non-oscillatory component of
the ground motion has been considered in the analysis of
tsunami generation. Usually a sudden uprise of a block of
ocean floor is used as the source (e.g. Sells,6 Kajiura, 4
Hammack2). Consequently, the high frequency components
of the ground motion has been ignored in consideration of
tsunamis.
At the recent conference on tsunamis, a however, it was
recognized that such high frequency components may be
important with regard to the tsunamis near the source, as
well as to the tsunami warnings for the locations far from
the source. For this reason, only the high frequency components of the ground motion are considered here.
The problem of propagation of gravity and acoustic
waves in a layer of water with a constant depth and a free
surface by a harmonic vertical oscillation of a block of rigid
bottom is considered in this paper. As the block of bottom
oscillates vertically the disturbances in the forms of displacement and pressure propagate through the fluid. At the
free surface, the disturbances form surface waves, the dispersions of which are uniquely governed by the free surface
boundary condition which is a function of the acceleration
of gravity, the water depth and the frequency of oscillation.
When surface water waves are considered, the fluid
medium is usually treated as an incompressible and irrotational fluid. Under this assumption, the analytical solution
which satisfies the free surface boundary condition is given
as a sum of only one propagating mode and an infinite
number of evanescent modes near the disturbance for a
given frequency of oscillation. 1 However, as water is actually
compressible, an otherwise non-propagating evanescent
mode changes to a propagating mode at a frequency higher
than a cut-off frequency as will be shown in this paper. In
other words, there can be more than one propagating surface waves for a given frequency. This peculiar phenomenon
0261-7277/82/020075-0852.00
1982 C M L P u b l i c a t i o n s
THEORETICAL ANALYSIS
(1)
2b
.I
Figure 1.
Definition of problem
75
1 b2
v2 -
e 2 at 2
(2)
= - - F ( w , h)
c = (Ko/Po)In
=h
(3)
rl = (3/ay)y=a
(4)
(5)
(.0 2
la2= w2
~ (Re0a)/> 0)
2to .
sin wb
(1 8)
law
(6)
(7)
(17)
We only consider time harmonic problems so that the displacement potential may be expressed by
(16)
where
gO + (02/Ot2)y=a ----0
(15)
(la-~)exp(ph)Cl-(la+~21exp(-lah)C2=Og !
(19)
to exp(--lah) sin wb
O)2
vV= --~f
(20)
(8)
]2 -- (.,o2/g
C2
--
to e x p ~ h ) sin wb
(21)
The function F(w,y) is obtained as
(bf)y
= t H=o( b 2 - x 2 ) ~ y
(9)
3f
co2
by
F(w,y) - - -
------f--
O(y-- h)
(10)
Through the Fourier inverse transformation defined by
equation (12), we obtain the functionf(x,y) as
oo
Since the bottom boundary condition (9) is inhomogeneous, it is difficult to solve the problem in the space
domain. However, the boundary condition becomes homogeneous in the wave number domain. Thus, the Fourier
wave number transformation is used to solve the problem.
Let F(w,y) be the Fourier transform o f f ( x , y ) so that
to r
f ( x , y ) = i2-~ )
--o
F(w,y)= i
exp(--iwx)f(x,y)dx
(11)
G(t,y)
--oo
f(x,y) = ~
o~
exp(iwx)F(w,y) dw
(12)
(24)
to
[G(x
f ( x , y ) = -i2rr
a2F
~ - - - w z F + - -3y
= 02
(131
+ b,y)
-- G(x -- b,y)]
(25)
(26)
dw
w(Tcosh h- si h h)
0,) 2
= 0
(27)
(28)
where ks, ko, kn and 3,n are real positive and given by
ks = ~ / c
ko = (P~ + k2s)w2
kn
~___ (/,,-2
2",1/2
(30)
(w) dw +
g(w) dw + ~
C_ko
g(w) dw
n=l
C_k n
~ 2 = - - g / a n tanpnh
w=O,+-ks,+-ko, + kn,+-ik"
+ f e w> w+
c-k~
+ ~.
f g(w)dw
Co
ck~
g(w) dw +
g(w) dw
Ckn
C_ ko
(n= 1,2 . . . . )
I w=+27ri
~,
n =N+
iA
?.., / \
G(~,y)= + 27ri ~
- k . -k,
-k s
ks
k',
(32)
ko
-tim
-ikn+ ,
R(+ iXn)
n=N+l
>k
kn
R(+-ikn)
1
-k o
(31)
n=l
r~O
R~
Iff
+
lCR
C_ks
+ ~"
ff
+
n= 0
C_kn
CO
-ikn+ 2
-iXn+ 3
+
~
Figure 2.
(33)
n=O
~s
The integration along the large semi-circle CR from R to
--R vanishes as R tends to infinity because,
77
f exp(iw~) dw
W 2
R..-..~,.~
CR
CR
f
~< lim
R--~
R dO
rr
exp(20) ~<R~Rlim
--= 0
f(x,y)
4~o
- .=1Z
k .(Zu.h +
lim
+
Co
Lf)
= r- 7ri R (0) + R ~
(+ kn)
n=O
(34)
R(+kn)=--
( n = 1,2 . . . . N )
R (+ iXn)
X~n(2/anh + sin2/anh) , (n = N +
1. . . . )
(35)
Substituting the results of the calculations given by
equations (35) into equation (33), the Green's function
G(~,y) is obtained in the following form:
4/ao cosh/aoy cosko~
ko~(2/aoh + sinh 2/aoh)
N 4/an COS/any coskn~
DISPERSION OF WAVES
-- n=l
~ k2n(2/anh + sin 2/anh)
~,
/.
or
6o2h
~/g)
~os-k~+k~sinkq~hj
} (36)
The upper signs of the double signs are for ~ > 0 and the
lower for ~ < 0 as before.
The function f ( x , y ) is now obtained from equation
(25) with the Green's function (36). The evaluation
o f f must be made for the three ranges o f x ; x > b , Ix [ < b ,
and x < - - b .
For the range of I x l > b , noting that
cos ko~ = Re [exp(--ikoO] and so on,
/ao sin kob cosh/aoy exp [--i(kox--7r/2)]
f ( x , y ) = + 4~o
k2(2/aoh + sinh 2/aoh )
4" n=lZ
-~ ~, /an sinh~"bcs/a"yexp(+X"x) t
n=N+l
~2n~5+~2~nh)
]
G(~,y) = +-Tri
2u.h)
n:N+l
~
r-*0
sin
(37)
- / a h tanh/ah
(39)
where
(40)
As seen in equations (39) and (40), for a given wave frequency w, the dispersions of gravity waves and acoustic
waves are governed by two non-dimensional parameters
o2h/g and gh/c 2. The dimensionless parameter 6o2h/g is the
Foude number and rewritten as hZ/(gh/o2). Since gh/w 2 is
proportional to the wavelength of surface gravity waves,
the Froude number co2h/g may be interpreted as the square
of the ratio of the water depth to the wave length of
surface gravity waves. The dimensionless parameter gh/c 2
is the square of the ratio of the phase velocity of surface
gravity waves to the velocity of acceleration waves.
Two limiting cases are derived from the general dispersion relation (39). For the case of an incompressible fluid,
gh/c 2 -~ O,/ah ~ wh, only the gravity wave mode exists. The
dispersion relation (39) coincides with the familiar dispersion relation of the gravity waves in an incompressible fluid
for this case,
W2h
-=
g
wh tanh wh
(4
Group
In consideration
of acoustic wave propagation in oceans,
the gravity effect is neglected, or g/w2h + 0. For this case
ph + i [02h2/c2 - (wh)ll
cos [W2h2/c2 - (wh)2]2 = 0
.6
.6
.4
.2
Vcloclty
I du
Fair
(3.4)
or
wJr+2h2,c2-((~j]12,
n=1,2,...
(43)
Figure 4.
...
n = 1,2,3
0n=-T-h3
The dispersion relations, wh/c versus kh, for the gravity
waves and the acoustic waves are shown in Fig. 3 for the
case of dgh/c = 0.15 which corresponds to h = 4.5 km. As
shown in Fig. 3, at low frequencies, wh/c < 1, only gravity
waves can propagate. However, acoustic waves start to
propagate in addition to gravity waves at high frequencies,
oh/c > 1.6.
For the case of h = 4.5 km, the critical frequency is
0.0777 Hz. The assumption of incompressible fluid results
in a longer wave length than the real wave length of the
gravity wave and permits no propagating acoustic waves.
The assumption of no gravitational effect gives lower cutoff frequencies
of acoustic waves compared to a real
situation.
By definition, the group velocity of a propagating wave
is given by
/
1/l
dw
c,=
dk
PROPAGATION
OF WAVE ENERGY
(46)
TLh
e=-
I+
TLh SJI
0 0 0
*ii
pP)Q
dxdydt
(47)
p=-p$=p&$
(48)
vector is given
Kh
of Gph)12/c= 0.15 or h = 4.5 km. Gravity waves propagate slower compared to acoustic waves. At frequencies
higher than the cut-off frequency, e.g. 0.0777 Hz for
h = 4.5 km, more than one wave propagates at the same
frequency but at different wave lengths. As can be seen in
Fig. 4, both gravity waves and acoustic waves in the ocean
are highly dispersive.
0.15
C=
(45)
cub
t
As seen in equation (43), acoustic wave modes can propagate only at frequencies higher than cut-off frequencies, w,
(2n - 1) c
I
0
A-@
A-Q
2n/w
I ,
cub
c
Figure 3.
Dispersion diagrams
1
e=---awk
4n2h
Since the displacement
s
0
2nlk
SI
f ;;
@2dy dx dt
(50)
02
potential
of a propagating
mode is
79
q5= +Q)
equation
IO-:
(51)
Gravity
(49) reduces to
e=-
Wave
&C=
Mode
+-
PO4
a)(v) dy
8hc2 s
0.15
= 2.0
(52)
function
,o-:
for the
Potential
p. sink0 b coshp,y
@o(Y) = 4to
Kinrtic
1
en = ,P.%
c2kz(2p,h
(54)
(55)
10-5
-4
lo
energy E, of the
(a > %)
+ sin 2&h)
3
E,C
Energy
(53)
Energy
by
TLh
j(=-
;pq2 dy dx dt
TLh .
0
(56)
of velocity is given by
q2 = z q;
I
.I
Id6
~~
(57)
K,
(60)
of water surface
(61)
.4
(62)
vo = 2Pgt:
(59)
v=GPg+
Figure 5.
(58)
.3
cub
80
I_/,
\
uf, sink,b
(63)
(64)
IO0.O
Wove
Modes
o.,o. . 2 . o
102
~Totol
(A-2)
Energy ( A - I )
Klnllic
and
=otentioI ( A - I
I
00
Figure 6.
I I II
II
i
I
i
m
|
m
I
I
|
II
II
|
m
I
m
m
i
l
I0.0
I|
II
II
II
I!
II
I|
##
sS
1.0
,,
Ill
0.01
I i
ii
i i
0.1
I I
I I
I I
t
t
I I
0.1
I I
h I
I I
1.0
IO.O
f (Hz)
Figure 7.
EoC
CONCLUSIONS
Elalltle(A-I
EC G
Po = phto6O 2
I0 I
I I
PRESSURE AT BOTTOM
~ , C = ~
.......
C=1400 m~
h=300 m
Acoustic
ACKNOWLEDGEMENTS
This work is entirely sponsored by the National Science
Foundation through a Grant No. CEE-8104429. The author
81
REFERENCES
1 Goda, Y. and Kikua, T. The generation of water waves with a
vertically oscillating flow at a channel bottom. TechnicalReport
No. 9, Port and Harbour Technical Research Institute, Ministry
of Transportation, Yokosuka, Japan, August, 1964
2 Hammack, J. L. A note on tsunamis: their generation and
propagation in an ocean of uniform depth, Journal of Fluid
Mechanics 1973, 60, part 4,769
82