This action might not be possible to undo. Are you sure you want to continue?
by
Frederick E. Camfield
SPECIAL REPORT NO. 6
FEB RUARY 1980
Approved for pub 1 ic re le a se;
distribution unlimited.
U.S. ARMY, CORPS OF ENGINEERS
COASTAL ENGINEERING
RESEARCH CENTER
Kingman Building
Fort Belvoir, Va. 22060
SR6
Reprint or republication of any of this material shall give appropriate
credit to the U.S. Army Coastal Engineering Research Center.
U.S. Army Coastal Engineering Research Center
Kingman Building
Fort Belvoir, Virginia 22060
For sale l>y the Superlntendent o! Documents, U.S. Oovcmmcnt Prlntlng omcc, Washington, D.C. 20402
UNCLASSIFIED
SECURITY CLASSIFICATION OF THIS PAGE (When Date Entered)
REPORT DOCUMENTATION PAGE
READ INSTRUCTIONS
BEFORE COMPLETING FORM
1. REPORT NUMBER
,2. GOVT ACCESSION NO. 3.
RECIPIENT'S CATALOG NUMBER
SR6
4. TITLE (and Subtil/a) S. TYPE OF REPORT a PERIOO COVERED
TSUNAMI ENGINEERING
Special Report
6. PERFORMING ORG. REPORT NUMe!OK
7. AUTHOR(e) 8. CONTRACT OR GRANT NUMBER(e)
Frederick E. Camfield
9. PERFORMING ORGANIZATION NAME AND ADDRESS 10. PROGRAM ELEMENT, PROJECT, TASK
Department of the Army
AREA a WORK UNIT NUMBERS
Coastal Engineering Research Center (CERENCD)
F31234
Kingman Building, Fort Belvoir, Virginia 22060
Il. CONTROLLING OFFICE NAME AND ACCRESS 12. REPORT DATE
Department of the Army February 1980
Coastal Engineering Research Center 13. NUMBER OF PAGES
Kingman Building, Fort Belvoir, Virginia 22060 222
14. MONITORING AGENCY NAME a ADDRESS(If dlllerent lrorn ControlllnS Otllce) IS. SECURITY CLASS. (olthle report)
UNCLASSIFIED
ISe. DECL AS SI FIC A TION/ DOWN GRAOI NG
SCHEDULE
16. DISTRIBUTION STATEMENT (ol lhle Report)
Approved for public release; distribution unlimited.
17. DISTRIBUTION STATEMENT (ol the abetract entered ln Block 20, Il dlllerenl from R eport)
18. SUPPL.EMENTARY NOTES
19. KEY WORCS (Continue on reveree sida Il neceuary 11nd ldenllfy by block numbsr)
Coastal engineering Mathematical models
Coastal structures Seismic safety
Edge waves Tsunamis
Flood frequencies Water waves
zo. A S S ~ A C T (QmtiDue .,.. ,..,.r_ eldiJ If,..,.,_..,. IIDd. ldenLify by block nurnber)
This report provides a source of stateoftheart information on tsunami
engineering. The report summarizes available information, identifies gaps in
existing knowledge, and discusses methods of predicting tsunami flooding . The
generating mechanisms of tsunamis and the method of determining the probabi li ty
of occurrence are given.
(continued)
DO
FORM
1 JAN 73
1473
EDinON OF 1 NOV 65 IS OBSOL.ETE
UNCLASSIFIED
SECURITY CLASSIFICATION OF THIS P AGE ( lrhen Data Enter ed)
IINCIASSTEJEQ
SECURITY CLASSIFICATION OF THIS PAGE(Whon Data Entarad)
Because of the limited data available on tsunamis, numerical methods are
commonly used to predict tsunami flooding of coastal areas. Finitedifference
equations are presented for simulating the propagation of tsunamis, but com
puter programs are omitted because of the continuing work in progress and the
availability of uptodate computer programs from other sources. Known mathe
matical solutions, for tsunamis approaching the shoreline and tsunamishoreline
interaction, are given to illustrate the effects of tsunamis and provide means
of verifying numerical results. The report discusses tsunamistructure inter
action and illustrates various types of damage caused by tsunamis .
2 UNCLASSIFIED
Së:CURITY CLASSIFICATION OF THIS PAGE(When Data Entered)
PREFACE
The high value and high utilization of the coastal zone require the
establishment of flood levels that may occur as the result of various
natural causes, and the consideration of preventive measures that can be
used to minimize lasses. This report, prepared as one of a series of
reports to be published to form a Coastal Engineering Manual, is con
cerned with the effects of tsunamis on the coastal zone. Another report
in the series, to be published separately, will be concerned with the
effect of storm surges. The work was carried out under the coastal engi
neering research program of the U.S. Army Coastal Engineering Research
Center (CERC).
The report was prepared by Dr. Frederick E. Camfield, Hydraulic
Engineer, under the general supervision of R.A. Jachowski, Chief, Coastal
Design Criteria Branch. The report was reviewed by Dr. F. Raichlen,
California Institute of Technology; Drs. H.G. Loomis and L.Q. Spielvogel,
Joint Institute of Marine and Atmospheric Research, National Oceanic and
Atmospheric Administration; Drs. R.W. Whalin and J.R. Houston, U.S. Army
Engineer Waterways Experiment Station; Dr. M. Fliegel, W. Bivins, and
L.G. Hulman, Nuclear Regulatory Commission; Dr. R.J. Geller, Stanford
University; and Dr. R.E. Meyer, University of Wisconsin. The author ex
presses his appreciation to the reviewers for their many helpful comments
and suggestions.
Sorne of the background work on tsunami engineering was partially
funded by the Nuclear Regulatory Commission.
Comments on this publication are invited.
Approved for publication in accordance with Public Law 166, 79th
Congress, approved 31 July 1945, as supplemented by Public Law 172, 88th
Congress, approved 7 November 1963.
3
Colonel, Corps of Engineers
Commander and Director
CONTENTS
CONVERSION FACTORS, U. S. CUSTOMARY TO METRIC (SI)
SYMBOLS AND DEFINITIONS
1 INTRODUCTION ..... .
1. Nature and Origin of Tsunamis.
2. Probability of Occurrence.
II THE GENERATION OF TSUNAMIS.
1. Subrnarine Earthquakes.
2. Volcanic Activity ...
3. Landslides and Subrnarine Slurnps.
4. Explosions ....
III MECHANICS OF GENERATION
1. Area and Height of Uplifting
2. Initial Wave Formation
IV TSUNAMI PROPAGATION . . . .
1. SrnallArnplitude Waves.
2. LongWave Equations ..
3. Distantly Generated Tsunamis
4. Nearshore Propagation ...
5. Computer Models ..
6. Nearshore Computer Models.
V TSUNAMIS APPROACHING THE SHORELINE.
1. Abrupt Depth Transitions ..
2. Linear Depth Transitions ...
3. Nonlinear Depth Transitions ..
4. Experimental Measurernents ...
5. Solitons and ShoalingInduced Dispersion
VI TSUNAMISHORELINE INTERACTION
1. Wave Reflection.
2. Shelf Resonance.
3. Reflection from Seaward Edge of Shelf.
4. Edge Waves
5. Refracted Waves and Caustics
6. MachStern Formation.
7. Bay and Harbor Resonance
VII TSUNAMI RUNUP AND INTERACTION WITH STRUCTURES
1. Tsunami Runup on a Shoreline ..... .
2. Interaction with ShoreProtection Structures
3. Other Shoreline Structures ..
4. Tsunami Surge on the Shoreline ...... .
4
Page
9
10
17
17
21
27
28
30
30
31
32
32
34
38
38
44
47
51
55
60
67
67
73
76
79
80
88
88
91
98
1@1
108
130
131
146
147
158
167
169
CONTENTSContinued
VIII TSUNAMI WARNING SYSTEM AND INSTRUMENTATION
1. The Tsunami Warning System.
2. Buman Response ..... .
3. Ionospheric Waves ... .
4. DeepOcean Tsunami Gages.
IX SUMMARY AND CONCLUSIONS .
LITERATURE CITED .
APPENDIX TSUNAMIS OCCURRING BETWEEN 1891 AND 1961
TABLES
Distribution of amplitude U
2 Values of horizontal water particle displacement, N, and
wave amplitude, U . ....
Page
193
194
196
198
198
199
201
217
93
98
3 Resonant edge wave parameters 103
4 Dimensions, periods of fundamental mode, and intensity of
secondary undulations of inlets of A l a s k ~ and British Columbia,
and of Puget Sound . 133
5 Allowable overtopping heights 162
6 . Drag coefficients . 178
7 Typology of tsunami events.
197
FIGURES
Oceanic zones of recent earthquake activity, showing association
with trench systems and island arcs. 18
2 Wave height versus tsunami magnitude. . . . 23
3 Principal fault systems and distribution of epicenters of major
· Alaskan earthquakes, 18981961 . . . . . . . . . . . 25
4 Mean annual occurrence of shallowfocus earthquake shocks for
the Aleutian and southeastern Alaska region. 26
5 Movement along faultlines . . 28
6 Wave record from Wake Island, showing arrivai of tsunami. 29
7 Horizontal motion normal to continental slope
35
5
CONTENTS
FIGURESContinued
8 Convergence of wave rays . .
9 Spherical coordinate system.
10 Coordinate system, rectangular coordinates
11 Coordinate system, spherical coordinates .
12 Graphical representation of the total transmission open
boundary condition. . . . . . . . . . . . . . .
13 Surface elevation contours 13,000 seconds after the 1964
Alaska earthquake . .
14 Position of variables.
15 Computation of pressure near the free surface.
16 Wave passing onto shelf ...
17 Wave reflection from a shelf
18 Transmitted wave angle e
2
versus incident wave angle e
1
19 Linear slope and shelf
20 Reflection and transmission coefficients
21 Slope and shelf .....
22 Reflection coefficients.
23 Reflection and transmission coefficients
24 Separation of solitons ...
25 Induced wave generation over a submerged bar
26 Wave enhancement . . . . . .
27 Solitary wave propagating over a slope onto a shelf.
28 Solitary wave propagating onto a shelf
29 Wave train propagating onto a shelf.
30 Wave reflection from a shoreline . .
6
Page
48
50
56
58
59
60
62
64
68
70
70
74
74
77
77
80
81
81
82
83
83
87
88
CONTENTS
FIGURESContinued
31 Shelf resonance.
32 Resonant amplification on a shelf.
33 Reflected waves on a shelf
34 Offshore profiles of edge waves.
35 Schematic of caus tic
36 1960 tsunami refraction, Hilo, Hawaii.
37 Trapping of generated tsunami.
38 Solution to equations (264) and (265).
39 Machstem formation, solitary wave
40 Plan view of inlet
41 Amplification factor versus relative harbor length
42 Wave radiation functions . . . . .
43 Frequency response of a fully open harbor.
44 Theoretical frequency response curves of harbors
45 Harbor with an entrance channel ..
46 Wavelength for Helmholtz resonance
47 Tsunami water levels in a bay.
Page
92
95
99
102
108
109
113
123
131
135
136
137
138
139
141
142
144
48 Response curve at point C of the Long B ~ a c h harbor madel . 146
49 1964 tsunami runup, Kodiak City, Alaska. 148
50 Tsunami runup at Kamaisi, Japan. 150
51 Tsunami runup at Hongo in Toni, Japan. 151
52 Tsunami runup at Ryoisi, Japan 152
53 Solitary wave runup. . . 153
54 Concrete seawa11 destroyed by 1960 tsunami, Shizukawa Bay,
Japan . . . . . . . . . . . . . . . . . . 160
7
55 Overtopping volumes.
CONTENTS
FIGURESContinued
Page
161
56 Suggested design for rehabilitated breakwater, Hilo, Hawaii. 163
57 Suggested design, typical nonovertopping barrier section,
Hilo, Hawaii. . . . . 163
58 Seawall cross sections 164
59 Cross sections of seawall, Yamada, Japan 165
60 Coconut palms near shoreline, Hilo, Hawaii 166
61 Grove of pandanus trees knocked dawn by 1946 tsunami on the
Island of Kauai, Hawaii . . 167
62 Dock damaged by 1964 tsunami at Crescent City, California. 168
63 Tsunami damage to railroad bridge on Wailuku River, Hilo,
Hawaii. . . . . . . . . . 170
64 Tsunami damage to railroad trestle on Kolekole Stream, Island
of Hawaii . . 170
65 Bridge damaged by 1960 tsunami at Mangoku, Japan 171
66 Definition sketch of surge on a dry bed. . 175
67 Determination of CD when flow passes under a structure
68 Large boulder moved by 1960 tsunami, Hilo, Hawaii ..
69 CM for twodimensional flow past rectangular bodies
70 Example plots of x versus t for abjects moved by tsunami
surge .
71 Building moved by tsunami surge.
72 Damage to oil tank farm at Whittier, Alaska.
8
180
184
187
188
189
191
CONVERSION FACTORS, U.S. CUSTOMARY TO METRIC (SI)
UNITS OF MEASUREMENT
U.S. customary units of measurement used in this report cru1 be converted
to metric (SI) units as follows:
Multiply
inches
square inches
cubic inches
fe et
square feet
cubic feet
yards
square yards
cubic yards
miles
squarv miles
knots
acres
footpounds
millibars
ounces
pounds
ton, long
ton, short
degrees (angle)
Fahrenheit degrees
by
25.4
2.54
6.452
16.39
30.48
0.3048
o. 0929
0.0283
0.9144
0.836
0.7646
1.6093
259.0
1.852
0.4047
1. 3558
1. 0197
28.35
453.6
0.4536
1. 0160
o. 9072
0.01745
5/9
x 1o3
mil limet ers
centimeters
To obtain
square centimeters
cubic centimeters
centimeters
met ers
square meters
cubic meters
met ers
square meters
cubic meters
kilometers
hectares
kilometers per hour
hectares
newton meters
kilograms per square centimeter
grams
grams
kilograms
metric tons
metric tons
radians
Celsius degrees or Kelvinsl
1
To obtain Celsius (C) temperature readings from Fahrenheit (F) readings ,
use formula: C = (5/9) (F 32).
To obtain Kelvin (K) readings, use formula: K = (5/9) (F 32) + 273.15.
9
SYMBOLS AND DEFINITIONS
A a coefficient
• area of uplifting
eprojected area of body normal to flow direction
a wave amplitude
• a coefficient
esemimajor axis of ellipse
einterfocal distance of coordinate ellipses
•a distance from the shoreline = ( /2  1) d
8
/S
elength of building in direction of flow
8n variable used in determining movement of the sea surface
a
1
amplitude of incident tsunami; amplitude in deeper water
a
2
amplitude of tsunami at head of bay or inlet; resonant amplitude
B mean width of a harbor or inlet
• a coefficient
Bj a variable used in determining wave amplitude
B
1
width of outer bay or inlet
B
2
width of inner bay or inlet
b a coefficient
ewidth of building transverse to direction of flow
esemiminor axis of ellipse
ewidth of breakwater opening; entrance width of bay, harbor, or inl et
c wave ce le ri ty
• a coefficient
cv
drag coefficient
Ce
a Mathieu function
Cp force coefficient
ca
group velocity
ch
Chezy roughness coefficient
CM
added mass coefficient
c maximum uplifted elevation
• a constant
10
D
F
f
SYMBOLS AND DEFINITIONSContinued
drift speed of the node or antinode of an edge wave
a coefficient used in determining wave reflection at the shoreline
a coefficient
focai depth of earthquake
a variable used in determining wave amplitude
water depth
eprojected dimension transverse ta direction of flow
average depth
average harbor depth
a length representative of water depth
depth of water at the point where a wave ray turns parallel ta
bottom contours
depth of water at toe of nearshore slope
depth of deeper water
depth of shallower water
energy
incident wave energy
reflected wave energy
transmitted wave energy
even Mathieu Transform
force
buoyant force
drag force
odd Mathieu Transform
friction factor
coriolis parameter 2SG cos 8
11
SYMBOLS AND DEFINITIONSContinued
fS(Z) probability of an astronomical tide of elevation Z
g acceleration due to gravity
H wave height
~ incident wave height
H(l) zeroarder Hankel function
0
Hr reflected wave height
Ht transmitted wave height
H wave height when the leading edge is at the shoreline
h surge height
• uplifting height
h average height of uplifting
h
8
surge height at the shoreline
hw wall height
• wetted height on a structure
I relative intensity of secondary undulations
j an integer used for increments
K R/h
8
Krz a constant
Kr coefficient of reflection = H r / ~
Kt coefficient of transmission = Ht/Hi
k an integer used for increments
ewave number = 2n"/L
k
0
wave nurnber at lowest mode of resonance (Helmholtz mode)
L distance normal to faultline over which vertical Earth movement
occurs
• wave length
1o length of bay, harbor, or inlet
1
0
length of harbor entrance channel
12
SYMBOLS AND DEFINITIONSContinued
Le effective length of harbor or inlet
Lf fault length
1
0
wavelength at lowest mode of resonance (Helmholtz mode)
1y longshore wavelength
1
1
wavelength in deeper water
1
2
wavelength in shallower water
t length of slope connecting sea bottom to a shelf
t
8
distance across a shelf
M Richter magnitude of earthquakes
•moment
M
0
momentum
m tsunami magnitude
N normalized horizontal water particle displacement
n Manning roughness coefficient
n(m) probability of tsunami with magnitude m being generated in any
given year
P(Z) probability of runup to elevation Z
p pressure
ea coefficient in wave refraction
Q flow rate under a wave
q a coefficient in wave refraction
evelocity of a water particle under a wave
R a coefficient
evertical height of runup above the stillwater leve! at the
shoreline
R average runup height at a shoreline area
Re radius of the Earth
Rs radius from the center of curvature to the shoreline
13
r
s
s
T
t
u
u
u*
v
SYMBOLS AND DEFINITIONS  Continued
radial distance
radius from the center of curvature to where the wave ray turns
parallel to the bottom contours
slope of sea bottom in direction of wave motion
slope of steep transition
slope of shelf
nearshore slope
even Mathieu Function
odd Mathieu Function
distance along a wave ray
wave period
bottom shear stress
natural period of harbor inside breakwater
nth mode of oscillation
component period of Earth vibration
period of bay or inlet
natural period of inlet with effective length, Le
ti me
time required for a wave to travel across a shelf
normalized wave amplitude
• (H/ d)(L/ d)
2
current or particle velocity in direction of wave motion
• veloci ty in the adirection (spherical coordinates)
current velocity of tsunami surge at the shoreline
a convection term
volume of water
14
v
w
w*
x
y
z
z
y
ô
~ ( x , t )
n
SYMBOLS AND DEFINITIONSContinued
horizontal velocity in direction transverse to wave motion
evelocity in the tdirection (spherical coordinates)
velocity in the vertical direction
a convective term
horizontal coordinate in direction of wave motion
distance between the shoreline and the point where a wave ray
turns parallel to bottom contours
horizontal coordinate in direction transverse to xdirection
maximum scarp height
depth of wave generation
a parameter used in wave reflection and transmission
vertical coordinate
dimensional constant
ecorrection term for harbor entrance channel length
angle in wave refraction between the incident wave ray and an
orthogonal to the contour of the sea bottom
angle of the beach slope
ewave ray separation distance
•angle of nearshore slope given in radians
edimensional constant
specifie weight of water
•an edge wave parameter
dimensionless amplitude
an incrementa! distance
a small value
dimensional constant
an arbitrary increment
eH/d
vertical movement of sea bottom
water surface elevation above still water at an arbitrary point
15
SYMBOLS AND DEPINITIONSContinued
n(x,t) vertical movement of sea surface
8 angle in polar coordinates
eangle of inclination of water surface at front of surge
•degrees latitude measured from the pole
p
angle of incident wave ray
ea constant
angle of transmi tted wave ray
ea constant
horizontal displacement of a water particle
density
edimensionless distance from the shoreline
cr d/L
~ longitude of a point
x/a
~ ( t ) variable used in determining movement of the sea surface
x dimensionless distance measured seaward from the shoreline
$ latitude of a point
$
1
wave radiation function
$
2
wave radiation function
n an arbitrary function
erotational speed of the Earth in radians per second
16
TSUNAMI ENGINEERING
by
Frederick E. CamfieZd
I. INTRODUCTION
The term tsunami is derived from two Japanese words: "tsu," meaning
harbor, and "nami," meaning wave. Tsunamis, or seismic sea waves, have
very long periods and are not easily dissipated. The waves may create
large surges or oscillations in bays or harbors which are not responsive
to the action of normal sea waves. In the original definition, the term
tsunami was applied to all large waves, including storm surges. However,
recent definitions have limited its application to waves generated by
tectonic or volcanic activity. Western literature previously referred
to these waves as tidal waves or seismic sea waves, but those terms have
generally been replaced by the term tsunami.
Tsunamis are primarily created by disturbances in the crust of the
Earth underlying bodies of water, and the resulting uplifting of the
water surface over a large area which forms a train of very longperiod
waves. The waves may have periods exceeding 1 hour, in contrast ta nor
mally occurring windgenerated sea waves which have periods less than 1
minute. When tsunamis are generated by volcanic activity or landslides,
the wave energy tends to spread along the wave crests and the tsunamis
affect mainly the areas near their source. Tsunami waves generated by
tectonic uplifting may travel across an ocean basin, causing great
destruction at locations far from their source.
Because of the potential destructive effects of tsunamis, it is
necessary to understand the mechanisms of their generation and propaga
tion, and ta be able to predict the extent of flooding and the effect of
wave forces in coastal areas subject to tsunami attack. Proper control
must be exercised over the use of such areas, and in designing structures
ta be placed in these areas. Also, sufficient warning of a tsunami
attack must be given ta people located in these areas, and procedures
must be established for an orderly evacuation when necessary.
This report discusses the prediction of tsunami effects in coastal
areas and attempts to provide guidance in determining the flooding
and wave forces at any particular location. The present knowledge of
tsunamis and the deficiencies in this knowledge are summarized.
1. Nature and Origin of Tsunamis.
Areas of seismic activity which could potentially generate tsunamis
are shawn in Figure 1. The major part of this activity occurs along the
boundaries of the Pacifie Ocean, with other regions of strong activity
primarily concentrated in the Caribbean and Mediterranean areas. Van Dom
(1965) indicates that the Japan Trench radiates detectable tsunamis at the
rate of about one per year. Lesser amounts of activity occur elsewhere.
17
 OCCASIONAL EARTHO\IAME9
ll!liliimi!l FREOUENT EARTHOUAKES
10 Ht+++t VOLCAHIC: ARCS
•o
 TRENCHES

=
ao• ao•
Figure 1. Oceanic zones of recent earthquake activity, showing
association with trench systems and island arcs.
Pacifie preponderance is apparent (from Van Dorn, 1965).
Tsunamis can be generated in any coastal area, including inland seas
ao•
and large lakes. Spaeth (1964) provides an extensive bibliography on
tsunamis. The Appendix summarizes the occurrence of tsunamis from 1891
to 1961, using Spaeth's data and sorne additional information from
Heck (1947), Ambraseys (1965), PararasCarayannis (1969), and Cox,
PararasCarayannis, and Calebaugh (1976). Part of Ambraseys' informa
tion has been omitted because of the lack of verification. Tsunamis
occurring between 1962 and the present are not listed because a complete
summary is not readily available.
Good records are available for more recently occurring tsunamis,
particularly in the present century; however, records of tsunamis in
past centuries are mostly based on accounts of personal observations.
The dates that tsunamis occurred have often been confused with the dates
on letters or other accounts rather than the date of the actual event.
There have also been many errors in interpreting these older accounts,
particularly when translating from one language to another. Soloviev
and Ferchev (1961) refer to the reports of an event in 1827 at the
Komandorskiye Islands, located between Kamchatka and the Aleutian Islands,
Alaska. A Russian expedition, under the command of F.P. Lutke, reported
the occurrence of an earthquake and noted that earthquakes were sometimes
accompanied by arise in water level. The original Russian report was
translated in French, then into.German, then into French, and back into
Russian again. The final translation indicated that a tsunami had
occurred along with the 1827 earthquake.
18
A similar instance of errors in interpretation and translation
occurred in the reports of an 1840 event at Santa Cruz, California.
Heavy rain and high waves caused by a storm resulted in considerable
damage. The collapse of buildings caused by flooding from the rain was
misinterpreted as an earthquake, and the waves as a tsunami. Holden
(1898) reported this as an earthquake and tsunami, when in fact neither
occurred.
Consideration must also be given to the fact that records based on
visual observations may not include all tsunamis which occurred. The
observers probably gave special notice only to those waves which caused
substantial flooding or large, rapid variations of the water level in
bays and harbors. At a location where the normal tidal range was of the
same order as the tsunami height, a tsunami occurring at a low tide stage
may have been given only passing notice, if noticed at all, while the
same tsunami occurring at a high tide stage would have been recorded as
a major tsunami. Likewise, the occurrence of a tsunami in conjunction
with high storm waves would have caused more flooding, and therefore,
may have been given more significance in the records than a tsunami
occurring during a relative calm.
Records of tsunamis in the Mediterranean and Middle East include
theories on the eruption of Thira (also known as Santorini) and the
tsunami on the coast of Crete that destroyed the Minoan Empire circa
1400 B.C. Factual accounts of tsunamis extend back at least 2,000
years. Accounts of tsunamis in Japan extend back at least 1,300 years.
In contrast, records of tsunamis originating in the ChilePeru coastal
areas only cover about 400 years (from 1562 to present), those originat
ing in Alaska about 200 years (from 1788), and those occurring in Hawaii
slightly more than 150 years (from 1813). Few records are available of
tsunamis occurring on the CaliforniaOregonWashington coastline. Holden
(1898) indicates tsunamis occurred at points on the California coastline
in 1812, with various occurrences at later dates, mainly recorded or
observed at San Francisco. Townley and Allen (1939) provide similar
information.
Knowledge of the action of more recent tsunamis can be helpful in
evaluating historical information. Although no record exists of major
tsunamis on Puget Sound in Washington State, the Puget Sound WeekZy (1866)
reported that a tide, the highest ever recorded, occurred at Port Townsend,
Washington, on 20 December 1866. The report stated, "The main street was
filled with drift logs, and the dwellers on lower floors were compelled
to elevate to the next story." Camfield's (1975) article on historical
accounts gives the date as 27 December 1866. Kelly (Seattle, Washington;
personal communication, 1979) also gives the year as 1866. Neither
Holden (1898) nor Townley and Allen (1939) report a tsunami occurrence
in 1866; both list a 26 December 1856 date, with no additional details,
which was probably an incorrect report of the 1866 event. The historical
accounts describing a graduai rise in water level indicate this was prob
ably a tsunami, but the origin is unknown.
19
Although tsunamis occur frequently in the Caribbean, they are much
less frequent in the North Atlantic Ocean. The only major recorded
tsunami along the east coast of the United States and Canada was the
tsunami which devastated the Burin Peninsula along Placentia Bay,
Newfoundland, in November 1929. At !east 26 lives were lost (Jaggar,
1929). The tsunami was enhanced by an exceptionally high tide and high
storm waves; otherwise, it may not have been of major proportions
(Hodgson and Doxsee, 1930). This tsunami was reported to have had a
height of 0.31 meter (1 foot) at Atlantic City, Néw Jersey (Murty and.
Wigen, 1976).
Stein, et al. (in preparation, 1980) report on earthquakes in eastern
Canada from Baffin Island to Newfoundland. For the 1929 Grand Banks
earthquake, which generated the tsunami, they give a magnitude of 7.2
as reported by Gutenberg and Richter (1965). Stein, et al. suggest that
the earthquakes in this area are associated with basement faults which
have been reactivated by the removal of Pleistocene glacial loads.
Earthquakes frequently qccur in the eastern United States. These
include a large earthquake that occurred in New England on 18 November
1755, shortly after the November 1755 earthquake near Lisbon, Portugal
(Reid, 1914), and the earthquake near Charleston, South Carolina, on
31 August 1886 (Taber, 1914). All of the earthquakes in the eastern
United States have occurred inland from the coastline. The probability
of an earthquake having an epicenter in a location that would cause a
tsunami, either on the coastline or in an estuary, cannot be determined
from available data. Brandsma, Divoky, and Hwang (1979) give probable
maximum waves for tsunamis at points near both the Atlantic and Pacifie
coasts of'the United States. Their results are based on mathematical
simulation of extreme events.
The only tsunamis of record that traveled across the North Atlantic
were those generated near Lisbon, Portugal, in 1755 and 1761. Both of
these were recorde.d on the south coast of England, as weil as in the
West Indies (Davison, 1936). For comparison, the 1755 tsunami had a
maximum rise of 2.4 meters (8 feet) at Penzance (England) and flowed over
the wharves and streets at Barbados (West Indies). In 1761, the sea rose
about 1.8 meters (6 feet) at Penzance and 1.2 meters (4 feet) at Barbados.
Other runup heights in 1755 were estimated at 4.9 meters (16 feet) on the
coast of Portugal, 18 meters (60 feet) at Cadiz (Spain), 1.8 meters (5.9
feet) ai Gibralter, 15 meters (50 feet) at Tangier (Morocco), 5.6 meters
(18 feet) at Madeira, 14.6 meters (48 feet) at Faial (the Azores), 2.5
meters (8.2 feet) at St. Ives (England), 3.7 meters (12 feet) at Antigua
(West Indies) 6.4 meters (21 feet) at Saba (West Indies), and the waves
overflowed the lowlands on the coasts of Martinique and other French
islands.
In general, good data are available for only a limited number of
tsunamis. A major gap in the data is tsunami heights in deep water.
Because of this gap, only l i ~ i t e d verification is available for numer
ical models for propagating tsunamis across large oceanic distances.
20
tf•
Because of the frequency of tsunamis occurring in the Pacifie Ocean,
a tsunami warning system has been developed for the inhabitants of
Pacifie coastal areas. A similar warning system has not been developed
for other areas.
2. Probability of Occurrence.
Where sufficient historical data are available on tsunami flood
levels, the probability of tsunami flooding at any elevation can be
determined by the same methods used for determining the probability of
floods on rivers. For a known period of record, the recorded flood
levels can be ranked from the largest to the smallest; i.e., the highest
flood level is ranked l, the next highest is ranked 2, and so on. Linsley,
Kohler, and Paulhus (1958) show that the probability of each flood level
is then given by
P(Z)
where
rn
n + 1
(1)
P(Z) the probability of flooding to the elevation Z in any year
z the elevation above a defined datum
rn the rank of the flood level
n the period of record in years
Houston, Carver, and Markle (1977) have determined the probability of
tsunami flood levels for the Hawaiian Islands. For recurrence intervals
greater than 10 years, i.e., P(Z) < 0.1, they give
(2)
where h
200
is the elevation of the maximum tsunami wave crest above
mean sea level (MSL) 200 feet (61 meters) shoreward of the coastline,
P(h
200
) the probability of a flood level occurring at elevation h
200
in any given year, and A and B the empirical coefficients which are
determined for each point on the coastline. Where sufficient historical
data were not available, they generated additional data using a mathe
matical model. The model data were multiplied by correction factors
and compared to historical data. This produced additional data at points
along the coastline where historical data were not available, and allowed
a determination of the empirical coefficients A and B at all coastal
points.
It should be noted that there is a probability of sorne error in the
predicted flood elevations based on available historical data. For
example, there is a 37percent probability that a 100year flood level
21
(i.e., a flood level with a recurrence interval of 100 years) will not
occur in any period of 100 years. Therefore, a 100year flood level
predicted from a 100year period of record may be too low. Also, there
is a 9.5percent probability that a 1,000year flood level will occur
at least once in any period of 100 years. Therefore, the predicted 100
year flood level, based on a 100year period of record, may be too high.
Confidence limits for the predicted flood levels can be obtained
using methods similar to those used for river flood levels. However,
rivers have a seasonal variation in flow, so a 1year time increment is
significant in that case. In the case of tsunamis, the 1year time
increment is a convenient means of measuring time, but there is no par
ticular relationship between this time increment and the generation of
tsunamis. Methods used for obtaining confidence limits for tsunami
flood levels should give the same results, regardless of the chosen
time increment.
Beard (1962) notes that there is a 5percent probability that the
magnitude of the difference between the real flood level and the pre
dicted flood level will be greater than or equal to twice the standard
error. Assuming there is an equal chance of the real flood level being
either greater than or less than the predicted value gives +2.5 and
2.5percent confidence limits.
Where no historical data are available, data may be constructed
entirely from a computer model by assigning magnitudes to various tsu
namis in the mathematical model, and by determining the probability of
generation for each tsunami magnitude. However, the results will not
have the same degree of accuracy.
An exact relationship between tsunami magnitude and earthquake
magnitude has not been determined . . Iida (1961) proposed that tsunamis
could be assigned a magnitude based on their energy (the energy of the
generated waves), with an increase in magnitude of 0.5 being equal to a
doubling of the energy. He also related the tsunami magnitude to the
maximum runup height in meters at the shoreline area experiencing the
strongest tsunami action (Iida, 1970). The relationship between the
runup height ~ a x and the tsunami magnitude rn is shown in Figure 2.
The dashlines show the range of the expected maximum runup, based on
Iida's data, due to differences in the characteristics of the individual
tsunamis and coastal areas.
Soloviev (i970) revised the definition of tsunami magnitude by relat
ing it. to the average runup height R (in meters) at the shoreline area
experiencing the strongest tsunami action. This tends to average out
any high runup heights related to a particular coastal feature, and should
be more representative of the actual tsunami energy. Soloviev does not
indicate the length of coastline to be used in the average, but does pro
vide an equation for the magnitude as
rn = log
2
(12' R)
(3)
22
80
60
50
40
30
'E
..
20
~
..
2'
Soloviev R
. ~ 15
a:
.1:.
10
1:'1
·;;
::z::
8
Q.
::1
c
::1
6 a:
5
4
3
2
6
Tsunami magnitude, m
Figure 2. Wave height versus tsunami magnitude.
As shawn in Figure 2, Soloviev's scale gives a more rapid increase in
maximum wave height than Iida's scale for a given change in tsunami mag
nitude. This indicates that an increase in magnitude on Soloviev's scale
would represent a greater increase in tsunami energy than an equivalent
increase in magnitude on Iida's scale.
Abe (1979) suggests that the tsunami magnitude can be represented
as a function of the average runup height and a constant which is
dependent upon the source region and the station where the tsunami is
rneasured. He shows that the magnitude, obtained by this rneans, can be
related to the seisrnic moment.
The probability n(rn) of a tsunami with magnitude rn being gener
ated in any given year in a specified generating area is given by the
ernpirical equation
n(rn)
23
bm
ae
( 4)
where the coefficients a and b are determined by a least squares
analysis of the available data for the generating area. To calculate
probabilities tsunamis may be placed in groups; e.g., a group of tsunamis
shown with magnitude 3.75 actually includes all tsunamis with magnitudes
from 3.5 to 4.0, etc. To analyze the probability of an individuaZ tsunami
having a magnitude greater than or equal to 3.5, the probabili ties would
be summe d
2
L:
j=o
n(3.75 + 0.5j)
n(3.75) + n(4.25) + n(4.75)
(5)
which would include all tsunamis with magnitudes from 3.5 to 5.0. I t
should be noted that the stress in rock cannat exceed sorne maximum value;
the rock will fracture when the stress reaches that value.
Abe (1975) and Geller (1976) show from empirical results that the
fault length of earthquakes is approximately equal to twice the fault
width. Using these results and the model of Haskell (1969), Ge11er gives
a maximum earthquake magnitude, M, of 8.22. Because of variations in
the assumed fau1t 1engthtowidth ratio, actual earthquake magnitudes
may exceed this value slightly; Ge11er lists a magnitude of 8.5 for the
1964 Alaska earthquake. However, as noted by Geller, the maximum magni
tude occurs because the conventional magnitude scale is saturated and
ceases to give a meaningful measure of the earthquake size .
It is assumed that tsunamis do not occur with magnitudes greatèr
than 5.0 where the tsunami magnitude has sorne relationship to earthquake
magnitude as mentioned previously. If tsunami magnitude is related to
seismic moment, defined by Kanamori (1972) as a function of rigidity,
fault area and average fault slip, Kanamori and Cipar (1974) indicate
that the 1960 Chilean earthquake had the largest seismic moment ever
reliably determined (2 x 1030 dynecentimeters).
The method for grouping tsunamis (eq. 5) has been uti lized by Houston
and Garcia (1974),using statistics for the entire trench along the Chilean
coast. Applying revised information for that particular generating area,
a major source of tsunamis in the western United States (Houston and
Garcia, 1978), a= 0.074 and b = 0.63. Taking the value rn= 3.5 for the
magnitude of a design tsunami (to be used for determining potential runup
in coastal areas), the probability for a tsunami with a magnitude of 3.5
or greater being generated in any given year is
n(3.5) = 0.074 [e0.63(3.75) + e0.63(4.00) + e0. 63(4.75)] (
6
)
which gives a value of 0.0166 or a recurrence interval of 60 years. For
a 412year period for the Chilean coast, the prediction would be seven
tsunamis of magnitude 3.5 or greater. This agrees with historical records
of tsunamis in this area.
24
Another major source of tsunamis in the western United States is the
Aleut i an Trench. Only relatively recent records exist for the area.
Analysis of these records by Houston and Garcia (1974), as revised in
Houston, et al. (1975b) and Houston and Garcia (1978), gives
n(m) = 0.113
0
·
71
m ( 7)
which is similar to the previous equation for the PeruChile Trench . The
probabi lity of tsunami occurrence is assumed to be uniform along the
trench. The distribution of recent earthquakes along the Aleutian Trench
i s shawn in Figure 3, and the mean annual number of earthquakes of any
given magnitude in Figure 4. The straight !ines in Figure 4 are not
accurate above an earthquake magnitude, M, of about 8.5 because of the
physical limits on allowable stresses in the rock forming the Earth's
crust. Also, the straight !ines in Figure 4 representing the occurrence
of earthquakes in Alaska and the world would intersect at an earthquake
magnitude of about 2.5, so the plotted lines should not be extrapolated
to values of earthquake magnitude less than those shawn.
EXPLANATION
Me pl tude
• U6.1
o i.llU
0 1.\l7.1
07.758..5
Fillodtircleo
"'hypoc•nur>50 Kil
Figure 3. Principal fault systems and distribution of epicenters of major
Alaskan earthquakes, le981961 (from Wilson, 1969; adapted from
Davis and Echols, 1962).
25
5
'
'
'
3
'
\
'
'
' \
. '
'
\
'
,.,
'
'
sœ <a
z
0.5
'
\
' 0
'
\
z
'
'
0
0.3
'
\
::Il
'
c
c
'
\
ct
'
c
'
0
'
\
«>
'
::E
'
0.1
'
'
0.05
Log
10
N=a+b(8M)
a=1.02
0.03
b= 0.98
0.01 .__ _ _,6 ____ _.7.__ ____ 8,__ _ _,
Magnitude, M
Figure 4. Mean annual occurrence of shallowfocus
earthquake shocks for the Aleutian and south
eastern Alaska region (from Wilson, 1969;
adaptcd from Berg, 1964) . Trends of World
and Japanese data are inserted for comparison.
Using equation (7), the probability of a tsunami with a magnitude of
3. 5 or greater is
n(3.5) = n(3.75) + n(4.25) + n(4.75)
(8)
which gives a value of 0 . 0174 for the Aleutian Trench. This value is
based on a relatively short period of data for large tsunamis only.
Dividing the trench into 12 segments gives the probability of 0.00145
for a tsunami of the given magnitude of 3.5 or greater to be generated
at any particular segment of the trench in any given year, assum1ng an
equal probability for each segment. The general equation for a
l ar segment of the trench becomes
n(i) = 0.0094 e
0
•
71
i
(9)
26
To determine the probability of runup of a given height at a given
location along the coastline, it is necessary to propagate tsunamis
across the ocean by numerical ~ e a n s frDm each segment of the trench for
all tsunami magnitudes (i.e., i = 2.0, 2.5, 3.0, 3.5, 4.0, 4.5, and
5.0). The wave train of each tsunami must be superimposed on segments
of the tidal cycle of an interval equal to the duration of the wave train.
This superposition must be made for each tidal segment of that interval
for a 1year period, and the probability of the resulting runup determined.
Tidal variations are discussed by Harris (in preparation, 1980). A cumu
lative probability can then be established for runup at a particular site.
Determining the probability of tsunami runup at a particular coastal
location for tsunamis generated in the Aleutian Trench area, would require
the numerical generation of 84 tsunamis (12 segments of trench and 7
intensities of each segment). As shown by Houston and Garcia (1974), each
runup value has an associated probability, and the totality of runup val
ues at a given shoreline point defines a probability distribution from
which the cumulative probability distribution, P
8
(Z), can be obtained
for runup greater than or equal to a particular value.
By approximating the probability fs(Z) of the astronomical tide by
a Gaussian distribution (Petrauskas and Borgman, 1971; Houston and Garcia,
1974), the probability of runup to a given elevation is given by
(10)
Probabilities for tsunami runup can then be determined at each coastal
point, combining the tsunami with the astronomical tide.
An analysis similar to that used for the Aleutian Trench could be
applied to tsunamis generated in other source areas. For the west coast
of the United States (excluding Hawaii), only the Aleutian Trench and the
PeruChi1e Trench appear to produce significant tsunami runup, although
Holden (1898) indicates sorne occurrence of tsunamis from sources along
the California coastline. Using numerical results obtained for tsunamis
generated along the Aleutian and PeruChile Trenches, Houston and Garcia
(1978) have determined probable 100 and 500year tsunami flood elevations
for the west coast of the continental United States.
II. THE GENERATION OF TSUNAMIS
Tsunamitype waves can be generated from a number of sources, includ
ing shallowfocus submarine earthquakes, volcanic eruptions, 1andslides
and submarine slumps, and explosions. Each of these sources has its own
generating mechanism, and the characteristics of the generated waves are
dependent of the generating mechanism. The tsunami waves which travel
long, transoceanic distances are normally generated by the tectonic
activity associated with shal1owfocus earthquakes. However, large waves
can be generated locally by the other generating mechanisms.
27
1. Submarine Earthquakes.
As shawn by !ida (1970), tsunamis are generated hy shallowfocus
earthquakes of a dipslip fault type; i.e., vertical EOtion upward on
one side of the fault and downward on the other side (Fig. 5). Shepard,
MacDonald, and Cox (1950) indicate that tsunamis which travel long dis
tances across the ocean are probaoly caused by unipolar disturbances.
(An example of a unipolar disturbance would be the uplift of a large
area of the sea floor where there is a net change in volume.) Waves
generated from a unipolar source decay much less rapidly with distance
than waves generated by a bipolar disturbance; i.e., a combination up
lifting and subsidence, or other apparent transfer of material on the
sea floor, without a net change in volume. Hammack and Segur (1974)
studied the propagation of waves bath experimentally and numerically.
They indicate that where there is a positive net change in volume (e.g.,
a unipolar uplifting of the sea floor), waves of stable form (solitons)
evolve, followed by a dispersive train of oscillatory waves. The number
and amplitude of the solitons depends on the initial generating mechanism.
The wave record for the 1964 tsunami at Wake Island (see Fig. 6) illus
trates this type of wave generation. Van Dorn (1965) discusses the
generating mechanism of the 1964 tsunami which originated in Alaska .
The ground motion was dipolar, having a positive pole (uplifting) under
the sea and a negative pole predominantly under the land. As the positive
pole was the main tsunamigenerating mechanism, this was equivalent to a
unipolar source.
Oip Slip Fouit
StrikeSiip Fouit
Figure 5. Movement along faultlines.
Heck (1936) indicates that horizontal motion of the sea floor does
not appear to generate large tsunamis. However, large "local" tsunamis
may be generated by horizontal motion. !ida (1970) shows that major
tsunamis (those that cause high water levels at many different coastal
locations) do not appear to occur as the result of deepfocus earthquakes
or the strikeslip fault type, i.e., horizontal motion along the fault
line (Fig. 5). A general expression for the lower limit of the earth
quake magnitude, M, of tsunamigenic earthquakes is given by !ida (1970)
as
M 6.3 + 0.005 Df (11)
28
Figure 6. Wave record from Wake Island, showing arrival
of tsunami (initial motion is positive and
remains above normal tide curve for more than
an hour) (from Van Dom, 1964).
based on tsunamigenic earthquakes in Japan, where Df is the focal depth
in kilometers and M the magnitude on the Richter scale. Tsunamis usu
ally do not occur for earthquake magnitudes less than that given by
equation (11), although a small number of tsunamis of lesser magnitude
have be.en associated wi th lesser magnitude earthquakes. It should be
noted that equation (11) does not consider the location of the earthquake
with respect to the coastline, the configuration of the coastline, and
possible local resonance effects. The Richter scale is given by
M = (logE  11.8)
1.5
where E is the earthquake energy in ergs.
Geller and Kanamori (1977) note that care must be taken when defining
earthquake magnitude. Richter (1958) gives higher values for earthquake
magnitudes than those listed by Gutenberg and Richter (1954). The differ
ence results from the relationships used to determine earthquake magni
tudes from surface wave magnitudes and body wave magnitudes.
Attempts have been made to define lower limits for earthquake magni
tudes associated with disastrous tsunamis. However, the definition of
"disastrous tsunamis" may be more a function of the location of the origin
and the population in. the adjacent coastal zone, rather than an analysis
of the actual waves generated. Also, the equations developed to define
these limits are based on limited data and do not fully consider coastal
configurations and resonant effects.
A tsunami generated from a dipslip fault source will have the
characteristics of being generated from aline source; i.e., the length
29
of the generating area is much greater than the width . When displace
ment occurs along a substantial length of faultline, the divergence of
the wave rays of the generated wave (i.e., the spreading of wave energy
along the wave crest) will be much less than for a wave generated from
a small source. For a "locally" generated wave, i.e., a wave generated
near the coastline under consideration, the main component of the wave
energy will travel perpendicular to the faultline and the energy per
unit length of wave crest would remain approximately constant for an
unrefracted wave.
2. Volcanic Activity.
Although most major tsunamis have been caused by shallowfocus earth
quakes, a small percentage have been caused by volcauic activity which
includes localized earthquakes, shoreline and submarine slumps, and
volcanic explosions. Examples of these are the volcanic activity of
April 1868 and November 1975 in Hawaii, with associated earthquakes off
the southeast coast of the island of Hawaii, and the August 1883 erup
tion and explosion of the island of Krakatoa near the Sunda Strait in
Indonesia. The explosion of Krakatoa destroyed an estimated 8 cubic
kilometers (1.92 cubic miles) of the island. Large shoreline subsidences
were associated with the eruptions and earthquakes on Hawaii.
Tsunamis with volcanic origins have the characteristics of waves
generated from a small source area. These waves spread geometrically
and do not cause large wave runup at locations distant from the source,
but may cause very large waves near the source. Also, there may be
refraction effects which trap waves along the coastline, or standing
edge waves may be generated along the coastline.
Both the 1868 and the 1975 tsunamis in Hawaii caused high waves at
points on all sides of the island of Hawaii as well as waves on the other
islands (PararasCarayannis, 1969; PararasCarayannis, International
Tsunami Information Center, persona! communication, 1975). The 1975
waves persisted for more than 4 hours at all points. Meyer (Department
of Mathematics, University of Wisconsin, persona! communication, 1975)
indicated that trapped waves may exist with many nodes around the island.
These trapped waves would gradually decay, leaking energy to the sur
rounding ocean.
3. Landslides and Submarine Slumps.
Landslides and submarine slumps can occur from various causes, but
are often associated with earthquakes. The waves generated by such events
will spread geometrically as they propagate from their source in an open
ocean, but can be very high near their origin. Waves can be particularly
high if they occur in a confined inlet, or if resonant or refraction
effects exist.
Examples of landslidegenerated waves have been reported by Miller
(1960) for Lituya Bay, Alaska, in 1853, 1874, 1936, and 1958. The 1958
30
wave reached an estimated maximum surge elevation of 530 meters (1,740
feet) on the opposite side of the bay, and generated a 61meterhigh
wave seaward in the bay. Waves were also generated by icefalls in
Y a k u t ~ t Bay, Alaska, in 1845 and 1905. Jorstad (1956), as referenced
by Wiegel (1964), reported on landslidegenerated waves in Tafjord,
Norway, in 1718, 1755, 1805, 1868, and 1934.
An example of a wave generated by a shoreline slump is given in
Berg, et al. (1970). A survey of the Valdez, Alaska, area after the
March 1964 earthquake showed that the water depth at the end of the
Valdez Dock had increased from 9 ta 37 meters (30 to 120 feet), destroy
ing the dock. Also, at a smallcraft harbor breakwater, the water depth
increased from 2.7 ta 27 meters (9 ta 90 feet), destroying the breakwater.
The owner of a fishing boat, heading toward the Valdez Narrows from the
open sea, reported a wave 10.7 ta 15 meters (35 ta 50 feet) high, in the
narrows, which dispersed after passing the narrows.
The first wave ta hit Valdez was generated by the s1ump of the
waterfront, and the second wave by the slump of a shoreline area sorne
distance away. After about 5 ta 6 hours, a third wave arrived, followed
more than 2 hours later by a fourth wave. These later waves apparently
resulted from sorne reflection or resonant effects within Prince William
Sound.
Ambraseys (1960) indicated that the tsunami of 9 July 1956 in the
Greek Archipelago was probably produced by a series of landslides on the
steep banks of a submarine trench. The wave had an amplitude of 30 meters
(100 feet) near its source. Striem and Miloh (1975) report that tsunamis
have probably been generated by slumping of the continental slope off the
coast of Israel. Van Dorn (1965) indicates that tsunamis generated from
this type of source appear to be fairly localized and will not be large
at long distances from the source. The generating mechanism is extremely
inefficient, and only about 2 percent of the potential energy of a falling
or sliding weight is converted into wave energy.
4. Explosions.
An explosion acts as an impulsivegenerating mechanism which generates
dispersive waves from a point source. Data from nuc1ear explosion Baker
at Bikini Atoll in 1946 show that the wave height is approximately in
versely proportional ta the radial distance from the point of origin;
i.e., Hr =constant where H is the height of the wave, and r is the
radial distance from the point source. At a radial distance equal ta
3Sd, where d is the water depth, the relationship changes slightly,
with the wave height decreasing less rapidly. Wilson (1963) discusses
data on wave dispersion.
The height of a wave generated by an explosion has been shawn ta be
dependent on the depth of the explosion charge. Van Dorn, Le Mehaute,
and Hwang (1968) show that two critical depths exist which will produce
the highest waves for any given explosive charge. The critical depths
are dependent on the charge yield, given in equivalent pounds of TNT.
Extensive material is avai1able on waves generated by explosions,
and will not be considered further here (see Smith, 1967).
31
III. MECHANICS OF GENERATION
The generation of large, transoceanic tsunamis results from the
displacement of water above the area of uplifted sea bottom associated
with a dipslip fault movement. Crustal displacement progresses along
a faultline from sorne initial source. BenMenahem (1961) developed a
method for determining the direction, speed, and length of rupture
propagating from the epicenter of a given earthquake by using recorded
seismic surface waves. Various analyses using this method, as reported
in Berg, et al. (1970) for the 1964 Alaskan tsunami give speeds from
3.0 to 3.5 kilometers (1.9 to 2.2 miles) per second for rupture propaga
tion and a rupture length of from 600 to 800 kilometers (370 to 500 miles) .
Because of the high speed of rupturing, it is generally assumed in analyz
ing wave generation that the total uplifting occurs instantaneously.
1. Area and Height of Uplifting.
Very little data are available on the size of the generating areas
and the height of uplifts for various tsunamis which have been recorded
at coastal points. After the 1964 tsunami generated in Alaska, extensive
surveys were undertaken in the area of origin (Plafker, 1965; Berg, et al.,
1970). These surveys included comparisons of tide levels at surviving
tide gages, establishment of previous tide levels by visual observation
and interviews with area residents, new hydrographie surveys in areas
previously surveyed, establishment of new elevations at hench marks, and
measurement of the displacement of sessile marine organisms. The uplifted
water area on the Continental Shelf was estimated as 1.1 x 10
1
1 square
meters (1.184 x 10
12
square feet). The potential energy of an incrementai
area of uplifting is proportional to h
2
, where h is the height of up
lifting. The average value of h
2
was estimated as 4.1 square meters
(44.1 square feet). The uplifted area in Prince William Sound was con
sidered to have a limited effect on the tsunami generation because of the
restricted connections between the sound and the shelf area.
An uplifting of the sea bottom will produce a vertical uplifting of
the overlying water. As a first approximation, it may be assumed that
the uplifting of the water surface equals the uplifting of the sea bottom.
The potential energy of the uplifted water is then given as
where
E
p
g
E
n h.
L pg A.h. ~
i=1 ~ ~ 2
the energy in ergs (footpounds)
(13)
the density of the seawater and is assumed to equal 1.0252
grams per cubic centimeter (1.989 slugs per cubic foot)
gravitational acceleration and is equal to 980.7 centimeters
(32.174 feet) per second squared
an incrementai area of uplifting
the height of uplifting over the incrementai area Ai
32
If the incrementa! areas are equal, i.e., A
1
equation (13) can be rewritten as
E
or, alternatively,
E
n h ~
pg n A. l: 7.
7. i=1 2n
Noting that the total area, A, is given by
A= n A.
1.
and that
h ~
h2
+ h ~
+ +
h2
(h2) avg
n
...J:..=
1 n
l:
i= 1
2n 2n 2
equation (15) becomes
E = pg
A (h2)aVfi.
2
Ayz, then
(14)
(15)
(16)
(17)
(18)
where (h2) is the average value of the square of the uplifted heights.
avg
For the 1964 Alaskan earthquake the height of uplifting varied con
siderably over the area of uplifting, and had a maximum in excess of 15
meters (49 feet) at a point near Montague Island (Malloy, 1964). The
tsunami had a calculated potential energy of 2.26 x 10
2
2 ergs (1.67 x 10
1
5
footpounds).
When using equation (18) it must be remembered that the average of
the height squared, (h
2
)avg. is not equal to the average height squared,
(havg)
2
. This is easily illustrated by the following example problem.
* * * * * * * * * * * * * EXAMPLE PROBLEM 1 * * * * * * * * * * * * * * *
GIVEN: An area of uplifting is divided into five equalsized areas of
~ x 101
2
square centimeters (2.48 x 109 square feet), with upliftings
of 30, 60, 90, 120, and 150 centimeters, respectively.
FIND:
(a) The value of
(havg)2'
(b) the value of
(h2)avg•
and
(c) the potential energy of the uplifted seawater.
33
SOLtrriON:
(a) h
30 + 60 + 90 + 120 + 150
90 centimeters
avg
5
(havg) 2
8,100 square centimeters (8. 72 square feet)
(b)
(h2) avg
30
2
+ 60
2
+
902
+ 120
2
+ 150
2
5
(h2) =
49
•
500
= 9,900 square centimeters (10.66 square feet)
avg 5
(c) From equation (16),
E
2
E (1.0252)(980.7)(5)(2.3 x 1012)
9
,9oo
2
E 5.72 x 10
19
gramsquare centimeters per second squared
5.72 x 10
1
9 ergs (4.22 x 1012 footpounds)
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
The horizontal motion along a rupture line may contribute very little
to tsunami generation. The maximum energy input from the horizontal
motion would occur when the rupture line is normal to the continental
slope. The motion along the rupture line, in that instance, would be
equivalent to a wedge moving a short distance through the water (see Fig.
7). Berg, et al. (1970) show that for a motion equivalent in magnitude
to that of the 1964 Alaskan earthquake, and acting normal to the continen
tal slope, the potential energy input to the resulting tsunami would have
been 3.12 x 1020 ergs (2.3 x 10
13
footpounds). This is less than 1.5
percent of the energy input from the vertical uplifting, and seems to
confirm Iida's (1970) analysis which showed that major tsunamis are asso
ciated with the dipslip fault type. In fact, the rupture line of the
1964 Alaskan earthquake was almost parallel to the continental slope, and
the energy input from the horizontal motion would have been negligible in
this case. In other cases, the contribution of horizontal motion may be
greater.
2. Initial Wave Formation.
Because of their long periods and corresponding long wavelengths,
the train of waves forrning a tsunami is taken to be shallowwater waves
at their origin, and propagates across the ocean as shallowwater waves.
The actual form of the wave train is determined by the initial generating
mechanism, i.e., the area of the uplifted sea bottom, the height of and
34
Equivalent displocement of woter surface
Figure 7. Horizontal motion normal to continental slope (scale exaggerated).
variation of the uplift within the area of uplift, and the depth of
water and coastal characteristics in the generating area. While ordi
nary sea waves are assumed to have a cnoidal shape as they approach a
shore (i.e., high crests and shallow troughs), the waves in a tsunami
may have various combinations of forms.
Visual observations of tsunamis have led to reports that the initial
wave was often a negative wave causing an initial drawdown of the water
level at the shoreline. Shepard, MacDonald, and Cox (1950) show that
s?me tide gage records indicate a small positive wave followed by a very
deep trough, the amplitude of the trough being about three times the
amplitude of the initial wave crest. This may have been misinterpreted
by observers who reported initial negative waves. However, Striem and
Miloh (1975) indicate that an initial drawdown may occur when the tsunami
is generated by slumping of the continental slope. Tsunamis may sometimes
produce waves with narrow, deep troughs and low, wide crests at the shore
'.· line, the opposite of the cnoidal waveform.
Wave records from Wake Island for the March 1964 tsunami (Van Dorn,
1964) show a positive surge with a period of 80 minutes (see Fig. 6).
There was a series of positive wave crests with the elevations of the
intervening troughs above the normal expected tide level. This was
followed by a series of crests and troughs with the elevations of the
troughs below the normal tide level. Using a shallowwater wave celerity
at the source and an average depth of approximately 100 meters (325 feet)
for the generating area, the period of the initial positive surge is
approximately equivalent to the time ~ e q u i r e d for the trailÎng edge of
the initial uplifted water surface to travel completely across the area
35
of generation. This indicates that the uplifted water surface at the
source formed a series of solitarytype waves. The multiple crest can
be accounted for by initial instabilities in the waveform caused by the
generating mechanism, and the effect of the varying bathymetry of the
ocean basin through which the wave passes. The lower waves following
the initial series of wave crests correspond to the expected oscillations
from a disturbance in the water surface as the disturbance is damped out.
Wilson, Webb, and Hendrickson (1962) showed that the height of a
tsunami at a coastal point near the source of generation could be given
as a first approximation by the empirical equation
log
10
H = 0.75 M 5.07
(19)
where H is the height in meters and M the Richter magnitude. Using
the value of M = 8.3 given by Berg, et al. (1970) for the March 1964
tsunami, H = 14.29 meters (46.9 feet). However, this is an empirical
relationship which does not completely account for the characteristics
of the generating mechanism or the coastline. Although equation (19)
might provide a first ruleofthumb estimate of wave heights, the actual
heights could be above or below that estimate. Determination of actual
heights would require computation by numerical or empirical means.
Wilson (1969) gives the relationship of Housner (1969) for the fault
length Lf in kilometers as
log
10
Lf = 0.87 M 4.44 (20)
g1v1ng a fault length for the March 1964 tsunami of Lf = 604 kilometers
(1.98 x 106 feet). This is within the range of estimates given in Berg,
et al. (1970) and approximates the length of the generating area, i.e.,
the length along the initial wave crest.
Wilson and T0rum (1968) give a relationship for the period T (in
minutes) of the primary tsunami (carrying maximum energy) as
0.625 M  3.31
(21)
1
For the March 1964 tsunami, this equation gives a period of T = 75.4
minutes (using the Richter magnitude M = 8.3). This is very close to
the period of the positive surge noted by Van Dorn (1964) at Wake Island,
and is equal to that period if the crest of the initial oscillatory wave
at the trailing edge of the surge is neglected.
The initial deformation of the water surface, for any tsunami, will
collapse into some system of waves which must be defined. The resulting
wave system depends on the shape of the seabed deformation and the water
depth above the deformation. The simplest means of analysis is to assume
36
that the water surface has an initial displacement equal to the seabed
displacement, that the initial displacement is not timedependent, and
then propagate the initial displacement outward from the generating area
using longwave equations (Brandsma, Divoky, and Hwang, 1975). Other
means of establishing the initial waves, of varying degrees of complexity,
are described by Wilson, Webb, and Hendrickson (1962) and in other sources.
Many of the mathematical representations of waves generated from
bottom uplifting are based on circular source regions; however, Levy and
Keller (1961) present one solution in terms of elliptic coordinates for
a source region which is more elongated than circular. This solution has
the form
where
I
and
n(r, 8, t)
I (k, 8)
[
2rrC ll/2
r1 Ao(k) /k(rct)+i(n/4) I(k, S)
k
1
1
2
cosh[k(Z + d)]
o ei'ITI4
(21r)
3
/
2
cosh(kd)
L:
n=o
+ So  cos
(
ka
n 2 ,
8) F (ka),l
n,no
2
Ihe terms Sen San are even and odd Mathieu functions, and
(22)
(23)
(24)
(ka/2) and (ka/2) are even and odd Mathieu transforms. The
variables are defined as n the wave height, r and 8 the coordinates
of a point in polar coordinates, t the time, C the wave celerity, k
the wave number, Ca the group velocity, Z
0
the depth of generation
lnegative downward and equal to d for bottom uplifting), d the water
depth, and a the interfocal distance of the coordinate ellipses.
Levy and Keller indicate that the veloci ty of the bot tom uplifting is
unimportant if the time of uplifting is small in comparison with the period
of the generated waves. This is generally true for tsunamis. They also
indicate that only the first few terms of I(k,8) may be important in the
solution, although a computer solution can sum a relatively large number
of terms. The limitations on the solution are that the solution was
derived for water of uniform depth, the initial wavelength (or wave period)
must be known, and the solution assumes that r is much greater than the
dimension (diameter or length) of the source.
37
Hwang and Divoky (1970, 1972) use a simplified monotonie displace
ment history to describe ground motion. They propose that, to a first
approximation, horizontal displacement of a sloping bottom can be repre
sented as purely vertical displacement.
Houston, et al. (1975b) use an ellipticalshaped generating area,
with an instantaneously displaced water surface, as input data for a
standard design tsunami in a numerical solution. They define the sur
face displacement as a modified elliptic paraboloid, having a parabolic
cross section parallel to the major axis of the ellipse, and a triangular
cross section parallel to the minor axis of the ellipse. The numerical
propagation of the wave uses the same procedure as used in Brandsma,
Divoky, and Hwang (1975). The potential energy of the uplifted water
surface for this type of surface displacement is given by
E
(
g)
b c
2
a 5/2
4 E
6
 J (a
2
 x
2
) dx
a a
4
o
(25)
where
x measured along the major axis of the ellipse
a the length of the semimajor axis
y measured along the minor axis of the ellipse
b the length of the semiminor axis
z the vertical direction upward from the undisturbed water surface
c the maximum uplifted elevation at coordinates (x = o, y= o,
z = c)
p the density of the seawater (taken as 1.026 grams per cubic
centime ter)
IV. TSUNAMI PROPAGATION
After determining the initial disturbance of the water surface, as
discussed in Section III, the propagation of the tsunami to nearby or
distant shorelines must be analyzed. Because tsunamis are longperiod
waves with long wavelengths in relation to both the water depth and the
wave height, longwave equations can be used.
1. SmallAmplitude Waves.
The simplest means of analyzing the wave motion, where the ratio of
the wave height to water depth, H/d, is small, is to use the following
smallamplitude solutions to the wave equations:
38
where
c2 tanh
u =
h [27T(Z + d)]
2 7T % _c_o_s . cos [ 27T ( f  f) J
Sln 
L
h [27T(Z + d)J
cos L
a __ ....;:;;... ____ =. sin
. h [27Td]
Sln 
L
C the wave celerity
L the wavelength
d the depth of the undisturbed water
(26)
(27)
(28)
u the horizontal velocity of a water particle in the direction
of the wave motion
a = the amplitude of the wave above the undisturbed water level
T the wave period
z the vertical distance of the particle from the undisturbed
water surface
x distance measured in the direction of wave motion
t time
the horizontal displacement of the water particle from its
undisturbed position
Tsunamis are shallowwater waves; i.e., the ratio of wavelength to
water depth, L/d, is very large. Therefore, equations (26), (27), and
(28) can be reduced to more basic smallamplitude, shallowwater equations .
Letting d/L + 0,
(29)
39
substituting equation (29) into equation (26) gives
c
2
= gL (2Tid)
2TI L
which gives smallamplitude, shallowwater wave celerity as
c = lgd
Because z < d, it will always be true that z/L+ 0 whenever d/L + O.
Therefore,
sinh
2
TIZ
L L
sinh 2Tid + 2Tid
L L
cash
2
TIZ +
L
cash
2
Tid +
L
Considering that
(30)
(31)
(32)
(33)
(34)
(35)
cash
2
TI(z + d) = cash
2
TIZ cash
2
Tid + sinh sinh
2
Tid (36)
L L L L L
and substituting equations (32) ta (35) into equation (36),
cash 2TI(z
1
+ d) = 1 +
(37)
The disturbed water surface elevation, n, at any point in relation ta
its undisturbed location is given by
n = a cos [ 2 TI (  J
Substituting equations (33), (37), and (38) into equation (27),
Noti ng that z/L+ 0,
1 +
U = 2 TI !}_
T
u=.!l.!!
d T
40
(38)
(39)
(40)
but, from the basic wave equation for all waves ,
so that
c = l:.
T
n gl/2
u=.!l/gd=
d dl/2
The water surface elevation can also be defined as
n = a
sin
[2TI (f
( 41)
(42)
( 43)
so that by substituting
(28)'
equations (33), (37), and (43) into equation
1 +
t;, n
( 44)
and noting again that z/L + 0,
t;,
nL
 2Tid
(45)
The wave energy, E, for a srnall arnp li tude wave is given by
E
pg H
2
L
8
(46)
where p is the density of the seawater. If energy is conserved between
two points for an unrefracted wave,
but as L = CT, and T is assurned to be constant, and as C
shallowwater wave, then
41
( 4 7)
( 48)
/gd for a
( 49)
which is the wellknown Green's Law. Noting that nmax
equations (42) and (45) can be written as
For the
Also,
agl/2
dl/2
aL
27Td
unrefracted wave, noting that
a2 H2
al Hl
jumaxl2
= (::)
( ~ 2 y/2
= GJ'' G: r
\umaxll
( ~ 1 y/2
lumaxl2
(::r
lumax\1
a, the ampl i tude ,
(50)
(51)
(52)
(53)
(54)
(55)
(56)
Equations (49), (54), and (56) provide a simple, firstorder solution
for the shoaling of an unrefracted, smallamplitude, shallowwater wave .
* * * * * * * * * * * * * EXAMPLE PROBLEM 2 * * * * * * * * * * * * * *
GIVEN: A long wave with a period of 20 minutes and a height, H, of
0.4 meter (1.31 feet) passes from a 1,000meter (3,280 feet) water
depth into a 500meter (1,640 feet) water depth. The wave is assumed
to be nondispersive.
42
FIND:
(a) The unrefracted wave height in the 500meter depth,
(b) the water particle velocity IUmaxl in each water depth, and
(c) the horizontal water particle motion l ~ m a x l in each water depth.
SOLUTION:
(a) H
1
= 0.4 meter, d
1
From equation (49),
1,000 meters, d
2
500 meters
Hsoo = (I,ooo)l/
4
0.4 500
1.189
H
500
= 0.4(1.189) 0.48 meter (1.56 feet)
(b) From equation (50),
Assuming a= H
1
/2 = 0.2 meter, at d
1
= 1,000 meters,
0.2(9.807)1/
2
(1 ,000) l/
2
From equation (54) where d
2
lumax lsoo
lumaxll,OOO
(
1 ,ooo)
3
/
4
500
(
1 000)
3
/
4
0.02 '
500
0.02 meter (0.065 foot) per second
500 meters,
0.034 meter (0.11 foot) per second
43
(c) Firs t s olving for L where d
1
= 1,000 meters and T
L CT = /gdT = /9.807 x 1,000 (20 x 60)
L 118,800 meters (73.8 miles)
From equation (51),
aL
2'ITd
0.2(118,800)
27r(l,OOO)
3.78 meters (12.4 feet)
From equation (56) where d
2
= 500 meters,
)
2
=
( ç;max \ d2
lç;max lso o
1 ç;max Il , o o o
(
1 ,000)
3
/
4
500
20 minutes,
l ç; 1 = 3.78
•
= 6.36 meters (20.9 feet)
1 max soo 500 }
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Soloviev, et al. (1976) compared solutions for tsunami amplitude
using equation (49) and a numerical integration method. Equation (49)
does not account for wave reflection from bottom slopes and results in
calculated wave amplitudes that are tao high. Also, equations (49),
(54), and (56) do not account for wave refraction, diffraction, or
dispersion; they cannat be used with any degree of accuracy when the
ratio of H/d becomes large. When waves travel long distances, it is
necessary to consider the curvature of the Earth, discussed later in
this section.
2. LongWave Equations.
To increase the accuracy of computations, the longwave equations
can be expressed in various forms of partial differentiai equations
which can be solved numerically. As given by Peregrine (1970) for two
dimensional waves propagating in water of constant depth, the equations
may be writt en as fol l ows in rectangular coordinates:
44
Linear equations:
élu
+
at
an
+
Clt
an
0
g ax ::
d au = o
Cl x
Finiteamplitude equations:
dU dU
 + u +
at ax
+ a[(d + n) u]
0
at ax
Boussinesq equations:
au au
+ u +
at ax
an .!. d2 a
2
u
g= t
ax 3 ax2
Cl [(d+ n) u] =
0
at ax
(57)
(58)
(59)
(60)
(61)
(62)
In addition, for waves traveling in only one direction, the Boussinesq
equations may be reduced to the KortewegdeVries equations which are
then wri tt en as
2
au+
2
(gd)l/2 au+
3
u au+ .!.d2
0
at ax ax 3 ax3
 (d)l/2
n  
g
u +
d3 a2 u
u
2
ax
2
4g  6 (gd) l/2
When considering the means of describing the propagation of long
period waves, the parameter, U, should be evaluated where U is
defined as
(63)
(64)
(65)
and sometimes referred to as the Stokes or Ursell parameter. The
importance of this parameter was first noted by Stokes (1847) when he
stated that the parameter must be small if his equations were to remain
valid for long waves.
45
Murty (1977) indicates that the value of (L/d)
2
is a measure of
frequency dispersion; the value of H/d is a measure of amplitude
dispersion. Murty points out that the linear longwave equations are
valid when U << 1. In this case H/d is small and amplitude dispersion
can be ignored.
When U is of the order one [U = 0(1)], both amplitude and frequency
dispersion are important. In this case Boussinesq or KortewegdeVries
equations should be used. Where U >> 1 amplitude dispersion dominates
the solution, and the finiteamplitude, nonlinear longwave equations
should be used. It should be emphasized that when U = 0(1) it is not
necessary that U ~ 1. Zabusky and Galvin (1971) show that the Korteweg
deVries equation accurately describes wave propagation for U < 800 in
sorne cases.
For tsunamis with very long periods (and therefore long wavelengths),
the condition that U << 1 is usually never satisfied. However, the error
which results from the use of the linear equations is quite small as long
as the value of H/d is small. The acceptable limit of the value for
H/d (i.e., the point where the error in the calculations becomes signif
icant) depends in part on the rate of shoaling of the wave, i.e., the
shoreward slope of the bottom topography.
* * * * * * * * * * * * * * EXAMPLE PROBLEM 3 * * * * * * * * * * * * * *
GIVEN: A tsunami has a period of 20 minutes and a wave height of 0.05
meter (0.16 foot) in a 1,000meter (3,280 feet) water depth.
FIND: The parameter U.
SOLUTION: From equation (31),
C /gd= 19.807 x 1,000 = 99 meters (325 feet) per second
L CT 99 x 20 x 60 = 118,800 meters
From equation (65),
u = o. 706
0.05 (118,800)
2
1,000 1,000
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
In more recent investigations, the parameter U given by equation
(65) has been redefined as U* where
46
U* ( ~ ) (n ~ 2
x
(66)
max
and
(67)
This is discussed by Peregrine (1970). He points out that nonlinear
terrns which were neglected in the linear equations (57) and (58) cause
a cumulative error that may become appreciable in a numerical solution
after a time given by
t
(68)
Where rapid shoaling occurs, i.e., where a wave passes over a large
change in water depth in a relatively short period of time, the accumu
lated error will be much smaller than for slow shoaling, where the wave
passes over the same change in water depth in a relatively long period
of time.
The finiteamplitude equations (59) and (60) are valid as long as
U* > 1
but generally become invalid after a finite time as the front face of
the wave steepens. The Boussinesq equations are also applicable where
U* > 1; i.e., where ( n x ) ~ a x < (H/d). Peregrine (1970) points out that
the Boussinesq approximation works well for values of H/d up to about
O.S. The Boussinesq or the KortewegdeVries equations are used for
waves approaching a shoreline where values of H/d become large.
Hammack (1973) gives the value of U* as
(69)
to describe a particular region of a complex waveform. However, the
value of U* would be expected to vary from region to region of the
waveforrn in this case. This variation would indicate that using a single
set of equations to describe a compZex waveform may lead to incorrect
results.
3. Distantly Generated Tsunamis.
When a tsunami travels a long distance across the ocean, the sphe
ricity of the Earth must be considered to determine the effects of the
47
tsunami on a distant shoreline. Waves which diverge near their source
will converge again at a point on the opposite side of the ocean. An
example of this was the 1960 tsunami whose source was on the Chilean
coastline, 39.5° S., 74.5° W. (PararasCarayannis, 1969). The coast of
Japan lies between 30° and 45° N. and about 135° to 140° E., a difference
of 145° to 150° longitude from the source area. As a result of the con
vergence of unrefraated wave rays, the coast of Japan suffered substantial
damage and many deaths occurred (U.S. Army Engineer District, Honolulu,
1960; Hirono, 1961). Figure 8 illustrates the convergence of the wave
rays due to the Earth's sphericity.
s o o ~ ~ ~ ~ ~  n ~   ~ ~ ~ ~ .  ~ ~   ~ ~ C ~ A ~ ~ ~ ~ ~ A ~   
45°
30°
15°
oo
15°
30°
45°
p
0
Figure 8. Convergence of wave rays.
Chao (1970) gives the equation for wave refraation in spherical
coordinates as
dx = _ 1_ (de) _ cos a. tan 1/J
ds C dw Re
where ray separation is defined by the equation
d
2
S + P dS + qS =
0
ds2 ds
and the coefficients p and q are defined as
p =  ..!. (de)  ..!._ sin a. tan 1/J
C ds R
e
48
(70)
(71)
(72)
q
1 dC
C dw
(
sin a )
2
(1 a2 c) ( sin 2a ) (1 a2 c)
R cos 1)J c ~  R2 cos 1)J c a1)Ja<t>
e e
+ ( ~ ) 2 (! ~ ) + (2 cos a tan lji) (! dC)
R c a1)J2 R c dw
e e
     + (cos a)  
1 [ (sin a)(l ac) ( 1 ClC ) ~
Re cos 1)J c a<t> c a1)J
(73)
(74)
(75)
where
R the radius of the Earth
e
<1> longitude of the point on the surface the wave ray is passing
through
1jJ latitude of the point
C celerity of the wave
s a measure of distance in the direction the wave is traveling
a the angle between the wave ray and a line of equal latitude
S a ray separation term
Using the spherical coordinate system shawn in Figure 9, Hwang and
Divoky (1975) give the linear longwave equations as
au =

g an
f +
v
at R ae
(J
(76)
e
and
av =
g
2..!1__
f
 u
at R sin 8
a<t>
(J
e
(77)
49
where
Re the radius of the Earth
e degrees latitude (measured from the pole)
degrees longitude (measured eastward)
u the velocity in the adirection
v the velocity in the
t time
fe the coriolis parameter = 2 Q cos e
Q the rotational speed of the Earth in radians per second
Figure 9. Spherical coordinate system.
For linear longwave equations, the acceleration component in the radial
direction is considered to be negligible. The continuity equation is
È!l =  1
at R sin
e
{
L [(d + n) u sin e] + L [(d + n) vl} +
e ae ap at
(78)
where ç is a timedependent vertical bottom displacement in the generat
ing area and equal to zero elsewhere.
For instantaneous bottom displacement, the initial wave is assumed to
be a water surface displacement equivalent to the bottom displacement,
and aç/at =O. If a wave is generated near a coastline, i.e., in rela
tively shallow water, the parameter U may initially be quite large.
For a wave propagating from the coastline, it is assumed that the
50
linear longwave equations can be applied to the initial propagation,
and that the resulting errors are of a size that can be accepted in the
calculations.
4. Nearshore Propagation.
The linear longwave equations may be used for the propagation of
waves from a shoreline, across an ocean basin, and up to an area near
another shoreline. It is also necessary to consider the propagation of
a tsunami toward a shoreline from a nearby generating area, or into the
nearshore area at a distant shoreline where the linearized longwave
equations will not provide solutions with sufficient accuracy. Peregrine
(1967) derived equations for threedimensional long waves in water of
varying depth (i.e., shoaling waves) which correspond to the Boussinesq
equations for solitary waves in water of constant depth. An expansion
is used similar to that used by Keller (1948).
The dimensional variables are defined with an *, and the dimension
less variables by the following equations:
x* Y* z*
x :::
' y
z :::
d d d
0 0 0
t t*
(!oY/
2
P*
' p
d
pg 0
u* v* w*
u :::
(gd ) l/2
' v ' w
(gd ) l/2 (gd ) l/2
0 0 0
where d
0
is a length representative of water depth; p the pressure;
the velocity u in the xdirection, v in the ydirection, w in the
zdirection; and the other variables are defined as before. Defining
where q is
where
Q,x
is
component of
q = (u
2
+ v
2
)
1
/
2
and Q =in qdz
d
velocity and Q the flow rate, the continui ty
ClQX ClQy
~ : : :
0  + +
Cl x a y at
the component of Q in the xdirection, and
Q in the ydirection.
Euler' s equations of mo ti on are
aq aq aq aq ap
ap =
0 + u + v+ w + +
at a x a y az a x a y
and
aw aw aw aw
lE.+
1 :::
0 + u + v+ w +
at a x a y az az
51
(79)
equation is
(80)
~
the
(81)
(82)
At the boundary z d,
ax ay
The variables n, p, q, and Q are expanded in the form
where E is the ratio of wave amplitude to depth H/d. The variable
w is expanded as
(83)
(84)
(85)
W = (J ( W O + EW l + E
2
W z + . . . ) ( 86)
where a is the ratio of water depth to wavelength d/L. The zero
arder solution gives
p z
0
The firstorder solution gives
aq
1
an
1
++
at
1
ax
1
an
1
a(du
1
) a ( dv
1
)
+ + = 0
at
1
ax
1
ayl
él(du
1
)
d(dvl) z cul +
wl

ax
1
ay
1
ax
1
The secondorder solution gives
a [a(du1 ) a(dv1) J
(xl, tl) z   +
ax
1
ax
1
ay
1
dv1)
ayl
z _a_ [a(du1) + a(dv
1
) J _ .!.. z
2
a [au
1
ëlv
1
)
ay
1
ax
1
ay
1
2 + ay
1
52
(87)
(88)
(89)
(90)
where n
2
(x
1
, t
1
) is an arbitrary function arising from integration. The
momentum equation is
an
2
aq
1
aq
1
an
2
+u +v++
at 1 ax
1
1 ay
1
ax
1
(91)
and the continuity equation is
an
2
a(Q
2
) x a(Q
2
) y
 + + = 0
at
1
ax
1
ay
1
(92)
Peregrine (1967) points out that secondorder terms will have first
order effects where t
1
. is not of sma11 value. He accounts for these
effects by incorporating secondorder terms into the firstorder variables.
Mei and Le Mehaute (1966) derived a solution for waves propagating
in one direction which gives the equations as
and
where
and
i:l + u i:l + (d + n) au
at ax ax
B
(93)
(94)
(95)
(96)
It can be seen that equation (94) has mixed derivatives, with respect to
x and t, where equation (63) has a thirdorder derivative with respect
to x only. Benjamin, Bona, and Mahony (1972) show that the Korteweg
deVries equations with mixed derivatives, such as equation (94), are
the preferred form for describing the behavior of long waves.
Street, Chan, and Fromm (1970) expanded on the work of Peregrine, and
for waves propagating in one direction give
au + u au + i:J. + ..!. d2 ~ + d ~ a
2
u + 1 a
2
d au (
97
)
at ax ax 3 atax2 ax atax 2 ax2 at
and
~ ~ + ~ x [ ( d + n) u]
0
(98)
53
The numerical solution of these equations gives results comparable to
experiments for varying bottom topography where H/d s 0.4 .
Using the work of Mei and Le Mehaute (1966), Madsen and Mei (1969)
developed characteristic equations which could be solved numerically.
Defining
dn = w and dw = a
dt dt
along the coinciding characteristics x
characte"ristics
constant, along the two distinct
(99)
and
(
Cd + ud) dn + ~ d2 ~ dw (
1
_ .!. d a
2
d) crdc
2
du
6 2 dS 12 ax dS + 2 ax2 6 dS
(
a ~ c2 )
+ crCd
2
da = Cd
0
+ ax
12 dS 2 6 u u
( 100)
where
(101)
and
C = ±C (x) along S = ( ~ ) = (t + J ~ ) = constant
o S
2
2 C
0
(x)
(102)
The KortewegdeVries equations provide a solution for wave propagation
in one direction only, i.e., for an unrefracted wave. The solutions gen
erated could be used to provide shoaling coefficients to obtain refracted
wave heights.
Alternative methods of obtaining solutions for refracted waves in two
dimensions are to use the linear longwave equations with additional terms
added to account for nonlinear effects, or to use solutions based on the
Boussinesq equations. Butler and Durham (1976) suggest a solution using
equations similar to those in a tidal hydraulic madel. The momentum
equations for the tidal madel are
54
au au au an
f v
TEx
0 + u + v+
dX 
+  =
()t ()t ()t
c d + n
(103)
av av
g h +
TB
u + v+ f u + ___!!jj_ =
0
()t Cl x Cl y Cl y c d + n
(104)
and the continuity equation is
h + L (du + nu) + L (dv + nv)
Clt Clx Cly
0 (lOS)
where the bottom stress is given by
T =
gu /u2
+ v2
Bx
c2
and
(106)
TBy
gv /u2
c2
+ v2 ( 107)
Chen, Divoky, and Hwang (1975) give numerical equations for the two
dimensional case based on the Boussinesq equations (see Sec. IV, 5).
S. Computer Models.
Solutions of the equations for long water waves are obtained by
numerical means. Leendertse (1967) gives the following method for solv
ing the linearized longwave equations by using a space staggered scheme
as shown in Figure 10. Taking the subscript n to indicate the value
at time t, the subscript n + l/2 to indicate the value at time t +
and the subscript n + l to indicate the value at time t + the compu
tations use alternate sets of equations at alternate time steps as
shown below. First u and n are calculated implicitly and v explic
itly at time t + then v and n implicitly and u explicitly at
time t + then u and n implicitly and v explicitly at time
t + etc. Calculating u at point (j + 1/2, k), n at (j,k),
and v at (j,k + 1/2), as defined in Figure 10, the calculations are,
at times t + t + t + ...
(108)
1 t { Cl [ y x x y }
nn  2 ax (d + n ) * u]n+l/2 + ay [ (d + ii ) v]n
(109)
v (an)
n 2 c un+l/2  2 t;S g Cly
n
(llO)
55
1 1
1
1
""!'
1

16
1
1
1
r

1Q 1o 1
y
l l
1 'T
1 1
L
I.
J1
x
j+l
+ Woter Level ( 1J}
o Depth ( d}
U Velocity ( u}
1 V Velocity (v}
Figure 10. Coordinate system,
rectangular coordinates.
and at times t + t + t + ...
..!. f u
vn+l/2  2 c n+l/2
1 (an)
2 g ay
n+l
..!. t f  1 t (an)
un+l/2 + 2 c vn+l  2 g ax n+l/2
(111)
(113)
These equations omit convective inertia terms, bottom effects, and any
forcing functions. The various terms used are defined as fo11ows:
x
n
56
(114)
(115)
(116)
(117)
(118)
= 1
v= 4 (vj,k1 / 2 + vj +l,k1/ 2 + vj , k+l/2 + vj+l , k+l/2)
{119)
The additional terms
and
require special computational procedures for n. These are described by
Leendertse (1967).
Hwang and Divoky (1975) show the same equations in spherical coordi
nates with u the velocity in the adirection, and v the velocity in
the as defined in Figure 9. They use a different approach
for the computation of the terms noted above which allows more direct
computation. Similar equations are shawn by Houston, et al. (1975b).
For the coordinate system in Figure 11, the equations given by Hwang and
Divoky (with coriolis terms added) are, at times t + ôt/2, t + 3ôt/2,
t + Sôt/2, . . .
u
n
g ôt (an) 1
+Iôtfv
e n+l/2 a n
n ôt {
1
a [ (d) u sin e]
n  2 Resin e Mas + n n+l/2
v
n
1 a e }
v]n
with terms such as an/ae, u, and v computed as before.
At times t + ôt, t + 2ôt, t + 3ôt,
v _ g ôt ( an) _!._ ôt f 
n+l/2 2 R sin a·
1
 2 a un+l/2
e n+
57
(120)
(121)
(122)
1 1
j1 XtiiXtX
1
l j x1l XtX·
8
j+l !11x+x
k1
0
Water Depth (dl
x Water Level {1))
u Velocity in the 9direction
11 Velocity in the
Figure 11. Coordinate system,
spherical coordinates.
 lit { l_ ( (Cf+n) <P
nn+l/2 2 Resin 8 a8
u sin 8]n+l/
2
1 d 8 )
+ [(d + n) v]n+l
(123)
u
n+l
g M (an) 1
un+l/2  2 ReM as n+l/2 + 2 fe vn+l
(124)
Again, the bottomfriction, forcing functions, and convective inertia
terms are ignored. Terms like are computed by averaging in
the 8direction, and by averaging in the cpdirection.
In addition to specifying equations of motion, boundary conditions
for the computational area must be established. Hwang and Divoky (1975)
use solid boundaries at coastlines and fictitious open boundaries at
edges of the computational area where it is necessary to truncate the
region of computation. At solid boundaries, complete reflection is
assumed. At open boundaries, the wave is assumed to travel without
change in form across the final space step, so that
nB  n2
nl  n2
with the terms defined in Figure 12.
58
(125)
8
(n + 1) ât
1 2
nât
s
1âS
BOUNDARY
Figure 12. Graphical representation of the total
transmission open boundary condition
(from Hwang and Divoky, 1975).
Houston and Garcia (1974) and Hwang and Divoky (1975) used numerical
techniques to obtain predicted wave heights for the 1964 tsunami originat
ing in Prince William Sound, Alaska. The uplifting and subsidence deter
mined from field surveys (Plafker, 1965; Berg, et al., 1970) was used as
the initial deformation of the water surface. Predicted wave heights
13,000 seconds (3.61 hours) after the time of generation are shawn in
Figure 13.
The assumptions that complete reflection occurs at a solid boundary
(i.e., at a shoreline) and that equation (125) will describe an open
boundary introduce errors into the computations which limit the length
of realtime records which can be simulated numerically. At a shoreline,
sorne amount of wave energy may be trapped so that complete reflection does
not occur. Wave trapping is discussed later in this report.
59
Figure 13. Surface elevation contours 13,000 seconds after the 1964
Alaska earthquake (from Hwang and Divoky, 1975).
The use of equation (125) at the open boundaries results in an error
in the computed wave heights at those boundaries. As computations are
carried out for the entire computational grid at each time step, the
error propagates through the grid at successive time steps. This error
will have the appearance of a wave reflected from the open boundary.
Shaw (1974, 1975; Shaw, Department of Engineering Science, State Univer
sity of New York, Buffalo, personal communication, 1977) suggests that
an outer boundary integral equation method can be used to eliminate the
error at the open boundaries. The outer boundary integral equation would
be used to determine the wave height, n, and its normal derivatives at
the boundary; those values would be used in the finitedifference solu
tions for the interior region.
6. Nearshore Computer Models.
For waves in the nearshore region, Peregrine (1967) developed finite
difference equations for the twodimensional case of a wave shoaling on
a beach. Defining
u
r,s
u(rl'.x, sl'.t) (r
nr,s n(rl'.x, sl'.t)
l, ... ,n s = l, ... ,n)
60
he calculated the values of u and n with a timestepping procedure,
first using an approximation to the continuity equation which gives
n*
1
, a provisional value of n
1
, by the equation
r
3
8+ r
3
8+
+ u
r
3
8
u  u
r+l ,8 r1 ,8
2t:.x
[
n  n
r+ 1,8 r1 ,8
2t:.x
+ a] = 0 (126)
Then, ur
38
+
1
is calculated from an approximation to the momentum equation
u
r+1 ,8+1
 u + u
r1 ,8+1 r+I ,8
4t:.x
 u
r1,8
u  2u + u  u + 2u  u
1
 (ax)
2
r+1,8+l r,8+l r1,8+1 r+1,8 r,8 r1,8
t:.x
2
t:.t 3
+ a2x ur+1,8+1  ur1,8+1  ur+1,8 + ur1,8
t:.xt:.t
(127)
Finally, the continuity equation is used again to give an improved value
for nr38+ 1
u u +u u
r+1,8+1 r1,8+1 r+1,8 r1,8
b. x
+ ur,8+1 + ur,8 [nr+1,8 nr1,8 +a]
2 2t:.x
0 (128)
Street, Chan, and Fromm (1970) and Chan, Street, and Fromm (1970)
extended Peregrine's work, using a MarkerandCell numerical technique
to obtain solutions for waves propagating in one direction. Where values
are known at time t, they compute the values of u and w at time
t + t:.t using the equations
u  u* t:. t:.t ( )
j+l/2,k j+ll2,k + gx t + t:.x Pj,k Pj+1,k
(129)
(130)
where the coordinate system is shown in Figure 14, p is pressure, gx
61
1
1
1
1
 T
1
1
1
'
Il X
j= 1 j=2 j =3 j=4 j=5
Figure 14. Position of variables.
and gz the components of gravitational acceleration, and u* and w*
convective terms. The terms on the right side of equations (129) and
(130) are taken at time t. Then, the pressures are recomputed from
the equation
P.
= !(Pj+l,k + Pjl,k + Pj,k+l + Pj,k1 )
k
+ RJ.,k
J, B
where
B
2 (1 + _1)
and
(
u*  u* w*  w*
R. = _ .!__ j+112,k j11z,k + j,k+l!z j,k112)
J,k
The convective terms are defined as shawn by Fromm (1968).
equation for
1
k is
J+l 2)
u* _ S  1 ( )
j+l/2,k  uj+l/2,k1 + 2 uj+l/Z,k2  uj+lf2,k
62
(131)
(132)
(133)
The
(134)
where
CL  1
uj+l/2,k uj1/2,k + 2 (uj3/2,k  uj+I/2,k)
(a.  1) 2
+ 2 (uj3/2,k
2
ujl/2,k + uj+l/2,k)
a. = uj+l/2 ,k
13 = w
j+l/2,k
Similar expressions can be developed for other convective terms.
(135)
(136)
(137)
For points near the free surface (Fig. 15), Chan and Street (1970a)
give the equation for Pj,k as
R. kl
J, :1
J
(138)
The freesurface position, as given by Chan, Street, and Fromm (1970)
at time t + is
nj =w. u. (nj+I njI) + nj(t)
J J 2 D.x tJ.t
(139)
where nj(t) is the elevation at time t, and Uj and Wj the horizontal
and vertical velocities, respectively, at the surface at time t + D.t.
For refracting waves propagating in two dimensions in the plane of
the water surface, Chen, Divoky, and Hwang (1975) give the equations below
using dimensionless expansions similar to those proposed by Peregrine
(1967). A timestaggered scheme is used, with the velocities and wave
amplitudes calculated explicitly at alternate time steps of tJ.t/2. The
amplitudes at ta + will be calculated using amplitude at ta 
and velocities at ta; then, the velocities at ta + will be calculated
using velocities at ta and amplitudes at ta + At time ta +
the amplitude is
(140)
63
(j, K+l)
(j, K1)
Figure 15. Computation of pressure near
the free surface.
where the j,k subscripts refer to positions in the plane of the still
water surface as shown in Figure 10, and ü and v the velocities
satisfying the linear longwave equations. Where the initial velocity
field is known, u and v can be computed at time ta + using values
of u and V at time ta and amplitudes at time ta + This gives
the equations
M
uj,k (nj+I,k
(141)
(142)
At time ta + the velocities u and v are given by
(143)
 (v  v. k )  Cn. k  n. k J (144)
j,k · j,k+1 J, 1 J, +1 J, 1
where the values on the right side of the equations are at previous
time steps as indicated.
64
Chen, Divoky, and Hwang (1975), using a stability criterion obtained
by Benjamin, Bona, and Mahony (1972), use a higher order solution for
the amplitude when
d < tu
(20 llt)
1
1
3
where the variables are expressed in dimensionless form. The solution
then becomes
lit
2lly
 lit (higher order derivative terms)
The higher order derivatives are approximated by central difference
equations as follows
2ü. k + 2ü. k ü. k]
J+l, J1, J2,
a3:y
1
[vj,k+2
2v. k
+
2v. k v. k ]
a0=

2(lly)
J' +1 J' 1 J• 2
a
2
d
[dj+1,k+1
 d  d + d. k ]
axa y 4llxlly
j1,k+1 j+1,k1 J1, 1
(145)
(146)
(147)
(148)
etc. Computed surface elevations were smoothed when one of the following
conditions was satisfied:
(a) A crest or trough has wave amplitude less than 25 percent
of the maximum wave amplitude at that instant;
(b) the local velocity component (u) or (v) has a different
sign from the average value of the surrounding four points;
(c) at a matching point where equations change from linear to
higher order equations.
Smoothing is accomplished by the average
65
(149)
where the values on the right side are before smoothing, and
where
T). k
J,
nj,k
12k + 4
n. k ( 1 + 4k) ( n . k + n . k + n . k 1 + n . k 1)
J, Jl, J+l, J,  J, +
 k(nj2,k + nj+2,k + nj,k2 + nj,k+2)
and k represents a weighting spline coefficient that varies from
(150)
(151)
0 to The influence from the surrounding points is controlled by
the values of (k). For the case k = 0, the equation reduces to Laplacian
interpolation.
To avoid numerical instability, Chen, Divcky, and Hwang (1975) imposed
the condition at matching points that
nmatching =
0
·
5
Cnzinear + nhigher arder)
Also, the partial derivative with respect to time was approximated by
~ = nj,k  nj,k
at M
where n is taken at time to + ôt/2 and n at ta  ôt/2, and
n . k
J,
For the open boundary condition previously mentioned (Fig. 12), the
finitedifference equation becomes
where
CM
nB,k  ~ (nB,k  nB1,k)
 ~ ~ Y [{Cd+ n) v}j,k+l {Cd+.n) v}j,kl]
ë = [g(d + n)] l/
2
(1 + 0.5 _n_)
d + n
(152)
(153)
(154)
(155)
Listings of typical computer programs for solutions of longwave
equations can be found in Brandsma, Divoky, and Hwang (1975) for linear
longwave equations, and in Chen, Divoky, and Hwang (1975) for Boussinesq
type equations.
66
V. TSUNAMIS APPROACHING THE SHORELINE
As a tsunami approaches a coastline, the waves are modified by the
various offshore and coastal features. Submerged ridges and reefs,
continental shelves, headlands, various shaped bays, and the steepness
of the beach slope may modify the wave period and wave height, cause
wave resonance, reflect wave energy, and cause the waves to form bores
which surge onto the shoreline.
Ocean ridges provide very little protection to a coastline. While
sorne amount of the energy in a tsunami might reflect from the ridge,
the major part of the energy will be transmitted across the ridge and
into the coastline. The 1960 tsunami which originated along the coast
of Chile is an example of this. That tsunami had high wave heights along
the coast of Japan, including Shikoku and Kyushu which lie behind the
South Honshu Ridge (Hirono, 1961).
1. Abrupt Depth Transitions.
An ocean shelf along a coastline may cause greater modification to
a tsunami than an ocean ridge. Waves may become higher and shorter, and
dispersion may occur. Lamb (1932) gave the equations for a single wave
passing over an abrupt change in water depth as shawn in Figure 16. He
considered only the case of a wave at a zero angle of incidence, i.e.,
6
1
= e
2
= O. The equations he derived are
H
dl/2

dl/2
r 1 2
H.=
dl/2 + dl/2
'[.,
1 2
(156)
Ht =
2di/2
H.
dl/2 + dl/2
'[.,
1 2
(157)
and
Ht
H
1
r
H."=
+
H.
(158)
'[., '[.,
or
Ht
H. + H
'[., r
where
Hi the incident wave height
Hr the reflected wave height
Ht the transmitted wave height
d
1
the initial water depth
d
2
the water depth under the transmitted wave
67
Reflected Wovt
Incident Wova
PLAN
z
Seo Surfoct
Bottom
Bottom
PROFILE
Figure 16. Wave passing onto shelf.
The equations predict that substantial reflection will occur when a wave
passes from deep water to shallow water, and also when a wave passes from
shallow water to deep water. It is assumed that no energy loss occurs,
and that a single incident wave splits into a single reflected wave and
a single transmitted wave. Taking Ei as the wave energy of the incident
wave, Er as the wave energy of the reflected wave, and Et as the wave
energy of the transmitted wave,
E H
2
L
r r r
==
(159)
(160)
68
and from equations (159) and (160)
E
1'
+
E.
z.,
which reduces to
(161)
(162)
Cochrane and Arthur (1948) extended Lamb's work to consider waves
approaching a shelf at varying angles of incidence. They give the ratio
of reflected wave height to incident wave height as
Hr = ~ cos 8
1
 v"(Ç" cos e
2
Hi ~ cos e
1
+ ~ cos e
2
(163)
for an abrupt change in water depth. The water depths d
1
and d2,
and the angles 8
1
and e
2
, are defined in Figure 16. This equation
also applies to a single wave with a reflected component and a trans
mitted component.
For a given incident wave angle e
1
, the value of e2 can be
determined using Snell's Law so that
(164)
Equation (163), as written, applies to shallowwater waves; wave disper
sion on the shelf is not considered. The solutions to equations (163)
and (164) are presented graphically in Figures 17 and 18, respectively.
The ratio of transmitted wave height Ht to the incident wave height
is given by
Ht = ___ z_Œ_d_
1
_c_o_s_e_
1
__ _
(165)
or, alternatively,
as before.
69
0.1 0.2 0.3 0.4 0.6 0.8 1 2 3 4 6 8 10
Figure 17. Wave reflection from a shelf Cafter Cochrane and
Arthur, 1948).
0.2 0.3 0.4 0.6 0.8 2 3 4 6 8 10
Figure 18. Transmitted wave angle 8
2
versus incident wave angle 8
1
•
70
20
* * * * * * * * * * * * * * EXAMPLE PROBLEM 4 * * * * * * * * * * * * * *
GIVEN: An incident wave with a height of 1 meter (3.28 feet) and a period
or30 minutes approaches a coastline through water 2,500 meters (8,200
feet) deep, and passes onto a shelf where the water depth is 100 meters,
at an angle of incidence e
1
= 30°
FIND:
(a) The angle at which the transmitted wave propagates onto the shelf,
(b) the height of the reflected wave, and
(c) the height of the transmitted wave.
SOLUTION:
(a) From equation (164)
sin
e2 ( : ~ f 2
sin
81
82
. 1 [ ( 100 )l/2
Sl.TI 2,500
sin
30"]
(b)
H
1'
H
r
e2
= 5.74°
From equation (163)
icÇ
cos
el
 v<Ç cos
e2
v'<Ç 81
+ I<Ç cos
e2
(Hi)
cos
h, 500 cos 30°  lïOO cos 5. 74 °
h,soo cos 30° + 1ïëi0 cos s. 74°
(c) From equation (158)
H. + H
z, r
( 1)
sin
1
(0.1)
0.626 meter (2.05 feet)
1 + 0.626
1.626 meters (5.33 feet)
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
When the initial angle of incidence e
1
> 0, the distance between
adjacent wave rays is different for the incident and transmitted waves.
As the energy equations are written for a unit length of wave crest,
from conservation of energy,
71
or, alternatively,
E.
1
E.
1
(166)
cos e2
E + E
t cos e
1
r
(167)
where bt is the distance between adjacent wave rays for the transmitted
wave , and bi the distance between adjacent wave rays for the incident
wave. Rewriting equation (167),
Et cos e
2
+
Ei cos e
1
which gives
E
r
E.
1
1 (168)
(169)
Not ing that Lt/Li = Ct/Ci = v'd
2
/d
1
and that Lr/Li = Cr/Ci = v'd
1
/d
1
= 1,
and substituting equations (163) and (165) into equation (169),
( ~ c o s
el
 lcÇ cos
02 )
2
+
1
~ c o s
el
+ lcÇ'" cos
e2
(170)
which reduces ta
( ~ c o s
~ c o s
a
1
+ ra; cos a
2
y =
1
e
1
+ lcÇ'" cos e
2
(171)
proving that the equations of Cochrane and Arthur conserve the energy of
the incident wave.
Cochrane and Arthur (1948) compared a calcu1ated value for a wave
from the 1946 tsunami, which reflected from the continental slope off
southern Oregon, with an actual recorded wave height at Hanasaki, Japan.
Using a rough approximation for the wave height at the top of the conti
nental slope, it was determined that the reflected wave arriving at
Hanasaki would have a height of 17 centimeters (0.56 foot). The observed
72
wave height, for an arrivai time equal to the calculated time for the
reflected wave, was in good agreement with the calculated wave height.
Figure 17 shows that for waves arriving with a higher angle of
incidence there will be sorne value of d
1
/d
2
for which equation (163)
predicts no reflected wave. For these same values of d
1
/d
2
, the pre
dicted transmitted wave height would equal the incident wave height.
Cochrane and Arthur note that reflected waves are normally of secondary ,
but not negligible, magnitude according to theory. At given stations,
convergence may cause reflected waves to be of primary magnitude, but
this occurs only in relatively few cases. Shepard, MacDonald, and Cox
(1950) note that the highest and most damaging waves at Napoopoo and
Hokeena, on the island of Hawaii, from the 1 April 1946 tsunami origi
nating in the Aleutian Islands, Alaska, were reflected waves from the
continental slopes of Japan and the Bonin Islands.
2. Linear Depth Transitions.
Cochrane and Arthur (1948) indicate that the length of a continental
slope, as well as the difference in water depth, should be considered in
the calculations. Dean (1964) considered the case of a wave normally
incident on a linear change in water depth shown in Figure 19. Defining
a parameter, zl, as
the transmission coefficient
and the reflection coefficient
K
r
H
r
H.
1
(172)
(173)
(174)
Dean found the results shown in Figure 20. As before, for the zero angle
of incidence assumed by Dean, when d
1
/d
2
< 1.0 the value of the reflec
tion coefficient Kr is negative. When d
1
/d
2
> 1.0 the value of Kr is
positive.
* * * * * * * * * * * * * * EXAMPLE PROBLEM 5 * * * * * * * * * * * * * *
GIVEN: An incident wave, which is 0.5 meter (1.64 feet) high and has a
peT"iod of 40 minutes, recedes from the coastline through water 100
meters deep and passes from the shal low water over a shelf into water
3,025 meters (9,925 feet) deep. The transition between the two water
73
Reflected Wove
Tronsmitted Wove
dz
Incident Wove
Figure 19. Linear s1ope and she1f.
,•
,.
Il/
l' li
i/ J
j, /
l/,1 1/
//;'
r: .:.=:.F·<1...
0 f=.= ..... ft_... "
0.01 0.02 0.05 0.1 0.2 0.5 10 20 50 100 200 500 1000
Figure 20. Ref1ection and transmission coefficients
(from Dean, 1964).
74
depths is a linear slope S = 0. 1 and the wave is at a zero angle of
incidence with the slope transition, i.e., el= o.
FIND:
(a) The height of the reflected wave,
(b) the height of the transmitted wave, and
(c) show that energy is conserved, i.e., that the total energy in
the reflected and transmitted waves equa1s the incident wave energy.
SOLUTION:
/g"ëÇ T
L
1
= /9.8 x 100 (40 x 60)
From equation (172),
75,100 meters
From
it is
z =
1
Figure
~ =
d2
found
K
r
H
r
47T x 100
0.167
75,100 x 0.1
20, where
_!QQ_ =
0.033 and z
1
0. 167
3,025
th at
0.62 (the negative sign indicates that the ref1ected wave
is TI radians out of phase with incident wave)
0.62 Hi = 0.62(0.5) = 0.31 meter (1.02 feet)
(b) From Figure 20,
Ht = 0.32 Hi = 0.32(0.5) = 0.16 meter (0.52 foot)
(c) y= pg = (1,026 ki1ograms per cubic meter) (9.8 meters per
second squared)
y = 10.055 ki1ograms per square meter  second squared
75
The energy per meter length of wave crest is
E.
Er
Lt
E =
t
y L 2
i  10,055(0.5) 75,100
8 8
2.36 x 10
7
kilogrammeters per second squared
2.36 x 10
7
y H2
L
10,055(0.31)
2
75,100
1' 1'
9.07 x 106 joules
8 8
c2
T
/g(Ç T
19.8 x 3,025 (40 x 60) 413,000 meters
y H; Lt
10,055(0.16)
2
413,000
1.33 x 10
7
joules
8 8
As Kr is negative, i.e., the reflected wave is out of phase, then
for energy to be conserved
Ei Er Et
E. E 2.36 x 10
7
 9.07 x 106 = 1.45 x 107 joules
1'
compared to the computed value of Et= 1.33 x 10
7
joules. The differ
ence results from the minor errors which occur using Figure 20.
3. Nonlinear Depth Transitions.
Kajiura (1963) investigated waves passing from deep water to shallow
water over the nonlinear slope profile shawn in Figure 21. The profile
is defined by the equation
where the effective slope length is given by
= 27T
n
(175)
(176)
L
1
is the wavelength at depth d
1
, L
2
is the wavelength at depth d
2
,
and n is an arbitrary small number 1n equation (175) which fits the
equation to the actual slope and determines the length of the slope in
equation (176).
The reflection
equation
coefficient obtained by Kajiura
sinh 
Hl' = L \ 2 l u
1
sinh
+ J
H.
76
is given by the
(177)
Figure 21. S1ope and shelf.
The solution of equation (177) is plotted in Figure 22. As shawn in the
figure, the reflection coefficient approaches zero as the slope length
~ approaches the length of the incident wave L
1
.
0.8
0.3
0.1
0.1 0.2 0.3 0.4 0.5 0.6 o. 7 0.8
J.
L;
Figure 22. Ref1ection coefficients (from Kajiura, 1963).
* * * * * * * * * * * * * * EXAMPLE PROBLEM 6 * * * * * * * * * * * * * *'
GIVEN: An incident wave which is 0.6 meter (1.97 feet) high and has a
përiod of 20 minutes, approaches the coastline through water 2,500
meters deep and passes onto a shelf where the water depth is 100 meters,
at a zero angle of incidence (8
1
= 0). The effective length of the
slope between the two water depths is ~ = 24,000 meters (14.9 miles).
It is assumed that the slope is defined by equation (175) and that
energy is conserved; i.e. , the total of the reflected and trans mi tted
wave energy equals the incident wave energy.
7_7
FIND:
(a) The height of the reflected wave ,
(b) the height of the transmitted wave, and
(c) the height of the reflected wave for a linear slope of the
same length.
SOLUTION:
C
1
T = .rg;ç T = 19.8(2,500) (20 x 60)
From
L
1
187,800 meters (116.7 miles)
Q, 24,000
r; = 187,800
Figure 22
H
1"
0.46
H:"=
1.
0.128
Hl" = 0.46 Hi = 0.46(0.6) = 0.28 meter (0.91 foot)
(b) From conservation of energy,
y H ~
Ll
10,055(0.6)
2
187,800
8.50 x 10
7
joules
E.
1.
1.
8 8
y H2
Ll
10,055(0.28)
2
187,800
E =
1"
1.85 x 10
7
joules
1"
8 8
E.  E = 6.65 x 10
7
joules
1. 1"
Lt = 37,600 meters (23.3 miles)
78
y H ~ Lt
Et =
8
6. 65 x 10 7 joules
(
8 x 6.65 x 10
7
)1/
2
Ht 10,055 x 37,600
Ht 1.19 meters (3.89 feet)
(c) The slope is defined by
s
(2,500  lOO)
24,000
From equation (172),
41fdl = 41f(2,SOO)
1.67
L
1
S 187,800(0.1)
From Figure 20, where
d/d2 = 25
H
2:
0.55
H.
1.
0.1
H 0.55 H. = 0.55(0.6)
=
0.33 meter (1.08 feet)
r 1.
which indicates that a linear slope gives a higher reflected wave
and 1ower transmitted wave than a slope defined by equation (176).
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
4. Experimental Measurements.
Bourodimos and Ippen (1968) obtained experimental results for waves
passing from deep water to shallow water over a slope where S = 0.125.
Their experimental curves for Kt and Kr as functions of d
1
/d
2
are
given in Figure 23. Gagnon and Bocco (1962) obtained measurements for
waves passing from shallow water to deep water at an abrupt transition
in depth. However, their results indicated a fairly constant value of
IKrl ~ 0.2, including the case whçre d
1
= d
2
(i.e., where no transition
occurs).
79
1.4
Kt
1.2
1.0
v +
0.8
0.6
Kr
0.4

0.2
0
0.45 0.50 0.60 0.70 0.80 0.90
( r
Figure 23. Ref1ection and transmission coefficients
(modified from Bourodimos and Ippen, 1968).
<;
S. So1itons and ShoalingInduced Dispersion.
For certain conditions, a wave will decompose into a train of waves.
Exampl es of this are shown in Figures 24 to 28. This train of waves will
consist of an initial wave having the highest amplitude, followed by a
fini t e number of waves of decreasing amplitude. Wave decomposition has
been investigated by Mason and Keu1egan (1944), Horikawa and Wiege! (195.9),
Benj amin and Feir (1967), Street, Burgess, and (1968), Madsen and
Mei (1969), Byrne (1969), Street, Chan, and Fromm (1970), Galvin (1970),
Zabusky and Galvin (1971), and Hammack and Segur (1974).
Benj amin and Feir (1967) discuss the stability of waves, and indicate
t hat t he waves will only be unstable if kd > 1.363, where k is the wave
number 2n/L. Whitham (1967) showed that equations governing extremely
graduai variations in wave properties are elliptic if kd > 1.363, and
hyperbolic if kd < 1.363. For tsunamis, where d/L << 1, the equations
wi l l be hyperbolic and the waves will be stable, at !east in a constant
water depth.
80
3 feet from
Gene rotor
6 5 4
T.ime (s)
0
0
0.25
c
.2
êi
>
CD
~
0.125::.
39 feet from
Generator
Time (s)
0
Figure 24. Separation of so1itons
(from Ga1vin, 1970).
0
1 /
~
2 ia i /
~
Figure 25. Induced wave generation over a
submerged bar (from Byrne, 1969).
81
c
.2
êi
>
CD
~
0.1
"7
d;"
0.2
0.1
"7
~
0
0.2
0.1
"7
~
(o) d
1
+f+ Slope 5=0.05
10 20
20 30 40
..!.
dl
(c)
Figure 26. Solitary wave propagating over a slope onto a
shelf (from Madsen and Mei, 1969).
82
J..==r===0 ____
"'10.25
d, 0.50
0.75
eosine Slope
did·= 0.15
t•: t(g/dl) 1/2
1.00 ....... '"""''...___.. _ _....__,
0 5 10 15 20 25 30 35 45 50 55 60 65 70 75
.!..
d,
Figure 27. So1itary wave propagating onto a she1f (from Street, Chan,
and Fromm, 1970).
0.25
'1 0.25
1
'0.50
0.75
r•=45
x
d.
T*=90
eosine Slope
d2/dl =0.15
t•:t(g/dl )1/2
Figure 28. Wave train propagating onto a she1f (from Street, Chan,
and Fromm, 1970).
* * * * * * * * * * * * * * EXAMPLE PROBLEM 7 * * * * * * * * * * * * * *
GIVEN: A tsunami with a wave period of 15 minutes travels through water
meters (9,840 feet) deep.
FIND: If the wave is stable.
SOLUTION: The wave ce1erity is given by
C = lgëi = v'9.807 x 3,000 = 171.5 meters (563 feet) per second
The wavelength is
L = CT = 171.5 x 15 x 60 154,350 meters (95.9 miles)
83
From this
d 3,000
I = 154,350
0.0194
and
kd = = 0.122
Therefore, kd < 1.363 and the wave is stable.
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Galvin (1970) investigated waves propagating through water of uniforrn
depth in a laboratory wave tank. He found that the initial generated
waves broke clown into several waves which are called solitons. Figure 24
illustrates an example where, for a water depth of 0.15 rneter (0.5 foot)
and a generator period of 5.2 seconds, each of the initial waves broke
clown into five solitons. Taking these waves as shallowwater waves, the
wavelength is approximately 6.4 rneters, and kd 0.15 which would indicate
that the waves are stable. However, it may be assumed that the generated
waves were not single waves, but rather a combination of
solitons. Galvin noted that if a group of such waves traveled over a
sufficiently long distance, the solitons would recombine into single
waves, separate again into solitons, etc. There are cornrnonly two or
solitons, but as many as seven could exist in sorne instances. If
a generated tsunami had the characteristics of a group of solitons, it
could appear differently at various coastal points, depending on the
distance from the generating area.
Zabusky and Galvin (1971) cornpared nurnerical and experimental
results for solitons, using the KortewegdeVries equations, for cases
where 22 U 777, where U is defined as (H/d)(L/d)
2
. They found
good cornparisons for slightly dissipative waves. Harnrnack and Segur
(1974) also studied numerical and experimental results. They found that
soliton generation is dependent on the net volume change in the body of
water. When the net volume of the initial wave system was positive
(e.g., from uplifting of the sea bottorn), solitons evolved followed by
a dispersive train of oscillatory waves. If the initial generating
rnechanisrn was negative everywhere (sea bottorn subsidence), no solitons
evolved.
Byrne (1969) made field observations of waves passing over a near
shore bar. He noted that a wave passing over a bar would sornetirnes
produce a second, trailing wave as shown schernatically in Figure 25.
As these additional waves developed near the shoreline, he was unable
to determine if such waves would recombine with the waves in the initial
wave train.
84
Masan and Keulegan (1944) investigated waves passing into a shallower
water depth, with an abrupt change in depth. The condition for instabil
ity obtained from their experiments was
(178)
where a
1
is the wave amplitude in the deeper water, L
1
the wavclength
in the deeper water, and d
2
the depth in the shallmver \vater. Thcir
results were confirmed by H o r i k a ~ a and Wiegel (1959), although in the
latter report thcrc is an apparent discrepancy in the presentation of the
results; the right side of equation (178) has been multiplied by 1.2.
* * * * * * * * * * * * * * EXAMPLE PROBLEM 8 * * * * * * * * * * * * * *
GIVEN: A tsunami with a period of 15 minutes passes from water 3,000
niëters deep onto a shelf \vhcrc the water depth is 200 meters (656 feet).
FIND: The maximum wave amplitude for a stable wave which will not
aëcompose into a train of waves.
SOLUTION: The wave celerity in deep water is
c
1
= rgcç = 19.807 x 3,000 = 171.5 meters per second
and the wavelength is
L
1
~ C
1
T = 171.5 x 15 x 60 = 154,350 meters (95.9 miles)
The condition for wave instability is given by equation (178) as
(alLI) 1/2 > 2d2
(a
1
x 154,350)
1
1
2
> 2 x 200
a
1
> 1.04 meters (3.40 feet)
Thus, waves with a deepwater amplitude less than 1.04 meters would not
decompose.
* * * * * * * * * * * * * * EXAMPLE PROBLEM 9 * * * * * * * * * * * * * *
GIVEN: A tsunami travels from a 3,000meter water depth into a 200meter
water depth. The wave period is 60 minutes.
FIND: The maximum wave amplitude for a stable wave which will not decom
pûse into a train of waves.
SOLUTION: The deepwater wave celerity is given as
C
1
= (gd
1
)
112
= (9.807 x 3,000) l/2 = 171.5 meters per second
85
and the wavelength is
L
1
= c
1
T = 171.5 x 60 x 60 = 617,400 meters (384 miles)
From equation (178), wave instability is given by
(alLl)l/2 > 2d2
(a
1
x 617,400)
1
/
2
> 2 x 200
a
1
> 0.26 meter (0.85 foot)
Waves with a deepwater amplitude less than 0.26 meter will not
decompose.
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Using the results of Mason and Keulegan (1944), the above examples
illustrate that the longer period tsunamis are much more likely to
decompose where the waves have the same height in the deep ocean.
Street, Burgess, and Whitford (1968) investigated solitary waves
passing from an initial water depth, over a steep slope, and into a
shallower water depth. They obtained results similar to those of other
investigators, showing that each wave changed from a single wave into a
train of several waves. In sorne instances, there was also a significant
increase in wave height. Defining the initial water depth as d1, the
shallower depth as d2, the initial wave height as Hi, and the wave
height in the shallower water depth as Ht, as the ratio d1/d2 increased,
relative wave height Ht/Hi reached a maximum value for any initial wave
height and then decreased. As
decreased, the maximum value
of became greater and occurred at a higher value of d
1
/d
2
. The
locus of the maximum values of wave enhancement, are shown in
Figure 29 with the results for the solitary wave experiments.
Madsen and Mei's (1969) numerical results for the propagation of
long waves give the results shown in Figure 26 for a solitary wave
passing over a slope and onto a shelf. The numerical results of Street,
Chan, and Fromm (1970) give the results shown in Figure 27 for a solitary
wave, and the results shown in Figure 28 for a train of waves. Goring
(1978) has also recently carried out experiments on solitary waves
propagating onto a shelf. His results are similar to those of Street,
Chan, and Fromm (1970) and Madsen and Mei (1969).
In all cases where a single wave produced a series of wave crests,
the first wave crest of the series was the highest. It may be presumed
that a number of initial wave crests will produce the same number of
groups of wave crests, each having a high initial wave followed by smaller
waves. The numerical work of Street, Chan, and Fromm (1970) for wave
trains is inconclusive in this regard as Figure 28 shows the additional
wave crests, but does not separate the waves into groups associated with
the initial crests.
86
3
.....
.......
',
,..,
    ~  L o c u s of Maximum
Wave Enhancement
........
...............
'
'
.......
.......
...........
......


 0.025
.... _

 0.050
Figure 29. Wave enhancement (from Street, Burgess, and
Whitford, 1968).
87
VI. TSUNAMISHORELINE INTERACTION
In addition to the shoaling of waves on the nearshore slope, a tsunami
may interact with a shoreline in a number of different ways, including
standing wave resonance at the shoreline, the generation of edge waves by
the impulse of the incident waves, the trapping of reflected incident
waves by refraction, and, as the reflected wave from the shoreline propa
gates seaward, the reflection of wave energy from an abrupt change in
water'depth at the seaward edge of a shelf. Also, a wave arriving at an
oblique angle to the shoreline may produce a Machstem along the shoreline.
All of the above interactions depend on wave reflection at the shoreline.
Tsunamis entering inlets and harbors may also produce resonant conditions
within the inlets and harbors. LeBlond and Mysak (1977) provide a general
discussion of edge waves and wave trapping.
1. Wave Reflection.
The reflection of an incident wave ray from a shoreline is illustrated
in Figure 30. The angle, a
1
, between the wave ray and aline normal to
a tangent to the shoreline will have the same value for the incident and
the reflected wave rays. For a steep nearshore slope, the reflected
wave will be in phase with the incident wave.
Line Normal ta Tangent
W a v e ~
Figure 30. Wave reflection from a shoreline.
88
Miche (1944) defined the wave reflection at a shoreline in terms of
a critical wave steepness, (H/L)c, which is given by
(179)
where S is the angle of the beach slope in radians. Complete reflection
will occur if the wave steepness, H/L, in deeper water is given by
!! < (!!)
L L
c
(180)
* * * * * * * * * * * * * * EXAMPLE PROBLEM 10 * * * * * * * * * * * * * *
GIVEN: A tsunami has a height of 0.5 meter and a period of 20 minutes
1n a 1,000meter water depth . The nearshore slope s
3
= 0.1 (S = 0.0997
radians).
FIND: If the wave is completely reflected at the shoreline.
SOLUTION: In the deeper water, the wave celerity, C, is
C lgd = 19.807 x 1,000 = 99 meters per second
L CT= 99 x 20 x 60 = 118,800 meters
H 0.5
I = ll8, 8oo
4.21 x 10
6
From equation (179)
(;sy/
2
= (2 x
c
<
c
thus , the wave is completely reflected at the shoreline.
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Wiege! (1964) indicates that where
Ï > (Ï)
c
(181)
89
that the reflection will be defined by
(182)
where cR is a coefficient of roughness and permeability which has a
value of cR= 0.8 for a smooth impervious beach. Various values of cR
were defined for rough slopes for shortperiod waves. However, the
effect of the slope roughness on longer period waves has not been
adequately determined.
* * * * * * * * * * * * EXAMPLE PROBLEM Il * * * * * * * * * * * * * *
GIVEN: A tsunami has a height of 0.5 meter and a period of 4 minutes
1n a 1,000meter water depth. The nearshore slope s
3
= 0.01 (S = 0.01
radians), and the slope is smooth and impervious.
FIND: The coefficient of reflection at the shoreline.
SOLUTION: In the deeper water, the wave celerity, C, is
C !gd= 19.807 x 1,000 = 99 meters per second
L CT= 99 x 4 x 60 = 23,800 meters (14.8 miles)
Ï =
= 2.10 x Io
5
From equation (179), where S is given in radians,
= (;sy/2 = (2 x'ITO.OI)l/2 sin2;o.OI)
c
(![) = 2.54 x 106
c
!:!. > (!:!.)
L L c
2.54 x 10
6
0.8 5
2.10 x 10·
0.097
which indicates a lowreflected wave height where the shoreline has
a very graduai slope.
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
90
2. Shelf Resonance .
Hidaka (1935a, 1935b) carried out a theoretical investigation of a
vertical wall at the shoreline, where the water depth at the wall was
d
8
, and the sea bottom sloped seaward. The depth d at any arbitrary
distance x from the shoreline is given by
d
(
x2)l/2
d 1 +
s a2
(183)
where the horizontal distance, x, is positive measured seaward from
the shoreline, x = 0 at the shoreline, and a is the distance from the
shoreline to the depth d = 1:2 d
8
• The depth variation defined by equation
(183) can be compared to a linear (constant) bottom slope, s
2
, between
the toe qf the nearshore slope (taken to be a vertical wall) and a point
at the distance x = a from the. shoreline. For the linear bottom slope,
s2,
d  d
s
x
or at a distance, a, from the shoreline
from which
a =
12 d  d
s s
a
( /2  1) d
The variables are shawn in Figure 31.
Defining the wave by the equation
(184)
(185)
s
( 186)
( 18.7)
Hidaka defined the surface elevation n above the undisturbed water as
(
27rt)
n = u cos T
(188)
and U a dimensionless amplitude obtained by dividing the amplitude at
any point by the amplitude at the shoreline (U = 1 at the shoreline),
T the wave period, and t time. Hidaka obtained a theoretical solution
for wave resonance on the sloping shelf defined by equation (183) using
Mathieu functions (see Blanch, 1964). The primary mode of oscillation
91
i.O ul

<;;:
Ill
0
x
Figure 31. Shelf resonance.
.s::.lll
"'E
Zo
1.0
for shelf is defined by the Mathieu function
e
1
) where
el = 7.51361. The period is given by
2a
3.2417 _a_
s
(189)
The second mode of oscillation is proportional to the Mathieu function
e
2
) where e
2
= 21.29863, and the period T
2
is
1.9254 _a_
%
(190)
The first and second modes are shown in Figure 31. The values obtained
by Hidaka for resonant periods are for a shelf extending a long distance
offshore; i.e., the shelf width i
8
>> L, where L is the wavelength of
the incident wave. These results have not been verified by other inves
tigators, but Hidaka's results of the variation of wave amplitude agree
very well with those obtained by Wilson (1972) for a constant (linear)
shelf slope, as shown in example problem 14.
92
To determine the variation of wave amplitude with respect to distance
from the shoreline, the equation for U is put in the form
d2u [ t t
dp2 + (1 + p2) l/2 + _(_l_+_p_2_) +
(191)
where p = x/a. This equation was solved by Hidaka using Stormer's method
(see Milne, 1953). The wave profile is defined in Table 1.
x/a
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9
Table 1. Distribution of amplitude U
(from Hidaka, 1935b).
First mode Second mode
u x/a u x/a u xia
l. 0000 2.0 0.7985 0.0 l. 0000 2.0
0.9813 2.1 0.7766 o. 1 0.9474 2. 1
0.9265 2.2 0.7432 0.2 0.7964 2.2
0.8391 2.3 0.6998 0.3 0.5668 2.3
o. 7244 2.4 0.6476 0.4 o. 2868 2.4
0. 5 889 2.5 0.5880 0.5 0.0115 2.5
0.4392 2.6 0.5224 0.6 0.2972 2.6
0. 2822 2.7 0.4520 0.7 0.5445 2.7
0.1239 2.8 0.3781 0.8 0.7346 2.8
0.0305 2.9 0.3018 0.9 0.8566 2.9
0.1766 3.0 0.2241 1.0 0.9070 3.0
0.3110 3.1 0.1462 1.1 0.8887 3.1
0.4313 3.2 0.0688 1.2 0.8096 3.2
0.5357 3.3 o. 0072 1.3 0.6807 3.3
0.6233 3.4 0.0810 1.4 0.5149 3.4
0.6936 3.5 0.1521 1.5 0.3256 3.5
0.7466 3.6 0.2197 1.6 0.1261 3.6
0.7827 3.7 0.2834 1.7 0.0716 3.7
0.8028 3.8 0.3428 1.8 o. 2572 3.8
0.8076 3.9 0. 3975 1.9 0. 4221 3.9
4.0 0.4473 4.0
u
0.5600
0.6665
0.7392
0. 7773
0.7818
0.7548
0.6995
0.6197
0.5200
0.4052
0.2800
0. 1493
0.0176
0.1110
0.2327
0.3443
0.4430
0.5267
0.5940
0.6436
0.6754
* * * * * * * * * * * * * EXAMPLE PROBLEM 12 * * * * * * * * * * * * * *
GIVEN: Water depth, d
8
, at the toe of a nearshore slope is 30 meters;
the distance a= 12,430 meters (7.72 miles). Complete reflection occurs
at the nearshore slope, and it can be assumed to behave as a vertical
slope.
93
FIND:
(a) The primary and secondary periods of oscillation, and
(b) the relative wave height of the wave at a distance one wavelength
from the shoreline in relation to the wave height at the shoreline, for
the second mode.
SOLliTION:
(a) d = 30 meters and a
s
From equation (189)
T
1
= 3.2417 _a_
;ger
s
3.2417
19. 807 ( 30)
From equation (190)
T = 1. 9254 _a_
2 licÇ
1.9254
19.807(30)
12,430 meters
2,350 seconds (39.2 minutes)
1,395 seconds (23.3 minutes)
(b) Both the first and second modes of oscillation are in the range
of tsunami periods which are likely to occur. Taking h
8
as the wave
height at the shoreline, Table 1 gives, for the second mode, a height
equal to 0.7818 hg where x/a= 2.4 or where x= 2.4 (12,430) = 29,800
meters (18.5 miles). The values in Table 1 and Figure 32 show that
this is approximately the distance between secondmode wave crests (one
wavelength).
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
* * * * * * * * * * * * EXAMPLE PROBLEM 13 * * * * * * * * * * * * * *
GIVEN: The water depth, d
8
, at the toe of a nearshore slope is 15
meters; the distance a 621 meters (2,039 feet). Complete reflection
occurs at the nearshore slope.
FIND: The primary and secondary periods of oscillation.
94
1.0
0.5
u 0
Exomple 4
0.5
5=0.001
1.0 0
0.5
X/a
Figure 32. amplification on a shelf.
SOLUTION: d
8
= 15 meters and a = 621 meters.
From equation (189)
T = 3.2417 _a_
1 ;gd"
s
T
1
= 3.2417
621
19.807(15)
From equation (190)
T = 1.9254 ___ a_
2
;gers
1. 9254
19. 807 (15)
166 seconds (2.77 minutes)
98.6 seconds (1.64 minutes)
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
A different means of calculating the amplitude u. which will also
account for refraction effects (i.e., the effect of a nonuniform offshore
bathymetry), is suggested by Wilson (1972). These equations are
N.
J+l
(BjDj  C) Nj  Uj
C + B.D.
1
J J+
95
(192)
uj+l
u . + 2C(N .
1
+ N .) (193)
J
J+
J
B.
2
(194)
J
[fl(b. l + b j)]
J+
llw
2
c
(4g)
(195)
D. b.d.
(196)
J J J
j = 1, 2, 3,
where bj and bj+l represent the distance between refracted wave
rays at stations j and j_+ 1, respectively, N the horizontal dis
placement of a water particle, and il the horizontal distance between
stations j and j + 1. For an unrefracted wave,
U
0
= 1 at the shoreline (as in the case of Hidaka) and N
0
= 0 at the
shoreline.
(197)
Looking at unrefracted waves on a constant shelf slope s
2
for the
same wave periods previously defined by equations (189) and (190),
T
n
K
a
n ;gd
s
where K
1
= 3.2417 and K
2
= 1.9254 as previously defined in equations
(189) and (190). From this
21T ;gd
s
and for the constant slope, from equation ( 186) ,
d
dsSz 52
_.!!...=
a
( 12  1)
ds
rz 1
96
(198)
(199)
(200)
Defining t:,. sa where s is an arbitrary increment,
dj = dj =
ds
+ jS sa
BjDj
/::,. sa sa
BjDj
s[ 1
s(/2 1)
+ j]
at the
jth
increment.
The equations of Wilson (1972) can then be expressed
Nj+l
t(<c// 1l •
j 
( •2<
S Kn 2( 12  1)
+
u. +
J
,2E )
Kn
2
( 12  1) N j 
1
+ j
+ 1)
s(12  1)
(201)
(202)
as
uj]
(203)
(204)
* * * * * * * * * * * * * EXAMPLE PROBLEM 14 * * * * * * * * * * * * * *
GIVEN: The water depth, d
8
, at the toe of a nearshore s1ope is 30
meters; the slope of the shelf is s
2
= 0.001. Complete reflection
occurs at the nearshore slope.
FIND: The wave profiles using the methods of Hidaka (1935b) and Wilson
cr 9 72) •
SOLUTION: From example problem 12,
(li  1) d
a= ____ ......::...s
s
( 12  1) 30
0.001
3.2417 ___ a_= 2,350 seconds
v'gd
s
1. 9254 ___ a_ = 1, 395 seconds
v' gd
s
12,430 meters
Exploring the second mode of oscillation as before, and using values of
t:,. = 0.1 a= 1,243 meters (4,078 feet)
97
the values of U obtained by Hidaka (1935b) are given in Table 1.
The values obtained by Wilson (1972) are given in Table 2. The wave
profiles are plotted in Figure 32.
Table 2. Values of horizontal water particle displacement,
N, and wave amplitude, U.
T2
=
1,395 s
w2
= 0.0045
ds = 30 rn
s
=
0.001
x/a N u x/a N u
0.0 0 1.0000 1.4 69.50 0.5928
0.1 38.79 0.9501 1.5 80.41 0.4001
0.2 70.95 0.8090 1.6 85.71 0.1865
0.3 94.23 0.5967 1.7 85.49 0.0336
0.4 107.48 0.3374 1.8 80.14 0.2465
0.5 110.56 0. 0571 1.9 70.33 0.4399
0.6 104.20 0.2190 2.0 56.92 0.6035
0.7 89. 82 0.4684 2.1 40.90 o. 7292
0.8 67.27 0.6729 2.2 23.31 0. 8117
0.9 44.66 0.8194 2.3 5.20 0.8484
1.0 18.17 0.9002 2.4 12.43 0.8391
1.1 8.15 0.9131 2.5 28.71 0.7862
1.2 32.47 0.8609 2.6 42.90 0.6941
1.3 53.29 0.7507
For the given conditions, Figure 32 shows that Wilson's method
produces almost the same results as those obtained by Hidaka; however,
Wilson's results would predict slightly less amplification over the
same distance.
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
In comparing the work of Hidaka (193Sa, 1935b) with that of Wilson
(1972), the numerical method proposed by Wilson for determining the wave
profile·is much more readily used for a particular shelf slope. Table 1
(Hidaka, 1935b) has the advantage of being a general solution. The value
of shelf slope, s
2
= 0.001, used in example problem 14 is typical of
slopes found on many continental shelves. Figure 32 shows that the two
methods produce comparable results. Wilson's method has the added feature
of considering wave refraction.
3. Reflection from Seaward Edge of Shelf.
Section V discussed the reflection of waves from an abrupt transition
in water depth. It was shawn that when a wave propagates seaward from
98
J
the shoreline sorne of the wave energy is reflected shoreward from the
transition in water depth at the seaward limit of the shelf. This is
further illustrated in Figure 33 where d
8
is the water depth at the
toe of the nearshore slope, d2 the water depth at the seaward limit of
the shelf, d
1
the water depth at the seaward limit of the steep tran
sition in water depth, sl the slope of the steep transition, s2 the
slope of the shelf, and s
3
the nearshore slope.
Wove Reflected Seoword wove Reflected Shoreword
from Shoreline from Oepth Transition
Figure 33. Reflected waves on a shelf.
The wave reflected shoreward from the steep transition may be
radians out of phase with the wave transmitted seaward across the tran
sition. However, the actual phase difference will depend on the geometry
of the shelf and transition, and the water depth. This was illustrated
in example problem S. For perfect reflection, the wave reflected from
the shoreline will be in phase with the initial wave incident on the
shoreline. The time, t , for the wave to travel the distance, •
from the steep to the nearshore slope will be the same as the
time required for the reflected wave from the nearshore slope to travel
back to the steep transition in depth. Therefore, where the wave re
flected from the transition is radians out of phase with the incident
wave, resonance will occur if
nT
2
where T is the incident wave period, and n = 1, 2, 3, ...
No ting th at C = ;gd where d
8
d d
2
, for a wave wi th a normal
angle of incidence,
dx
c
99
(205)
(206)
where x is measured seaward from the toe of the nearshore slope .
By definition
or defining dx,
This gives
dd = s2
dx
dx
and, for a constant shelf slope,
Substituting equation (210) into equation (205) gives
T
8 Cdi/2  dJ12)
n s2 gl/2
(207)
(208)
(209)
(210)
(211)
where T is a resonant wave period where the reflected wave and incident
wave are rr radians out of phase., and n = 1, 2, 3, . . . Equation (211)
provides a first approximation for the resonant wave periods.
* * * * * * * * * * * * * * EXAMPLE PROBLEM 15 * * * * * * * * * * * * * *
GIVEN: The water depth, d
8
, at the toe of a nearshore slope is 30
meters. The width of the shelf, ~
8
, is 30,000 meters (18.6 miles)
and the water depth, d
8
, at the the seaward edge of the shelf is
60 meters (196.9 feet).
FIND: The resonant wave periods for the shelf.
SOLUTION: The slope of the shelf, s
2
, for a constant slope is given by
60  30
30,000
100
0.001
From equation (211),
T
8
( d ~ / 2  dy2)
gll2
' n
n
s2
8
(601/2  301/2)
n 0.001 (9.807)
1
1
2
T
Tl
5,800 seconds (96. 7
T2
2,900 seconds (48.3
T3
1,933 seconds (32.2
T4
1,450 seconds (24.2
etc.
1, 2' 3, . . .
5,800
n
' n
minutes),
minutes),
minutes),
minutes),
n
n =
n =
n =
1' 2' 3, . ..
1
2
3
4
* *
* * * * * * * * * * * * * * * * * * * * *
* * * * * *
* * * * * *
* *
Nagaoka (1901) considered the possibility of currents parallel to the
coast acting as boundaries which would reflect waves. He speculated that
the currents would act as quasielastic boundaries. In this case waves
generated near a shoreline could be trapped between the shoreline and an
offshore current, creating a resonant condition between two boundaries.
4. Edge Waves.
The impulse of incident waves reflecting from the shoreline may
generate edge waves in the longshore direction. These edge waves, the
trapped mode of longshore wave motion, have wave periods which will be
longer than the incident wave periods; standing edge waves will have
peaks and nodes at points along the shoreline, although edge waves may
be either standing or progressive waves. Guza and Bowen (1975) indicate
that experimental results confirm the work of Galvin (1965) and Bowen
and Inman (1971) which show that incident waves that are not strongly
reflected will not excite edge waves visible at the shoreline.
Guza and Inman (1975) have defined the water surface profile of edge
waves in the seaward direction using the dimensionless wave amplitude,
U, and a dimensionless distance, x, in the seaward direction given as
x
w
2
x
g tan S
(212)
where w is the radian frequency ( 2 ~ / T ) of the edge wave, x the dis
tance measured from the shoreline in the seaward direction, and S the
angle of the nearshore slope (tanS= S). The water surface profile is
given in Figure 34 which shows that higher modes of standing edge waves
101
will have peaks and nades in the seaward direction in addition to the
peaks and nades in tne longshore direction.
u
1.0
0.8  Edge woves
0.6
 Reflected normolly
incident wove
0.4
0.2
0
0.2
0.4
0.6
Figure 34. Offshore profiles of edge waves
(from Guza and Inrnan, 1975).
Guza and Davis (1974) carried out a theoretical investigation of the
mechanism of edge wave generation by norrnally incident, shallowwater
waves on a constant beach slope. They define the longshore wavelength
of the edge wave by the longshore wave nurnber, ky. given by
where
k
y
2
~ : = ( ~ : ) g(2n + Î) tanS
(213)
(2n + 1) S « 1 (214)
Ly is the wavelength of the edge wave, ~ y the period of the edge wave,
and S the angle of the nearshore slope 1n radians. Guza and Davis
attribute the generating mechanism to a nonlinear interaction between
the incident wave and a pair of progressive edge waves with frequencies
defined by w
1
and w
2
where w = 2TI/T and
(215)
i.e., the incident wave frequency is equal to the surn of the two edge
wave frequencies. The two edge waves have the same wavelength, but
propagate in opposite directions along the shoreline. Therefore, the
edge wave pair forrns a standing wave. This standing wave will always
have a frequency equal to onehalf the incident wave frequency (a period
twice the incident wave period) even though the frequencies of the edge
wave pairs may vary, as shawn in Table 3. Where the frequencies of the
two progressive edge waves forrning the pair are different, the nades and
antinodes of the standing wave will move in the direction of the edge
wave with the higher frequency (shorter period). Defining the edge
wave with the lower frequency by
102
Table 3. Resonant edge wave parameters.
Nl N2
(l)l (1)2 ky
c K
0 0 0.5 0.5 0.25 1.68 x
102
13
0 1 0.366 0.634 0.134 4.40 x
103
51
0 2 0.309 0.691 0.095 2.28 x
103
100
0 3 0.274 o. 726 0.075 1.44 x
103
160
1 1 0.5 0.5 0.083 1.56
x 103
180
1 2 0.427 0.563 0.063 0.88 x
103
330
1 3 o. 396 0.604 0.052 0.56
x 10
3
520
2 2 0.5 0.5 0.05 0.52
x 103
610
2 3 0.458 0.542 0.041 0.36
x 103
810
3 3 0.5 0.5 o. 035 0.28 x
103
1,200
(l)l = (0.5  p) (l)
(216)
where p is a variable given as 0 < p < 0.5, and the edge wave with the
higher frequency by
(1)2 = (0.5 + p) (l)
(217)
the drift speed, ca, of the nodes and antinodes of the standing wave
is given by
ca = p ~ (218)
ky
where w is the radian frequency of the incident wave and ky the wave
number, 2n/Ly, of the edge wave.
Munk, Snodgrass, and Gilbert (1964) note that because of coriolis
splitting, in general, the frequency of edge waves moving left (looking
seaward) exceeds the frequency of waves moving to the right in the
Northern Hemisphere. In the Southern Hemisphere the opposite would be
true. Therefore, for a uniform straight coastline, the higher frequency
edge waves would display a preference for moving in a particular direction.
Guza and Davis (1974) obtained values for resonant edge wave parameters.
Corrected values were presented by Guza and Bowen (1975), and values of
the parameters are given in Table 3 for various modes of resonance. The
parameters shown in Table 3 are in dimensionless form. The parameter K
defines a critical value of incident wave amplitude by
a2 w k2(w w )1/2
_c_= Y 12
K
(219)
\}
103
where c is a coupling coefficient reevaluated by Guza and Bowen and
given in Table 3 in dimensionless form, ac the dimensional critical
incident wave amplitude, v the dimensional kinematic viscosity, and
the values N
1
and N2 define n for w
1
and w
2
, respectively.
The dimensionless value of ky used in Table 3 and equation (219) is
given by
gB
k* 
y (w*) 2
(220)
where ky is the dimensional wave number of the edge wave defined by
equation (213) and w* the dimensional radian frequency, 2TI/T, of the
incident wave. The dimensionless values of w
1
and w
2
used in Table
3 and equation (219) are given by
w*
1
w*
2.
where wr and w2 are the dimensional values. Table 3 shows that the
number of edge wave pairs would increase as the incident wave amplitude
increases, i.e., the primary pair (N
1
= N
2
= O) would be excited while
the wave is still sorne distance from the shoreline and the other pairs
would be excited closer to the shoreline as the wave amplitude increases.
Therefore, the primary pair of edge waves would experience the greatest
growth.
* * * * * * * * * * * * * * EXAMPLE PROBLEM 16 * * * * * * * * * * * * * *
GIVEN: A tsunami with a period of 20 minutes approaches the shoreline
on a constant shelf slope s
2
= 0.001 (B = 0.001). It is assumed that
the nearshore slope is steep enough for the wave to reflect strongly
from the shoreline.
FIND:
(a) The wave periods and wavelengths of the first three edge wave
pai rs (N
1
= 0
1
, N
2
= 0), (N
1
= 0, N
2
= 1), and (N
1
0, N
2
= 2), and
(b) the wave amplitudes necessary to excite the first three edge
wave pairs.
SOLUTION:
(a) From Table 3
0) w
1
= 0.5 w
~ = 0.5 ~ = 0.5
2
TI
T
1
T 20 x 60
104
(20 x 60)
0 . 5
2, 400 seconds (40 minutes)
0.5 w
T
1
= 2,400 seconds
2rr
o. 366 w = o. 366 r
= ___ T ___ = 20 x 60
Tl 0.366 0.366
T
1
= 3,280 seconds (54.6 minutes)
= ___ T ___ = 20 x 60
T2 0.634 0.634
T
2
= 1,893 seconds (31.5 minutes)
2) w
1
= .0.309 w
=r ___ T ___ = 20 x 60
Tl 0.309 0.309
T
1
= 3,880 seconds (64.7 minutes)
= ___ T ___ = 20 x 60
T
2
0.691 0.691
T
2
= 1,737 seconds (28.9 minutes)
The wavelength of the first edge wave pair is given by equation (213)
where n = 0 éN
1
0, N
2
= 0), ..
g(2n f3
k  (
2
7T )
2

1
 = 6.99 x 10'+
y 2»4_0_0 9.807(1) 0.001
and Ly = 8,990 meters (5.6 miles) for both edge waves.
For the second edge wave pair, N
1
= 0 for w
1
and N
2
1 for
k  ( 2rr )2 1
y  3, 280 9. 8_0_7_(1)_0_._0_0_1
3. 74 x 10'+
10 5
and Ly 16,780 meters (10.4 miles) for the first edge wave
2
k  ( ~ )
1
 3.74 x 10
4
y  1,893 9.807(3) 0.001 
and Ly = 16,780 meters for the second edge wave. Note that the two
edge waves of a pair have the same length as indicated by the single
value for the dimensionless wave nurnber, ky, in Table 3.
For the third edge wave pair, N
1
= 0 for w
1
and Ly
(
21T )
2
__ 1 __
ky = 3,880 9.807(1) 0.001
2.66 x 10
4
23,500 meters (14.6 miles) for the first edge wave
(
21T )
2
1
ky= 1,737 9.807(5) 0.001
4
2.66 x 10
and Ly = 23,500 meters for the second edge wave.
2 for
(b) The incident wave amplitude needed to generate the first edge
wave pair is given by equation (219) as
where K is given in Table 3 and v = 1.5 x 10
2
stokes (square centi
meters per second) (1.6 x ros square feet per second) = 1.5 x ro
6
square meters per second, and w = 2TI/1,200
a = (13 x 1.5 x 106)
112
c 21T
1,200
0.061 meter (0.20 foot)
For the second edge wave pair
ac= el x
1.5 x 106)1/2
21T
1,200
0.12 meter (0.39 foot)
For the third edge wave pair
ac = (JOO x ~ ~ 5 x ro6)
112
\ 1,200
0.17 meter (0.56 foot)
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
106
Guza and Bowen (1975) investigated edge waves generated by incident
waves at sorne arbitrary angle of incidence, a
1
, with the shoreline
(see Fig. 30). Defining a parameter
y = sin a
1
tan S (221)
the longshore wave numbers, k
1
and k
2
, of the primary edge wave pair
(N
1
= 0, N
2
= 0) are given as
kl
w2
(l + 2y) (222)
4g tan
s
and
k2 =
w2
4g tan
s
( 1  2y)
(223)
when the angle of incidence, a
1
, is small, and where w is the radian
frequency of the incident wave. Where a
1
= 0, equations (222) and (223)
reduce to equation (214). The standing edge wave where a
1
> 0 will pro
gress along the shoreline, and the drift speed, ca, of a node or anti
node of the primary edge wave pair is now given as
yw
ca=
2ky
(224)
Gallagher (1971) shows that an increase in the angle of incidence,
a
1
, will produce greater edge wave energy at higher frequencies (shorter
periods).
Guza and Bowen (1976) discuss the height of the edge waves occurring
along a coastline. They show that the maximum edge wave amplitude at the
shoreline is theoretically three times the amplitude of the incident wave
for a straight coastline. Gallagher (1971) indicates that energy would
be lost because of bottom friction and the disperison caused by irregu
larities in the coastline. Guza and Bowen (1976) indicate that edge wave
growth is limited by radiation of energy to deep water and by finite
amplitude demodulation; i.e., as the edge waves increase in height their
natural frequency increases and no longer matches the forcing frequency.
From equation (213) and the work of Munk, Snodgrass, and Gilbert (1964)
relating trapped modes to leaky modes, it can be seen that leaky modes
(i.e., edge waves radiating energy to deep water) will only occur on
steep nearshore slopes . These nearshore slopes are very short in com
parison to the tsunami wavelength, and are not of concern here. The
edge waves associated with the tsunami are assumed to occur over the
wider and flatter shelf slope shown in Figure 33.
A progressive edge wave moving along a coastline may be reflected
from an obstacle such as a large headland. Guza and Bowen (1975) demon
strate that this could produce a standing edge wave with higher amplitudes
near the obstacle. Reflection could also occur from a depth discontinuity
such as a submarine canyon in the manner described in Section VI, 3.
107
5. Refracted Waves and Caustics.
When very long waves such as tsunamis arrive at a shoreline, a
substantial amount of wave energy will be reflected seaward from the
shoreline. These ref!ected waves will interact with the bottom topogra
phy, and will refract as they travel seaward. Refraction diagrams of
these waves show a tendency for the waves to turn parallel to the shore
line as they move into deeper water. When a shelf slopes away from the
shoreline, and extends a sufficient distance seaward, the waves may be
turned back shoreward (see Fig. 35). The line tangent to the wave rays
where they turn shoreward is a caustic. The wave rays will not cross
the caustic, and the wave energy tends to be trapped, although sorne wave
energy will leak across the caustic (Chao, 1970; Chao and Pierson, 1970;
Pierson, 1972).
Figure 35. Schematic of caustic (uniform bottom slope).
Chao (1970) and Chao and Pierson (1970) investigated higher frequency
waves trapped by a caustic. They demonstrate that lower frequency (longer
period) waves will form caustics closer to the coastline, and that waves
with frequencies above sorne maximum value will propagate seaward into
deep water. For tsunamis, only the lower frequency waves are significant.
As the wave rays are not normal to the shoreline, different parts of
the wave crest would arrive at the shoreline at different times. Where a
coastline is irregular, parts of a wave crest reflected from one section
of coastline may be refracted and trapped so that they coincide with an
incident wave on another section of coastline. Palmer, Mulvihill, and
Funasaki (1965) illustrated the effects of wave trapping at Hilo, Hawaii,
where the reflected wave rays were turned by refraction so that they
arrived simultaneously at a point inside Hilo Harbor (see Fig. 36).
The case of wave energy being trapped by refraction can be most
easily illustrated for a long, shallowwater wave on a straight section
of shoreline, with sorne water depth, d
8
, at the toe of the shoreline
slope, and with a constant shelf slope extending seaward. It is assumed
that the wave reflects from the shoreline slope and refracts on the shelf.
108
Figure 36. 1960 tsunami refraction, Hi1o,
Hawaii (after Palmer, Mu1vihi11,
and Funasaki, 1965).
Appiying Sneii's Law where the wave maves through sorne incrementai
distance, with an incrementai change in water depth d, and assuming
that there is a shaiiowwater wave so that the ceierity C is given by
c = (gd) l/2
then the incrementai refraction of a wave ray is defined by
sin(a. + da.) = (d +d dd)
112
sin a.
or, squaring both sides of equation (226),
sin
2
(a. + da.)
sin
2
a.
109
(d + dd)
d
(225)
(226)
(227)
Now, if the numerator of the left side of equation (227) is expanded,
the term can be written
sin
2
(a + da) =
1
[1  cos 2(a + da)]
2
1
[1  cos(2a) cos(2da) + sin(2a) sin(2da)]
2
+ 2 sin a cos a(2 sin da cos da)]
But, where da ~ 0,
cos da ~ 1
sin da ~ da
Then equation (228) can be written as
sin
2
(a +da) = l [1 (cos
2
a sin
2
a)(l) + 4 sin a cos a da]
2
= l [2 sin2a + 4 sin a cos a da]
2
and equation (227) becomes
which reduces to
2 sin
2
a + 4 sin a cos a da
2 sin
2
a
(d + dd)
d
1 + 2 cot a da
1 + dd
d
2 cot a da = dd
d
(228)
(229)
(230)
(231)
(232)
Now, integrating along the wave ray from the shoreline to the point where
it turns parallel to the shoreline, taking ~ as the water depth where
the ray is parallel to the shoreline and ~ as the distance from the
shoreline at that point, a
1
as the initial direction of the wave ray at
the shoreline (Fig. 35), and noting that a= n/2 radians for a straight,
uniform coastline at the point where the wave ray turns parallel to the
bottom contour,
110
(ddp ddd lrr/2
),l 2 cot a da
s al
which when integrated gives
2 fu(sin
which reduces to
fu d fu d 2 fu sin
al
p s
Taking the antilogs
d
d
8.
p
sin
2
al
But cp
ds
+ Sxp where s is the bottom
x = 
p S sin
2
a
1
slope,
d
s
s
(233)
(234)
(235)
(236)
so
(237)
To compute the coordinate parallel to the shoreline of a point on the
wave ray, note that
tan a =
dy
dx
The x coordinate is given by
d sin
2
a d
s s
x=.
S sin
2
a
1
S
or, differentiating equation (240) with respect to a,
2 d sin a cos a da
dx =
s
(238)
(239)
(240)
substituting equation (240) into equation (238), equation (238) becomes
2 d
dy = 
8
 sin a cos a tan a da
S sin
2
a
1
111
(241)
Collecting terms and integrating
l
y/2 2 ds f rr /2
dy =  sin
2
a da
0
S sin
2
a
1
a
1
This gives
y
 =
2
__ 2____,d..=.s_
S sin
2
a
1
sin
4
2a
1
]
(242)
(243)
1hese equations are limited to the particular case of a long, straight
coastline, but may provide a first approximation for solutions on sorne
sections of continental shelves. Refraction diagrams would be required
to obtain exact solutions for irregular coastlines. If the waves travel
for long distances over a shelf, it may be desirable to use wave refrac
tion equations in spherical coordinates such as the equations given by
Chao (1970) (see Sec. IV, 3).
* * * * * * * * * * * * * * EXAMPLE PROBLEM 17 * * * * * * * * * * * * * *
GIVEN: A wave ray reflects from a straight shoreline at an initial angle
rr/4 radians. The water depth at the toe of the shoreline slope
= 30 meters and the shelf at the toe of the shoreline slope has a
uniform seaward slope S
2
= 0.003.
FIND: The distance the wave ray will travel away from the shoreline, and
distance along the shoreline to the point where the reflected wave
ray will impinge upon the shoreline.
SOLUTION:
From equation (237)
d d
____ __;:_s___ s
30
0.003 sin
2
t
30
0.003
Xp 10,000 meters (6.214 miles)
From equation (243)
y
2
r.=
2
y
2
2 d [ a sin
4
2a
1
]
_____ s;::;.___ _4rr  _21 + 
S sin
2
a
1
T]
25,700 meters (15.97 miles)
112
For this example, the shelf needs to extend 10,000 meters from the
toe of the shoreline slope to have the wave ray turn parallel to the
bottom contours. The wave ray which was reflected from the shoreline
will impinge upon the shoreline again at a point 51,400 meters (31.94
miles) along the coast, provided the wave is trapped.
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
For the particular case of a shallowwater wave on a straight section
of coastline and uniform shelf slope, given by equations (237) and (243)
and illustrated in Figure 35, as the angle a
1
decreases the distance of
the caustic from the shoreline increases and the distance y between the
point of reflection and the point where the wave ray impinges again on
the coastline also increases.
When the tsunami energy becomes trapped between a caustic and a
coastline, the energy will tend to propagate along the coastline. This
will excite longshore edge waves along the coastline, and may substan
tially increase observed wave heights. When the coastline is irregular,
the trapped waves may concentrate their energy at particular coastal
points. An investigation of the wave rays using the usual wave refraction
techniques will define the caustic locations, and the locations of any
coastal points where energy concentrates.
Tsunamis generated in coastal areas may have part of their energy
trapped along the coastline, as waves radiating away from a source area
may become trapped within a caustic in the same manner as reflected waves.
For a wave ray originating within the coastal area, d
8
is the water
depth at the point of origin, Xp the distance seaward from the point of
origin, and a
1
the angle between the wave ray and the orthogonal to the
bottom contours as before. This is illustrated in the following example
problem and in Figure 37.
Figure 37. Trapping of generated tsunami.
113
* * * * * * * * * * * * * * EXAMPLE PROBLEM 18 * * * * * * * * * * * * * *
GIVEN: A tsunami is generated by uplifting along a faultline located on
a coastal shelf (Fig. 37). The faultline is oriented so that the angle
between the coastline and the faultline is 8 = 40°. The faultline
extends to the outer edge of the shelf which is 100 kilometers (62.14
miles) from the toe of the shoreline slope. The shelf has a uniform
slope seaward of S = 0.003 and a 30meter water depth at the toe of the
shoreline slope. The uplifting is uniform along the faultline so that
it acts as a line source.
FIND: The percentage of the wave energy trapped on the shelf.
SOLUTION: The depth of water at the outer edge of the shelf is
d d
8
+ Sx = 30 + 0.003(100,000)
330 meters (1,083 feet)
Looking first at wave rays traveling seaward from the faultline, at the
inner end of the faultline, using equation (237),
d d
s s
~
s sin
2
al
s
30 30
 
0.003 sin
2
40° 0.003
xp = 14,200 meters (8.83 miles)
At the outer end of the faultline, taking d
8
as the depth of water
at the faultline, i.e., where the wave ray originates
330 330
0.003 sin
2
40° 0.003
Xp = 156.23 kilometers (97.08 miles)
which is beyond the 1imits of the she1f. Noting that Xp varies
1inear1y with d
3
, from proportionality the 1ength ~ of the part
of the fau1tline (of 1ength Lf) contributing trapped energy is
Lf(100  14.2)
(100 + 156.23  14.2)
~ = 0.355 Lf
or 35.5 percent of the energy generated seaward in the examp1e is
trapped.
114
Consider now the wave rays generated shoreward . If the wave ray
generated from the point at distance 2 along the faultline is con
sidered for the straight, uniform section of shoreline (Fig. 37), the
wave ray will reflect from the shoreline and be directed at the angle
a
1
at the position shown in the figure. This wave ray will turn
parallel to the shoreline (i.e., parallel to the bottom contours) at
the edge of the shelf. Therefore, 35.5 percent of the energy gener
ated shoreward in the example will also be trapped. Note that the
energy generated shoreward would have a tendency to form a Machstem
along the shoreline if S is greater than 45°.
A caustic, by definition, is a line tangent to a family of wave
rays. For the waves generated seaward, a caustic will be formed by
those wave rays refracted back to the shoreline (i.e., the trapped
wave rays), after they have reflected from the shoreline. For the
wave rays generated landward from the faultline, a caustic is formed
after reflection as shown by the dashline in Figure 37.
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
As illustrated in Figure 37, wave rays trapped on a shelf may diverge
apart. These wave rays may reconverge at various points along the coast
line, producing high waves at the points of convergence.
Chao (1970), Chao and Pierson (1970), and Pierson (1972) discuss the
case of shortperiod waves reflected from a shoreline, where severa! wave
crests exist between the point of reflection and the caustic. The wave
rays follow similar paths to those discussed above, but the wave crests
propagating shoreward from the caustic will interact with the wave crests
propagating seaward f r om the coastline, producing many peaks and nodes
between the coastline and the caustic.
For the straight coastline shown in Figure 35, the traveltime, t,
along the wave ray between the shoreline and the caustic can be easily
determined. Where the distance s is measured along the wave ray,
but
so that
ds
c =
dt
d
~
s
cos a
rOt dt = fOXp _dx __
J l c cos a
Also, from shallowwater assumptions,
c = /gd= lg(d
8
+ Sx)
115
(244)
(245)
(246)
(24 7)
and from equation (240),
dx 2 d
8
sin a da
 = 
s sin
2
a
1
(248)
cos a
Using equation (239) to define x and substituting equations (247) and
(248) into equation (246), it becomes
lrr/2 2
d
sin çx dx
t =
s
a S sin
2
al
lg
t
(d sin2 a
;s)
l
s s
+
\j s
S sin
2
a
1
(249)
Collecting terms,
(250)
which gives
t (251)
Shen and Meyer (1967) and Shen (1972) give a solution for curved
coastlines. For a circular arc, wave trapping can be defined using the
equations
tan a
rde =
±(n2r2 _ c2)l/2 (252)
dr
and
n tanh [n(:
2
) d]
1
(253)
where
r the radius of curvature of a contour line (taking circular
bottom contours to define the shelf around the coastline)
e the coordinate angle in polar coordinates of a point along
the wave ray at radius r
c a constant
n a variable along the wave ray
d the water depth at radius r
1, 2, 3 ... defines an integer number of wave crests around
a circular island
116
Taking the radius at the shoreline as R
8
, and the radius wher e the
wave ray turns parallel to the bottom contours as rp• at the shoreline
(254)
(255)
1 (256)
Where the wave ray turns parallel to the bottom contours
0 (257)
c
(258) n "'
r
p
Substituting into equation (253),
(259)
where d is the water depth at radius rp. Equation (256) can be solved
to determine c for any integer value n
2
satisfying the equation, and
equation (259) can then be solved to obtain a value for rp corresponding
to each value of c which will provide a solution. Shen and Meyer (1967)
indicate that a number of caustics may exist. For a circular island, reso
nance will occur between the shoreline and the caustics for wave periods
defined by integer values of n
2
for which solutions exist.
Equations (252) and (253) were derived for dimensionless variables
where the dimensional values of length had been divided by sorne horizontal
length scale. Camfield (1979) gives the following development to express
the solution in terms of several dimensionless parameters. Shen and Meyer
(1967) infer that the radius rp of the caustic (i.e., where the wave ray
turns parallel to the bottom contours) is an appropriate length scale.
Using an asterisk (*) to define dimensional values, the radius of curva
ture rp is normalized so that its dimensionless value is
r = 1
p
The derivation of the equations also assumes that
vertical length scale = E
horizontal length scale
117
(260)
(261)
so the vertical dimensions in equations to (259) are to
represent dimensional values divided by rp E, where the term rp E is
the vertical length'scale. Shen (1972) taKes EM = 1. It is assumed
E + o, where 0 < o << 1, so that M must be large. Shen defines M as
M
which gives
(w*)2(horizontal length scale)
g
(262)
For longperiod waves (e.g., tsunamis) where the period T is large,
the caustic radius rp must be large in arder for M to be large. In
general, the solution is for cases where the shoreline radius R; (and
therefore the caustic radius rp) js much greater than the wavelength.
Using Shen's work, equation (236) now reduces to
1 (263)
Shen has defined M = n
2
/c so that E
reduces to
1/M = c/n
2
. Equation (263) further
c2 tan2 a
1
)l/
2
+ c2 tan2 a
1
)l/
2
ct;] R;
tanh  =
tan
2
a tan
2
a c
2
R* r*
l l s p
(264)
in its dimensional form, where ct; is the dimensional depth at the toe
of the shoreline slope, the dimensional·radius of curvature of the
shoreline, and rp the dimensional radius to the point where the wave
ray turns paralle1 to the bottom contours. Also the dimensional form of
equation (259) becomes
d*]
c tanh
r* c
p
1 (265)
where d* is the water depth at radius rp.
Consider first a concave coastline, such as a large bay, where the
radii r:p and R; are measured from the center of curvature offshore,
and rP. < R;. In the limiting case, a concave coastline would forma
closeo circular basin with radius R;. Therefore, all wave rays could
obviously be trapped by this type of coastline.
118
Where a
1
~ 90° the wave rays become trapped very close to the shore
line so that rp ~ R;. As a
1
becomes smaller, rp becomes smaller so
that the caustics are farther from the shoreline . As a
1
~ 0, rp ~ 0,
meaning that a wave ray reflected along an orthogonal to the shoreline
will pass through the center of curvature.
Equations (264) and 265) are used to investigate wave rays at any
angle a
1
. From equation (265), c ~ 1 as tanh [ n ~ d*/Crp c)] ~ 1. To
determine the limiting values of n
2
which will provide solutions, it
may be noted that, in equation (264),
R*
s > 1
r*
p
(266)
From equation (265), as c ~ 1, nz ~ oo, From the definition that 1/M œ T2
(eq. 262), and that 1/M = c/n
2
, 1t can be seen that waves would be trapped
where T ~ 0, which is a restatement of the fact that all waves would be
trapped where the coastline is concave; e.g., a large, circular bay.
Finding the caustic location, i.e., the radius, rp, when n ~ oo and T ~ 0
is of interest. From equation (264), when n ~ oo, and a
1
> 0,
which then gives
and as c ~ 1,
which reduces to
~
_/ l/2 ]
(
1 + c
2
tan
2
a
1
) n
2
d*
tanh tan2 al c ~ R; ~ 1
(r*) . = R* sin a
p ~ n s 1
R*
~ ~
r*
p
(267)
(268)
(269)
(270)
Where the angle a
1
is known, defining the angle between the reflected
wave ray at the shoreline and the normal to the shoreline (see Fig. 35),
the wave energy will always be trapped between the radius, r;, defining
the caustic and the radius, R;, defining the shoreline if the concave
shoreline extends a sufficient distance. Where the wave period becomes
longer, it will be trapped closer to the shoreline.
It is of interest to note that equation (270) provides a solution
independent of water depth or shelf slope. Equation (270) defines the
distance from the center of curvature to a chard across a circular arc,
where a
1
is the angle between the chard and a radius drawn to the end
119
of the chord. This defines the path of an unrefracted wave ray (the
expected result when T + 0), and therefore verifies equations (264) and
(265) for a concave coastline such as a large bay.
For a convex coastline (e.g., a circular island), the wave rays
which would probably be trapped are those where a
1
is large, i.e.,
the wave rays most nearly parallel to the shoreline. Letting a
1
+ n/2,
tan a
1
+ ~ so that the term
(271)
From equation (264)
(
n ~ d;) R;
c tanh  =
c R* r*
s p
(272)
Substituting equation (265) in equation (272) above,
tanh ( n ~ d;) = R; tanh ( n ~ d*) (273)
c R; r:p c r:p
as rp > R; for a convex coastline, then R;/rp < 1. Therefore,
d* d;
r* > R* (274)
p s
as a condition of wave trapping on a convex coastline. This means that
the slope of the shelf must be greater than sorne minimum value defined
by d ~ / R ~ in order to have a caustic, i.e., to have waves trapped on the
shelf. This is necessary in order to have the rate of curvature of the
wave ray exceed the rate of curvature of the bottom contours, a necessary
condition of wave trapping. Where a circular island has a small radius,
R;, in relation to the water depth at the shoreline, d$, there is a
greater probability of the wave rays spiraling off into deep water than
there would be for an island with a large radius.
The minimum and maximum values of n
2
which will produce solutions
can be found as follows:
From equation (264), where Ici + ~
c tanh ( n ~ d;) = R;
c R* r*
s p
tan a
1
> 0 + o
(275)
120.
where o is sorne small value. But as c ~ oo,
tanh
and from equation (275)
(
n2 d*) n2 d*
2 s 2 s
 ~   
c R* c R*
s s
(R*) 2
s
d* r*
s p
From equation (265), noting, that for c ~ oo,
it is found that
[
n ~ d*] n ~ d*
tanh  ~ 
r* c r* c
p p
n ~ d*
;; = 1
p
(276)
(277)
(278)
(279)
(280)
Substituting the value of
a uniform slope
r*
p
from equation (278), and noting that for
d* d* + (r*  R*) S
s p s 2
(281)
equation (280) is then
n
4
(d* R* S  (d*)
2
]  n
2
(R*)
2
S + (R*)
2
= 0
2 s s 2 s 2 s 2 s
(282)
Equation (282) is a quadratic equation for n ~ which provides minimum
and maximum values given by
R*
(n2) 2.
s
rmn
R* s  d*
s
2
s
(283)
(R*)
2
S
s 2
 R*
d* s
(n2) ~ a x
s
R* s  d*
(284)
s 2 s
121
At the minimum value of n
2
,
d*
n2 2.. =
2 R*
s
and at the maximum value of n
2
,
d*
n2 s
2 R*
s
d*
s
R*S
s 2
(285)
(286)
For a solution to exist for given values of and the maximum
value of defined by equation (285) must be less than or equal
to 1, so the minimum value of s
2
for wave trapping to occur is given by
which reduces to
d*
s
d*
1 __ s_
R; s
2
d*
1

8
 < 0.5
S R*
2 s
(287)
(288)
The parameter d;/CS
2
R;) is a shelf parameter which .determines wave
trapping. A continuous band of solutions exists for equations (264) and
(265), for values of n
2
between the minimum and maximum values defined
by equations (283) and (284). Solutions are not limited to integer values
of n
2
, which define resonant periods for a circular island.
Solutions for equations (264) and (265) are plotted in Figure 38.
Equation (264) is plotted on lines of constant equation (265) is
plotted on lines of constant S
2
). Solutions for trapped waves are
obtained where the two families of curves intersect. From equation (288)
it is seen that solutions will only exist where d;/CR; S
2
) From
equation (262) note that
21T M
gT2 = 21T rp
(289)
122
t'
N
w
1.0
0 .8
R*
s
rr o.s
0.4
0.2
01

2 3 4 5 6 7 8 9 10
c
a. a, =5o
Il 12
__ l
;·::=t:·:::_:::.;t :::... ··::::: +
13 14 15 16 17
Figure 38. Solution to equations (264) and (265).
ï:__:=
.;
oc;""'" ]"
18 19 20
t'
N
0.3
IJ!.
.
'S·U' ..
Ill
.
r "·' •
... , ..
·.· .. i
· ,,. 1 .,ë,,, L. . ··1 1·. ,,: J .J, .1 ... ; '' "'.;,o;.
20
c
b. a
1
= 10°
Figure 38. Solution to equations· (264) and (265).Continued
1.0
0.9
0.8
0.7
0.6
R*
rj 0.5
t'
1'J
0.4
Vl
0 .3
0. 2
0. 1
c
c. al = 15o
Figure 38. Solution to equations (264) and (265).Continued
! 1
0.9 .

1  +  +  + · H ~ 11
Figure 38 . Solution to equations (264) and (265).Continued
126
1.0
0.9
0.8
0.7
0.6
R*
s
110.5
0.4
0.3
0.2
0.1
0

·
 
f.l,. l
1t
. :
.;< L . ,..

1 :i)\ rr. ' 1(1 t t
.....
1
i
"· T 1 { _ j 
H2

,; .P<
fo'
j1'
•t ; ' · _  .,._
.
_,_
1 

v/  C!
O.
,. :'

 '
1"
 
Ir 1 0
;;
 'it
rr 
, ,..
,_ ,il
! • : 1': 1 '"'
t
   
Il R
r. Ill
... .A.
1
. 1\
1
U lW
_f
 11
H

.
' '"()
    
t ·:\ yV;r . "!4 J"
::
 . 1 
H

ri' . ·
. ..
·ttt :P< . ' [1 't t J
·
t' ,. ' c
H
.
'
. 
1
,. l '
,.
·

l '+ i j r:·i + ·
"


;a
.l

1J ' ........ 1 H r:·
T
"'
; . 
'' 1 ,_ _   . 
+
f"
+
"
t

'
j
N
J.l
i 1
1
' L.[{J\tr
... i
i
1
.
1 •. 
.. t
1

__._
lb" 
1
k r1
.1±llJ.li'! 1
.
jl
"'
··
,_.
Il'''
l±r
L
 Il !
1
•
11 1
i
j
1
II. .
!
tt1 Ir'  1  +l i  il i i1
1 
11 ff
!tt
•
1;t
:tt wr :: .ri ;;
·jHI
__ _ ·f:f .. ·'· If '
  i·
1 2 3 4 5 6 7 8 9 10
Figure 38 . Solution to equations (264) and (265).Continued
127
and as M n
2
/c, as previously defined, this can be rewritten as
(290)
which reduces to
21T d* n (R*)\n2 d*) s 2 1 s 2 s
gT
2
= 21TC r*
p s
(291)
Lines of constant 21T n
2
/(gT
2
) are also plotted in Figure 38. The
minimum period of the trapped waves is defined where a line of constant
21T n
2
/(gT
2
) is tangent to the line where
is constant for
the given values of s
2
, and Solutions for equations (264)
and (265) at greater values of R;/rp (at longer wave periods) define
trapped waves. Solutions for equations (264) and (265) for smaller
values of (at longer wave periods) define the damping zone dis
cussed by Lozano and Meyer (1976).
Solutions of equations (264) and (265) for values of n
2
, defined
by equation (286), define caustics at the inner limit of the trapped
wave zone near the shoreline (i.e., where rp These solutions
will give the maximum trapped wave periods, Tmax• but the solutions
tend to break down at this point as the parameter, U, defined by
equation (66) as U = (H/d)(L/d)
2
, becomes very large (U >> 1). However,
the minimum trapped wave period and the outer limit of the trapped wave
zone can be approximated using equations (264) and (265).
Figure 38 shows that, for varying values of a
1
, the m1n1mum trapped
wave period will increase as a
1
decreases. This is expected since
shorter period waves, at lower values of a
1
, tend to pass into deep
water and are not trapped.
The theoretical solutions given by equations (264) and (265) are for
the case of a coastline approximated by a circular arc. In the limiting
case, this will approach a straight coastline. Solutions for irregular
coastlines must be obtained by numerical methods. The theoretical solu
tions presented here can be used to verify numerical methods used for
more complex solutions for irregular coastlines. An example of a numeri
cal solution is given by Houston, Carver, and Markle (1977) using a
finiteelement numerical model developed by Chen and Mei (1974).
* * * * * * * * * * * * * * EXAMPLE PROBLEM 19 * * * * * * * * * ·k * * * *
GIVEN: A curved section of coastline is convex, with a radius of curva
ture = 100,000 meters (62.14 miles), and the depth at the toe of
the shoreline slope = 30 meters. A tsunami reflects from the shore
line slope and refracts over a shelf where the bottom slope of the shelf
128
is s
2
= 0.003. The angle between the reflected wave ray and the
orthogonal to the shoreline a
1
= 45°.
FIND:
(a) The minimum shelf slope required for trapped waves to exist,
(b) the radius,
0
rp, defining the outer limit of the trapped wave
zone where a
1
= 45 , and
(c) the minimum period of waves trapped by refraction, where a
1
45° .
SOLUTION:
(a) From equation (288), for wave trapping to exist
d*

8
 < 0.5
s
2
R;
d* 2(30)
s
2
> 2 _!!_ =
0.0006
R; 100,000
(b) The shelf parameter is
d*
s
=
S R*
2 s
30
0.003(100,000)
30
300
0.1
From Figure 38, for a
1
45°, the minimum wave period is at
where
and
0.068
n
2
d*
~ = 0.71
R*
s
R* 100,000
0. 71 d! = 0. 71 30
s
n
2
= 48.65
R*
~ = 0.93
r*
p
129
2,367
which gives
R*
r* = =
p 0.93
100,000
 = 107,530 meters (66.8 miles)
0.93
The width of the trapped wave zone from the shoreline to the outer limit
is given by
r* 
R* 107,530  100,000 = 7,530 meters (4.68 miles)
p 8
d*
n2
d;
n2
(c)
s
0.068 or T
2
=
gT2
0.068 g
T2
= 13,747
9.81(0.068)
T = 117.2 seconds (1.95 minutes)
for the minimum trapped wave period.
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
6. MachStem Formation.
Figure 38 illustrates solutions for trapped waves for angles a
1
45°.
Perroud (1957) showed that a
1
= 45° defines a critical angle for wave
reflection. When a
1
< 45° regular reflection occurs, i.e., the wave
reflects in a manner described in Section VI, 1 and S. When a
1
= 45° the
end of the wave crest at the shoreline turns perpendicular to the shore
line (see Fig. 39). Regular reflection no longer occurs when a
1
> 45°.
Perroud showed that, for a
1
> 45°, the incident wave produces two
components. The first is a reflected wave, lower than the incident wave,
and with the angle, a
2
, between the reflected wave ray and the normal
to the shoreline defined by a
2
< a
1
• The second component is a Mach stem
which moves along the shoreline in the direction of the longshore compo
nent of the incident wave, growing in size as it progresses along the
shoreline. Figure 39 shows the initial growth of a Mach stem along a
vertical wall for the critical angle a
1
= 45°.
Experimental measurements by Perroud (1957) show that the Mach stem
has a profile at the shoreline similar to the profile of the incident
wave, giving the Mach stem the appearance of a large wave moving along
the shoreline. The Mach stem remains attached to the shoreline end of
the incident wave crest, so its speed of propagation, along the
shoreline is given as
c =
c
sin a
1
where C is the celerity of the incident wave near the shoreline.
130
(292)
Refleeted Wove Crest
Figure 39. Machstem formation, solitary wave. Lines of equal
surface elevation above still water normalized to unit
incident wave amplitude (after Perroud, 1957).
Chen (1961) studied Machstem development using a range of values
for the nearshore slope. He showed that where the angle of the nearshore
slope S < 60°, and a
1
> 55°, the Mach stem formed a breaking wave along
the shoreline. The relationship between the wavelength and the slope
length was not considered in this case, and was not varied during the
experiments.
lt was generally found by Perroud (1957) and Chen (1961) that the
incident waves neither reflected from the shoreline nor formed Mach stems
when a
1
> 70°. This would be the case for an incident wave traveling
nearly parallel to the section of shoreline; e.g., a wave entering an
inlet with a gradually varying cross section.
7. Bay and Harbor Resonance.
When a bay or harbor is very long in relation to the tsunami wave
length, the tsunami may cause resonance if a natural mode of oscillation
of a bay or harbor corresponds to the period of the tsunami. Murty,
Wigen, and Chawla (1975) have tabulated the approximate periods of inlets
on the Pacifie coast of North and South America based on the formula
(293)
131
where T
1
is the primary period, Lb the length of the inlet, and
da the average depth of the inlet. Values of length, depth, width,
period, and relative intensity of secondary oscillations of the water
level, as given by Murty, Wigen, and Chawla, for inlets on the coast of
Alaska and British Columbia, and for Puget Sound, are given in Table 4.
These values are only approximate because variations in inlet cross
section, restricted entrances, and the effects of branched inlets are
not considered.
Referring to the work of Nakano (1932) which showed secondary undu
lations to be proportional to the length of an inlet, Lb, and inversely
proportional to the width, B, and to d ~ /
2
, Murty, Wigen, and Chawla
(1975) proposed that the relative intensity, I, of the secondary undu
lations could be given as
I (294)
Values of I for inlets of Alaska, British Columbia, and for Puget Sound
are shawn in Table 4. Inlets with higher relative intensities, I, would
be expected to excite larger amplitudes of oscillation. As indicated by
Murty, Wigen, and Chawla, sorne hays which have small ratios of Lb/B have
large secondary oscillations. They point out that equations (293) and
(294) are based on a onedimensional theory which is not valid for law
ratios of Lb/B, and that transverse motion is important in these cases.
Fukuuchi and Ito (1966) consider a tsunami passing from a larger bay
or inlet into a smaller inlet. Where the larger inlet has a width B
1
,
and the width narrows to a width B
2
in the smaller inlet (see Fig. 40),
they give the amplitude a
2
at the head of the smaller inlet as
(295)
where a is the incident tsunami amplitude in the larger inlet, T the
per1od o} the tsunami, and T
1
the period of the smaller inlet as given
by equation (293). The maximum amplitude a
2
will occur when T
1
/T =
1, 3, 5 ... while the minimum amplitude would be at T
1
/T = 2, 4, 6 ...
etc. Equation (295) would predict very high values of a
2
/a
1
where B
1
/B
2
is large. This is not consistent with the work of other investigators.
Ippen, Raichlen, and Sullivan (1962) carried out a hydraulic madel
investigation of an inlet connected to an "infinite ocean." The ocean
was simulated in a wave basin, using wave absorbers to minimize reflected
132
Table 4. Dimensions, periods of fundamental mode, and intensity of secondary
undulations of inlets of Alaska and British Columbia, and of Puget
Sound (from Murty, Wigen, and Chawla, 1975).
In let
Alaska
Tarr Inlet
Glacier Bay
Muir Inlet
Lynn Canal
Gastineau Canal
Taku Inlet
Stephens Passage
Tracy Arm
Endicott Arm
'Frederick Sound
Thomas Bay
Tenakee Inlet
Peril Strai t
Bradfi e ld Canal
Ernest Sound
Behm Canal West
Bell Arm
Burroughs Bay
Behm Canal East
Rudyerd Bay
Boca de Quadra
carroll Inlet
Georse Inlet
British Coli.Uilbia
Portland Canal
Cllservatory Inlet
Hastings Arm
Alice Arm
IChutzeymateen Inlet
Work Channel
Prince Rupert Inlet
Douglas Channel
Kildala Arm
Gardner Canal
Surf Inlet
Laredo Inlet
5heep Passage
Musse! lnlet
Spiller Channel
Roscoe Inlet
Cousins Inlet
Cascade Inlet
Dean Channel
Kwatna Inlet
South Bentinck Arm
Rivers Inlet
loklses Inlet
~ ·
length
(km)
Ill
35
146
18
133
43
44
80
20.
64
71
80
72
ll3
22
56
44
22
ll5
76
19
25
54
19
83
19
91
22
39
33
46
43
12
26
Ill
24
37
46
26
da·
mean depth
(km)
0.220
0.215
0.360
0.040
0.295
0.270
0.260
0.165
0.150
0.140
0.210
0.310
0.425
0.420
0.170
0.245
0.130
0.225
0.255
0.385
0.240
0.120
0.240
0.045
0.330
0.175
0.275
0.220
0.295
0.275
0.255
0.135
0.070
0.250
0.420
0.345
0.240
0.295
.0.200
133
Period,
Tl
(min)
159
51
164
61
165
56
58
133
35
ll5
104
97
74
174
36
76
82
31
153
82
26
49
74
60
97
31
ll7
32
48
42
61
79
31
35
ll5
28
51
57
39
B,
mean width
(km)
5.6
3.5
6.6
1.3
13.0
1.8
3.3
22.2
2.8
3.2
4.0
5.7
5.2
3.5
0.9
1.3
1.6
1.4
2.2
2.2
1.3
1.0
2.0
1.2
3.5
1.5
1.9
0.9
1.5
1.5
1.9
1.1
0.8
1.1
2.4
2.0
2.2
3.0
0.9
~ / 8
19.8 192
10.0 100
22.1 102
13.8 1,725
10.2 64
23.9 170
13.3 100
3.6 54
7.1 122
20.0 382
17.8 185
14.0 81
13.8 50
32.3 119
24.4 348
43.1 355
27.5 587
15.7 147
52.3 406
34.5 144
14.6 124
25.0 601
27.0 230
15.6 1,634
23.7 125
12.7 173
47.9 332
24.4 236
26.0 162
22.0 153
24.2 188
39.1 788
15.0 810
23.6 189
46.3 170
12.0 59
16.8 143
15.3 95
28.9 323
Table 4. Dimensions, periods of fundamental mode , and intensi ty of secondary
undulations of in lets of Alaska and British Columbia, and of Puget
Sound (from Murty, Wigen, and Chawla, 1975) .Continued
Lz,>
da,
Period, 8,
Io/B
Inlet length mean depth
Tl
mean width
io/B
d3/2
(km) (km) (min) (km)
a
Smith Inlet 33 0.270 43 1.3 25.4 181
Mereworth Sound 19 0.090 43 0.4 47.5 1, 759
Belize In let 52 0.255 69 1.1 47.3 367
Nugent Sound 24 0.075 59 o. 7 34.3 1,669
Seymour Inlet 67 0.420 70 1.7 39.4 145
Drury Inlet 22 0.040 74 1.3 16.9 2,112
Knight Inlet 130 0.295 161 3.0 43.3 270
Ca11 Inlet 28 0.135 51 1.5 18.7 377
Loughborough In! et 35 0.190 54 1.7 20.6 249
Bute Inlet 76 0.510 72 3. 7 20.5 56
Toba Inlet 37 0.390 40 2.6 14.2 58
Jervis Inlet 89 0.495 85 3.2 27.8 80
Howe Sound 43 0.225 61 7.0 6.1 57
Vancouver Island
British Cohunbia
Holberg Rupert In let 44 0.165 73 1.4 31.4 469
Quatsino Sound 59 0.150 103 2.2 26.8 461
Neroutsos Inlet
Forward Inlet 11 0.030 43 1.1 10.0 1,925
Klaskino Inlet 11 0.035 40 0.7 15.7 2,398
Ououkinsh Inlet 14 0.085 32 1.2 11.7 472
Port Eliza 11 0.050 33 o. 7 15.7 1,404
Espinosa Inlet 14 0.215 20 1.3 10.8 108
Nuchali tz Inlet 15 0.025 64 1.3 11.5 2,909
Tahsis Inlet 29 0.120 56 0.9 32.2 775
Cook Channel 31 0.150 54 1.9 16.3 281
Tlupana Inlet
Zuciarte Channel 48 0.220 69 1.5 32.0 310
Mechalat Inlet
Sydney Inlet 20 0.080 48 1.3 15.4 681
Shel ter In let 19 0.115 38 1.3 14.6 374
Herbert Inlet 23 0.100 49 2.0 11.5 364
Pipestem Inlet 9 0.045 29 o. 7 12.9 1,351
Effingham Inlet 17 0.095 37 1.2 14.2 485
Alberni Inlet 69 0.145 122
1.3 53.1 962
Saanich Inlet 23 0.180 37 2.5 9.2 120
Puget Sound
Puget Sound 111 0.165 184 6.0 18.5 276
Hood Canal 102 0.110 207 2.5 40.8 1,118
Possession Sound 70 0.090 157 3.7 18.9 700
Saratoga Passage
134
f
Bz
'
Lb
Figure 40. Plan view of inlet.
1
1
waves. It is assumed in this case that B
1
+ oo. The experimental results
of Ippen, Raichlen, and Sullivan are shown in Figure 41, where k is the
wave number 2 ~ / L . The results are for a fixed inlet width and varying
wavelength, the·variation in the curves illustrating the dependence of
the results on the ratio of wavelength to inlet width, particularly for
short, wide inlets.
Each curve in Figure 41 was obtained by varying the inlet length for
a fixed wavelength. The results were dependent upon the efficiency of
the wave filters and wave absorbers used in the experiments. Using
equation (293) to define T
1
,
(296)
Therefore, Figure 41 shows that the maximum amplification occurs where
T
1
/T < 1. Ippen, Raichlen, and Sullivan did not explore the amplifica
tion for shorter period waves; i.e., T
1
/T ~ 3, 5, 7 ... etc. The maxi
mum amplification occurring where T
1
/T < 1 is equivalent to resonance
for a longer inlet. It can be assumed, therefore, that the inlet has an
effective length Le extending into the open sea; i.e., since a node
does not exist at the entrance, Le > ~ and the effective primary period,
T le, is
The length Le, is defined by equation (297) if it is assumed that
T
1
e/T = 1 where maximum amplification occurs.
(297)
Nishimura, Horikawa, and Shuto (1971) carried out similar experiments
for an inlet with the entrance partially closed by a breakwater. They
also found that the inlet had an effective length greater than the actual
length. Ippen, Raichlen, and Sullivan (1962) and Nishimura, Horikawa, and
135
2.2
1.8
1.4
02
a,
1.0
Absorber No. 1
Filter No. 1
Depth=25.71 cm
0.6 Harbor width fixcd
œ
k=4.907
b=5.72 cm
e k=4.290
Harbor length varies
()
k=4.157
0 k=4.119
e k=3.427
00
0.5 1.0 1.5
klb
Figure 41. Amplification factor versus rel ative
harbor length (from Ippen, Raichlen,
and Sullivan, 1962).
Shuto all indicate that the effective length may be determined by the
r atio of inlet width to inlet length.
Nishimura, Horikawa, and Shuto reported that variations in opening
width at the mouth of an inlet did not affect the effective length. How
ever, they investigated a halfharbor width and assumed symmetry would
produce the same resonant motion in the half harbor as it would in a full
harbor. Ippen and Goda (1963) indicated this would not be true because
the half harbor has a asymmetric entrance onehalf the width of the
centered entrance of the full harbor. Ippen and Goda showed that the
harbor entrance width determined the value of wave radiation functions
whi ch are used to determi ne water surface elevations.
For a fully open inlet or harbor (see Fig. 40) , Ippen and Goda defined
r esonant amplification (the ratio of an amplitude in the harbor to the
amplitude at the closed harbor entrance) as
1
(298)
where ljJ
1
and ljJ
2
are wave radiation functions given in Figure 42. The
resonant amplification would occur where TLe/T = 1 as before. The func
t ions shawn in Figure 42 apply to all harbor openings, where b is the
width of the opening.
136
1.2
1.0
O.B
~
~ 0.6
0
.:;.
0.4
0.2
/
'2
/
v
/
/
1 v
1(,/
0.2 0.4
/
/_,
0.6
v
,/
0.8
'!!>
2
~ 1
 ~
/
,..,.,
/
/
..
"'2
rr...,

~
1.0 1.2 1.4
Figure 42. Wave radiation functions
(from Ippen and Goda, 1963).

~
1.6
* * * * * * * * * * * * * * EXAMPLE PROBLEM 20 * * * * * * * * * * * * * *
GIVEN: A ful1y open inlet has a width, B, given by B = 0.194 Lb, where
~ is the length of the in1et. The incident wave1ength L = 258.
FIND: The resonant amplification in the inlet.
SOLUTION:
kB 27T(O. 04L)
0.1257
2 2L
From Figure 42, where b = B for a fully open inlet
ljll 0.12
and
From equation (298),
a2 1
al [ (cos kLb  lji
2
sin kLb)
2
+ ljl
1
sin
2
kLb] l/2
137
1
0.24 sin{21T
0
·
04
})
2
+ (0.12)2 sin2(21r
0
·
04
0.194 0.194
8.16
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Ippen and Goda (1963) compared theoretical and experimental results
of a fully open harbor (Fig. 43). They also obtained theoretical results
for partially closed harbors with both symmetric and asymmetric entrances
(Fig. 44). Their experimental results for partially closed harbors gen
erally showed that amplification factors were less than those predicted
theoretically. However, comparisons between experiments and theory were
only obtained for higher modes of oscillation.
9 Wove Period (s)
...
'1!e:
N
C!
"'
ID ... 10
0 0 ci 0
... ...
0 0
8
Theory
Harbor
7
6
•
j
.
.
.f5
1jJ
0
8/Lb= 0.194
4
3
2
•
klb
Figure 43. Frequency response of a fully open harbor
(from Ippen and Goda, 1963).
138
20
18
16
14
12

0
..... 10
N
0
8
6
4
2
0
8
7
6
 0
..... 5
N
0
4
3
2
0 2 3 4 6
o. Norrow Harbor
2 3 4 6
b. Wi de Harbor
b/8
·0.01
0.1
 1.0
x Point of predicted
amplification
b/8
·0.01
0.1
 1.0
9
1 8 9
Figure 44. Theoretical frequency response curves of harbors
(from Ippen and Goda, 1963).
139
10
10
 0
......
N
0
0
.......
N
0
9
8/Lb= 1.0
€
b/8 :0.1
bl!t
0
6
x Point of predicted
amplification
 E:: 0.5 ( symmetrical )
· E: 0.75
 E: 0.95 (most asymmetric)
5
4
3
2
0 2 3
klb
c.Squore Harbor with Asymmetric Entronce
4 5 6
2
8/Lb: 2.0
b/8 : 0.1
x Point of predicted
amplification
 E : 0.5 ( symmetrical)
· E: 0.75
 E : 0.95 (most osymmetric l
3 4 5
klb
d. Wide Harbor with Asymmetric Entronce
Figure 44. Theoretical frequency response curves of harbors
(from Ippen and Goda, 1963).Continued
140
6
Miles (1972) indicates that for waves passing from a Continental
Shelf into a harbor, where the dimensions of the harbor and the entry
channel are small compared to the local wavelength of the tsunami, the
response of the harbor is essentially restricted to the Helmholtz mode;
i.e., the lowest mode of resonance. The harbor undergoes a pumping
motion where the water leve! in the harbor is assumed to rise and fall
uniformly across the total area of the harbor (Carrier, Shaw, and Miyata,
1971). The water passing through the entry channel is assurned to have a
high velocity, represented as kinetic energy; the water in the harbor has
a much lower velocity, and the rise and fall of the water level in the
harbor is represented as potential energy.
Carrier, Shaw, and Miyata (1971) show that the wave number, ka, for
Helmholtz resonance is represented
k
a
(299)
where La is the length of the entrance channel (Fig. 45). The terrn
ln (ka b/2) in the denominator represents the effect of energy
radiation from the seaward end of the entrance channel (Rayleigh, 1945;
Miles, 1948). Equation (299) is restricted to very limited cases, and
Figure 46 shows a comparison of equation (299) and the results of Miles
(1971) for a harbor with a zerolength entrance channel (La= 0). Figure
46 shows that equation (299) will generally predict resonant wavelengths
that are tao short (and therefore predicted resonant periods with values
lower than the actual resonant periods).
1
1
b B
Figure 45. Harbor with an entrance channel.
141
t'
+:
N
3.011 u:rn1. w. l iù1l!t111i=iJ 11m ! 11 til 11l! ll lllillll
2.0
N
::::
.0 1··
i  i ,l
'fr+""9 .......... ..,,
... •
++ 1Hi+H1++
t; :·
01 '
' CD 1.5
l+l+++lHl+Hl+Hi li 1W11+1++Hl1l

IUIJn
J!!: l:!l": ll+l++l il+II+I++Jl+l+"' . ;...j..j.L.fl1l i +U+TI1Il
1.0
O.
1 rn
0 H±HjH !+Htttt !ffi 1 t1
0.01 tHtHI I IHII!·IH!IIH!!!J
0.02 0.03 0.04
::l;! :, j:::j::: .... lw+  _ ,
: jo.
t i_tH iHI·JJI·J Hilil lttttj 1
r• ',;; +llrrr:m:m:nmm1=11+1
i'i!
0.06 0.08 0.1
b/8
0.2 0.3 0.4
1 11
:[j! !!i: !::!
0.6 0.8 1.0
Figure 46. Wavelength for Helmholtz resonance (centered harbor entrance, entrance length La= 0) .
Carrier, Shaw, and Miyata (1971) suggest an approximate method for
determining resonant wavelengths, for harbors with entrance channels,
which will be more generally applicable than equation (299). Their
method assumes that the resonant wavelength, L
0
(L
0
= 0), for an equiva
lent harbor of the same dimensions but having no entrance channel (L
0
=
0), can be obtained. Correcting an error which appears in Carrier, Shaw,
and Miyata, the resonant wavelength for the harbor with an entrance chan
nel is then given by the equation
(300)
The resonant wavelength where L
0
results (see Fig. 46).
0 can be obtained using Miles' (1971)
For a harbor with an entrance channel (Fig . 45), Miles (1971) indi
cates that narrowing the entrance width or increasing the length of the
entrance channel will significantly increase the response of the harbor
to the Helmholtz mode, which may dominate tsunami response. This narrow
ing or lengthening also has the effect of decreasing the resonant fre
quency (Carrier, Shaw, and Miyata, 1971). Carrier, Shaw, and Miyata point
out that lengthening the entrance channel to a harbor also increases the
frictional resistance so amplification factors for a very long entrance
channel may be significantly reduced (although the resonant frequencies
would still be less than for a harbor without an entrance channel; i.e.,
where L
0
= 0).
Seelig, Harris, and Herchenroder (1977) present a numerical means for
analyzing harbors responding to the Helmholtz mode of resonance. The
method uses a RungeKuttaGill technique where
(301)
hb is the surface elevation of the water in the harbor above sorne arbi
trary fixed datum, Q the flow rate through the entrance channel, and
Ab the area of the harbor (Ab= Lb B). The governing differentiai
equation is
where
~ _
1
9' ( 1 1 ) 2
     Q  g I (h  h )  I F
dt 2 A
2
A
2
g b s g
be sc
1
g
143
(302)
(303)
A
0
is the crosssectional area of flow through the entrance channel at
any point X between the seaward end at X
3
and the harbor end Xb
(A
0
, therefore, being a function of X), Aba the crosssectional area
at the bay end, A
80
the crosssectional area at the sea end, hs the
height of the sea level above the arbitrary fixed datum, and F defined
as the total bottom friction in the entrance channel. A sample compu
tation for a tsunami entering a bay is given in Seelig, Harris, and
Herchenroder (1977) (Fig. 47). It can be seen that the peak water levels
in the bay occur slightly after the peak water levels just seaward from
the entrance channel. Also, the peak water levels were slightly lower
in this case.
~
0
~
0.75
0.50
0.25
0.25
0.50
0.75
0
Inlet length = 1 22 m
Jnlet depth = 7.3 m
Inlet width = 24 m
Bay oreo=4.6 x J0Sm2
Tsunami at boy entronce
 Woter levet in boy
0.5
Time (hr)
Figure 47. Tsunami water levels in a bay (tide excluded)
(after Seelig, Harris, and Herchenroder, 1977).
Miles (1971) found that he could transform his equations for wave
induced oscillations in a harbor to an integral equation equivalent to
the equation formulated by Lee (1969, 1971). Lee expresses the governing
equations for wave oscillations in an arbitraryshaped harbor as
(304)
and
a
2
f(x,y) + a
2
f(x,y) + k2 f(x,y)
ax
2
ay
2
0
(305)
144
where
z
ag cosh [k(z + d)]
cosh(kd)
(306)
and f(x,y) is a wave function to be deterrnined. Equation (305) is the
Helmholtz equation. The following boundary conditions are assurned:
(a) f(x,y)/an = 0 along all fixed boundaries where n is
in the normal direction to the boundary.
(b) The harbor does not affect the wave system where
(x2 + y2)l/2 + oo; i.e.' at large distances from the harbor
entrance.
Lee determines the value of the unknown wave function f(x,y) by
deterrnining the function f
1
(x,y) in the open sea and the function f
2
(x,y)
in the harbor, then rnatching the functions at the harbor entrance; i.e.,
the wave amplitude and the slope of the water surface must be the sarne
for f
1
(x,y) and f
2
(x,y) at the entrance.
The function f2(x,y) at sorne position (x,y) within the harbor is
defined by a line integral !
8
taken around the harbor boundary in a
counterclockwise direction giving
f
2
(x, y)
(307)
where HJ
1
) is a zeroorder Hankel function of the first kind, f
2
(x
0
,y
0
)
the function at a boundary point (x
0
,y
0
), and r the distance between
the boundary point (x
0
,y
0
) and the interior point (x,y).
The wave function in the open sea is represented by the surn of three
functions
(308)
where fi(x,y) is the known incident wave function, fp(x,y) the reflected
wave function, and fra(x,y) the wave function for the wave radiating
seaward from the harbor entrance. The reflected wave function is deter
rnined from the incident wave function for total reflection. The radiated
wave function is deterrnined as
f.,.a(o,y) =  .!_ J H(l) (kr) a [f
2
(o,y
0
)] d
.L 2so an s
(309)
145
where the line integral !
8
is taken across the harbor entrance, y is
measured along the coastline (across the harbor entrance), and x is
measured normal to the coastline.
Lee (1969) expressed equations (307) and (309) in matrix form and
solved them numerically. Figure 48 is an example of his experimental
results for a small laboratory mode!_ of an arbitraryshaped harbor.
7 r       r       r      ~       ~     ~       ~     ~       ~
6
5
4
2
 Arbitrory shoped
harbor theory
• Experiment
6
Figure 48. Response curve at point C of the Long Beach harbor
mode! (from Lee, 1969).
Chen and Mei (1974) have developed a finiteelement numerical mode!
which can be used to study water leve! oscillations in a harbor. Houston
(1976, 1977) applied Chen and Mei's mode! to studies of Los Angeles and
Long Beach harbors.
VII. TSUNAMI RUNUP AND INTERACTION WITH STRUCTURES
The arriva! of a tsunami at a shoreline may cause an increase in water
leve! as much as 30 meters or greater in an extreme case. Increases of
10 meters (32.8 feet) are not uncommon. The large increase in water ·
leve!, combined with the surge of the tsunami, can impose powerful forces
on shore protection structures and on structures located near the shore
line. Structures may be seriously damaged or destroyed by the tsunami.
Damage may be caused by strong currents produced by waves overtopping the
structures, by the direct force of the surge produced by a wave, by the
146
hydrostatic pressure created by flooding behind a structure combined with
the loss of equalizing forces at the front of a structure due to extreme
drawdown of the water level when the waves recede, and by erosion at the
base of the structure. Major damage may also be caused by debris carried
forward by the tsunami in the nearshore area.
To determine the potential damage to structures located along a shore
line, the probable increase in water level caused by the tsunami, i.e.,
the runup height, must be estirnated. Estimates of tsunami runup are also
needed for flood zone planning along the shoreline, and for operation of
the tsunami warning system to evacuate people from endangered areas.
1. Tsunami Runup on a Shoreline.
The height of a tsunami will vary from point to point along a
coastline. The nurnerical rnodels for prediction of tsunami height at
the shoreline, i.e., the elevation of water at the shoreline due to the
tsunami, must be applied to a sufficient nurnber of points along the
shoreline to determine this variation. When the variation is large
between adjacent points, calculations for tsunami heights should be
carried out at additional shoreline points between those points. After
the height of the tsunami at a point along the shoreline has been deter
rnined, the vertical runup height at that point can be estirnated.
When the tsunami height along a section of coastline is relatively
constant, and the variations in onshore topography are relatively rninor,
the runup height may be assurned to be constant along that section of
coastline as a first approximation. Variations in tsunami height and
shoreline topography will actually cause sorne variation in runup charac
teristics along any section of coastline. An example of how extrerne
this variation can be is given by Shepard, MacDonald, and Cox (1950) for
Haena, on the Island of Kauai, Hawaii, where there was a gentle rise of
water level on the western side of the bay, but less than 1 mile to the
east, waves rushed onshore, flattening graves of trees and destroying
houses. An exarnple of the variation in runup height is given by Wilson
and T ~ r u r n (1968) for Kodiak City, Alaska (Fig. 49). The mean runup
height at Kodiak City was a little more than 6 rneters (20 feet) above
mean lower law water (MLLW), with variations from about 5 ta 8 rneters
(17 to 27 feet). Because these variations are difficult to predict, the
predicted runup heights may contain substantial errors. Where tsunamis
of a known height have produced variations in runup at a particular sec
tion of coastline, the higher heights should norrnally be used for conserv
ative design.
It should be noted that the characteristics of the waves may vary from
one wave to another at the sarne coastal point. Shepard, MacDonald, and
Cox (1950) cite a case in Hawaii where the first waves came in so gently
that a man was able to wade through chesthigh water ahead of the rising
water. Later waves were so violent that they destroyed houses and left
a line of debris against trees 150 meters (500 feet) inland.
147
Figure 49. 1964 tsunami runup, Kodiak City, Alaska (contours in feet);
heavy line is maximum flood level (from Wilson and T0rum, 1968).
An .added complication, which is an important consideration in comput
ing runup heights, is the possibility of storm waves occurring simultane
ously with the tsunami. The prediction of maximum runup heights would
require the consideration of joint probabilities of tsunamis and storm
waves, as well as the probability of a high tidal stage. The probability
of a high tide, tsunami, and storm waves occurring simultaneously may
appear to be small; however, such an event did occur in Newfoundland in
1929 (Hodgson and Doxsee, 1930).
Because a tsunami has a very long period relative to storm waves, it
causes an apparent variation in water depth over a long distance. Storm
waves riding on top the tsunami will have a wave celerity corresponding
t o the depth (including tsunami height) at any particular point. If two
storm waves are otherwise equivalent (e.g., the same period and wave
148
height), and one is at the crest of the tsunami while the other is at
the leading edge, the storm wave at the tsunami crest will have a higher
celerity (U.S. Army Engineer District, Honolulu, 1960). Therefore, the
tsunami can cause one storm wave to overtake and superimpose itself on
another storm wave, producing higher waves at the shoreline.
Storm waves alone may be more severe than a tsunami at sorne exposed
coastal points. Shepard, MacDonald, and Cox (1950) refer to the
Kalaupapa Peninsula, on the Island of Molokai, Hawaii, where the 1946
tsunami left driftwood at elevations slightly more than 2 meters (7 or 8
feet) above normal sea level, but winter storms had left driftwood 6
meters above the same datum plane. A combination of a winter storm and
a large tsunami could be very destructive.
Houston and Garcia (1974) assume that tsunami runup on a shoreline
will have a runup height (vertical rise) equal to the wave height at the
shoreline. This assumption is based on the idea that a tsunami will act
like a rapidly rising tide. The assumption was compared with a few cases
where both height and runup data were available. For those cases, which
included the 1960 tsunami at Hilo, Hawaii, that produced a borefronted
wave, the predicted value of runup equal to wave height at the shoreline
compared well with the maximum runup measured in the area. Nasu's (1934)
data for a tsunami occurring in Japan also indicate that the total runup
was about equal to the wave height at the shoreline at many locations.
Wiegel (1965) reports that maximum runup elevations above MLLW at Crescent
City, California, were equal to or greater than the maximum wave height
(cresttotrough) at the Crescent City tide gage for the 1952, 1960, and
1964 tsunamis. Magoon (1965) indicates that the 1964 tsunami at Crescent
City had an elevation of about 6 meters above MLLW along a substantial
length of shoreline, and that the line of maximum tsunam; inundation
generally followed a contour at that elevation. While the assumption
that maximum runup heights will equal the tsunami height at the shoreline
provides an initial estimate, this assumption cannot always be used with
accuracy. The effects of ground slope, wave period, and the possible
convergence or divergence of the runup must be considered.
The results of Nasu (1934) indicate that the tsunami height at the
shoreline and the runup height are dependent on the configuration of the
coastline. At Kamaisi, Japan (Fig. 50), on the north side of a bay, the
runup height was actually somewhat less than the wave height at the shore
line, equal to slightly more than 3 meters. At Hongo (Fig. 51), at the
head of a bay, the tsunami flowed directly up a canyon along a streambed,
and the maximum runup height was about 11 meters (36 feet). At Ryoisi
(Fig. 52), the tsunami intruded into a small inlet opening onto the main
bay, flowed up a canyon along a streambed and highway, and reached a
maximum runup height equal to about 10.5 meters (34 feet). The wave at
Kamaisi was probably traveling parallel to the shoreline as it flooded
into the bay. The wave at Hongo and Ryoisi was probably traveling in a
direction oriented directly along the axis of the canyons as the surge
came onshore.
149
Contour intervol in meters;
heovier numerols ore woter
heights in centimeters;
Smoller numerols ore ground
elevations in meters (meosured
from the sa me dotum)
0 50
SCALE
100
MET ERS
200
Figure 50. Tsunami runup at Kamaisi, Japan (after Nasu, 1934).
Iwasaki and Horikawa (1960) show that the period of the waves will
be a major factor in determining maximum inundation levels. The waves
from the tsunami which struck the coast of Japan on 24 May 1960 had
periods of about 60 minutes, while the 1933 tsunami reported on by Nasu
(1934) had wave periods of about 12 minutes. The 1960 tsunami did not
forma bore or a spilling front, and the water level gradually gained
height over the entire surface of the bays where it was observed. For
the longer period waves of the 1960 tsunami, the orientation of the bays
appeared to have no influence on the runup heights; the height of the
150
Contour intervol in meters;
heovier numerols ore woter
heights in centimeters
SCAl€
0 50 100
c=w+ 
IIETERS
200
Figure 51. Tsunami runup at Hongo in Toni, Japan (after Nasu, 1934) o
runup was equal to (or sometimes less than) the height of the wave at
the shoreline.
Tsunamis at a shoreline could be categorized into three types of
waves: nonbreaking waves (i.e., a tsunami which acts as a rapidly rising
tide); waves which break far from the shoreline and become fully developed
bores before reaching the shoreline; and waves which break near the shore
line and act as partially developed bores which are not uniform in height.
In addition, there are sorne cases where reflected waves become bores after
reflecting from a shoreline.
For the nonbreaking wave, the assumption that the runup height equals
the wave height at the shoreline may be reasonable and possibly even con
servative. Field observations (e.g., Nasu, 1934) indicate that the runup
151
Contour intervol in meters;
heovler numerols ore woter
heights in centimeters
ICAU
50 100 200
Mi ..
....
.
.
.
..
Figure 52. Tsunami runup at Ryoisi, Japan (after Nasu, 1934).
height is sometimes less than that value. To analyze the runup of
breaking waves and fully developed bores, where maximum runup heights
have been observed to be much higher than the wave or bore height at
the shoreline, it is necessary to consider the actual form of the runup.
Using solitary waves, Camfield and Street's (1967) experimental
results for an 8° nearshore slope fronted by a slope S
2
= 0.01 indicated
that the runup takes the initial form of a horizontal water surface at
an elevation equal to the wave height at the shoreline (Fig. 53), and
that the higher runup on the slope washes up the slope at a shallow depth.
Results for plunging breakers on 4° and 8° slopes fronted by a slope
S
2
= 0.01, and on 4°, 8°, and 12° slopes fronted by a slope s
2
= 0.03,
indicated similar runup characteristics. The higher, shallow runup may
cause sorne flooding, but would not be expected to otherwise cause damage
152
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Distance (ft)
Figure 53. Solitary wave runup (from Camfield and Street, 1967).
because of the shallow depth. Also, this higher, shallow runup may not
be representative of prototype runup. O'Brien (1977) points out that a
fraction of the uprush percolates into a natural, permeable beach. This
percolation tends to partially dissipate the shallow part of the runup
observed on the impermeable model beach.
Kononkova and Reihrudel (1976) studied the runup of solitary waves
on uniform slopes which were apparently fronted by a horizontal tank
bottom. For nearshore slopes less than 8°, their results were comparable
to those of Camfield and Street (1967). For nearshore slopes greater
than 8°, they found runup values higher than the wave height at the shore
line.
Miller (1968) gives results for borelike waves which act as surge
runup on a shoreline. He shows that the runup in this case also takes
the initial form of a horizontal water surface at an elevation equal to
the wave height at the shoreline, and that the higher runup flows up the
slope as a thin sheet. Miller comments that, "ln the later stages of
runup, the form of the wave was of a thin, fastmoving greatly elongated
wedge."
The experimental work of Camfield and Street (1967), Miller (1968),
and Kononkova and Reihrudel (1976) was for flat, uniform slopes with no
convergence of the wave crest. In general, the experiments show that
for flatter slopes (less than 8°) the runup height appears equal to or
less than the wave height at the shoreline. For steeper slopes, the
runup height increases as the slope increases, and the ratio of runup
height to wave height at the shoreline appears to reach a maximum value
for vertical walls. However, the higher runup on the steeper slopes
appears to have a relatively shallow depth.
Sorne attempts have been made to develop theoretical solutions.
Freeman and Le Mehaute (1964) give a formula for surge runup as
R
u
2
(1 + A) (1 + 2A)
s
153'
(310)
where
R the vertical height of runup above the stillwater level
us the current velocity of the surge at the shoreline
f the friction factor
S the ground slope
g gravitational acceleration
A a coefficient
Adapting the work of Keulegan (19SO), they obtain a maximum value of
A= O.S. Using a value for us given by
u
s
1
 (gh ) 1/2
A s
where hs is the surge height at the shoreline, taking the friction
factor, f, as
f
8g
c ~
(311)
(312)
where Ch is the Chezy coefficient, and using the maximum value for A,
equation (310) reduces to
R 6
hs = 32g (313)
1 + C2s
h
As Ch varies with depth, this equation would predict that the relative
runup R/hs varies between a prototype and madel unless proper roughness
scaling is used. Because Ch decreases with increasing roughness, the
relative runup would decrease as the roughness increases. Also, as slope
increases, the relative runup increases. As the slope approaches infinity
(a vertical wall), the relative runup R/hs = 6. This value is .somewhat
higher than experimentally obtained values. Camfield and Street (1967,
1968) obtained values of relative runup between 4.S and S.O from solitary
wave experiments for breaking waves running up on a vertical wall. Equa
tion (313) does not consider the effects of wave period.
Freeman and Le Mehaute (1964) noted that coefficient A in equation
(310) should be somewhat less than O.S. Kishi and Saeki (1966) inèicate
that the value of A decreases as the slope decreases, which is consist
ent with Freeman and Le Mehaute that the value of A depends on the form
of the wave at the shoreline.
154
It is also necessary to account for the dependence of Ch on the
varying height, h, of the surge traveling up the onshore slope. Noting
that, for uniform flow, Ch can be related to the Manning roughness
coefficient n by
(314)
in metric units (the right side of eq. 314 is multiplied by 2.22 for
the footpoundsecond system of units), and that a plot of h
1
/
3
versus
h for 0 ~ h ~ h
8
will give an average value of h ~ 0.75 h
8
. It is pro
posed equation (314) can then be written
0.91 hl/
3
s
in metric units. This allows equation (310) to be rewritten as
~ = _1_ ( 1 + A) ( 1 + 2A)
h
8
2A
2
8gn2
1
+ o.91 A
2
s h
1
/
3
s
(315)
(316)
in metric units (the coefficient 0.91 on the right side is equal to 2.02
in the footpoundsecond system of units). Kishi and Saeki give a log
log plot for A versus S, with A= 0.25 when S = 0.03, and A= 0.04
when S = 0.07. Values of A were only obtained for that range of slopes.
Also, the effect of wave period on the results was apparently not inves
tigated.
Camfield and Street's (1967) laboratory results for borelike solitary
waves running up a 4° slope (S = 0.0699), fronted by a slope S = 0.01,
give a total relative runup R/H
8
of 3.3 for a value of h
8
= 0.061 meter
(0.2 foot) on a smooth aluminum slope. Using a value of n = 0.01 in
equation (316), and using a value of A= 0.4 suggested by Kishi and
Saeki (1966), R/h
8
would have a calculated value of 2.67, which is
close to the measured value. Kishi and Saeki obtain similar results for
rough slopes. As previously mentioned, the runup values of Camfield and
Street (1967) include a shallow flooding which may not be an accurate
prediction of prototype conditions. If only the g r e a t ~ r water depths
were coitsidered, such as shown in Figure 53, then the measured value of
R/h
8
"' 1. O.
It should be noted that the above equations assume a uniform slope.
For runup on a shoreline where the slope varies, it would be necessary
to use a numerical solution to determine the limits of the runup. Freeman
and Le Mehaute have carried out numerical calculations for slopes S ~ 0.1,
but present no results for very flat slopes. Very little data exist to
verify such equations or to determine their full range of application.
155
The solut ion of equation (316) is very dependent on a correct choice
of the roughness coefficient . Onl y very limited data are presently
available for estimating values of the roughness coefficient n. For
prototype conditions, the "roughness" may consist of groves of trees or
subdivisions of houses. Also, the roughness elements, e.g., trees and
houses, may be moved by the waves.
Bretschneider and Wybro (1976) investigated the effect of bottom
friction on tsunami inundation by using the Manning n to describe the
roughness of the onshore slope. Although this is not entirely correct
(the Manning relationship was developed for uniform flow), it provides
a simple means of investigating the effects of roughness on the limits
of inundation. It was shown that decreasing the Manning roughn.ess coef
ficient, n, from n = 0.025 (long grass with brush) ton = 0.015 (short,
eut grass and pavement) could increase the distance required for dissipa
tion of the surge by 160 percent (from 670 to 1,770 meters or 2,200 to
5,800 f eet in the example used, where h
8
= 10 meters or 33 feet).
Bretschneider and Wybro also demonstrated that a bore would be dissipated
faster than a tsunami acting as a rapidly rising tide.
Chan , Street, and Strelkoff (1969) and Chan and Street (1970a, 1970b)
use a modi f ied Marker and Cell (SUMMAC) numerical finitedifference tech
nique for cal culating the wave runup of solitary waves on a 45° slope and
on a ver tical wal l . Their results compared well with the experimental
results of Street and Camfield (1966), but their numerical method was
not applied t o wave runup on the shoreline for flatter slopes.
(1969) developed a numerical method based on finite elements. However,
he provides onl y limited results for simulating waves in laboratory
channels, and the results depend on the choice of a bottomfriction fac
tor and an artificial viscosity.
Spielvogel (1975) developed a theoretical solution for tsunami runup
based on !he wave or surge height at the shoreline, h
8
, and the wave
height, H, at the point where the leading edge of the wave is at the
shoreline. This relates the runup to the rate of shoaling just before
the wave reaches the shoreline, and effectively includes the influence
of the bottom slope and the wave period. Replotting Spielvogel's results
into a more usable form gives the equation
=
h
s
h
.....
8
 0.8
H
(317)
Equation (317) indicates thatAthe higher values of relative runup, R/h
8
,
occur when the values of h
8
/H are the 1owest. Spielvogel indicates that
equation (317) correct for 3.74 > h
8
/H > 2.12, limited application
where 2.1 2 > h /H > 1.76, and is invalid where h
8
/H < 1.76. This latter,
invalid case wguld be a nearshore bore or breaking wave.
In addi tion to considering wave runup, it is necessary to consider
the drawdown of the water when the wave trough arrives at the shoreline.
156
Not as much attention has been given to wave rundown ; however, the
drawdown of the water level may result in the seaward collapse of sea
walls, result in damage to ships in a harbor, or expose seawaterintake
pipelines. It should also be noted that a graduai increase in water
level, with very low velocity currents, may be followed by a sudden
withdrawal of water producing very strong currents.
During the 1946 tsunami in Hawaii, waves at Hanamaulu Bay rose 2. 7
meters at a breakwater and wharf, but the water receded to a level 5.6
meters below normal sea l evel between waves (Shepard, MacDonald, and Cox,
1950). Most of the damage was caused by the violent withdrawal of the
water.
The rundown elevations will depend on the wave train generated at
the tsunami source. For the 1946 tsunami, the tide gage record at
Honolulu, Hawaii, indicated sorne very narrow, deep wave troughs with
the initial troughs having greater amplitudes than the initial crests.
Consideration must also be given to the current velocities of the
runup. Ishimoto and Hagiwara (1934) investigated the large 1933 tsunami
at Kamaisi, Japan, and estimated current velocities with a maximum value
of 1 meter per second . Houston and Garcia (1974) estimated that small
tsunamis in southern California acting as rapidly rising tides would
have maximum current velocities of about 0.5 meter per second . The
current velocity for the 1933 tsunami, which was about doubl e the veloc
ity estimated by Houston and Garcia for small tsunamis, destroyed sorne
buildings when the water depth reached a height of 2 meters (6.15 feet).
Water overflowing a coastal barrier will have a current velocity
determined by the difference in height between the top of the barrier
and the ground level behind the barrier, as well as the quantit y of
water overtopping the barrier, rather than acting like a rapidly rising
tide. The barrier will also limit the height of the runup; however,
large drain openings must be provided to prevent water levels from
building up behind the barrier if it is overtopped by successive waves.
Magoon (1965) cites one example south of Crescent City, California,
during the 1964 tsunami where water flowed over narrow coastal dunes.
The quantity of water overflowing the dunes was insufficient in sorne
instances to fill the low areas to landward, reducing the resulting
runup height.
Where the slope is very long compared to the wavelength, and friction
effects must be considered, it can be seen that for low velocities the
retarding effect of the slope roughness (deceleration) may prevent the
water from rising to a runup height equal to the wave height at the shore
line (i.e., drawdown will start at the shoreline, reversing the direction
of flow). As previously noted, the currents associated with the rundown
might have much higher velocities than the currents associated with the
runup. No estimates are available for the rundown currents.
157
~ u r g e runup on a dry bed will have a much higher velocity than the
values given by Houston and Garcia (1974) for a tsunami which acts like
a rapidly rising tide. Keulegan (1950) gives
u = 2(gh)
1
/
2
(318)
where h is the surge height at any point and u the water velocity at
the same point. Fukui, et al. (1963) give a lower value of velocity as
u = 1.83(gh)
1
/
2
(319)
The higher value would be conservative.
2. Interaction with Shore Protection Structures.
Breakwaters and seawalls may provide coastal areas protection from
tsunamis. When a tsunami occurs, breakwaters may decrease the volume of
water flowing into a harbor and onto the coastline. Proper placement of
breakwaters may also decrease wave heights by changing the natural period
of an inlet discussed in Section VI, 7. However, breakwaters may also
affect the resonant period of a harbor so that wave heights are increased,
and seawalls may reflect waves within a harbor. A high seawall along a
coastline may prevent flooding of the backshore areas.
A tsunami may damage shore protection structures; therefore, care
must be exercised in the design of the structures. Numerous instances
of tsunamis damaging or destroying protective structures have been
recorded. The 1946 tsunami in Hawaii overtopped and breached the break
water at Hilo, removing 7.25metric ton (8 tons) stones to a depth 0.9
meter (3 feet) below the water surface along nine sections of the break
water crest with a total length of over 1,800 meters (6,000 feet) (U.S.
Army Engineer District, Honolulu, 1960). Matuo (1934) refers to the case
of an earthern embankment at Yosihama on the northeast coast of Honshu,
Japan, which had been constructed to protect a section of coastline. The
1933 Sanriku tsunami overtopped the embankment, and it was swept away,
flush with the original ground level.
Iwasaki and Horikawa (1960) investigated areas along the northeast
coast of Honshu after the 1960 tsunami. They indicated that a sea dike
at Kesennuma Bay failed during the 1960 tsunami because the water from
the incident waves, which had overtopped the dike, caused extensive
erosion receding at a gap in the dike. The receding water gradually
widened the gap. They also noted that a quay wall at Ofunato failed
because of scouring of the backfilling, and that a quay wall constructed
of reinforced concrete sheet piles at Hachinohe collapsed due to a lack
of interlocking strength after backfilling was washed away.
Iwasaki and Horikawa also indicated that receding water may seriously
scour the seaward base of a revetment or seawall. The combination of this
158
scouring and the increased hydrostatic pressure from initial overtopping
may cause failure. The concrete seawall along a highway between Hadenya
and Mitobe on Shizukawa Bay (Fig. 54) collapsed seaward. Similar fail
ures occurred along a highway on Onagawa Bay and along a quay wall at
Kamaishi, Japan. Magoon (1962) noted that approximately 2 meters of
sand was scoured at the seaward toe of a steel sheetpile retaining wall
at Crescent City, California, in 1960 which contributed to its partial
failure. Also, a woodpilemooring dolphin was destroyed as a result of
the loss of sand at its base. Matuo (1934) mentions a concrete retaining
wall which was overturned seaward by the 1933 Sanriku tsunami.
The damage from the 1960 tsunami in Hawaii is evidence of the erosive
force of a tsunami. Concrete seawalls 0.9 meter high were washed out and
a gully about 3 meters deep and 27 meters wide was washed into a highway
along the shoreline at Hilo, extending inland about 18 meters. Large
stones from a seawall, weighing up to 20 metric tons (22 tons), were
carried inland (Eaton, Richter, and Ault, 1961). Shepard, MacDonald,
and Cox (1950) mention a case where water overtopping sand dunes eut a
channel about 30 mèters wide and about 5 meters (15 feet) deep.
Tsunamis will not always produce the maximum forces on a structure.
A concrete seawall protected the buildings at the Puu Maile Hospital at
Hilo during the 1946 tsunami. The seawall was undamaged by the tsunami,
but a few months later storm waves destroyed parts of the wall and damaged
the lower floor of the hospital (Shepard, MacDonald, and Cox, 1950).
Matuo (1934) reports on a dynamometer located on a breakwater at
Hatinohe harbor, Japan, during the 1933 Sanriku tsunami. The dynamometer
was located 0.76 meter (2.5 feet) below the level of the water surface at
the time of arrivai of the tsunami. The recorded maximum pressure was
38,300 newtons per square meter (800 pounds per square foot) for a wave
with a height of 3.2 meters (10.5 feet) and a period of 6 minutes.
Nasu (1948) developed sorne empirical criteria for the stability of
breakwaters based on the geometrie shape of the breakwater. For a break
water with a seaward slope of 1:2.5 and a landward slope of 1:2, he gives
h + 0.89b
u 2 < __;;_V ___ _
0.0358
(320)
for the condition of geometrie stability, where u is the current veloc
ity in meters per second, hv the height in meters of the vertical seg
ment of the face of the breakwater against which the current acts, and
b t he top width of the breakwater in meters.
Kaplan (1955) gives an empirical equation for the volume of overtop
ping of a seawall at ·the shorel ine. This equation can be rewri tt en as
v
21.65(Kh h)
3
s ü)
159
(321)
1'
(j\
0
Figure 54. Concrete seawall destroyed by 1960 tsunami, Shizukawa Bay, Japan
(from Iwasaki and Horikawa, 1960).
where V is the quantity of water overtopping the wall in cubic meters
per meter or cubic feet per foot length of wall, h
8
the wave height at
the shoreline in meters or feet, hw the wall height in meters or feet,
and
K
R
hs
(322)
where R would be the vertical height of wave runup on a similar wall
high enough to prevent overtopping.
Wiegel (1970) gives the following empirical equation for overtopping
volume in cubic meters per meter length of wall
h
t2(1 21Tt h,_,)3/2
V=0.287 t . h cos w dt
1 2 s T
(323)
where h
8
is the total wave height in meters (cresttotrough) of the
wave at the shoreline, T the wave period, t
1
the point in time where
overtopping begins, and t
2
the time when overtopping ends. As the
wall height, hw, is measured in meters from the sea level at the time
the tsunami occurs, it varies but its lowest value (i.e., the greatest
overtopping) would occur when the sea level is at the highest tidal stage.
Values for overtopping are shown in Figure 55 .
=
c;
J:
~
en
'0
e
~
~
...
c
 ~
~
0
'0
..
§
~
10,000
8,000
6,000
4,000
3,000
2,000
1,000
800
600
400
300
200
100
0.1 0.2 0.3 0.4 0.6 0.8 1.0 3 4 6 e •o
Elevation of Tsunami Crest obove Seowoll, (ml
Figure 55. Overtopping volumes
Cafter Wiegel, 1970).
161
Based on stability tests carried out in a hydraulic madel, Karnel
(1967) developed suggested breakwater sections for Hilo, Hawaii (Figs.
56 and 57). Allowable overtopping heights are given in Table S.
Table S. Allowable overtopping heights
(after Karnel, 1967).
Allowable height
of overtopping (rn)
Slope of 9.1rnetric ton 18.2rnetric ton
harborside of barrier Armor stones Armor stones
1:2 0.3 1.0
1:2.5 0.8 1.2
1:3 1.3 1.5
1:3.5 1.6
1:4 2.0
1:4.5 2.3
1:5 2.4
1:6 2.7
1:7 2.9
Iwasaki and Horikawa (1960) show typical cross sections of seawalls
at locations on the northeast coast of Honshu (see Fig. 58). In sorne
instances, such as in fishing ports or harbor areas, it is undesirable
to have high seawalls directly on the waterfront. The seawall at Yarnada
(Fig. 59) is in two stages. A law seawall along the waterfront allows
access to the water; a higher seawall, set back from the shoreline,
protects the town from higher waves. Figures 58 and 59 show that the
seaward toe of a wall is protected by rubble to deter scouring. Also,
the area behind the top of the shoreline wall, such as at Yarnada, is
paved to prevent saturation and erosion of the backfill rnaterial.
The protection provided by a breakwater depends on its location and
the width of the navigation channel through the breakwater. Iwasaki,
Miura, and Terada (1961) ran madel tests for breakwaters in Kesennurna
Bay. They discovered that a breakwater at the rnouth of the bay would
substantially reduce wave heights in the bay for all wave periods tested.
As expected, the greatest reduction in wave height occurred when the area
of the breakwater opening was the least. When the ratio of the breakwater
opening area to the crosssectional area of the bay was equal to about
0.1, the wave height was reduced to about 0.25 times the height which
would occur without the breakwater. Surprisingly, when the breakwater
was rnoved to the rnouth of Kesennurna harbor in the madel, at the inner
end of the bay, the breakwater had alrnost no effect in reducing wave
heights. The location of the breakwater would be expected to affect the
resonant periods of the bay and the harbor. Therefore, care should be
exercised in placing a breakwater in any bay or harbor.
162
•
..J
..J
:::1!
=
z
0
;:
~
IIJ
..J
IIJ
SWOE
70
fL tif FOR JO. 'l'OH STOHfS
fL t:IO FOR »TOH STOHfS
60
0, !XImHG AHO RfHAIIIUTATr:D
SARi fR
20
DISTANCE FROM CENTERLINE 1 fil
HARBORSIDE
60 70 50
(Dimensions ahown in feet)
HOT'E: CROSSHATCHEO AREA REPAESENTI
THE EXISTIHG BARRIER.
Figure 56. Suggested design for rehabilitated breakwater, Hilo, Hawaii
(from Kamel, 1967).
S1osid6 Horborsid6
i
_.s20lb tl2ton Stonu/
1
(Dimension shawn in feet)
Figure 57. Suggested design, typical nonovertopping barrier
section, Hilo, Hawaii (from Kamel, 1967).
16 3
Seaside
o. Seowoll ot Toro
Seoside
b. Seowoll at Yoshihomo
Figure 58. Seawall cross sections (from Iwasaki and Horikawa, 1960).
Dimensions in meters.
16 4
Section A
   ~
El +2 80 1
9
·
90
H:W.L·""all R•2$0
L.WL:!:O
Section B
+5.50(1896)
El.+ 4.80
Figure 59. Cross sections of seawall, Yamada, Japan (from Iwasaki and
Horikawa, 1960). Dimensions in meters.
Caution is also necessary when placing seawalls in a harbor area.
A seawall may cause waves to reflect into the harbor. It was determined
at Hilo, Hawaii, that a seawall might aggravate surge conditions within
the harbor (U.S. Army Engineer District, Honolulu, 1960).
In sorne instances, trees may offer sorne protection against a tsunami
surge. Groves of trees alone or as supplements to shore protection struc
tures may dissipate tsunami energy and reduce surge heights. Groves of
coconut palms (Fig. 60) may withstand a tsunami surge but may be sheared
off by debris carried forward by the tsunami. Other types of trees may
be easily uprooted and flattened. Figure 61 shows a grove of pandanus
165
t'
0'\
0'\
' Figure 60. Coconut palms near shoreline, Hilo, Hawaii (from Matlock, Reese, and Matlock, 1962).
Figure 61. Grove of pandanus trees knocked down by
1946 tsunami on the Island of Kauai, Hawaii
(from Shepard, MacDonald, and Cox, 1950).
trees which were knocked over in parallel rows by the 1946 tsunami in
Hawaii (Shephard, MacDonald, and Cox, 1950). Reid and Taber (1919) noted
that palm trees were uprooted by the 1918 Puerto Rico tsunami. Shepard,
MacDonald, and Cox (1950) indicated that dense thickets of hau trees pro
vided effective shields in many places during the 1946 tsunami in Hawaii.
Matuo (1934) calculated that trees could be broken by water velocities
of 2 meters per second or greater, but did not analyze specifie types of
trees. He indicated that trees broken off by higher velocities may add
debris to the surge and increase the damages resulting from the surge.
Magoon (1965) indicates that a buildup of debris in front of a structure
may increase its effective area. This would result in an increased drag
force, and may cause the entire structure to be swept away by the tsunami.
3. Other Shoreline Structures.
Damage from a tsunami may occur to structures located at the shore
line or along river channels near the shoreline. In 1964, a dock at
Crescent City, California (Fig. 62), was damaged when the water elevation
increased to 2 meters above the deck elevation, uplifting a large lumber
barge moored to the dock (Wilson and T ~ r u m , 1968). The tsunami surge at
Seaside, Oregon, destroyed a bridge over the Necanicum River and a rail
raad trestle over Neawanna Creek. Shepard, MacDonald, and Cox (1950)
167
l'
0"\
(X)
.....;t "
....
Figure 62. Dock darnaged by 1964 tsunami at Crescent City, California
(photo by U.S. Coast Guard).
illustrate the damage to the railroad bridge on the Wailuku River (Fig .
63) and the railroad trestle on Kilekole Stream (Fig. 64) caused by the
1946 tsunami in Hawaii. Iwasaki and Horikawa (1960) show a case of
Mangoku, Japan, where a bridge support (Fig. 65) slumped almost ·1 meter
due to the heavy scouring of the channel bottom.
4. Tsunami Surge on the Shoreline.
The determination of the runup he.ight of a tsunami is discussed in
Section VII, 1. After the runup height of a tsunami has been established,
the effects of this runup on structures and other objects located near
the shoreline must be determined. When the tsunami acts as a rapidly
rising tide, the resulting incident current velocities are relatively
low, and most initial damage will result from buoyant and hydrostatic
·forces and the effects of flooding. Shepard, MacDonald, and Cox (1950)
noted that in many instances the withdrawal of the water occurred much
more rapidly than the runup and flooding. In sorne instances, damage may
result from the higher current velocities associated with the withdrawal.
These velocities would be on the order of those normally associated with
an incident surge. More concern is therefore given to a tsunami which
approaches the shoreline as a bore.
When the tsunami forms a borelike wave, the runup on the shoreline
has the form of a surge on dry ground. This surge shouZd not be confused
with the bore approaching the shoreZine, as different equations govern
the motion and profile of the surge. Miller (1968) noted from laboratory
observations that a bore approaching a shoreline exhibits a relative
steepening of the bore face just before reaching the shoreline, and that
this is followed by a flattening of the face of the surge on the dry
slope. The current velocities associated with the surge are proportional
to the square root of the surge height, and approximations of the current
velocities can be obtained from equations (318) and (319), with equation
(318) providing the more conservative result. For a surge height approach
ing 5 meters, the estimated current velocity would be about 14 meters (46
feet) per second. When the tsunami runup acts as a high velocity surge
169
Figure 63. Tsunami damage t o r ailroad bri dge on Wailuku
River, Hi lo, Hawaii (photo by Shigeru Ushijima;
fr om Shepa:rd, and Cox, 1950).
Figure 64. Tsunami damage to r ail r oàd trestle on Kolekole
Stream, Island of Hawaii (from Shepard, MacDonald,
and Cox, 1950) .
17 0
t'
....J
t'
Figure 65. Bridge damaged by 1960 tsunami at Mangoku, Japan (from Iwasaki and Horikawa, 1960) .
of water across the ground, five types of forces may result from the
surging water:
(a) Buoyant forces caused by partial or total submergence
in the surging water. When water or water pressure intrudes
under a structure, the buoyant force tends to lift the structure
from its foundations. Vehicles and other large items may also
be lifted up into the surging water.
(b) Surge forces caused by the leading edge of the surge
impinging on a structure. This leading edge has the appearance
of an elongated wedge, and the force of the surge on a structure
gradually increases as a function of the increase in surge height.
The buoyant force also increases as a function of surge height,
so that a structure may be carried forward by the leading edge
of the surge, or may be destroyed in place if the surge force is
high enough and the buoyant force is not sufficient to lift the
structure from its foundations.
(c) Drag forces caused by the high velocity of the surging
water, where the water level is relatively constant. These
forces will displace buildings or other items in the direction
of the current, and the high velocity flow may cause severe
erosion of the ground and damage waterfront structures by scour
ing material at the base of the structure.
(d) Impact forces caused by buildings, boats, or other
material carried forward by the surging water. These forces
may either destroy other structures on impact or create momentum
which, when added to other forces, will move a structure in the
direction of the current.
(e) Hydrostatic forces caused by partial or total submer
gence of structures by the tsunami. This can result in cracking
or collapse of a structure or wall.
a. Buoyant Forces. Buoyant forces are defined by the weight of the
displaced water when abjects are partially or totally submerged. For
saltwater, taking the density p = 1.026 grams per cubic centimeter (1.99
poundseconds squared per foot
4
), the buoyant force is
FB = pg V (324)
where V is the displaced volume of water: This assumes water intrudes
under the structure.
* * * * * * * * * * * * * * EXAMPLE PROBLEM 21 * * * * * * * * * * * * * *
GIVEN: A house occupies a floor area of 225 square meters (2,422 square
feet). Calculations to predict tsunami runup have indicated a probable
surge depth of 2 meters at that location. It is assumed that the flow
of water will be at a constant depth around the house.
172
FIND: The buoyant force on the house.
SOLUTION: The buoyant force is given by
pg v
1,026 kilograms per cubic meter
(9.81 meters per second squared) (225 square
meters) (2 meters)
4.53 x 10
6
kilogrammeters per second squared
4.53 x 106 newtons (1.02 x 106 pounds)
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
* * * * * * * * * * * * * * EXAMPLE PROBLEM 22 * * * * * * * * * * * * *
GIVEN: .An empty oil storage tank is 3 meters high and 6.1 meters in
diameter. Assume that the tank has a mass of 3,180 kilograms (7,000
poundmass), and that it is filled to a depth of 2.5 meters (8.2 feet)
with oil having a specifie gravity of 0.88 (density p = 880 kilograms
per cubic meter). The tsunami water depth is 1.8 meters.
FIND:
(a) The buoyant force on the tank, and
(b) the force holding the tank in place.
SOLUTION:
(a) The buoyant force is given by
5.29 x 105 kilogrammeters per second squared
FB 5.29 x 10
5
newtons (1.19 x 105 pounds)
(b) The force holding the tank in place is the mass, M, of the
empty tank multiplied by g plus the volume of oil multiplied by its
density and g, so that
F = 3,180(9.81) +
F = 6.62 x 105 newtons (1.49 x 105 pounds)
173
It can be seen that very little reserve force remains to resist drag
forces from the surge. With a lower level of oil in the tank, the
buoyant force could overcome the mass of the tank and the oil plus
the strength of any structural anchorages.
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
In Shepard, MacDonald, and Cox's (1950) discussion of the 1946
tsunami in Hawaii, a house at Kawela Bay on Oahu was floated off its
foundation and deposited in a canefield 61 meters inland, leaving break
fast cooking on the stove and dishes intact on shelves. Many other
houses were also gently floated from their foundations, and sorne houses
could be moved back to their original foundations with very little repair
work required. Damage caused by buoyant forces was the result of build
ings being deposited on uneven ground, the fact that sorne buildings had
weak structures and broke apart when lifted from their foundations, and
minor damage from the breaking of water pipes and electric lines.
In many instances, where tsunamis act like rapidly rising tides, the
current velocity associated with the waves is very low, so that the major
damages are similar to those discussed above. Shepard, MacDonald, and
Cox (1950) mentioned instances of people wading through chesthigh water
to escape from the tsunami.
b. Surge Forces. Cross (1967) showed that the force per unit length
of vertical wall, from the leading edge of a surge impinging normally to
the wall could be given as
F (325)
where F is the force in newtons per meter of width, h the surge height
in meters, u the surge velocity in meters per second, and Cp a force
coefficient defined by
C = (tan 8)
1
•
2
+ 1
F
(326)
where 8 is the inclination of the water surface of the surge shown in
Figure 66; tan 8 is given by the equation
tan e (327)
where Ch is the Chezy roughness coefficient, Z the distance from the
leading edge, and b given by
b
..!_ du _ S
g dt
17 4
(328)
9 cos 13 S=gS
( for smoll 13 )
Surge Direction ~
g
on {3
Figure 66. Definition sketch of surge on a dry bed
(slope exaggerated).
where S is the bottom slope (negative upward), and du/dt the accelera
tion term for flow under the tip of the surge.
Substituting equations (326), (327), and (328) into equation (325)
F = ..!_ pgh2
2
If it is assumed that
+ [ ( ~ + ..!. du _ s \ 1. ·z +
1
]
C ~ h g dt )
du
 = gS
dt
(329)
(330)
i.e., the acceleration is equal to the influence of gravity acting along
the slope (see Fig. 66), then b ~ 0 and
or, defining u from equation (318) and assuming that the value is
relatively constant under the surge
where Ch varies with depth and is given by
17 5
(331)
(332)
(333)
in the footpoundsecond system of units, or
(334)
in the meterkilogramsecond system of units, and n is Manning's rough
ness coefficient.
Substituting equations (330), (332), and (334) into equation (329)
for metric units and collecting terms
(335)
It must be remembered that many approximations have been used in this
solution.
The coefficient Cp accounts for both inertial forces and drag
forces. It may be noted that when e = 0, and the velocity remains con
stant, the force F is simply the hydrostatic force plus a drag force
(per meter width) where Cv = 2.
* * * * * * * * * * * * * * EXAMPLE PROBLEM 23 * * * * * * * * * * * * *
GIVEN: A surge with a maximum height of 2.5 meters impacts normally
against the vertical side of a building. The Manning roughness
n = 0.1, and it is assumed that the surge velocity,
u = remains relatively constant and that the surge profile
remains constant.
FIND: The surge force per meter of building width as a function of
sürge height.
SOLUTION: The surge velocity is determined from the maximum surge
height," so that
u = = 219.81(2.5) = 9.9 meters per second
The surge force is given by equation (335) as
F =! pgh
2
+
+ 1] pu
2
h
2 h1/3
F =! (1 026)(9.8l)h
2
+ [(
4
(
9
·
81
) (O.l)
2
)
1
•
2
+ 1] 1,026(9.9)
2
h
2 , h1/3
F = 5,033 h
2
+
+ 1] 100,560 h
176
For various values of h, the force, F, is tabulated below:
h, meters 0.5 1.0 1.5 2.0 2.5
F, newtons per meter 73,100 138,300 203,900 270,800 339,500
NOTE:Calculations will show that Cp> 1 at the maximum s u r g ~ height
(where the rate of change of surge height ~ 0). This indicates that
the calculated value is conservative for design purposes. It can be
seen that the hydrostatic pressure component of the force is a rela
tively small part of the total force.
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
As indicated by example problem 23 and shawn in Figure 66, there is
a graduai rise in water level at the front of the surge, although this
change in water level appears to occur rapidly with respect to time
because of the forward velocity of the surge. A surge on a dry bed has
a much flatter front than a bore approaching a shoreline. This is seen
in laboratory tests. The buoyant force of the leading edge of the surge
tends to lift abjects into the surging water, and the force of the surge
will then carry these abjects forward.
Wilson and T ~ r u m (1968) report on the case at Seward, Alaska, of a
tsunami surge overtaking a pickup truck being driven from the shoreline.
The truck was swept up by the surge and carried forward like a surfboard
into nearby woods.
The water velocity near the leading edge of a surge is relatively
high, and the height of the leading edge is relatively law (i.e., the
buoyant force is law). Therefore, it is possible that the surge force
may destroy a structure before the buoyant force lifts it into the flow.
c. Drag Forces. The velocity of the water in the surge produced
by the tsunami runup creates a drag force which tends to move a structure
in the direction of the surge. If the velocity is assumed to remain
relatively constant under the surge, i.e., acceleration is negligible
and its effects can be ignored, then it can be assumed that the inertia
or mass coefficient, CM, approaches zero so that the drag force in
newtons is
(336)
where
p the density of seawater = 1.026 grams per cubic centimeter
1,026 kilograms per cubic meter
CD a coefficient of drag, depending on the body (Table 6)
A the projected area of the body normal to the direction of
flow in square meters
u the velocity of the water in meters per second
177
Table 6. Drag coe
ff' .
1c1ents.
Object L/d Reynolds number
cv
Circuler cylinder
1 10
5
0.63
5 10
5
0.74

OJ
00
10
5
1. 20
00
>5 x
10
5
0.33
Square cylinder
0]
00
3.5 x 10
4
2.0
<>à
00
t
10
4
to 105
1.6
Rectangular flat plate 1 >10
3
1.1
( totally submerged)
1
5 >10
3
1.2
d
20 >10
3
1.5
_l_
00
>10
3
2.0
NOTE.L The height of a submerged cylinder, or the length of the
flat plate.
d The projected dimension shown, or the width of the flat
plate.
17 8
To determine the drag force in pounds, p is in units of poundseconds
per foot
4
, the area in square feet, and the.velocity in feet per second.
The coefficient of drag, CD, is dimensionless and retains the same
value as in the kilogrammetersecond system.
Tabulated values of drag coefficients are generally not available for
freesurface flow at high Reynolds numbers. Therefore, existing tables
of drag coefficients must be used to establish maximum coefficients to
ensure safe design. Table 6 gives examples of drag coefficients.
Hallermeier (1976) discusses the importance of the parameter, u
2
/(gd),
where d is the projected horizontal dimension of the structure trans
verse to the direction of flow. Where this parameter approaches unity
there are strong unidirectional freesurface flow effects. In that case,
the coefficients of drag, CD, given in Table 6 may be too low. Individ
ual model tests would be required to determine a more exact interaction
between the tsunami and the structure.
For cases where flow does not overtop a structure, and where there
is no underflow, the flow may be treated as flow around an "infini tel y
long" structure where the ground and the free surface define the bounda
ries of a layer of fluid. For example, flow around a vertical cylindrical
storage tank would be treated as flow around an infinitely long cylinder
in order to obtain a drag coefficient.
In cases where there are overtopping and underflow, the ratio of
lengthtowidth for the structure should be determined. This ratio should
then be used for determining the coefficient of drag.
For a situation in which there is either overflow or the
coefficient of drag can be determined by using an approximation. Assume
that the depth of flow around the structure is twice the actual depth,
and that the height of the structure is equal to twice the wetted height.
Then obtain a coefficient of drag as if there were both underflow and
overflow (see Fig. 67). An example of this type of calculation follows:
* * * * * * * * * * * * * * EXAMPLE PROBLEM 24 * * * * * * * * * * * * * *
GIVEN: A flatsided structure is 14 meters wide and normal to the
direction of flow. The structure is 3. 5 meter·s (11. 5 feet) high and
supported on columns so there is a 1.5meterhigh (4.9 feet) open spate
under the base of the structure. The tsunami surge has a depth of
2.5 meters, giving a wetted height on the structure equal to 1.0 meter
(3.3 feet).
FIND:
(a) The coefficient of drag of the structure, and
(b) the coefficient of drag of a similar structure located at ground
level with no underflow.
17 9
SOLUTION :
v

h'
!
J w
___1
Flow Direction
l
Tsunami surge flowing post elevated structure
h = surge height
hw = wetted height on structure
L = length meosured perpendicular to the
sections shawn above and below
L/d = L/(2hwl
~      1 1 1
 2hw 2h
Flow Direction
j 1
Equivalent body used for determining coefficient of Drag Co
Flow is ossumed symmetricol about the doshline
Figure 67. Determination of CD when flow
passes under a structure.
(a) It can be assurned, for purposes of deterrnining the coefficient
of drag, that the structuré is equivalent to a structure 14 meters
wide and 2.0 rneters high with both underflow and overflow (Fig. 67).
From Table 6, for a flat plate normal to the flow direction where
L/d = 7, the coefficient of drag C D ~ 1.25.
(b) The structure is higher than the depth of flow so there would
be neither underflow or overflow. This corresponds to an infinitely
high structure where L/d = oo, From Table 6, CD 2.0.
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
180
* * * * * * * * * * * * * EXAMPLE PROBLEM 25 * * * * * * * * * * * * *
(This example is taken from an actual situation which occurred at Seward,
Alaska, in 1964; see Wilson and T ~ r u m , 1968.)
GIVEN: A 104.5metric ton (230,000 pounds) railroad locomotive was over
turned by a tsunami surge. The surge was assumed to have a depth of
1.83 meters. The clear space under the locomotive was approximately
0.91 meter (3 feet) and the length of the locomotive body was 12.5
meters (41 feet). The width between the rails was 1.52 meters (5 feet)
and the width of the locomotive body was 3.05 meters (10 feet). The
surge was assumed to act normal to the side of the locomotive.
FIND: The overturning force on the locomotive.
SOLUTION: The buoyant force is given by equation (324) as
pg v 1,026(9.81) (1.83 0.91) (3.05) (12.5)
FB 3.53 x 10
5
newtons (78,700 pounds)
As indicated previously, the coefficient of drag can be determined by
doubling the wetted height and assuming both underflow and overflow
for a flat surface 1.83 meters high and 12.5 meters long. Interpolating
in Table 6 for a flat plate for L/d = 6.8 gives
CD = 1.24
The velocity can be obtained from equation (318), so for h
u = 2 ~ = 219.81(1.83) = 8.47 meters per second
From equation (336), the drag force is
F = 1,026(1.24)(1.83 0.91)(12.5) (
8
·
47
)
2
D 2
F = 5.24 x 105 newtons (1.17 x 105 pounds)
D
1.83 meters
which will act against the side of the locomotive at a distance, Z,
above the ground, given as
z = 0.91 + (1.83  0.91)
2
Z 1.37 meters (4.5 feet)
181
The downward force from the mass of the locomotive is the mass, m,
times gravitational acceleration, g, or
F mg= 104,500 kilograms (9.81 meters per second squared)
F 1.025 x 10
6
newtons (2.3 x 10
5
pounds)
Taking overturning moments about a rail, the center of mass of the
locomotive is equidistant from the two rails, or 0.76 meter (2.5 feet)
from the rail . The buoyancy and drag forces produce overturning moments
(+) and the mass of the locomotive a restraining force(). Summing
moments
M FB(O. 76) + FD Z  F(O. 76)
M 3.53 x 10
5
(0.76) + 5.24 x 105(1.37)  1.025 x 106(0.76)
M 2.07 x 10
5
newtonmeters (1.48 x 10
5
footpounds)
indicating that the overturning moments are greater than the restraining
moment. Therefore, the locomotive will be overturned.
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
* * * * * * * * * * * * * * EXAMPLE PROBLEM 26 * * * * * * * * * * * * *
GIVEN: A platform, 3 meters above ground level, is supported by square
columns with 14 by 14centimeter (5.5 by 5.5 inches) cross sections.
A tsunami creates a surge with a depth of 2.44 meters (8 feet) under the
platform. The surge acts normal to the sides of the columns, which are
rigidly fixed at ground level.
FIND: The moment of the surge force about the base of a column.
SOLUTION: To determine the coefficient of drag, the columns may be con
sidered as infinitely long columns, and from Table 6, Cv = 2.0. From
equation (318)
u = 2 ~ = 219.81(2.44) = 9.79 meters (32.1 feet) per second
The drag force on a column is given by equation (336) as
C
A u
2
1,026(2) (2 .44) (0 .14) (
9
·
79
)
2
P D Z = 2
3.36 x 10
4
newtons
182
The velocity is assumed to be equal over the 2.44meter depth so that
the resultant drag force acts 1.22 meters (4 feet) above ground l evel .
The moment is then
M 3.36 x 10
4
(1.22)
4.1 x 10
4
newtonmeters (2.93 x 104 footpounds)
on each of the columns.
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
As indicated in example problem 25 drag forces and surge forces can
act in conjunction with buoyant forces. The buoyant forces can lift
buildings from their foundations, and the surge or drag forces can slam
them into such things as trees or other structures. Buildings that are
firmly attached to their foundations to resist the buoyant forces must
also have sufficient structural strength to withstand the drag forces
acting against them. The drag forces can be lessened by constructing a
building on an elevated platform sorne distance above the ground. In sorne
instances, the first floor of a building may be designed to be carried
away by the tsunami, thereby reducing the forces on the building and
protecting the higher floors. However, this may be an expensive solution
and has the undesirable feature of adding debris to the water.
The high velocity of a tsunami surge can also damage structures by
scouring material near the structures' foundations. Shepard, MacDonald,
and Cox (1950) noted numerous instances of severe erosion caused by the
1946 tsunami in Hawaii. At Haena Bay, a sand beach eroded and sand was
deposited 1.2 meters deep across a highway. A section of shoreline at
Moloaa, was eut back about 21 meters (70 feet). At Kalaupapa, the back
wash from the tsunami undermined a raad. Other instances of erosion
were also noted. Erosion and deposition of surface material are quite
common when severe tsunamis occur. Imamura (1942) gives an example from
1707, when a tsunami washed away layered sediments which had covered an
old ricefield. Conversely, the Earthquake Research Institute (1934)
reported instances of ricefields being covered with sand by the 1933
Sanriku tsunami. Instances of deposition of sand are also indicated by
Shepard, MacDonald, and Cox (1950) in Hawaii, and by Reid and Taber (1919)
in Puerto Rico. Shepard, MacDonald, and Cox noted that dense stands of
grass prevented or greatly diminished ground erosion during the 1946
tsunami in Hawaii.
d. Impact Forces. The high velocity of a tsunami surge will sweep
large quantities of material forward with the surge. This material may
include automobiles, trees, petroleum tanks, buildings, debris from build
ings, or other materials in the path of the surge. A large boulder moved
by the 1960 tsunami at Hawaii is shawn in Figure 68. In higher latitudes,
when tsunamis occur during the winter, the material may include large
quantities of broken ice.
183
t'
00
Figure 68.
. !
Large boulder moved by 1960 tsunami, Hilo, Hawaii (from Matlock, Reese, and Matlock, 1962) .
Impact forces from material carried forward by the current are not
as easily analyzed as other forces. The drag force will initially
accelerate material which is swept up into the current. The velocity
of forward motion of such material depends on the distance the material
has moved, the ratio of the drag force to the actual mass of the abject
in motion, and the resistance created by the abject dragging against
the ground or impacting and grinding against other abjects.
Analyzing the effects of a structure impacting with another structure
also requires knowledge of the rigidity of the structures and the angle
of impact. If the flat side of one structure impacts with the flat side
of a second structure, the impact force is spread over a wide area. How
ever, if a corner of the first structure impacts with the flat side of
the second structure, the force is concentrated and there will be a
greater tendency to crush the impacting structures. It should be remem
bered that if a structure is partially flooded, the water within the
structure becomes a part of the mass of the structure.
Considering an abject being swept forward from a stationary position
by a moving fluid such as a tsunami surge, the velocity of the fluid,
u, with respect to the ground is assumed to be constant, and the velocity
of the abject, ub, with respect to the ground varies as the abject is
accelerated. The veloci ty, u'b, of the obj ect approaches the veloci ty,
u, of the fluid after the abject has moved sorne distance (i.e., the
velocity of the abject approaches sorne terminal velocity). The force
accelerating the body is a combination of drag forces and inertia forces,
and is given by the equation
where
p
A
t
F
(337)
the coefficient of drag
the density of water
the crosssectional area of the abject transverse to
the direction of motion
the velocity of the water with respect to the abject
at any instant in time
the inertia or mass coefficient
the volume of water displaced by the abject
ti me
185
For a structure or any other large object floating in the water, the
mass, rn, of the object is equal to the displaced mass, pV, of the
water. This mass may vary as w a ~ e r gradually floods the interior of a
structure, but for the analysis presented here the mass will be assumed
constant. From Newton's second law
F
(338)
At any instant in time the magnitude of the deceleration of the fluid
wi th respect to the object is equal to the magnitude of the acceleration
of the ground with respect to the object (which is equal to the accelera
tion of the object with respect to the ground), i.e., where u is assumed
constant,
d ~
dt
so equation (337) becomes
2
dub (u  ub) dub
F = pV = CD pA   C pV 
dt 2 M dt
or, rearranging terms,
d ~ =
dt
(339)
(340)
(341)
For an object moving a short distance, the coefficients CD and CM
will be assumed constant. This is not entirely correct (e.g., the value
of Cv will vary as a function of velocity), but will be assumed as
approximately correct for a short distance. A constant, a, can then
be defined by
Cif
a=
2V(l + CM)
(342)
Substituting equation (342) into equation (341) and rearranging terms give
d ~
a dt =
(u  ~ ) 2
Integrat ing equation (343),
1

u ~
186
1
u
(343)
(344)
whi ch reduces to
ub = u  cmt + 1
u
(345)
defining the velocity of the object at any time t.
The distance, x, traveled by the object as a function of time can
be determined by noting
Substituting equation (345) into equation (346) and integrating give
which gives
x = ut  l fu (aut + 1)
Cl
(346)
(34 7)
(348)
Typical drag coefficients are given in Table 6. The coefficient of
added mass, CM, can be estimated for a rectangular structure by using
the results of Riabouchinski (1920) as given by Brater, McNown, and Stair
(1958) (Fig. 69). The values in Figure 69 are for irrotational flow
without separation, and the formation of a wake behind the structure
would be expected to modify these values. Individual madel tests would
be required to obtain exact values. Example solutions of equation (348)
are shown in Figure 70.
0.2 0.4 0.6 0.8 1.0
o/b
2 4 6 8 10
Figure 69. CM for twodimensional flow past rectangular
bodies (irrotational flow with no separation)
(from Riabouchinski, 1920).
187
Ois tance,
1
(ml
Figure 70. Example plots of x versus t for abjects moved by tsunami surge.
* * * * * * * * * * * * * EXAMPLE PROBLEM 27 * * * * * * * * * * * * * *
GIVEN: A tsunami surge is 5 meters high .at the shoreline. A building
rüeated at the shoreline is swept forward a distance of 6.1 meters
and impacts with another building. The building is rectangular, 12
meters (39.4 feet) wide and 6 meters (19.7 feet) deep in the direction
of flow, and is submerged to a 3meter depth as it is carried forward
(see Fig. 71). The velocity of the surge is approximated as u = 14
meters per second.
FIND:
(a) The time required for the building to impact with the other
building,
(b) the force accelerating the building at the moment of impact, and
(c) the momentum of the building at the moment of impact.
SOLUTION:
(a) The submerged crosssectional area of the building, transverse
to the direction of the surge, is given as
A= width x submerged depth = 12.0 x 3.0 = 36 square meters
and the submerged volume (the displaced water) is
V = width x length x depth = 12.0 x 6.0 x 3.0 = 216 cubic meters
188
r"'
Building impocts with \
.,.,
.,.,
Building set
into motion
Building moving
second build,..in...;g;._ __ i\
\
\
\
\
\
F F F
\
t = 0 t = 1. 0 second
x = 0 x = 1. 7 meters
t = 2. 02 seconds
x = 6.1 meters
\
\
\
\
F = 9.12xlos newtons F = 5,45xl0
5
newtons F = 3.59xl0
5
newtons
Mo=O Mo=
7.05x10
5
newtonseconds
u = 14 meters/second
Cv= 1.13
CM= 3.5
A "' 36 meters
2
V = 216 meters
3
Mo=
1.16xl0
6
newton
·seconds
Figure 71. Building moved by tsunami surge.
The coefficient of drag can be approximated by assuming the side of
the building is a flat plate. To determine an equivalent flat plate
using Table 6, assume that the submerged depth for underflow and
overflow (a totally submerged plate) is twice the depth of the
building, or
.!: = 12.0
d 2 x 3.0
2.0
and from Table 6
From Figure 69, where
a ~ =
b = 12.0 °'
5
th en
and equation (342) gives
1.13 x 36
0.021
2 x 216(1 + 3.5)
189
...
The relationship between distance and time is shawn in Figure 70,
which gives, for x= 6.1 meters,
t = 2.02 seconds
(b) From equation (345),
u 14
~
= u  14 
aut + 1 (0.021
x
14
x
2. 02) + 1
~
5.22 meters (17.1 feet) per second
Substituting equations (342) and (341) into equation (338),
F pVa (u  ~ )
2
F 1,026 x 216 x 0.021(14  5.22)
2
3.59 x 10
5
kilogrammeters per second squared
F 3.59 x 105 newtons (8.1 x 10
4
pounds)
(c) Momentum, M
0
, at impact is
M
0
= ~ x mass
taking the mass of the building equal to the mass of the displaced
water for a partially submerged building which is floating (the mass
includes water within the building),
mass = pV = 1,026 kilograms per cubic meter x 216 cubic meters
and the momentum is
M
0
ub x mass = 5.22 x 1,026 x 216
1.16 x 106 kilogrammeters per second
(2.56 x 105 poundseconds)
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Magoon (1965) indicates that substantial damage occurred at Crescent
City during the 1964 tsunami as a result of debris impacting on structures.
This debris included logs, automobiles, and baled lumber. The impact
forces either destroyed the loadcarrying capacity of walls, or caused
bending or breaking of light columns.
Wilson and t ~ r u m (1968) discussed some instances of impact damage
resulting from the 1964 Alaskan tsunami. Figure 72 illustrates the damage
at the Union Oil Company tank farm at Whittier. Buildings and larger
tanks were able to withstand the force of the tsunami; however, smaller
190
t'
\.0
t'
····.;:·· ····r='· ··' ·n
/ \ / \ \ \ lUSC!I ., !itL CG. "1
/ \ 1 \ 1 \
1 \ 1 Buo•l5 1 \ \ :
1 4106
1
\ SCAAP 1 \  \170 FEU WHAA, 1
1 \ 53 480 BBLS 1 SWEPT IWAYCOIIIPL.ETf LOSS
, UI\III.CAJ(8[A[PAIA(0 ! t • \ \
\ 53,4BOBBL. / \ / 1 \
\ / . ', // \ \ :
',, ,//. '........ ____ ..... \ 1
· 1 \
1
,
l' '
1 \
1 }9}6
D
Pu.IUlMQUSf
COMP\..11'1 ....... P\.Atf
D
D
u
801\.(lt • GAA&G[
CC,IIIIIl.(f( IU1t'0UT .,_,.LA(f
COOII[P&Gf & lllllllii[MC:JSf
COMPLET[ IUANOUl lllf IlL AC[
• •
1
1
1
1
1:
1
1
!< •< "
0
'• UfoiO<; 01\. CCY•A•n OF 1
.. .. .. ... .. ·· .. .. .. •. .. •.
Figure 72. Damage to oil tank farm at Whittier, Alaska (from Wilson and 1968;
data from Union Oil Company of California).
tanks were carried forward by the surge and impacted with other tanks.
Sorne of the larger tanks were apparently set into motion by the impact,
and most of the tanks were ruptured. A resulting fire destroyed the
tank farm.
Wilson and T ~ r u m also mentioned the problem of a smallcraft harbor
located immediately in front of a developed shoreline at Kodiak City.
The boat harbor contained a large number of fishing boats and yachts
which were carried into the adjacent waterfront business area by the
tsunami, adding substantially to the damage. Van Dorn (1965) notes that
harbor regulations could be instituted requiring ships large enough to
damage harbor structures to stand clear of a harbor in the event of a
tsunami warning. In the case at Kodiak City there was only about 30
minutes between the tsunami warning and the arriva! of the first large
wave crest of the tsunami. (Spaeth and Berkman, 1972). However, when
tsunamis are generated from distant sources there may be enough time to
clear the harbors.
An interesting example of impact forces is reported by Wilson and
T ~ r u m (1968). During the 1964 tsunami a house was washed out to sea
near Point Whiteshed. The house was swept more than 12 miles along the
shoreline, carried into the harbor at Cordova, and rammed the dock,
destroying the end of the dock.
e. Hydrostatic Forces. Hydrostatic forces are normally relatively
small compared to surge and drag forces. The hydrostatic force on a
wall, per foot width of wall, for a water depth h is
F (349)
As seen in example problem 4, the hydrostatic force would probably not
exceed 10 to 20 percent of the drag force at higher water levels, and
would appear tq be relatively insignificant at lower water levels.
Once the initial surge has passed a structure, assuming that water
levels are equal on all sides of the structure, the hydrostatic force
will not contribute to the motion or potential motion of the structure.
However, this force can cause cracking of exterior walls and interior
flooding of the structure.
Magoon (1962) indicated that the flooding caused by a tsunami can
saturate the fil! behind a retaining wall. Combined with the large
drawdowri of the water leve! which may occur at the seaward toe of a wall
during the withdrawal of a tsunami wave, large hydrostatic forces on the
wall may result. It is believed that this contributed to the partial
failure of a retaining wall at Crescent City, California.
There was an unusual occurrence at the abandoned Kahuku Airfield on
Oahu during the 1946 tsunami (Shepard, MacDonald, and Cox, 1950). Blacks
192
of pavement were tilted in circular areas 1 to 1. 5 meters (3 to 5 feet)
across, apparently as a result of hydraulic pressure from water penetrat 
ing into the sand under the pavement when the tsunami flooded the area.
The higher pressure under the pavement has not been explained, but could
have resulted from water trapped in the sand during a rapid withdrawal
of the tsunami.
f. Other Hazards. When considering the total effects of a tsunami
surge, additional hazards should be considered in addition to the actual
forces of the surge. Sorne of these are listed below:
(a) Contamination from debris carried in the surging water;
(b) effects of flooding, including spoilage of goods and
materials, shorting of electrical lines and transformers, and
contamination of water supplies with saltwater;
(c) fire and explosion from the impact and rupturing of
petroleum tanks or containers of chemicals (see Fig. 72); and
(d) release of poisonous gas or toxic materials from ruptured
containers.
VIII. TSUNAMI WARNING SYSTEM AND INSTRUMENTATION
Cox (1964) discusses the development of the tsunami warning system
in Japan. Local, informai warning systems operated sporadically for
centuries, and a formai tsunami warning system was recommended as early
as the late 19th century. The Japan Meteorological Agency organized the
present Japanese system in 1941.
Spaeth and Berkman (1972) discuss the early history of the seismic
sea wave (tsunami) warning system in the United States. The need for a
warning system in the United States was recognized following the 1 April
1946 tsunami generated in the Aleutian Islands. That tsunami caused
heavy damage and resulted in the loss of many lives in Hawaii, particu
larly at Hilo.
The present tsunami warning system was organized by the former U.S.
Coast and Geodetic Survey (now the National Ocean Survey, National Oceanic
and Atmospheric Administration NOAA). Tsunami detectors were designed
and installed at tide stations to alert personnel of forthcoming tsunamis.
The first detector was installed at Honolulu in 1947. Meetings to discuss
implementation of the warning system throughout the Pacifie coastal areas
were held in 1948. The tentative communication plan for the warning
system was approved in 1948. Initially, the warning system supplied
information to civil authorities in Hawaii and to military bases through
out the Pacifie. In 1953, Civil Defense Agencies of California, Oregon,
and Washington were added to the agencies receiving tsunami warning
information, and the system has expanded since that tirne.
19 3
Wei gel (1 974) notes that management of the tsunami warning system
was transferred to the Environrnental Research Laboratori es, NOAA , in
1971, then to the National Weather Service in 1973.
1. The Tsunami Warning System.
The Intergovernmental Oceanographi e Commission (l OC) maintains an
International Coordination Group for t he Tsunami Warning System in the
Pacifie. Member countries are Canada, Chi le, China, Ecuador, France,
Guatemala, Japan , Korea, New Zealand, Peru , Phil ippines , Thailand,
U.S.S.R., and the United States. The warning system is based on 51
tide stations and 32 seismograph stat ions (1976 lOC dat a ). The tsunami
warning system i n the United States is based at t he Honolulu Observatory
and receives dat a from 18 seismograph s t ations and 16 tide states in the
United States (including Alaska and Hawaii , but excluding Pacifie Ocean
territories) , as well as the remaining 14 seismograph and 35 tide sta
tions of the Tsunami Warning System in the Paci f ie (including Pacifie
Ocean territories). Tsunami warnings for the Uni ted St ates are based
primarily on the stations in the United States and on thos e stations
located in both North and South America. Stat ions i n the Pacifie Ocean
territories of t he United States and t hose in t he remainder of the
Pacifie provide additional informati on on the t sunami s ; but generally,
tsunamis that would create a hazard on t he coast lines of the United
States do not arrive from those direct ions .
The tsunami warning system functi ons best for dis t antly generated
tsunamis, i .e ., t sunamis where the ar r i va i t ime i s several hours after
the initial wave generation. However, the warning s ys t em can also alert
the population to the possibility of t sunami generat ion from nearby
seismic activi ty. As the period of the tsunami varies from several min
utes to approximately 30 to 40 minut es , cont i nuously recording tide gages
are required . Modern, digital tide gages whi ch pr ovide a tide height
every several minute s will not provide suffici ent data for recording
tsunamis.
An alarm at tached to the seismogr aph at Hono lulu Observatory is
triggered by the arrivai of seismic waves, initiat ing activity in the
Tsunami Warning System. The 1964 Alas ka earthquake began at 0336 G.m.t.
(Spaeth and Berkman, 1972), and the alarm sounded at Honolulu Observatory
at 0344 G.m.t. (G.m.t. times are used to provide a uniform time at all
points in the system.) After the alarm, i nquir ies are sent to various
seismic obs ervatories in the system to obt ain sei smic r eadings. After
receiving and evaluating init1al dat a , a deci sion is made as to whether
or not an advi s ory bulletin should be i ssued. I f there i s a possibility
of tsunami generation, an advisory bullet in i s sent to dissemination
agencies i n the warning system.
Char t s showing tsunami tyaveltimes between various coa stal points
and t sunami generating areas have been prepared . Thes e charts are
used to pr edict the arrivai time of a potent ial tsunami at the various
coastal points after the epicenter of the earthquake has been determined.
194
Inquiries are sent to tide stations near the earthquake epicenter to
determine if tsunami waves have been generated. If there is reasonable
cause to assume that a tsunami may have been generated (based on seismic
records or observations at tide stations), an advisory bulletin is issued
giving estimated times of arrivai (ETA's) of the tsunami at various
coastal points.
The Honolulu Observatory continues to receive reports and issue
bulletins to Civil Defense agencies as information becomes available.
Various agencies are kept advised of the status of the tsunami until an
allclear bulletin is issued. If a tsunami was generated, various report
ing stations continue to report water levels to the Honolulu Observatory
until all information has been obtained from the tide stations of the
Tsunami Warning System in the Pacifie.
Cox (1964) indicates that tsunami warnings, as distinct from advisory
bulletins, are issued only under the following circumstances:
(a) Unusual sea level disturbances having tsunami character
istics are recorded at one or more of the warning system tide
stations scattered about the Pacifie.
(b) No reply is received from a tide station in a critical
recording position in response to a query from the Honolulu
Observatory after the occurrence of an earthquake large enough
to trigger the seismograph alarm.
(c) An earthquake occurs whose epicenter is in or on the
borders of the Pacifie Ocean in such a location that a tsunami
generated there would not arrive at any tide stations sufficiently
in advance of its arrivai at a particular shoreline to allow warn
ing that shoreline.
It is the responsibility of agencies receiving tsunami warnings to
disseminate information to the civilian population. This may be clone
through broadcasting news media (radio and television), by police and
civil defense personnel, or by sounding signais on sirens. Schank (1978)
discusses the public warning system in the State of Hawaii. Broadcasters
in the State are linked together in a system called "CivAlert." If it
is considered probable that a tsunami will cause property damage and loss
of life, a tsunami alert is given. The first alert is given 2 hours
before the estimated ETA by sounding an alert on all of the sirens in
the State. CivAlert broadcasts information and instructions. The alert
is repeated at 1 hour and then a half hour before the estimated ETA. The
alert is extended by police, fire, forestry, and Civil Air Patrol person
nel. Maps of potential tsunami hazard areas are included in the county
telephone directories in Hawaii to define areas which should be evacuated.
Cox (1978) considered the cost of false alarms, i.e., tsunami alerts
when no significant tsunami occurs, g1v1ng the cost of false alarms at
$264,000 per year (1977) in Hawaii alone. The cost of false alarms must
195
be balanced against the high cast of casualties to the local population
if a tsunami occurs without sufficient warning.
Cox and Stewart (1972) discuss the particular problems of providing
tsunami warnings to areas near a tsunami source. Tsunami warnings must
be timely to be effective. To reduce the hazards on particular coast
lines, a policy of regional evaluation was adopted in 1966. Adams (1978)
discusses the local tsunami warning system used in Hawaii. Seismic
detection instruments have been placed in police and fire stations and
similar locations. Seismic activity with a magnitude which may generate
a tsunami triggers an alarm. A decision is then made at the local level
as to whether or not the local population should be alerted and evacuated.
Various investigators have called attention to phenomena occurring
just before a tsunami. These include the unusual feeding habits of
fish before a tsunami due to the presence of large quantities of bottom
adherent diatoms in the upper layer of the sea (Suyehiro, 1934) and
increases in the Earth's magnetic field preceding an earthquake (Moore,
1972). The study of such phenomena has not been developed enough to be
included in a forma! tsunami warning system.
2. Human Response.
Spaeth and Berkman (1972) note that the response of a local population
to a tsunami warning may be slow unless the population is well trained to
respond. Approximately 30 minutes after the Alaska earthquake of 1964,
the U.S. Fleet Weather Central at Kodiak Naval Station received ward of
a large tsunami at Cape Chiniak, Alaska, and had the Armed Forces Radio
Station broadcast a tsunami warning. Military and government personnel
promptly evacuated the endangered areas. Although reasonably prompt,
the evacuation of the city of Kodiak was not as well carried out (there
were eight deaths at Kodiak). The first large tsunami wave crest arrived
at Kodiak about 30 minutes after the warning, so a prompt response to the
warning was essential.
Haas (1978) separates tsunamis into the following four types (summa
rized in Table 7);
(a) Type I. Shoreline slumping, earthslides, and large rock
and ice falls coïncident with the earthquake. Large waves gener
ated onto the shoreline almost immediately.
(b) Type II. Very heavy Earth temblors can be felt by the
local population for a period up to severa! minutes. The tsunami
arrives within 10 minutes.
(c) Type III. Noticeable Earth shocks felt by the local popu
lation for a period up to severa! minutes. Severe Earth temblors
not present. Tsunami arrives within 30 minutes.
(d) Type IV. No local Earth shocks. The tsunami is generated
at a distant source.
196
Table 7. Typology of tsunami events Cafter Haas , 1978) .
Tsunami Physical Approximate time Maximum credible
type elues for evacuation preventive action
I Visible Less th an 1 Almost none
slumping minute
or sliding
II Severe 5 to 10 minutes Ambulatory persans
Earth can be evacuated
temblors
III Noticeable 15 to 30 minutes Sorne persans can be
Earth evacuated
shocks
IV None 45 minutes to Most persans can be
12 hours evacuated and up to
75 percent of all
movable property
Haas notes that while no effective warning can be given for a type
tsunami, the possibility of warnings for types II, III, and IV will
depend on the education of the public and the effectiveness of warning
systems. Evacuation for a type II tsunami requires prompt response by
the population based on their individual sensing of strong earthquake
shocks, and little time is available for an organized warning system to
operate. Therefore, there is almost total reliance on prior education.
For a type III tsunami, public education alone is insufficient because
the physical evidence of a possible tsunami is not as strong. A reliable
local warning system is needed to alert the public, and the population
must be educated to respond to the alert. For a type IV tsunami, a large
warning system, such as the Tsunami Warning System in the Pacifie, is
required.
Weller (1972) cites a number of instances of human response during
the 1964 earthquake and tsunami in Alaska. At Seward, the initial slump
ing of the waterfront gave warning to the residents of the town, and most
people evacuated the low areas; but 11 people were killed by a wave 9 to
12 meters (30 to 40 feet) in height. At the village of Kaguyak on Kodiak
Island, residents moved to high ground when they observed the initial
signs of a tsunami; but three people were killed when they returned to
low areas before the arrival of the largest wave. At Ouzinkie, on Spruce
~ s l a n d , the residents evacuated the town when they observed the initial
development of wave action offshore, and there was no loss of life.
Many lives are lost either because sorne residents do not respond to
visible signs of a possible tsunami, such as at Seward, or residents
197
return tao saon to law areas, such as at Kaguyak. Spaeth and Berkman
(1972) note that severa! people were killed at Crescent City, California,
because they returned to a law area before the arrivai of the largest
wave. A large amount of damage at Crescent City resulted from the failure
to remove vehicles, including a gasoline tank truck, from the endangered
area.
3. Ionospheric Waves.
To evaluate the possibility of a tsunami being generated by an earth
quake, it is desirable to have information about the source mechanism of
the earthquake, i.e., whether the earthquake is a dipslip type or a
strikeslip type. If the earthquake is a strikeslip type, it may be
assumed that a large transoceanic tsunami will not be generated, and
tsunami alerts can be canceled at ali locations except those near the
epicenter. If the earthquake is a dipslip type, there is a high proba
bility that a tsunami may have been generated, and additional information
must be obtained from tide stations.
Van Dorn (1965) indicated that a dipolar barometric wave in the
atmosphere was associated with the dipolar ground motion of the 1964
Alaska earthquake , and that this raised a possibility for early tsunami
prediction. Row (1972) discusses the atmospheric waves associated with
the Alaska earthquake in greater detail. Row indicates that there is
bath an earlyarriving pressure disturbance, associated with seismic
waves in the Earth, and the latearriving disturbance (propagating at
about 300 meters per second) previously mentioned by Van Dorn. The late
arriving disturbance is associated with the tectonic deformation and, at
distances far from the source, would arrive weil in advance of gravity
waves traveling through the ocean. For example, at a distance of 5,000
kilometers (3,100 miles) from the source, a tsunami traveling across the
ocean at a speed of 200 meters per second (447 miles per hour) will
arrive approximately 2 hours 20 minutes after the atmospheric wave.
Pressure disturbances also propagate through the ionosphere. Row
(1972) discussed the possible association between ground motion and
ionospheric waves. Furumoto (1970) reports on the use of a lümegahertz
Doppler recording of Rayleigh waves to estimate the initial phase of the
source of the 11 August 1969 Kuril Islands earthquake. He notes that
this provides a rapid approach to source mechanism estimation. The
Doppler shift associated with the ionospheric waves can be monitored at
relatively law cast (Furumoto, 1970). Murty (1977) provides further
discussion on ionospheric effects.
The use of atmospheric waves to estimate the ground motion of a
tsunami source requires further investigation. However, this method may
be useful for the Tsunami Warning System.
4. DeepOcean Tsunami Gages.
In addition to supplying information for the Tsunami Warning System,
the tide stations provide records of tsunami heights and periods. Unfor
tunately, the local topography distorts the tsunami recorded on tide gages
near the coastline, and these records do not provide information on the
198
deepocean form of the tsunami. Therefore, a means of recording tsunamis
in the open ocean is needed.
Vitousek (1961) proposed placing permanent instrument packages in
the ocean, connected to abandoned transpacific telegraph cables. A
system of this type would provide deepocean data, and would also provide
additional useful information for the Tsunami Warning System because of
the direct connection with the gage. However, Vitousek and Miller (1970)
indicate that cableconnected systems 'would be expensive, and that the
cost of laying special cables would be unrealistic.
Vitousek and Miller discuss four possible methods of measuring a
tsunami in the open ocean: (a) Freedrop recoverable instrument package,
(b) an undership instrument, (c) an underbuoy instrument, and (d) the
cableconnected instrument previously discussed by Vitousek. Shinmoto
and Vitousek (1978) give details of an airdeployable freedrop tsunami
gage which can be emplaced quickly after a tsunami occurs.
While sorne deployment of openocean gages has been carried out,
experience in operating such gages is limited. Future use of such gages
is required to determine their practicality and reliability as part of
the Tsunami Warning System, and to obtain openocean tsunami data.
IX. SUMMARY AND CONCLUSIONS
The potentially high value of property in the coastal zone and the
intensive development of such land for both private and public use
require that careful consideration be given to the possibility of cata
strophic flooding of areas of the coastal zone, in or near seismologi
cally active regions, by tsunamis. Small variations in predicted flood
levels may affect property worth millions of dollars, and may have
substantial effects on flood insurance premiums and permits for utiliza
tion of property, so it is necessary to have as high a degree of accuracy
as possible in defining flood zones, e.g., the 100year flood level.
Also, large powerplants are typically located at low elevations because
of pumping requirements for cooling water, and port facilities are neces
sarily located near the shoreline, so that welldesigned protection is
required for highcost facilities.
Available data on tsunami inundation come from visual observations
(including posttsunami surveys) and from tide gage records. Data are
generally only available for a few occurrences, and only at specifie
coastal areas. Sorne data can be obtained from historical accounts, but
such data are dependent on incomplete persona! observations, usually by
untrained observers. Openocean data on tsunamis, needed for verification
of numerical investigations, are almost nonexistent.
Numerical data are used to supplement the available field data on
tsunami flood levels. Numerical procedures have been developed that
allow the simulation of a tsunami source, the generation and propagation
of the tsunami waves across the ocean, and the interaction of the tsunami
and coastal topography. Procedures have also been developed to simulate
tsunami flooding shoreward of the coastline. The numerical results, which
are compared to the more limited field measurements for verification,
199
provide the additional data needed to construct tsunami flood level
maps for various probabilities of recurrence.
Numerical procedures can also be verified by comparing with theoreti
cal results for idealized cases. Theoretical solutions exist for wave
r efraction at coastlines with uniform topography, waves passing over
mathematically defined transitions from deep water to shallow water, etc.
Deviations between the numerical results and the theoretical solutions
indicate the degree of accuracy where the numerical procedures are applied
to more complex topography.
A continuing program of gathering field data on tsunamis in the open
ocean and coastal inundation by tsunamis is needed. Because of the long
periods of time between the occurrence of tsunamis, the accumulation of
data for particular coastal points is very slow. It is necessary to
maintain tide gages with the capability of recording tsunamis, and also
to have standby plans with designated personnel to obtain field observa
tions immediately after tsunamis occur. It is also desirable to maintain
a st andby capability for dropping instrument packages into the open ocean
immediately after a tsunami occurs. This latter capability requires the
maintenance of gages and associated instrument packages in operating
condition over long periods of time, and the maintenance of a system for
placing the instrument packages quickly and on short notice, including
the periodic testing of the system by placing and recovering the instru
ments. An airdropped system is probably the most practical for this
purpose.
Also, continuing improvements are needed in the numerical procedures
for simulating tsunamis. A particular area of possible improvement is
t he treatment of boundaries of the computational grid. Errors in the
wave reflection from solid boundaries, and errors at openocean boundaries
where the waves must pass completely through the boundary, propagate
through the computational grid at each succeeding time step. These errors
grow with increasing time so that the solution is not accurate for long
per iods of real time. It becomes necessary to use large time steps to
r educe computational errors, and consequently to use a course grid, i.e.,
to use long real distances between grid points. This smooths out the
topographical variations so that wave scattering caused by small topograph
ical features is not properly accounted for. Because of the limited field
data available, the numerical solutions cannat always be verified and
adjusted to match field data.
Improvements in the numerical simulation of tsunami generation are
al so desirable. However, this requires both accurate data on real tsunami
generating mechanisms, and openocean tsunami data so that errors in
simulating tsunami generation can be separated from errors in the simula
tion of nearshore propagation.
Continued research should be carried out in areas such as shelf
resonance. In particular, theoretical solutions are needed for simple
topography to provide verification for numerical procedures where field
data do not exist. At the present time, the various effects on tsunami
propagation cannat be adequately separated in the computational procedure
as t he available data are mainly from tide gages and visual observations
of maximum inundation levels.
200
LITERATURE CITED
ABE, K., "Reliable Estimation of the Seismic Moment of Large Earthquakes, "
Journal of Physics of the Vol. 23, 1975, pp. 381390.
ABE, K., "Size of Great Earthquakes of 18371974 Inferred from Tsunami
Data," Joùrnal of Geophysical Vol. 84, No. B4, Apr. 1979,
pp. 15611568.
ADAMS, W.M., "Relationship of Instruments and Policy in the Hawaii Warning
System," Proceedings of the Symposium on Department of Fisher
ies and the Environment, Ottawa, Canada, 1978, pp. 205217.
AMBRASEYS, N.N., "The Seismic Sea Wave of July 9, 1956, in the Greek
Archipelago," Journal of Geophysical Vol. 65, No. 4, Apr.
1960, pp. 12571265.
AMBRASEYS, N.N., "Data for the Investigation of Seismic SeaWaves in
Europe," Monograph No. 29, European Seismological Commission, Interna
tional Union of Geodesy and Geophysics, Paris, France, Nov. 1965.
BEARD, L.R., "Statistical Methods in Hydrology," U.S. Army Engineer
District, Sacramento, Sacramento, Calif., Jan. 1962.
BENJAMIN, T. B., BONA, J. L., and MAHONY, J .J., "On Model Equations for
Long Waves in Nonlinear Dispersive Systems," Philosophical Transactions
of the Royal Series A, Vol. 272, 1972, pp. 4778.
BENJAMIN, T.B., and FEIR, J .E., "The Disintegration of Wave Trains on
Deep Water," Journal of Fluid Vol. 27, Part 3, 1967, pp.
417430.
BENMENAHEM, A., "Radiation of Seismic SurfaceWaves from Finite Moving
Sources," Bulletin of the Seismological Society of Vol. 51,
No. 3, July 1961, pp. 401435.
BERG, E., "The Alaska Earthquake; Its Location and Seismic Setting,"
Proceedings of the 15th Alaska Science American Association
for the Advancement of Science, 1964, pp. 218232.
BERG, E., et al., "Field Survey of the Tsunamis of 28 March 1964 in
Alaska, and Conclusions as to the Origin of the Major Tsunami," Report
No. HIG702, Hawaii Institute of Geophysics, Honolulu, Hawaii, Jan,
1970.
BLANCH, G., "Mathieu Functions," Handbook of Mathematical
Ch. 20, National Bureau of Standards, U.S. Government Printing Office,
Washington, D.C., 1964.
BOURODIMOS, E.L., and IPPEN, A.T., "Wave Reflection and Transmission in
Channels of Variable Section," Proceedings of the llth Conference on
Coastal American Society of Civil Engineers, Ch. 13, 1968,
pp. 195212.
201
BOWEN, A.J., and INMAN, D.L., "Edge Waves and Crescentic Bars," Journal
of Geophysical Research
3
Vol. 76, No. 36, Dec. 1971, pp. 86628671.
BRANDSMA, M., DIVOKY, D., and HWANG, LS., "Seawave A Revised Madel
for Tsunami Applications," Final Report, National Science Foundation,
Washington, D.C., Apr. 1975.
BRANDSMA, M., DIVOKY, D., and HWANG, LS., "Tsunami Atlas for the Coasts
of the United States," Report No. TC486, U.S. Nuclear Regulatory
Commission, Washington, D.C., Nov. 1979.
BRATER, E.F., McNOWN, J.S., and STAIR, L.D., "Wave Forces on Submerged
Structures," Journal of the HydrauUcs Division
3
Vol. 84, No. HY6,
Nov. 1958.
BRETSCHNEIDER, C.L., and WYBRO, P.G., "Tsunami Inundation Prediction,"
Proceedings of the 15th Conference on Coastal Engineering
3
American
Society of Civil Engineers, Ch. 60, 1976, pp. 10061024.
BUTLER, H.L., and DURHAM, D.L., "Applications of Numerical Modeling to
Coastal Engineering Problems," Proceedings of the 19?6 A ~ y Numerical
Analysis and Computers Conference
3
Sept. 1976, pp. 471508.
BYRNE, R.J., "Field Occurrences of Induced Multiple Gravity Waves,"
Journal of Geophysical Research
3
Vol. 74, No. 10, May 1969, pp. 2590
2596 (copyrighted by American Geophysical Union).
CAMFIELD, F.E., "Wave Trapping of a Coastline Approximated by a Circular
Arc," Proceedings of the Symposium on Long Waves in the Ocean
3
Depart
ment of Fisheries and the Environment, Ottawa, Canada, 1979, pp. 183
187.
CAMFIELD, F.E., and STREET, R.L., Stanford University, Department of
Civil Engineering, Palo Alto, Calif., unpublished data, 1967.
CAMFIELD, F.E., and STREET, R.L., "The Effects of Bottom Configuration
of the Deformation, Breaking and Runup of Solitary Waves," Proceedings
of the llth Conference on Coastal Engineering
3
American Society of
Civil Engineers, Ch. 11, Sept. 1968, pp. 173189.
CAMFIELD, T.W., "The 'Great Flood' of 1866," Port Townsend Leader
3
Summer
Supplement, Port Townsend, Wash., 1975.
CARRIER, G.F., SHAW, R.P., and MIYATA, M., "Channel Effects in Harbor
Resonance," Journal of the Engineering Mechanics Division, Vol. 97,
No. EM6, Dec. 1971, pp. 17031716.
CHAN, R.K.C., and STREET, R.L., "A Computer Study of FiniteAmplitude
Water Waves," Journal ofComputational Physics
3
Vol. 6, Aug. 1970a,
pp. 6894.
CHAN, R.K.C., and STREET, R.L., "Shoaling of FiniteAmplitude Waves on
Plane Beaches," Proceedings of the 12th Coastal Engineering Conference3
American Society of Civil Engineers, 1970b.
202
CHAN, R. K. C., STREET, R. L., and FROMM, J. E., "The Digital Simulation of
Water WavesAn Evaluation of SUMMAC," Proceedings of the Second Inter
national Conference on Numerical Methods in Fluid Department
of Civil Engineering, Stanford University, Palo Alto, Calif., 1970.
CHAN, R.K.C., STREET, R.L., and STRELKOFF, T., "Computer Studies of
FiniteAmplitude Water Waves," Technical Report No. 104, Department of
Civil Engineering, Stanford University, Palo Alto, Calif., June 1969.
CHAO, YY., "The Theory of Wave Refraction in Shoaling Water, Including
the Effects of Caustics and the Spherical Earth," Report No. TR707,
New York University, Geophysical Sciences Laboratory, New York, N.Y.,
June 1970.
CHAO, YY., and PIERSON, W.J., "An Experimental Study of Gravity Wave
Behavior Near a Straight Caustic," Report No. TR7017, New York
University, Geophysical Sciences Laboratory, New York, N.Y., Dec. 1970.
CÀEN, H.S., and MEl, c.e., "Oscillations and Wave Forces in an Offshore
' Harbor (Applications of the Hybrid Finite Element Method to WaterWave
Scattering)," Report No. 190, Massachusetts Institute of Technology,
Cambridge, Mass., 1974.
CHEN, M., DIVOKY, D., and HWANG, LS., "Nearfield Tsunami Behavior,"
Final Report, National Science Foundation, Washington, D.C., Apr. 1975.
CHEN, T.C., "Experimental Study on the Solitary Wave Reflection Along a
Straight Sloped Wall at Oblique Angle of Incidence," TM 124, U.S. Army,
Corps of Engineers, Beach Erosion Board, Washington, D.C., Mar. 1961.
COCHRANE, J.O., and ARTHUR, R.S., "Reflection of Tsunami," Journal of
Marine Vol. 7, No. 3, 1948, pp. 239251.
COX, C., and STEWART, H.B., "Technical Evaluation of the Seismic Sea Wave
Warning System," The Great Alaska Earthquake of 1964., Oceanography and
Coastal National Academy of Sciences, National Research
Council, Washington, D.C., 1972, pp. 229245.
COX, D.C., "Tsunami Forecasting," Report HIG6415, Hawaii Institute of
Geophysics, Honolulu, Hawaii, Aug. 1964.
COX, D.C., PARARASCARAYANNIS, G., and CALEBAUGH, J.P., "Catalog of
Tsunamis in Alaska," Report SEl, Wôrld Data Center A, Solid Earth
Geophysics, National Oceanic and Atmospheric Administration, Rockville,
Md., Mar. 1976.
CROSS, R.H., "Tsunami Surge Forces," Journal of the Harbors,
and Coastal Engineering Division, Vol. 93, No. WW4, Nov. 1967, pp. 201
231.
DAVIS, T.N., and ECHOLS, C., "A Table of Alaska Earthquakes, 17881961,"
Report No. 8 (UAGRl31), University of Alaska Geophysical Institute,
Fairbanks, Alaska, 1962.
203
DAVISON, C. , Great Thomas Murby & Co., London, 1936.
DEAN, R.G., "Long Wave Modification by Linear Transitions," Journal
of the Waterways and Harbors Vol. 90, No. WWl, Feb. 1964,
pp. 129.
EARTHQUAKE RESEARCH INSTITUTE, "Reports on the Tsunami of 1933 on the
Sanriku Coast, Japan," Bulletin of the Earthquake Research
Tokyo Imperial University, Tokyo, Japan, Vol. 1, Part II (7), 1934.
EATON, J.P., RICHTER, D.H., and AULT, W.U., "The Tsunami of May 23, 1960,
on the Island of Hawaii," Bulletin of the Seismological Society of
Vol. 51, No. 2, Apr. 1961, pp. 135157.
FREEMAN, J.C., and LE MEHAUTE, B., "Wave Breakers on a Beach and Surges
on a Dry Bed," Journal of the HyéJ:r>auZics Vol. 90, No. HY2,
Mar. 1964, pp. 187216.
FROMM, J.E., IBM Research Report RJ531, IBM Research Laboratory, San Uose,
Cal if., 1968.
FUKUI, Y., et al., "Hydraulic Study on Tsunami," Coastal Engineering in
Vol. 6, 1963, pp. 6782.
FUKUUCHI, H., and ITO, Y., "On the Effect of Breakwaters Against Tsunami,"
Proceedings of the lOth Conference on Coastal American
Society of Civil Engineers, Ch. 47, 1966, pp. 821839.
FURUMOTO, A.S., "Ionospheric Recordings of Rayleigh Waves for Estimating
Source Mechanisms," Tsunamis in the Pacifie Ch. 9, EastWest
Center Press, Honolulu, Hawaii, 1970, pp. 119134.
GAGNON, M.L., and BOCCO, M.V., "The Effect of a Gradual Change of Depth
on a Train of Surface Waves," S.M. Thesis, Department of Civil Engi
neering, Massachusetts Institute of Technology, Cambridge, Mass., 1962.
GALLAGHER, B., "Generation of Surf Beat by NonLinear Wave Interactions,"
Journal of Fluid Vol. 49, Part 1, 1971, pp. 120.
GALVIN, C.J., "Resonant Edge Waves on Laboratory Beaches," Transactions
of the American Geophysical Vol. 46, 1965, p. 112.
GALVIN, C.J., "FiniteAmplitude, Shallow WaterWaves of Periodically
Recurring Form," Proceedings of the Symposium on Long University
of Delaware, Newark, Del., 1970, pp. 132 (also Reprint 572, U.S. Army,
Corps of Engineers, Coastal Engineering Research Center, Fort Belvoir,
Va., NTIS AD 754 868).
GELLER, R.J., "Scaling Relations for Earthquake Source Parameters and
Magnitudes," Bulletin of the Seismological Society of Vol. 66,
No. 5, Oct. 1976, pp. 15011523.
204
GELLER, R.J., and KANA.MORI, H., "Magnitudes of Great Shallow Earthquakes
from .1904 to 1952," Bulletin of the Seismological Society of
Vol. 67, No. 3, June 1977, pp. 587598.
GORING, D.G., "Tsunamis The Propagation of Long Waves Onto a Shelf,"
W.M. Keck Laboratory Report No. KHR38, California Institute of
Technology, Pasadena, Calif., Nov. 1978.
GUTENBERG, B., and RICHTER, C.F., Seismicity of the Earth and Associated
2d ed., Princeton University Press, Princeton, New Jersey,
1954.
GUTENBERG, B., and RICHTER, C.F., Seismicity of the Earth and Associated
Hafner, New York, 1965.
GUZA, R.T., and BOWEN, A.J., "The Resonant Instabilities of Long Waves
Obliquely Incident on a Beach," Journal of Geophysical Vol.
80, No. 33, Nov. 1975, pp. 45294534.
GUZA, R.T., and BOWEN, A.J., "FiniteAmplitude Edge Waves," Journal of
Marine Vol. 34, No. 1, 1976, pp. 269293.
GUZA, R.T., and DAVIS, R.E., "Excitation of Edge Waves by Waves Incident
on a Beach," Journal of Geophysical Vol. 79, No. 9, Mar. 1974,
pp . 12 8 5 12 91.
GUZA, R.T., and INMAN, D.L., "Edge Waves and Beach Cusps," Journal of
Geophysical Vol. 80, No. 21, July 1975, pp. 29973012
(copyrighted by American Geophysical Union).
HAAS, J.E., "Human Response to the Tsunami Warning System," Proceedings
of the Symposium on Department of Fisheries and the Environ
ment, Ottawa, Canada, 1978, pp. 224235.
HALLERMEIER, R.J., "Nonlinear Flow of Wave Crests Past a Thin Pile,"
Journal of the Harbors and Coastal Engineering
Vol. 102, No. WW4, Nov. 1976, pp. 365377.
HAMMACK, J.L., "A Note on Tsunamis: Their Generation and Propagation
in an Ocean of Uniform Depth," Journal of Fluid Vol. 60,
Part 4, 1973, pp. 769799.
HAMMACK, J.L., and SEGUR, H., "The KortewegdeVries Equation and Water
Waves, Part 2, Comparison with Experiments," Journal of Fluid
Vol. 65, Part 2, 1974, pp. 289314.
HARRIS, D.L., "Tides and Tidal Datums," U.S. Army, Corps of Engineers,
Coastal Engineering Research Center, Fort Belvoir, Va. (in preparation,
1980) .
HASKELL, N.A., "Elastic Displacements in the NearField of a Propagating
Faul t," Bulletin of the Seismological Society of Vol. 59, 1969,
pp. 865908.
205
HECK, N.H., Princeton University Press, Princeton, N.J., 1936.
HECK, N. H. , "List of Seismic Sea Waves," BuUetin of the Seismo logical
Society of Vol. 37, No. 4, Oct. 1947.
HEITNER, K.L., "A Mathematical Model for Calculation of the Runup of
Tsunamis," Dissertation, Earthquake Engineering Research Laboratory,
California Institute of Technology, Pasadena, Calif., 1969.
HIDAKA, K., "A Theory of Shelf Seiches," The Memoirs of the Imperial
Marine Kobe, Japan, Vol. 6, No. 1, Dec. 1935a, pp. 911.
HIDAKA, K., "Seiches Due to a Submarine Bank," The Memoirs of the Imperial
Marine Kobe, Japan, Vol. 6, No. 1, Dec. 1935b, pp. 18.
HIRONO, T., "The Chilean Earthquake of 1960," Journal of Tokyo
Geographical Society, Tokyo, .Japan, Vol. 70, No. 3, 1961, pp. 2031
(in Japanese with English abstract).
HODGSON, E.A., and DOXSEE, W.W., "The Grand Banks Earthquake, November
18, 1929," Earthquake Vol. 2, Nos. 1 and 2, 1930, pp. 7281.
HOLDEN, E. S., "A Catalogue of Earthquakes on the Pacifie Coast, 1769 to
1897," Smi thsonian Miscellaneous Collections 1087, Smithsonian Insti tu
tian, Washington, D.C., 1898.
HORIKAWA, K., and WIEGEL, R.L., "Secondary Wave Crest Formation," Series
89, Issue 4, Wave Research Laboratory, Institute of Engineering Research,
University of California, Berkeley, Calif., Feb. 1959.
HOUSNER, G.S., "Engineering Estimates of Ground Shaking and Maximum
Earthquake Magnitude," Proceedings of the Fourth World Conference on
Earthquake 1969.
HOUSTON, J.R., "Long Beach Harbor Numerical Analysis of Harbor Oscilla
tions, Report 1, Existing Conditions and Proposed Improvements,"
Miscellaneous Paper H7620, U.S. Army Engineer Waterways Experiment
Station, Vicksburg, Miss., Sept. 1976.
HOUSTON, J.R., "Los Angeles Harbor Numerical Analysis of Harbor Oscilla
tions," Miscellaneous Paper H772, U.S. Army Engineer Waterways
Experiment Station, Vicksburg, Miss., Feb. 1977.
HOUSTON, J.R., CARVER, R.D., and MARKLE, D.G., "TsunamiWave Elevation
Frequency of Occurrence for the Hawaiian Islands," Technical Report
H7716, U.S. Army Engineer Waterways Experiment Station, Vicksburg,
Miss., Aug. 1977.
HOUSTON, J.R., and GARCIA, A.W., "Type 16 Flood Insurance Study: Tsunami
Predictions for Pacifie Coastal Communities," Technical Report H743,
U.S. Army Engineer Waterways Experiment Station, Vicksburg, Miss.,
May 1974.
206
HOUSTON, J.R., and GARCIA, A. W., "Type 16 Flood Insurance Study: Tsunami
Predictions for the West Coast of the Continental United States,"
Technical Report H7826, U.S. Army Engineer Waterways Experiment
Station, Vicksburg, Miss., Dec. 1978.
HOUSTON, J.R., et al., "Probable Maximum Tsunami Runup for Distant Seismic
Events," Appendix 2.4B, Amendment 23, NORCONP1 PSAR, Puerto Rico Water
Resources Authority, Puerto Rico, May l975a.
HOUSTON, J .R., et al., "Effect of Source Orientation and Location in the
Aleutian Trench on Tsunami Amplitude Along the Pacifie Coast of the
Continental United States,'' Research Report H754, U.S. Army Engineer
Waterways Experiment Station, Vicksburg, Miss., July l975b.
HWANG, LS., and DIVOKY, D., "Tsunami Generation," Journal of Geophysical
Vol. 75, No. 33, Nov. 1970, pp . 68026817.
HWANG, LS., and DIVOKY, D. , "A Numerical Model of the Major Tsunami," The
Great Alaska Earthquake of 1964, Oceanography and Coastal Engineering,
National Academy of Sciences , National Research Council, Washington,
D.C., 1972, pp. 191210.
HWANG, LS., and DIVOKY, D., "Numerical Investigations of Tsunami
Behaviour," Final Report , National Science Foundation, Washington, D.C.,
Mar. 1975.
IIDA, K., "Magnitude, Energy, and Generation Mechanisms of Tsunamis and
a Catalogue of Earthquakes Associated with Tsunamis, " Proceedings of
the lOth Pacifie Science Congress Symposium, International Union of
Geodesy and Geophysics, Monograph No. 24, 1961, pp. 718.
IIDA, K., "The Generation of Tsunamis and the Focal Mechanism of Earth
quakes," Tsunamis in the Pacifie Ocean, Ch. l, EastWest Center Press,
Honolulu, Hawaii, 1970, pp. 318.
IMAMURA, A., "An Interpretation of the Reexposure of an Old Rice Field
Buried in the Ground," Zisin, Vol. 14, No. 6, 1942, pp. 149153 (in
Japanese).
IPPEN, A.T., and GODA, Y., "Wave Induced Oscillations _in Harbors: The
Solution for a Rectangular Harbor Connected to the OpenSea," Report
No. 59, Massachusetts Institute of Technology, Hydrodynamics Laboratory,
Cambridge, Mass., July 1963.
IPPEN, A.T., RAICHLEN, F., and SULLIVAN, R. K., Jr., "Wave Induced Oscil
lations in Harbors: Effect of Energy Dissipators in Coupled Basin
Systems," Report No. 52, Massachusetts Institute of Technology, Hydro
dynamics Laboratory, Cambridge, Mass., July 1962.
ISHIMOTO, M., and HAGIWARA, T., "The Phenomenon of Sea Water Overflowing
the Land," Bulletin of the Earthquake Research Institute, Tokyo Imperial
University, Tokyo, Japan, Suppl ementary Vol. l, 1934, pp. 1723.
207
IWASAKI, T., and HORIKAWA, K., "Tsunami Caused by Chile Earthquake in
May, 1960 and Outline of Disasters in Northeastern Coasts of Japan,"
CoastaZ Engi neering in Vol. III, 1960, pp. 3348.
IWASAKI , T., MIURA, A., and TERADA, S., "On the Effect of the Breakwater
in Case of Tsunami, Part 1 (The Model Experiment of Tsunami in Kesennuma
Bay) , " The TechnoZogy Reports of the Tohoku Sendai, Japan,
Vol. 25, No. 2, 1961, pp. 123135.
JAGGAR, T.A., "A Big Atlantic Earthquake," The VoZcano Dec. 1929.
JORSTAD, F.A., "Fjellskredet ved Tjelle; et 200ars minne," Vol.
80, No. 6, 1956, pp. 323333.
KAJIURA, K., "On the Partial Reflection of Water Waves Passing Over a
Bottom of Variable Depth," Proceedings of the lOth Pacifie Science
Congress on Tsunami International Union of Geodesy and
Geophysics, Monograph No. 24, 1963, pp. 206230.
KAMEL, A., "Stability of RubbleMound Tsunami Barrier Hilo Harbor,
Hawaii," Technical Report No. 2792, U.S. Army Engineer Waterways
Experiment Station, Vicksburg, Miss., Aug. 1967.
KANAMORI, H., "Mechanism of Tsunami Earthquakes," Physics of the Earth
and PZanetary • Vol. 6, 1972, pp. 346359.
KANAMORI, H., and CIPAR, J.J., "Focal Processes of the Great Chilean
Earthquake, May 22, 1960," Physics of the Earth and PZanetary
Vol. 9, 1974, pp. 128136.
KAPLAN, K., "Generalized Laboratory Study of Tsunami Runup," TM60, U.S.
Army, Corps of Engineers, Beach Erosion Board, Washington, D.C., Jan.
1955 .
KELLER, J.B., "The Solitary Wave and Periodic Waves in Shallow Water,"
Communications i n Pure and AppZied Vol. 1, 1948, pp. 323
339.
KEULEGAN, G.H., "Wave Motion," Engineering Ch. 11, John Wiley
and Sons, Inc., New York, 1950, pp. 711768.
KISHI, T. , and SAEKI, H. , "The Shoal ing, Breaking and Runup of the
Solitary Wave on Impermeable Rough Slopes,".Proceedings of the lOth
Conference on CoastaZ American Society of Civil Engineers,
Ch. 21, 1966, pp. 322348.
KONONKOVA, G.E., and REIHRUDEL, A.E., "Experimental Study of Solitary
Tsunami Waves," Proceedings of the Tsunami Research Bulletin
15, Royal Society of New Zealand, 1976.
LAMB . H., 6th ed., Dover Publications, Inc., New York, 1932.
208
LEBLOND, P.H., and MYSAK, L.A., "Trapped Coastal Waves and Their Role in
Shelf Dynamics," Mcœine Modeling
3
The Sea
3
Ch. 10,. Vol. 6, 1977, pp.
459495.
LEE, J .J., "Wave Induced Oscillations in Harbors of Arbitrary Shape,"
Report No. KHR20, W .N. Keck Laboratory of Hydraulics and Water
Resources, California Institute of Technology, Pasadena, Calif., Dec.
1969.
LEE, J.J., "Wave Induced Oscillations in Harbors of Arbitrary Geometry,"
Journal of Fluid Mechanics
3
Vol. 45, 1971, pp. 375393.
LEENDERTSE, J.J., "Aspects of a Computational Model for LongPeriod
WaterWave Propagation," Memorandum RM5294PR, Rand Corporation,
Santa Monica, Calif., May 1967.
LEVY, D .M., and KELLER, J. B., "Water Wave Production by Point Sources
and Extended Sources," Proceedings of the lOth Pacifie Science Congress
on Tsunami Hydrodynamics
3
International Union of Geodesy and Geophysics,
Monograph No. 24, 1961, pp. 162166.
LINSLEY, R.K., Jr., KOHLER, M.A., and PAULHUS, J.L.H., Hydrology for
Engineers
3
McGrawHill Book Company, New York, 1958.
LOZANO, . C. , and MEYER, R. E . , "Leakage and Response of Waves Trapped by
Round Islands," The Physics of Fluids
3
Vol. 19, No. 8, Aug. 1976,
pp. 10751088.
MADSEN, O.S., and MEI, c.e., "The Transformation of a Solitary Wave Over
an Uneven Bottom," Journal of Fluid Mechanics
3
Vol. 39, Part 4, 1969,
pp. 781791.
MAGOON, O.T., "The Tsunami of May 1960 as it Affected Northern California,"
Arnerican Society of Civil Engineers, presented at the Hydraulics
Division Conference, Aug. 1962 (unpublished).
MAGOON, O.T., "Structural Damage by Tsunamis," Proceedings of the Coastal
Engineering Specialty Conference
3
Arnerican Society of Civil Engineers,
Ch. 4, 1965, pp. 3568.
MALLOY, R.J., "Crustal Uplift Southwest of Montague Island, Alaska,"
Science
3
Vol. 146, 1964, pp. 10481049.
MASON, M.A., and KEULEGAN, G.H., "A Wave Method for Determining Depths
Over Bottom Discontinuities," TM5, U.S. Arrny, Corps of Engineers,
Beach Erosion Board, Washington, D.C., May 1944.
MATLOCK, H., REESE, L.C., and MATLOCK, R.B., "Analysis of Structural
Damage from the 1960 Tsunami at Hilo, Hawaii," Report DASA1268,
Structural Mechanics Research Laboratory, University of Texas, Austin,
Tex., Mar. 1962.
209
MATUO, H., "Estimation of Energy of Tsunami and Protection of Coasts,"
Bulletin of the Earthquake Research Tokyo Imperial Univer
sity, Tokyo, Japan, Mar. 1934, pp. 5564.
MEl, C. C., and LE MEHAUTE, B., "Note on the Equations of Long Waves Over
an Uneven Bottom," Journal of Geophysical Vol. 71, No. 2,
Jan. 1966, pp. 393400.
MICHE, R., ''Movement Ondulatores de Mers en Profondour Cons," Ann. Ponts
Vol. 114, 1944.
MILES, J. W., "Coupling of a Cylindrical Tube to a Half Space," Journal
of the Acoustical Society of Vol. 20, No. 5, 1948, pp. 652664.
MILES, J.W., "Resonant Response of Harbors: An EquivalentCircuit
Analysis," Journal of Fluid Mechanics, Vol. 46, Part 2, 1971, pp. 241
265.
MILES, J.W., "Wave Propagation Across the Continental Shelf," Journal of
Fluid Vol. 54, Part 1, 1972, pp. 6380.
MILLER, D.J., "Giant Waves in Lituya Bay, Alaska," Professional Paper
354c, U.S. Geological Survey, Washington, D.C., 1960.
MILLER, R.L., "Experimental Determination of Runup of Undular and Fully
Developed Bores and an Examination of Transition Modes and Internai
Structure," Technical Report No. 8, Department of Geophysical Sciences,
University of Chicago, Chicago, Ill., June 1968.
MILNE, W.E., Numerical Solution of Differential Equations, John Wiley &
Sons, Inc., New York, 1953.
MOORE, G.W., "Magnetic Disturbances Preceding the Earthquake," The Great
Alaska Earthquake of Seismology and National Academy of
Sciences, National Research Council, Washington, D.C., 1972, pp. 518525.
MUNK, W., SNODGRASS, F., and GILBERT, F., "Long Waves on the Continental
Shelf: An Experiment to Separate Trapped and Leaky Modes," Journal of
Fluid Vol. 20, Part 4, 1964, pp. 529554.
MURTY, T.S., "Seismic Sea Wavès Tsunamis," Bulletin 198, Department of
Fisheries and the Environment, Fisheries and Marine Service, Ottawa,
Canada, 1977.
MURTY, T.S., and WIGEN, S.O., "Tsunami Behavior on the Atlantic Coast of
Canada and Sorne Similarities to the Peru Coast," Proceedings of the
Tsunami Research Royal Society of New Zealand, Bulletin 15,
1976, pp. 5160.
MURTY, T.S., WIGEN, S.O., and CHAWLA, R., "Sorne Features of Tsunamis on
the Pacifie Coast of ·south and North America," Series No. 36, Marine
Sciences Directorate, Department of the Environment, Ottawa, Canada,
1975.
210
NAGAOKA, H., "On Destructive Sea Waves (Tsunami)," Proceedings of the
Tokyo MathematicoPhysical Vol. 2, No. 2, 1901, pp. 126136.
NAKANO, M., "Preliminary Note on the Accumulation and Dissipation of
Energy of the Secondary Undulations in a Bay," Proceedings of the
PhysicoMathematical Society of Vol. 14, 1932, pp. 4456.
NASU, N., "Heights of Tsunamis and Damage to Structures," Bulletin of
the Earthquake Research Tokyo Imperial University, Tokyo,
Japan, Vol. 1, Mar. 1934, pp. 218226.
NASU, N., "Local Phenomena of Tsunami, Part 2," Bulletin of the Earthquake
Research Tokyo Imperial University, Tokyo, Japan, Vol. 26,
Nos. 14, 1948, pp. 2735.
NISHIMURA, H., HORIKAWA, K., and SHUTO, N., "On the Function of Tsunami
Breakwaters," Coastal Engineering in Tokyo, Japan, Vol. 14, 1971,
pp. 6372.
O'BRIEN, M.P., Discussion of "Similitude in Coastal Engineering" by
B. Le Mehaute, Journal of the and Ocean
Vol. 103, No. WW3, Aug. 1977, pp. 393400.
PALMER, R.Q., MULVIHILL, M.E., and FUNASAKI, G. T., "Hilo Harbor Tsunami
Madel  Reflected Waves Superimposed," Proceedings of the Coastal
Engineering Santa Barbara Specialty American Society of
Civil Engineers, Ch. 2, 1965, pp. 2131.
PARARASCARAYANNIS, G., "Catalog of Tsunamis in the Hawaiian Islands,"
Report WDCAT 692, ESSA Coast and Geodetic Survey, Boulder, Cola.,
May 1969.
PEREGRINE, D.H., "Long Waves on a Beach," Journal of Fluid
Vol. 27, No. 4, 1967, pp. 815827.
PEREGRINE, D.H., "Long Waves in Two and Three Dimensions," Proceedings of
the Symposium on Long University of Delaware, 1970, pp. 6390.
PERROUD, P.H., "The Solitary Wave Reflection Along a Straight Vertical
Wall at Oblique Incidence," University of California, Institute of
Engineering Research, Berkeley, Calif., Series 99, Issue 3, Sept. 1957.
PETRAUSKAS, C. , and BORGMAN, L. E. , "Frequencies of Crest Heights for
Random Combinations of Astronomical Tides and Tsunamis Recorded at
Crescent City, California," Technical Report HEL 168, Hydraulic
Engineering Laboratory, University of California, Berkeley, Calif.,
Mar. 1971.
PIERSON, W.J., "Wave Behavior Near Caustics in Models and in Nature,"
Waves on Academie Press, New York, 1972, pp. 163180.
211
PLAFKER, G. , "Tectonic Deformation Associated with the 1964 Alaska
Earthquake," Vol. 148, June 1965, pp. 16751687.
PUGET SOUND Seattle, Wash., 31 Dec. 1866 (newspaper) .
RAYLEIGH, LORD, The Theory of Vol. II, Ch. XVI, Reprint of 1896
ed., Dover Publications, Inc., New York, 1945.
REID, H.F., "The Lis bon Earthquake of November 1, 1755," BuUetin of the
Seismological Society of Vol. 4, No. 2, June 1914, pp. 5380.
REID, H.F., and TABER, S., "The Puerto Rico Earthquake of 1918, Report
of the Earthquake Investigation Commission," H.D. 269, 66th Cong.,
lst Sess., U.S. Government Printing Office, Washington, D.C., 1919.
RIABOUCHINSKI, D., "Sur la Resistance des Fluides," International Congress
of Mathematics, Strasbourg, France, 1920, pp. 568585.
RICHTER, C.F., Elementary W.H. Freeman & Sons, San Francisco,
Cali f ., 1958.
ROW, R. V. , "Atmospheric Waves," The Great Alaska Earthquake of
Seismology and National Academy of Sciences, Washington, D.C.,
pp. 494507.
SCHANK, R.E., "Hazard Reduction and the Mitigation of Tsunami Effects
Through Effective Public Warning in Hawaii," Proceedings of the
Symposium on Department of Fisheries and the Environment,
Series No. 48, 1978, pp. 252254.
SEELIG, W.N., HARRIS, D.L., and HERCHENRODER, B.E., "A Spatially Inte
grated Numerical Madel of Inlet Hydraulics," GITI Report 14, U.S. Army
Coastal Engineering Research Center, Fort Belvoir, Va., and U.S. Army
Engineer Waterways Experiment Station, Vicksburg, Miss., Nov. 1977.
SHAW, R.P., "The Outer Boundary Integral Equation Method Applied to
Wave Scattering in an Infinite, Locally Inhomogeneous Medium," Report
No. HIG746 (NOAAJTRE126), Hawaii Institute of Geophysics and Joint
Tsunami Research Effort, University of Hawaii, Honolulu, Hawaii, Sept.
1974.
SHAW, R.P . , "An Outer Boundary Integral Equation Applied to Transient
Wave Scattering in an Inhomogeneous Medium," Journal of Applied
Vol. 42, No. 1, Mar. 1975, pp. 147152.
SHEN, M.C . , "Wave Resonance Near Shores," Waves on Academie
Press, New York, 1972, pp. 123161.
SHEN, M. C. , and MEYER, R. E. , "Surface Wave Resonance on Continental and
I s land Slopes," Technical Summary Report No. 781, Army Mathematics
Research Center, University of Wisconsin, Madison, Wise., Sept. 1967.
212
SHEPARD, F. P. , MacDONALD, G. A. , and COX, D. C. , "The Tsunami of Apri 1 1,
1946," Bulletin of the Scripps Institution of Oceanography, Vol. 5,
No. 6, 1950, pp. 391528.
SHINMOTO, D.Y., and VITOUSEK, M.J., "An Inexpensive AirDeployable
Midocean Tsunami Gauge," Proceedings of the Symposium on Tsunamis,
Department of Fisheries and the Environment, Series No. 48, 1978,
pp. 242244.
SMITH, R.A., "Annotated Bibliography on Water Waves Caused by Exp1osions
1946 to 1966," DASIAC Special Report 58, DASA Information and Analysis
Center, Apr. 1967.
SOLOVIEV, S.L., "Recurrence of Tsunamis in the Pacifie," Tsunamis in the
Pacifie Ocean, Ch. 11, EastWest Center Press, Honolulu, Hawaii, 1970,
pp. 149164.
SOLOVIEV, S. L. , and FER CHEV, M. D. , "Summary of Da ta on Tsunamis in the
USSR," Bulletin of the Council for Seismology, Academy of Sciences of
the U.S.S.R., No. 9, 196i, pp. 2355 (translated by W.G. Van Campen,
Hawaii Institute of Geophysics).
SOLOVIEV, S.L., et al., "Materials for Preliminary Tsunami Zonation of
KurileKamchatka Coast Based on Hydrodynamica1 Calculations," Proceed
ings of the Tsunami Research Symposium, Royal Society of New Zealand,
Bulletin 15, 1976, pp. 2938.
SPAETH, M.G., Annotated Bibliography on Tsunamis, Monograph No. 27,
International Union of Geodesy and Geophysics, Paris, France, July 1964.
SPAETH, M.G., and BERKMAN, s.e., "The Tsunamis as Recorded at Tide
Stations and the Seismic Sea Wave Warning System," The Great Alaska
Earthquake of 1964, Oceanography and Coastal Engineering, National
Academy of Sciences, Washington, D.C., 1972, pp. 38110.
SPIELVOGEL, L.Q., "SingleWave Runup on Sloping Beaches," Journal of
Fluid Mechanics, Vol. 74, Part 4, 1975, pp. 685694.
STEIN, S., et al., "Earthquakes Along the Passive Margin of Canada
Induced by Déglaciation," Geophysical Research Letters (in preparation,
1980).
STOKES, G.G., "On the Theory of Oscillatory Waves," Transactions of the
Cambridge Philosophical Society, Vol. 8, 1847, pp. 441455.
STREET, R.L., BURGESS, S.J., and WHITFORD, P.W., "The Behavior of Solitary
Waves on a Stepped Slope," Technical Report No. 93, Department of Civil
Engineering, Stanford University, Palo Alto, Ca1if., Aug. 1968.
213
STREET, R.L., and CAMFIELD, F.E., "Observations and Experiments on
Solitary Wave Deformation," Proceedings of the lOth Conference on
Coastal American Society of Civil Engineers, Ch. 19,
1966, pp. 284301.
STREET, R.L., CHAN, R.K.C., and· FROMM, J.E., "The Numerical Simulation
of Long Water Waves: Progress on Two Fronts," Tsunamis in the Pacifie
Ch. 30, W.M. Adams, ed., EastWest Center Press, University of
Hawaii, Honolulu, Hawaii, 1970, pp. 453473.
STRIEM, H. L., and MILOH, T., "Tsunamis Induced by Submarine Slumpings
off the Coast of Israel," Report IALD1102, Israel Atomic Energy
Commission, Israel, July 1975.
SUYEHIRO, Y., "Sorne Observations on the Unusual Behaviour of Fishes Prior
to an Earthquake," Bulletin of the Earthquake Research Tokyo
Imperial University, Tokyo, Japan, Vol. 1, 1934.
TABER, S., "Seismic Activity in the Atlantic Coastal Plain Near Charleston,
South Carolina," Bulletin of the Seismological Society of Vol.
4, No. 3, Sept. 1914, pp. 108160.
TOWNLEY, S.D., and ALLEN, M.W., "Descriptive Catalog of Earthquakes of
the Pacifie Coast of the United States 1769 to 1928," Bulletin of the
Seismological Society of Vol. 29, No. 1, Jan. 1939, pp. 1297.
U.S. ARMY ENGINEER DISTRICT, HONOLULU, "Hilo Harbor, Hawaii, Report on
Survey for Tidal Wave Protection and Navigation," Nov. 1960.
U.S. ARMY ENGINEER DISTRICT, HONOLULU, "The Tsunami of 23 May 1960 in
Hawaii  Final Post Flood Report," 1960.
VAN DORN, W.G., "Source Mechanism of the Tsunami of March 28, 1964 in
Alaska," Proceedings of the Ninth Conference on Coastal
American Society of Civil Engineers, Ch. 10, 1964, pp. 166190.
VAN DORN, W.G., "Tsunamis," Advances in Vol. 2, Academie
Press, New York, 1965 (also published as Scripps Institution of
Oceanography Report, Federal Clearinghouse No. AD457729, Jan. 1965).
VAN DORN, W.G., LE MEHAUTE, B., and HWANG, LS., "Handbook of Explosion
Generated Water Waves, Volume I  StateoftheArt," Report No. TC130,
Tetra Tech, Inc., Pasadena, Calif., Oct. 1968.
VITOUSEK, M.J., "Proposed MidOcean Tsunami Gage and Oceanography
Instrument System," Proceedings of the Tsunami Meetings Associated
With the lOth Pacifie Science International Union of
Geodesy and Geophysics, 1961, pp. 131133.
VITOUSEK, M.J., and MILLER, G., "An Instrumentation System for Measuring
Tsunamis in the the Deep Ocean," Tsunamis in the Pacifie Ch. 16,
EastWest Center Press, Honolulu, Hawaii, 1970, pp. 239252.
214
WEIGEL , E.P. , Tsunami! National Oceanic and Atmospheric Administration ,
Vol. 4, No . 1, Jan. 1974.
WELLER, J.M., "Human Response to Tsunami Warnings," The Great Alaska
Earthquake of 1964
3
Oaeanography and Coastal Engineering
3
National
Academy of Sciences, National Research Council, Washington, O.C.,
1972, pp. 222228.
WHITMAN, G.B., "NonLinear Dispersion of Water Waves," Journal of Fluid
Meahanias
3
Vol. 27, Part 2, 1967 , pp. 339412.
WIEGEL, R.L., Oaeanographiaal Engineering
3
PrenticeHall, Englewood
Cliffs, N.J., 1964.
WIEGEL, R.L., "Protection of Crescent City, California from Tsunami
Waves," Report for the Redevelopment Agency of the City of Crescent
City, Crescent City, Calif., Mar. 1965.
WIEGEL, R.L., "Tsunamis," Earthquake Engineering
3
Ch. 11, R.L. Wiegel,
ed., PrenticeHall, Englewood Cliffs, N.J., 1970, pp. 253306.
WILSON, B.W., "Long Waves Generated by Nuclear Explosions," Appendix I,
Investigation of Long Waves and Their Effects on the Coastal and
Harbor Environment of the Lower Chesapeake Bay, Vol. II, National
Engineering Science Co., Report No. DASA1355, Apr. 1963.
WILSON, B.W., "Earthquake Occurrence and Effects in Ocean Areas,"
Technical Report CR. 69.027, U.S. Naval Civil Engineering Laboratory,
Port Hueneme, Calif., Feb. 1969.
WILSON, B.W., "Estimate of Tsunami Effect at San Onofre Nuclear Generating
Station Units 2 and 3, California," Southern California Edison Co.,
Los Angeles, Calif., Dec. 1972.
WILSON, B.W. , and T ~ R U M , A. , "The Tsunami of the Al as kan Earthquake, 1964;
Engineering Evaluation," TM25, U.S. Army Coastal Engineering Research
Center, Washington, D.C., May 1968.
WILSON, B.W., WEBB, L.M., and HENDRICKSON, J .A., "The Nature of Tsunamis;
Their Generation and Dispersion in Water of Finite Depth," Technical
Report No. SN 572, National Engineering Science Co., Aug. 1962.
ZABUSKY, N.J., and GALVIN, C.J., "ShallowWater Waves, the Korteweg
deVries Equation and Solitons," Journal of Fluid Meahanias
3
Vol. 47,
Part 4, 1971, pp. 811824.
215
APPENDIX
TSUNAMIS OCCURRING BETWEEN 1891 AND 1961
Date Source
30 July 1891 Lerdo, ~ l e x i c o
29 Nov. 1891 Seattle, Washington
16 May 1892 Marian as Islands
1893 Gree ce
1894
22 Mar. 1894 Nemuro, Japan
Apr. 1894 Lokris, Gree ce
10 June 1894 • Turkey
9 Jan. 1896 Kashima Sea
15 June 1896 Sanriku, Japan
5 Aug. 1897 Tohoku District, Japan
21 Sopt. 1897 Sulu Sea
29 Nov. 1897 West Indies
1898 Gree ce
22 Jan. 1899 Southwest Peloponessus,
Greece
Sept. 1899 Yakutat Bay, Alaska
30 Sept. 1899 Banda Sea
7 Oct. 1899 Kyushu
9 Aug. 1901 Rikuchu, Japan
26 Feb. 1902 El Salvador
5 Ju1y 1902 Thessaloniki, Gree ce
25 June 1904 Kamchatka
1905
8 Sept. 1905 Calabria, Ital y
31 Jan. 1906 Ecuador  Columbia
Remarks
Tidal wave of considerable height at head of the Gulf
of California.
Water in Lake Washington surged on to the beach 2 feet
above the mark of the highest water and 8 feet above
the lake stage on that date.
Min or tsunami at Guam.
Wave 4 feet high at Bizerte, Tunisia; sorne damage.
Tsunamis associated with these earthquakes were
generally small.
Large tsunamis induced by earthquakes at Constantinople
(Istanbul).
Slight tsunami in Japan.
Much damage and loss of li fe, devastated ports along
the northeast coast of Japan. Runup 80 feet high at
Shirahama. Variation in terres trial magnetism
observed at Sendai preceding the tsunami. Maximum
height of 30 feet at Napoopoo, Hawaii.
Tsunami.
Severe damage at Tacloban, Philippines (western shores
of Basilan Island).
Large tsunami at Montserrat.
Tsunami reported at Marathos.
Waves generated did little damage.
Large tsunami on south coast of Ceram, Indonesia.
Tsunami in Tagonoura, Japan.
Tsunami at Hilo, Kai1ua, and Keauhou, Hawaii; minor
damage.
Tsunami in El Sai v ad or and Guatemala.
Saloniki Harbor flooded by waves.
Wave at Avachinskaya Bay.
Tsunami at Bizerte, Tunisia; believed min or.
Possible tsunami.
l'laves observed in Hawaii; maximum height 3. 6 meters
observed at Hilo; no damage.
217
TSUNAMIS OCCURRING BETWEEN 1891 AND 1961Continued
Date Source Remarks
19 Aug . 1906 Valparaiso, Ch ile Recorded in Hawaii and Japan; negligible along Chilean
coast; some damage in Hawaii.
15 Sept. 1906
14 Jan. 1907
14 Apr. 1907
23 Oct. 1907
20 Sept. 1908
28 Dec. 1908
30 Jan. 1911
26 Feb. 1913
14 Mar. 1913
11 Oct. 1913
12 Jan. 1914
26 May 1914
7 Aug. 1915
1 Jan. 1916
1 May. 1917
26 June 1917
15 Aug. 1918
Sept. 1918
11 Oct. 1918
Nov. 1918
4 Dec. 1918
30 Apr. 1919
S May 1919
20 Sept. 1920
Dampier Strait
Jamaica
Mexico
Calabria, Ital y
Puna, Hawaii
Messina, Italy
Philippine Islands
South Island, New
Zealand
Sangi, East Indies
Near east end of New
Guinca
Sakuraj ima, Japan
North coast of New
Guinea
Ionian Sea
New Britain
Kermadec Islands
Tonga Islands
Southern Mindanao,
Philippine Islands
Kuril Islands
Puerto Rico
Copiapo, Ch ile
North of Vava 'u, Tonga
New Britain
New Hebrides Islands
Tsunami in New Guinea.
Tsunami generated, main damage at Kingston.
Thirtyfoot wave at Acapulco, Mexico.
Tsunami recorded at Messina and Ca tania.
Weak tsunami.
Heavy damage along shoreline; breakwater destroyed at
Messina.
Seismic waves in Lake Bombon washed away severa!
villages; sorne loss of life; waves 2.5 ta 3 meters
high on shoreline.
Small tsunami.
Tsunami generated.
Weak tsunami.
Wave with 10foot amplitude caused serious damage ta
small boats in Kagoshima Harbor.
Probable tsunami.
Wav'e height about 5 feet on Greek coast.
Water level in Rabaul Harbor fell 15 feet and rose
ag ain rapidly; causeway washed out.
Weak tsunami in Hawaii.
Fortyfoot wave in Samoa; wave recorded at Honolulu.
Wave swept coast from Lebak ta Glan; height estimated
at 24 feet at sorne points.
Tsunami at Uruppu Island, 24 men killed on
Sirnusirijimi; minor damage at Hilo, Hawaii.
Tsunami caused fata1i t ies and damage at Point
Borinqucn and Aguadilla; also damage at Mayaguez.
Tsunami observed at Futami Harbor, Bonin Islands.
Tsunami he igh t 5 met ers at Port of Cal der a where i t
caused damage.
Observed in Hawaii.
Wave similar ta 1 January 1916.
Slight tsunami at Samoa.
218
TSUNAMIS OCCURRING BETWEEN 1891 AND 1961Continued
Date Source
18 Dec. 1920 Strait of Otranto
11 Nov. 1922 At a cuma, Ch ile
3 Feb. 1923 East Kamchatka
13 Apr. 1923 Kamchatka
1 Sept. 1923 Kwanto, Japan
9 Jan. 1924
14 Apr. 1924 South China Sea
16 Mar. 1926 Tonga Island
16 Sept. 1926 Solomon Islands
18 Nov. 1926 Saint Pierre and
Miquelon
7 Mar. 1927 Tango, Japan
26 June 1927 Black Sea
11 and 12 Black Sea
Sept. 1927
Nov. 1927 California
28 Dec. 1927 Kamcha t ka, U.S.S.R.
25 Apr. 1928 Ne ar Piraeus, Gree ce
H June 1928 Mexico
17 Jan. 1929 Cumana, Venezuela
6 Mar. 1929 Aleutiari Islands
26 May 1929 Queen Charlotte
Islands, Canada
18 Nov. 1929 Grand Banks
3 Oct. 1931 Solomon Islands
3 June 1932 Jalisco, Mexico
18 June 1932 J ali seo, Me xi co
Remarks
Wave height 10 feet on Al banian coast.
Tsunami destructive in many places along the coast of
Chile; minor damage in Hawaii.
Tsunami in Hawaii; maximum amplitude 15 feët at Hilo; one
pers on killed, mu ch damage; tsunami on coast of Kamchatka.
Tsunami observed in Hawaii.
Tsunami hi t towns on shore of Sagami Bay; maximum height
of 26 feet at Atami.
Tsunami on French coast (Atlantic coast); possibly a
storm surge.
Small tsunami at Agno, Pangasman, caused min or damage ..
Wave swept Palmerston Island, 300 miles northwest of
Raratonga; one person killed; all buildings swept away
except the chur ch.
Tsunami at Guadalcanal.
Possible minor tsunami.
Height of tsunami about 5 feet on Sea of Japan coastline.
Small tsunami recorded in Crime a.
Small tsunami recorded in Crime a.
Sixfoot wave at Surf; 0.24 inch high at L a ~ T o l l a .
Minor tsunami recorded at Hilo.
Waves 7 fe et high on north coast of Crete.
Waterfront damaged at Puerto Angel, Mexico.
Many boats wrecked.
Tsunami measured in Hawaii (maximum ampli tude 1 foot at
Hilo).
Fourfoot wave at Que en Charlotte City.
Tsunami hit Newfoundland; damage and loss of life on
Burin Pi)ninsula.
Eighteen native villages destroyed on San Christobal
Island; approximately 50 persons killed.
Railroad track swept away between Cayutlan and
Manzanillo, Mexico.
Small tsunami .
219
Date
22 June 1932
26 Sept. 1932
3 J.lar. 1933
14 Feb. 1934
27 Oct. 1936
3 Nov. 1936
28 and 29
May 1937
6 Mar. 1938
22 Mar. 1938
19 May 1938
23 t.lay 1938
5 Nov. 1938
10 Nov. 1938
1 May 1939
2 Aug. 1940
5 Dec. 1941
24 Aug. 1942
6 Apr. 1943
7 Dec. 1944
27 Nov. 1945
1 Apr. 1946
23 June 1946
Aug. 1946
12 and 13
Nov. 1946
TSUNAMIS OCCURRING BETWEEN 1819 AND 1961Continued
Source Remarks
Jalisco, Mexico Small tsunami, sorne damage.
Hierissos, Gree ce Small tsunami noted in Gulf of Orphano.
Sanriku coast, Japan Immense damage; 3,022 people killed; 8,831 houses
destroyed; 8,180 vessels wrecked.
China Se a Tsunami at San Esteban, Philippine Islands.
Li tuya Bay, Alaska Maximum runup of 400 feet.
Off Kinkazan, Japan Mi nor tsunami on Japan coast.
Blanche Bay,
New Britain
Volcanic eruption. Highest waves 14 to 16 feet from
crest to trough.
Solomon Islands Minor tsunami.
Queen Charlotte Mi nor tsunami.
Islands, Canada
Macassar Strai t Mi nor tsunami.
Ibaraki, Japan Small tsunami.
Off Iwaki, Japan Tsunami observed in nort hern Japan.
Off south coast of Mi nor tsunami.
Alaska
Near Ogasima, Japan Tsunami observed along coast of Sea of Japan; sorne damage.
Northern part of Tsunami on coast of Hokkaido.
Sea of Japan
Panama  Costa Rica Slight tsunami at Punta Arenas, Costa Rica; height of
O. 75 foot.
Near Lima, Peru
Chile
Japan (Kumanonoda)
Arabian Sea
Aleutian Islands,
Alaska.
British Columbia,
Canada
Dominican Republic
~ 1 a t u a Island,
Kurile Islands
Slight tsunami recorded in Peru.
Small tsunami at Valparaiso; ampli tu des more th an 3 fe et.
Ampli tude less than 2 feet in Aleutians; waves along
Pacifie coast of Japan 2. 5 to 5 meters high ne ar the
Kii Peninsula.
Heavy damage on coast of Pakistan and India; tsunami at
Karachi and Bombay where there was damage and loss of li fe;
two new islands appeared in the Arabi an Se a.
Damage and loss of life in Alaska; heavy damage and many
li v es !ost in Hawaï i.
Bot tom of Deep Bay sank from 9 to 84 fe et; waves
generated in Georgia Strai t flooded fields and highways.
Town of Matanzas badly damaged and abandoned; more than
100 persons ki lled; mi nor damage on coast of Hai ti.
Volcanic explosion. Tsunami generated.
220
TSUNAMIS OCCURRING BETWEEN 1891 AND 1961Continued
Date
21 Dec. 1946
6 Oct. 1947
Feb. 1948
Sept. 1948
22 Aug. 1949
20 Oct. 1949
29 Dec. 1949
5 Oct. 1950
23 Oct. 1950
14 Dec. 1950
21 Aug. 1951
4 Mar. 1952
17 Mar. 1952
19 1952
13 July 1952
Sept. 1952
4 Nov. 1952
31 May 1953
11 Aug. 1953
10 Sept. 1953
14 Sept. 1953
26 Nov. 1953
12 Dec. 1953
18 Jan. 1955
19 Apr. 1955
5 Mar. 1956
Source
Hons hu, Japan
(Nanka i Earthquake)
Kononi,
Gree ce
Sea of Crete
Tonga Island
Queen Charlotte
Islands, Canada
Solomon Sea
Philippine Islands
Costa Rica
Guatemala
Guerrero, Mexico
Kona, Hawaii
Tokachi Oki
Earthquake, Japan
South of Hawaii
East of Mindanao,
Philippines
Near New llebrides
in Reef
Paramushir Island,
Kurile Islands
Near Dominican
Republic
lonian Islands
Cyprus
Suva Earthquake
BosoOki, Japan
Peruvian E;;rthqu3.ke
Central Chi le
North coast of
Hokkaido, Japan
Remarks
Damage on south coast of Japan (1\'akayama Prefect ure) ;
1,500 people killed on Shikoku Island.
Large wave on coast of Methoni, believed caused by a
submarine lands 1 ide.
Eightfoot wave.
Mi nor tsunami in Hawaii.
Twofo.ot wave at Ketchikan, Alaska.
Minor (very slight) tsunami in Hawaii.
Tsunami kil led one pers on near Mercedes.
Very slight tsunami in Hawaii.
Very slight tsunami in Hawaii.
Very slight tsunami in Hawaii.
Cliff collapsed near Napoopoo creating a 12foot wave
which destroyed a boat dock.
Tsunami on Pacifie coast of Hokkaido from Nemuro to
Hidaka (main damage in Ki ri ttapu and Tokotan); mass
of ice accompanied the tsunami.
Slight tsunami at Kalapana.
Small tsunami.
Small tsunami.
Submarine volcano eruption at Myojin Reef (400 kilo
meters south of Tokyo, Japan). Tsunami generat ed;
largest wave 24 September, presumed to have destroyed
hydrographie research vessel.
East Kamchatka Earthquake. Large tsunami; maximum
height of 12 feet at Hilo, Hawaii.
Very slight tsunami; ampli tude of 0. 2 foot at Puerto
Plata.
Tsunami.
Series of Waves at Paphus, Cyprus (no major damage ).
Tsunami in Fiji Islands; amplitude of 0.7 foot.
Tsunami had maximum height of 3 met ers at Choshi.
Wave amplitude of 3.2 feet at Talara, Peru.
Tsunami caused damage at LaVela, Venezuela.
Slight Tsunami in Hawaii.
221
Date
30 1956
9 Ju1y 1956
2 Nov. 1956
9 .Mar. 1957
28 July 1957
19 Jan. 1958
9 July 1958
6 Nov. 1958
22 Jan. 1959
7 Feb. 1959
4 May 1959
18 Aug. 1959
13 Jan. 1960
29 Feb. 1960
20 Mar. 1960
22 May 1960
20 Nov. 1960
16 Jan. 1961
27 Feb. 1961
TSUNAMIS OCCURRING BETWEEN 1891 AND 1961Continued
Source
Ne ar Kamchatka
Greek Archipelago
Volos, Greece
Aleu tian Islands
Me.xico
Ecuador
Lituya Bay, Alaska
Iturup Earthquake,
Pacifie coast,
U.S.S.R.
Remarks
Wave believed to be caused by submarine
landslides; wave 100 feet high near source.
Fourfoot wave.
Tsunami waves in Hawaii; heavy damage.
Tsunami recorded at Acapulco and Salina Cruz.
Tsunami damaged Esmeraldas and Quayaquil; sorne
deaths.
Giant wave from rockfall; wave runup estimated
at 1,740 feet at· one point.
Waves 3 to 5 meters high in South Kurile Islands
(referred to as EtorofuOki Earthquake by
Japanese).
East coast of Honshu Very faint at Miyako.
Northern Peru Tsunami recorded at Talara.
Near Kamchatka Slight tsunami in Hawaii and the Aleutian Islands.
Hebgen Lake, Montana Waves in lake from earthquake.
Southern Peru Sma11 tsunami struck Ancon.
Morocco Tsunami at Agadir, Morocco.
Coast of Japan Fivefoot wave locally.
Chile Ofunato, Shizukawa, and Kiritappu heavily damaged
in Japan; damage at Hilo, Hawaii.
Coast of Peru Damage and deaths along Peruvian coast.
Ibaraki Oki Small tsunami.
Earthquake
HiugaNada, Japan Tsunami on coast of Miyazaki prefecture occurred
at low tide; height small.
tr U. S. GOVERNMENT PRINTING OFF I CE : 1980 622  607
222
Camfield, Frederick E.
Tsunami engineering 1 by Frederick E. Camfield.  Fort Belvoir, Va.
: U. S. Coast al Engineering Research Center, 1980.
222 p. : ill ; 27 cm.  (Special report  U.S. Coastal Engineering
Research Center ; no. 6)
Cover title.
Bibliography : p. 201.
This report provides a source of stateofthe·art information on
tsunami engineering. The report summarizes available information,
identifies gaps in existing knowledge, and discusses methods of pre
dicting tsunami flooding. The generating mechanisms of tsunamis and
the method of determining the probability of occurrence are given.
1. Coastal engineering. 2. Offshore structures. 3. Edge waves.
4. Math.ematical models. Tsunamis. S. Water waves. I. Title. II.
Series: U. S. Coast al Engineering Research Center. Special report no.
6.
TC203 .US81sr no. 6 627
Camfield, Frederick E.
Tsunami engineering 1 by Frederick E. Camfield. Fort Belvoir, Va.
: U. S. Coast al Engineering Research Center, 1980.
222 p. : Hi ; 27 cm.  (Special report  U.S. Coastal Engineering
Research Center ; no. 6)
Cover title.
Bibliography : p. 201.
This report provides a source of stateof theart information on
tsunami engineering. T.he report summarizes available information,
identifies gaps in existing knowledge, and discusses methods of pre
dicting tsunami flooding. The generating mechanisms of tsunamis and
the method of determining the probability of occurrence are given.
1. Coastal engineering. 2. Offshore structures. 3. Edge waves.
4. Mathematical models. Tsunamis. S. Water waves. I. Ti tle. II.
Series: U.S. Coastal Engineering Research Center. Special report no.
6.
TC203 .US81sr no. 6 (27
Camfield, Frederick E.
Tsunami engineering 1 by Frederick E. Camfield.  Fort Belvoir, Va.
: U. S. Coastal Engineering Research Center, 1980.
222 p. : ill ; 27 cm.  (Special report  U.S. Coastal Engineering
Research Center ; no. 6)
Cover title.
Bibliography : p. 201 •
This report provides a source of stateoftheart information on
tsunami engineering. The report summarizes available information,
identifies gaps in existing knowledge, and discusses methods of pre
dicting tsunami flooding. The generating mechanisms of tsunamis and
the method of determining the probability of occurrence are given.
1. Coastal engineering. 2. Offshore structures. 3. Edge waves.
4. Mathematical models. Tsunamis. S. Water waves. I. Title. II.
Series: U. S. Coast al Engineering Research Center. Special report no .
6.
TC203 .U581sr no. 6 627
Camfield, Frederick E.
Tsunami engineering 1 by Frederick E. Camfield. Fort Belvoir, Va .
: U. S. Coast al Engineering Research Center, 1980.
222 p. : ill ; 27 cm.  (Special report  U. S. Coast al Engineering
Research Center ; no. 6)
Cover title.
Bibliography : p. 201.
This report provides a source of state·oftheart information on
tsunami engineering. The report summarizes available information,
identifies gaps in existing knowledge, and discusses methods of pre
dicting tsunami flooding. The generating mechanisms of tsunamis and
the method of determining the prob;;bility of occurrence are given.
1. Coast al engineering. 2. Offshore structures. 3. Edge waves.
4. Mathematical models. Tsunamis. S. Water waves. I. Title. II.
Series: U.S. Coastal Engineering Research Center. Special report no.
6.
TC203 .US81sr no. 6 627
Camfield, Frederick E.
Tsunami engineering 1 by Frederick E. Camfield.  Fort Belvoir, Va.
: U.S. Coastal Engineering Research Center, 1980.
222 p. : ill ; 27 cm.  (Special report  U. S. Coast al Engineering
Research Center ; no. 6)
Cover title.
Bibliography : p. 201.
This report provides a source of stateoftheart information on
tsunami engineering. The report summarizes available information,
identifies gaps in existing knowledge, and discusses methods of pre
dicting tsunami flooding. The generating mechanisms of tsunamis and
the method of determining the probability of occurrence are given.
1. Coastal engineering. 2. Offshore structures. 3. Edge waves.
4. Mathematical models. Tsunamis. 5. Water waves. I. Title. II.
Series: U. S. Coast al Engineering Research Center. Special report no.
6.
TC203 .U581sr no. 6 627
Camfield, Frederick E.
Tsunami· engineering 1 by Frederick E. Camfield.  Fort Belvoir, Va.
: U.S. Coastal Engineering Research Center, 1980.
222 p. : ill ; 27 cm.  (Special report  U. S. Coastal Engineering
Research Center ; no. 6)
Cover title.
Bibliography : p. 201.
This report provides a source of stateoftheart information on
tsunami engineering. The report summarizes available information,
identifies gaps in existing knowledge, and discusses methods of pre
dicting tsunami flooding. The generating mechanisms of tsunamis and
the method of determining the probability of occurrence are given.
1. Coast al engineering. 2. Offshore structures. 3. Edge waves.
4. Mathematical models. Tsunamis. 5. Water waves. I. Title. II.
Series: U. S. Coastal Engineering Research Center. Special report no.
6.
TC203 .U581sr no. 6 (27
Camfield, Frederick E.
Tsunami engineering 1 by Frederick E. Camfield.  Fort Belvoir, Va.
: U. S. Coastal Engineering Research Center, 1980.
222 p. : ill ; 27 cm.  (Special report  U. S. Coast al Engineering
Research Center ; no. 6)
Cover title.
Bibliography : p. 201 •
This report provides a source of stateoftheart information on
tsunami engineering. The report summarizes available information,
identifies gaps in existing knowledge, and discusses methods of pre
dicting tsunami flooding. The generating mechanisms of tsunamis and
the method of determining the probability of occurrence are given.
1. Coast al engineering. 2. Offshore structures. 3. Edge waves.
4. Mathematical models. Tsunamis. 5. Water waves. I. Title. II.
Series: U. S. Coast al Engineering Research Center. Special report no.
6.
TC203 .tr581sr no. 6 627
Camf ield, Frederick E.
Tsunami engineering 1 by Frederick E. Camfield.  Fort Belvoir, Va.
: U. S. Coast al Engineering Research Center, 1980.
222 p. : ill ; 27 cm.  (Special report  U. S. Coastal Engineering
Research Center ; no. 6)
Cover title.
Bibliography : p. 201 .
This report provides a source of stateoftheart information on
tsunami engineering. The report summarizes available information,
identifies gaps in existing knowledge, and discusses methods of pre
dicting tsunami flooding. The generating mechanisms of tsunamis and
the method of determining the probability of occurrence are given.
1. Coastal engineering. 2. Offshore structures. 3. Edge waves.
4. Mathematical models. Tsunamis. 5. Water waves. I. Title. II.
Series: U. S. Coast al Engineering Research Center. Special report no.
6.
TC203 .U581sr no. 6 627
.l
Reprint or republication of any of this material shall give appropriate credit to the U.S. Army Coastal Engineering Research Center.
U.S. Army Coastal Engineering Research Center Kingman Building Fort Belvoir, Virginia 22060
For sale l>y the Superlntendent o! Documents, U.S. Oovcmmcnt Prlntlng omcc, Washington, D.C. 20402
UNCLASSIFIED
SECURITY CLASSIFICATION OF THIS PAGE (When Date Entered)
REPORT DOCUMENTATION PAGE
1. REPORT NUMBER
READ INSTRUCTIONS BEFORE COMPLETING FORM ,2. GOVT ACCESSION NO. 3. RECIPIENT'S CATALOG NUMBER
SR6
4. TITLE (and Subtil/a) S. TYPE OF REPORT
a
PERIOO COVERED
TSUNAMI ENGINEERING
7. AUTHOR(e)
Special Report
6 . PERFORMING ORG. REPORT NUMe!OK
8.
CONTRACT OR GRANT NUMBER(e)
Frederick E. Camfield
9. PERFORMING ORGANIZATION NAME AND ADDRESS 10. PROGRAM ELEMENT, PROJECT, TASK
AREA
Department of the Army Coastal Engineering Research Center (CERENCD) Kingman Building, Fort Belvoir, Virginia 22060
Il. CONTROLLING OFFICE NAME AND ACCRESS
a
WORK UNIT NUMBERS
F31234
12. REPORT DATE
Department of the Army Coastal Engineering Research Center Kingman Building, Fort Belvoir, Virginia
14. MONITORING AGENCY NAME
February 1980
13. NUMBER OF PAGES
22060
IS.
222
SECURITY CLASS. (olthle report)
a
ADDRESS(If dlllerent lrorn ControlllnS Otllce)
UNCLASSIFIED
ISe. DECL AS SI FIC A TION/ DO WN GRAOI NG SCHEDULE
16. DISTRIBUTION STATEMENT (ol lhle Report)
Approved for public release; distribution unlimited.
17. DISTRIBUTION STATEMENT (ol the abetract entered ln Block 20, Il dlllerenl from R e port)
18. SUPPL.EMENTARY NOTES
19. KEY WORCS (Continue on reveree sida Il neceuary 11nd ldenllfy by block numbsr)
Coastal engineering Coastal structures Edge waves Flood frequencies
zo.
ASS~ACT
Mathematical models Seismic safety Tsunamis Water waves
(QmtiDue .,.. ,..,.r_ eldiJ If,..,.,_..,. IIDd. ldenLify by block nurnber)
This report provides a source of stateoftheart information on tsunami engineering. The report summarizes available information, identifies gaps in existing knowledge, and discusses methods of predicting tsunami flooding . The generating mechanisms of tsunamis and the method of determining the probabi li ty of occurrence are given. (continued)
DO
FORM 1 JAN 73
1473
EDinON OF 1 NOV 65 IS OBSOL.ETE
UNCLASSIFIED
SECURITY CLASSIFICATION OF THIS P A GE ( lrhen Data Enter ed)
IINCIASSTEJEQ
SECURITY CLASSIFICATION OF THIS PAGE(Whon Data Entarad)
Because of the limited data available on tsunamis, numerical methods are commonly used to predict tsunami flooding of coastal areas. Finitedifference equations are presented for simulating the propagation of tsunamis, but computer programs are omitted because of the continuing work in progress and the availability of uptodate computer programs from other sources. Known mathematical solutions, for tsunamis approaching the shoreline and tsunamishoreline interaction, are given to illustrate the effects of tsunamis and provide means of verifying numerical results. The report discusses tsunamistructure interaction and illustrates various types of damage caused by tsunamis .
2
U NCLASSIFIED
Së:CURITY CLASSIFICATION OF THIS PAGE(Wh e n Data Entered)
is concerned with the effects of tsunamis on the coastal zone. Hulman. R. National Oceanic and Atmospheric Administration. The author expresses his appreciation to the reviewers for their many helpful comments and suggestions.A. Houston.PREFACE The high value and high utilization of the coastal zone require the establishment of flood levels that may occur as the result of various natural causes. Meyer. Colonel. Stanford University. 79th Congress. H. and Dr. Approved for publication in accordance with Public Law 166. Camfield. and the consideration of preventive measures that can be used to minimize lasses.J.S. Drs. The report was reviewed by Dr.G.S. U. Frederick E. as supplemented by Public Law 172. Coastal Design Criteria Branch. approved 7 November 1963. Another report in the series. Raichlen. under the general supervision of R. prepared as one of a series of reports to be published to form a Coastal Engineering Manual. Corps of Engineers Commander and Director 3 . Dr. California Institute of Technology. This report. M.E. Nuclear Regulatory Commission. Chief. Loomis and L.W. R. Sorne of the background work on tsunami engineering was partially funded by the Nuclear Regulatory Commission. Jachowski. Joint Institute of Marine and Atmospheric Research. and L. The work was carried out under the coastal engineering research program of the U. W. Army Engineer Waterways Experiment Station. Dr.G. will be concerned with the effect of storm surges. F. Geller. R. Hydraulic Engineer. Whalin and J. Comments on this publication are invited. Bivins. Drs. Army Coastal Engineering Research Center (CERC).Q. 88th Congress.R. University of Wisconsin. Spielvogel. to be published separately. Fliegel. approved 31 July 1945. The report was prepared by Dr.
4. 3. Probability of Occurrence. . . 2. 4. . THE GENERATION OF TSUNAMIS. . . S. 9 10 17 17 21 II 27 28 30 30 31 III 32 32 34 IV 38 38 44 47 51 55 60 V TSUNAMIS APPROACHING THE SHORELINE. Volcanic Activity . . Tsunami Surge on the Shoreline . . Nearshore Computer Models. 2. Shelf Resonance. Computer Models . 2. 6. Distantly Generated Tsunamis 4. 7. . . Nature and Origin of Tsunamis. . . Bay and Harbor Resonance TSUNAMI RUNUP AND INTERACTION WITH STRUCTURES 1. 5. LongWave Equations . . Area and Height of Uplifting 2. . . . . 4. 3. 3. .CONTENTS Page CONVERSION FACTORS. MECHANICS OF GENERATION 1. Linear Depth Transitions . Initial Wave Formation TSUNAMI PROPAGATION . Subrnarine Earthquakes. 5. CUSTOMARY TO METRIC (SI) SYMBOLS AND DEFINITIONS 1 INTRODUCTION . 1. Reflection from Seaward Edge of Shelf. Solitons and ShoalingInduced Dispersion VI TSUNAMISHORELINE INTERACTION 1. Other Shoreline Structures . 67 67 73 76 79 80 88 88 91 98 1@1 108 130 131 VII 146 147 158 167 169 4 . . . 4. MachStern Formation. . Refracted Waves and Caustics 6. Nonlinear Depth Transitions . . . . . . Nearshore Propagation . . 2. . Landslides and Subrnarine Slurnps. 2. . Interaction with ShoreProtection Structures 3. . Edge Waves 5. . 1. 1. . . Experimental Measurernents . 3. . Wave Reflection. . SrnallArnplitude Waves. Tsunami Runup on a Shoreline . U. . Abrupt Depth Transitions . 1. 2. Explosions .
7 Typology of tsunami events. . . 18 23 25 26 28 29 35 5 Movement along faultlines . . . 18981961 . Buman Response . . 2. . U . . . IX APPENDIX TSUNAMIS OCCURRING BETWEEN 1891 AND 1961 TABLES Distribution of amplitude U 2 Values of horizontal water particle displacement. and of Puget Sound . . 7 Horizontal motion normal to continental slope 5 . . Ionospheric Waves . showing arrivai of tsunami. . . . 3 4 Principal fault systems and distribution of epicenters of major . and intensity of secondary undulations of inlets of Alask~ and British Columbia. and 93 98 103 133 162 178 197 5 6 . . . . · Alaskan earthquakes. DeepOcean Tsunami Gages. 6 Wave record from Wake Island. Allowable overtopping heights N. . The Tsunami Warning System. Mean annual occurrence of shallowfocus earthquake shocks for the Aleutian and southeastern Alaska region. Drag coefficients . . 2 Wave height versus tsunami magnitude. periods of fundamental mode. FIGURES Oceanic zones of recent earthquake activity. 3 4 Resonant edge wave parameters Dimensions. LITERATURE CITED . . . SUMMARY AND CONCLUSIONS . . showing association with trench systems and island arcs.CONTENTSContinued Page 193 194 196 198 198 199 201 217 VIII TSUNAMI WARNING SYSTEM AND INSTRUMENTATION 1. 4. 3. . wave amplitude. . . .
Page 48 50 56 58 59 60 62 64 68 70 70 74 74 e2 versus incident wave angle e1 77 77 80 81 81 82 83 83 87 88 6 . . .000 seconds after the 1964 Alaska earthquake . . . . . Computation of pressure near the free surface. spherical coordinates . Coordinate system. Reflection and transmission coefficients Separation of solitons . Position of variables. Spherical coordinate system. Solitary wave propagating over a slope onto a shelf. . . Surface elevation contours 13. . rectangular coordinates Coordinate system. Graphical representation of the total transmission open boundary condition. Wave reflection from a shoreline . . . .CONTENTS FIGURESContinued 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Convergence of wave rays . . . . . Wave reflection from a shelf Transmitted wave angle Linear slope and shelf Reflection and transmission coefficients Slope and shelf . . . . . Reflection coefficients. Solitary wave propagating onto a shelf Wave train propagating onto a shelf. . . . Induced wave generation over a submerged bar Wave enhancement . . . . . Wave passing onto shelf . . . .
Frequency response of a fully open harbor. Schematic of caus tic 1960 tsunami refraction. solitary wave Plan view of inlet Amplification factor versus relative harbor length Wave radiation functions . . . Japan. Hilo. . . . . Tsunami runup at Hongo in Toni. Theoretical frequency response curves of harbors Harbor with an entrance channel . . Response curve at point C of the Long B~ach Page 92 95 99 102 108 109 113 123 131 135 136 137 138 139 141 142 144 harbor madel . Trapping of generated tsunami. Machstem formation. . . . . Japan Solitary wave runup. Wavelength for Helmholtz resonance Tsunami water levels in a bay.CONTENTS FIGURESContinued 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 Shelf resonance. Japan . 146 148 150 151 152 153 160 1964 tsunami runup. Shizukawa Bay. . . . Resonant amplification on a shelf. . . Tsunami runup at Kamaisi. Concrete seawa11 destroyed by 1960 tsunami. . Solution to equations (264) and (265). . . . Kodiak City. . Hawaii. Japan. . Tsunami runup at Ryoisi. Alaska. . . 7 . Reflected waves on a shelf Offshore profiles of edge waves.
CONTENTS FIGURESContinued Page 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 Overtopping volumes. Suggested design for rehabilitated breakwater, Hilo, Hawaii. Suggested design, typical nonovertopping barrier section, Hilo, Hawaii. . . . . Seawall cross sections Cross sections of seawall, Yamada, Japan Coconut palms near shoreline, Hilo, Hawaii Grove of pandanus trees knocked dawn by 1946 tsunami on the Island of Kauai, Hawaii . . Dock damaged by 1964 tsunami at Crescent City, California. Tsunami damage to railroad bridge on Wailuku River, Hilo, Hawaii. . . . . . . . . . Tsunami damage to railroad trestle on Kolekole Stream, Island of Hawaii . . Bridge damaged by 1960 tsunami at Mangoku, Japan Definition sketch of surge on a dry bed. . Determination of
CD
161 163 163 164 165 166 167 168 170 170 171 175 180 184 187 188 189 191
when flow passes under a structure
Large boulder moved by 1960 tsunami, Hilo, Hawaii . .
CM
for twodimensional flow past rectangular bodies x versus t for abjects moved by tsunami
Example plots of surge .
Building moved by tsunami surge. Damage to oil tank farm at Whittier, Alaska.
8
CONVERSION FACTORS, U.S. CUSTOMARY TO METRIC (SI) UNITS OF MEASUREMENT U.S. customary units of measurement used in this report cru1 be converted to metric (SI) units as follows: Multiply inches square inches cubic inches fe et square feet cubic feet yards square yards cubic yards miles squarv miles knots acres footpounds millibars ounces pounds by 25.4 2.54 6.452 16.39 30.48 0.3048 o. 0929 0.0283 0.9144 0.836 0.7646 1.6093 259.0 1.852 0.4047 1. 3558 1. 0197 28.35 453.6 0.4536 1. 0160 o. 9072 0.01745 5/9
x
To obtain mil limet ers centimeters square centimeters cubic centimeters centimeters met ers square meters cubic meters met ers square meters cubic meters kilometers hectares kilometers per hour hectares newton meters 1o3 kilograms per square centimeter grams grams kilograms metric tons metric tons radians Celsius degrees or Kelvinsl
ton, long ton, short
degrees (angle) Fahrenheit degrees
1To
obtain Celsius (C) temperature readings from Fahrenheit (F) readings , use formula: C = (5/9) (F 32). K = (5/9) (F 32)
+
To obtain Kelvin (K) readings, use formula:
273.15.
9
SYMBOLS AND DEFINITIONS
A a coefficient • area of uplifting eprojected area of body normal to flow direction wave amplitude • a coefficient esemimajor axis of ellipse einterfocal distance of coordinate ellipses •a distance from the shoreline = ( /2  1) d8 /S elength of building in direction of flow variable used in determining movement of the sea surface amplitude of incident tsunami; amplitude in deeper water amplitude of tsunami at head of bay or inlet; resonant amplitude mean width of a harbor or inlet • a coefficient a variable used in determining wave amplitude width of outer bay or inlet width of inner bay or inlet a coefficient ewidth of building transverse to direction of flow esemiminor axis of ellipse ewidth of breakwater opening; entrance width of bay, harbor, or in l et wave ce le ri ty • a coefficient
a
8n
a1 a2 B
Bj
B1 B2 b
c
cv drag coefficient
Ce Cp
a Mathieu function force coefficient
ca group velocity ch
CM
c Chezy roughness coefficient added mass coefficient maximum uplifted elevation • a constant
10
SYMBOLS AND DEFINITIONSContinued drift speed of the node or antinode of an edge wave a coefficient used in determining wave reflection at the shoreline
D
a coefficient focai depth of earthquake a variable used in determining wave amplitude water depth eprojected dimension transverse ta direction of flow average depth average harbor depth a length representative of water depth depth of water at the point where a wave ray turns parallel ta bottom contours depth of water at toe of nearshore slope depth of deeper water depth of shallower water energy incident wave energy reflected wave energy transmitted wave energy even Mathieu Transform
F
force buoyant force drag force odd Mathieu Transform
f
friction factor coriolis parameter 2SG cos 8 11
harbor.SYMBOLS AND DEFINITIONSContinued fS(Z) g H ~ probability of an astronomical tide of elevation Z acceleration due to gravity wave height incident wave height zeroarder Hankel function reflected wave height transmitted wave height wave height when the leading edge is at the shoreline surge height • uplifting height average height of uplifting surge height at the shoreline wall height • wetted height on a structure relative intensity of secondary undulations an integer used for increments H(l) 0 Hr Ht H h h h8 hw I j K R/h8 a constant coefficient of reflection Krz Kr = Hr/~ = Ht/Hi Kt k k0 coefficient of transmission an integer used for increments ewave number = 2n"/L wave nurnber at lowest mode of resonance (Helmholtz mode) distance normal to faultline over which vertical Earth movement occurs • wave length length of bay. or inlet length of harbor entrance channel L 1o 10 12 .
SYMBOLS AND DEFINITIONSContinued Le effective length of harbor or inlet fault length wavelength at lowest mode of resonance (Helmholtz mode) longshore wavelength wavelength in deeper water wavelength in shallower water length of slope connecting sea bottom to a shelf distance across a shelf Richter magnitude of earthquakes •moment momentum tsunami magnitude normalized horizontal water particle displacement Manning roughness coefficient probability of tsunami with magnitude given year probability of runup to elevation Z pressure ea coefficient in wave refraction flow rate under a wave a coefficient in wave refraction evelocity of a water particle under a wave a coefficient evertical height of runup above the stillwater leve! at the shoreline average runup height at a shoreline area radius of the Earth radius from the center of curvature to the shoreline m being generated in any Lf 10 1y 11 12 t t8 M M 0 m N n n(m) P(Z) p Q q R R Re Rs 13 .
Le t ti me time required for a wave to travel across a shelf u u normalized wave amplitude • (H/ d)(L/ d) 2 current or particle velocity in direction of wave motion • veloci ty in the adirection (spherical coordinates) current velocity of tsunami surge at the shoreline u* a convection term volume of water v 14 .Continued r radial distance radius from the center of curvature to where the wave ray turns parallel to the bottom contours s slope of sea bottom in direction of wave motion slope of steep transition slope of shelf nearshore slope even Mathieu Function odd Mathieu Function s T distance along a wave ray wave period bottom shear stress natural period of harbor inside breakwater nth mode of oscillation component period of Earth vibration period of bay or inlet natural period of inlet with effective length.SYMBOLS AND DEFINITIONS..
t) vertical movement of sea bottom water surface elevation above still water at an arbitrary point n 15 .SYMBOLS AND DEFINITIONSContinued v horizontal velocity in direction transverse to wave motion evelocity in the tdirection (spherical coordinates) velocity in the vertical direction a convective term horizontal coordinate in direction of wave motion distance between the shoreline and the point where a wave ray turns parallel to bottom contours w w* x y horizontal coordinate in direction transverse to xdirection maximum scarp height depth of wave generation a parameter used in wave reflection and transmission z z vertical coordinate dimensional constant ecorrection term for harbor entrance channel length angle in wave refraction between the incident wave ray and an orthogonal to the contour of the sea bottom angle of the beach slope ewave ray separation distance •angle of nearshore slope given in radians edimensional constant y specifie weight of water •an edge wave parameter dimensionless amplitude an incrementa! distance ô a small value dimensional constant an arbitrary increment eH/d ~(x.
t) 8 vertical movement of sea surface angle in polar coordinates eangle of inclination of water surface at front of surge •degrees latitude measured from the pole angle of incident wave ray ea constant angle of transmi tted wave ray ea constant horizontal displacement of a water particle p density edimensionless distance from the shoreline d/L x/a cr ~ longitude of a point variable used in determining movement of the sea surface dimensionless distance measured seaward from the shoreline latitude of a point wave radiation function wave radiation function an arbitrary function erotational speed of the Earth in radians per second ~(t) x $ $1 $2 n 16 .SYMBOLS AND DEPINITIONSContinued n(x.
and ta be able to predict the extent of flooding and the effect of wave forces in coastal areas subject to tsunami attack. sufficient warning of a tsunami attack must be given ta people located in these areas. In the original definition. Tsunamis. the term tsunami was applied to all large waves. The major part of this activity occurs along the boundaries of the Pacifie Ocean. However. This report discusses the prediction of tsunami effects in coastal areas and attempts to provide guidance in determining the flooding and wave forces at any particular location. Van Dom (1965) indicates that the Japan Trench radiates detectable tsunamis at the rate of about one per year." meaning wave." meaning harbor. Lesser amounts of activity occur elsewhere. 1. Tsunamis are primarily created by disturbances in the crust of the Earth underlying bodies of water. and the resulting uplifting of the water surface over a large area which forms a train of very longperiod waves. with other regions of strong activity primarily concentrated in the Caribbean and Mediterranean areas. 17 . The waves may have periods exceeding 1 hour. INTRODUCTION The term tsunami is derived from two Japanese words: "tsu. the wave energy tends to spread along the wave crests and the tsunamis affect mainly the areas near their source. Proper control must be exercised over the use of such areas. and procedures must be established for an orderly evacuation when necessary. but those terms have generally been replaced by the term tsunami. Also. and "nami. recent definitions have limited its application to waves generated by tectonic or volcanic activity. have very long periods and are not easily dissipated. Western literature previously referred to these waves as tidal waves or seismic sea waves. The present knowledge of tsunamis and the deficiencies in this knowledge are summarized. in contrast ta normally occurring windgenerated sea waves which have periods less than 1 minute. Areas of seismic activity which could potentially generate tsunamis are shawn in Figure 1. Tsunami waves generated by tectonic uplifting may travel across an ocean basin. Because of the potential destructive effects of tsunamis. or seismic sea waves.TSUNAMI ENGINEERING by Frederick E. When tsunamis are generated by volcanic activity or landslides. and in designing structures ta be placed in these areas. CamfieZd I. The waves may create large surges or oscillations in bays or harbors which are not responsive to the action of normal sea waves. including storm surges. it is necessary to understand the mechanisms of their generation and propagation. Nature and Origin of Tsunamis. causing great destruction at locations far from their source.
A Russian expedition. The final translation indicated that a tsunami had occurred along with the 1827 earthquake. Tsunamis occurring between 1962 and the present are not listed because a complete summary is not readily available. Soloviev and Ferchev (1961) refer to the reports of an event in 1827 at the Komandorskiye Islands. Part of Ambraseys' information has been omitted because of the lack of verification. Lutke. 1965). records of tsunamis in past centuries are mostly based on accounts of personal observations.. then into French.German. particularly when translating from one language to another.. located between Kamchatka and the Aleutian Islands. Oceanic zones of recent earthquake activity. 18 . reported the occurrence of an earthquake and noted that earthquakes were sometimes accompanied by arise in water level..TRENCHES •o ao• ao• = ao• Figure 1. particularly in the present century. The dates that tsunamis occurred have often been confused with the dates on letters or other accounts rather than the date of the actual event. then into. Good records are available for more recently occurring tsunamis. There have also been many errors in interpreting these older accounts. using Spaeth's data and sorne additional information from Heck (1947). however. OCCASIONAL FREOUENT EARTHO\IAME9 EARTHOUAKES ARCS ll!liliimi!l 10 Ht+++t VOLCAHIC: . The original Russian report was translated in French. and Cox. Tsunamis can be generated in any coastal area. under the command of F.P. Pacifie preponderance is apparent (from Van Dorn. Ambraseys (1965). and back into Russian again. Alaska. including inland seas and large lakes. showing association with trench systems and island arcs. The Appendix summarizes the occurrence of tsunamis from 1891 to 1961. PararasCarayannis (1969). PararasCarayannis. and Calebaugh (1976). Spaeth (1964) provides an extensive bibliography on tsunamis.
Townley and Allen (1939) provide similar information. and the waves as a tsunami. occurred at Port Townsend. Records of tsunamis in the Mediterranean and Middle East include theories on the eruption of Thira (also known as Santorini) and the tsunami on the coast of Crete that destroyed the Minoan Empire circa 1400 B. Heavy rain and high waves caused by a storm resulted in considerable damage. The collapse of buildings caused by flooding from the rain was misinterpreted as an earthquake. with various occurrences at later dates. which was probably an incorrect report of the 1866 event. the occurrence of a tsunami in conjunction with high storm waves would have caused more flooding. records of tsunamis originating in the ChilePeru coastal areas only cover about 400 years (from 1562 to present). Few records are available of tsunamis occurring on the CaliforniaOregonWashington coastline. California. a tsunami occurring at a low tide stage may have been given only passing notice." Camfield's (1975) article on historical accounts gives the date as 27 December 1866. "The main street was filled with drift logs. The observers probably gave special notice only to those waves which caused substantial flooding or large. those originating in Alaska about 200 years (from 1788).000 years. Factual accounts of tsunamis extend back at least 2. Kelly (Seattle. Knowledge of the action of more recent tsunamis can be helpful in evaluating historical information. Consideration must also be given to the fact that records based on visual observations may not include all tsunamis which occurred. 1979) also gives the year as 1866. personal communication. if noticed at all.C. Holden (1898) indicates tsunamis occurred at points on the California coastline in 1812. Washington. rapid variations of the water level in bays and harbors. when in fact neither occurred.A similar instance of errors in interpretation and translation occurred in the reports of an 1840 event at Santa Cruz. 19 . the Puget Sound WeekZy (1866) reported that a tide. may have been given more significance in the records than a tsunami occurring during a relative calm. In contrast. both list a 26 December 1856 date. the highest ever recorded. The report stated. and therefore. Accounts of tsunamis in Japan extend back at least 1. The historical accounts describing a graduai rise in water level indicate this was probably a tsunami. on 20 December 1866. Holden (1898) reported this as an earthquake and tsunami. mainly recorded or observed at San Francisco. but the origin is unknown. At a location where the normal tidal range was of the same order as the tsunami height. Neither Holden (1898) nor Townley and Allen (1939) report a tsunami occurrence in 1866. Washington. Although no record exists of major tsunamis on Puget Sound in Washington State. while the same tsunami occurring at a high tide stage would have been recorded as a major tsunami. and those occurring in Hawaii slightly more than 150 years (from 1813).300 years. and the dwellers on lower floors were compelled to elevate to the next story. with no additional details. Likewise.
et al.2 as reported by Gutenberg and Richter (1965). The probability of an earthquake having an epicenter in a location that would cause a tsunami.31 meter (1 foot) at Atlantic City. suggest that the earthquakes in this area are associated with basement faults which have been reactivated by the removal of Pleistocene glacial loads. otherwise.9 feet) ai Gibralter. only li~ited verification is available for numerical models for propagating tsunamis across large oceanic distances. Néw Jersey (Murty and. 1930). Because of this gap.7 meters (12 feet) at Antigua (West Indies) 6. Earthquakes frequently qccur in the eastern United States. cannot be determined from available data. shortly after the November 1755 earthquake near Lisbon. The only major recorded tsunami along the east coast of the United States and Canada was the tsunami which devastated the Burin Peninsula along Placentia Bay. 1976). Newfoundland.6 meters (18 feet) at Madeira. Portugal (Reid.2 meters (4 feet) at Barbados.d on the south coast of England. For comparison.4 meters (21 feet) at Saba (West Indies). Both of these were recorde. in November 1929. Their results are based on mathematical simulation of extreme events. it may not have been of major proportions (Hodgson and Doxsee. and the waves overflowed the lowlands on the coasts of Martinique and other French islands. they give a magnitude of 7. The only tsunamis of record that traveled across the North Atlantic were those generated near Lisbon. In 1761. and the earthquake near Charleston. Stein.Although tsunamis occur frequently in the Caribbean. and Hwang (1979) give probable maximum waves for tsunamis at points near both the Atlantic and Pacifie coasts of'the United States. et al. 1. 1914). Brandsma.5 meters (8. 2. 1914). 18 meters (60 feet) at Cadiz (Spain). either on the coastline or in an estuary. This tsunami was reported to have had a height of 0. (in preparation. 15 meters (50 feet) at Tangier (Morocco). Stein. 14.9 meters (16 feet) on the coast of Portugal. Wigen. Divoky. they are much less frequent in the North Atlantic Ocean. Ives (England). 3.2 feet) at St. Other runup heights in 1755 were estimated at 4.8 meters (6 feet) at Penzance and 1. All of the earthquakes in the eastern United States have occurred inland from the coastline. 5. the sea rose about 1.8 meters (5. A major gap in the data is tsunami heights in deep water.4 meters (8 feet) at Penzance (England) and flowed over the wharves and streets at Barbados (West Indies). the 1755 tsunami had a maximum rise of 2. In general. good data are available for only a limited number of tsunamis. which generated the tsunami. The tsunami was enhanced by an exceptionally high tide and high storm waves. in 1755 and 1761. At !east 26 lives were lost (Jaggar. For the 1929 Grand Banks earthquake.6 meters (48 feet) at Faial (the Azores). These include a large earthquake that occurred in New England on 18 November 1755. 1936). 1980) report on earthquakes in eastern Canada from Baffin Island to Newfoundland. South Carolina. 1929). on 31 August 1886 (Taber. 20 . Portugal. as weil as in the West Indies (Davison.
Because of the frequency of tsunamis occurring in the Pacifie Ocean.. P(h ) the probability of a flood level occurring at elevation h200 200 in any given year. For a known period of record. a tsunami warning system has been developed for the inhabitants of Pacifie coastal areas.e. 2. the probability of tsunami flooding at any elevation can be determined by the same methods used for determining the probability of floods on rivers. The model data were multiplied by correction factors and compared to historical data. It should be noted that there is a probability of sorne error in the predicted flood elevations based on available historical data. This produced additional data at points along the coastline where historical data were not available. Where sufficient historical data were not available. they give (2) where h 200 is the elevation of the maximum tsunami wave crest above mean sea level (MSL) 200 feet (61 meters) shoreward of the coastline.. and A and B the empirical coefficients which are determined for each point on the coastline. they generated additional data using a mathematical model. Carver. P(Z) < 0. the highest flood level is ranked l. A similar warning system has not been developed for other areas. the next highest is ranked 2. Kohler. and Markle (1977) have determined the probability of tsunami flood levels for the Hawaiian Islands. i. and so on.e. there is a 37percent probability that a 100year flood level 21 . For recurrence intervals greater than 10 years. the recorded flood levels can be ranked from the largest to the smallest. Probability of Occurrence. Linsley. For example. and allowed a determination of the empirical coefficients A and B at all coastal points. tf• Where sufficient historical data are available on tsunami flood levels. i.1. and Paulhus (1958) show that the probability of each flood level is then given by P(Z) where P(Z) the probability of flooding to the elevation the elevation above a defined datum the rank of the flood level the period of record in years Z in any year rn n + 1 (1) z rn n Houston.
with an increase in magnitude of 0. Soloviev (i970) revised the definition of tsunami magnitude by relating it. regardless of the chosen time increment.. there is a 9. a flood level with a recurrence interval of 100 years) will not occur in any period of 100 years. Methods used for obtaining confidence limits for tsunami flood levels should give the same results.(i. Also. An exact relationship between tsunami magnitude and earthquake magnitude has not been determined . Where no historical data are available. a 100year flood level predicted from a 100year period of record may be too low. rivers have a seasonal variation in flow. The dashlines show the range of the expected maximum runup. the predicted 100year flood level.5 being equal to a doubling of the energy.5. Assuming there is an equal chance of the real flood level being either greater than or less than the predicted value gives +2. and should be more representative of the actual tsunami energy. based on a 100year period of record. may be too high. .000year flood level will occur at least once in any period of 100 years.e.5percent probability that a 1. However. data may be constructed entirely from a computer model by assigning magnitudes to various tsunamis in the mathematical model. The relationship between the runup height ~ax and the tsunami magnitude rn is shown in Figure 2. Iida (1961) proposed that tsunamis could be assigned a magnitude based on their energy (the energy of the generated waves). Confidence limits for the predicted flood levels can be obtained using methods similar to those used for river flood levels. but there is no particular relationship between this time increment and the generation of tsunamis.5percent confidence limits. However. Therefore. and by determining the probability of generation for each tsunami magnitude. This tends to average out any high runup heights related to a particular coastal feature. He also related the tsunami magnitude to the maximum runup height in meters at the shoreline area experiencing the strongest tsunami action (Iida. due to differences in the characteristics of the individual tsunamis and coastal areas. but does provide an equation for the magnitude as rn = log 2 (12' R) 22 (3) . 1970). In the case of tsunamis.and 2. so a 1year time increment is significant in that case. Beard (1962) notes that there is a 5percent probability that the magnitude of the difference between the real flood level and the predicted flood level will be greater than or equal to twice the standard error. to the average runup height R (in meters) at the shoreline area experiencing the strongest tsunami action. the results will not have the same degree of accuracy. the 1year time increment is a convenient means of measuring time. based on Iida's data. Soloviev does not indicate the length of coastline to be used in the average. Therefore.
Abe (1979) suggests that the tsunami magnitude can be represented as a function of the average runup height and a constant which is dependent upon the source region and the station where the tsunami is rneasured.80 60 50 40 30 'E ~ 2' . can be related to the seisrnic moment. m Figure 2. obtained by this rneans. Soloviev's scale gives a more rapid increase in maximum wave height than Iida's scale for a given change in tsunami magnitude. ::1 a: 3 2 6 Tsunami magnitude.1:.~ . The probability n(rn) of a tsunami with magnitude rn being generated in any given year in a specified generating area is given by the ernpirical equation n(rn) ae bm ( 4) 23 .. ::z:: 1:'1 Q.. c ::1 20 15 10 8 6 5 4 Soloviev R a: . This indicates that an increase in magnitude on Soloviev's scale would represent a greater increase in tsunami energy than an equivalent increase in magnitude on Iida's scale. ·. Wave height versus tsunami magnitude. As shawn in Figure 2. .. He shows that the magnitude.
.0. Using these results and the model of Haskell (1969).074 and b = 0.75) + n(4.75 + 0.074 [e0. Taking the value rn= 3. M. fault area and average fault slip. as noted by Geller.5 to 4. It is assumed that tsunamis do not occur with magnitudes greatèr than 5.g. The method for grouping tsunamis (eq. a major source of tsunamis in the western United States (Houston and Garcia. actual earthquake magnitudes may exceed this value slightly.5 or greater. 1978). the prediction would be seven tsunamis of magnitude 3. Abe (1975) and Geller (1976) show from empirical results that the fault length of earthquakes is approximately equal to twice the fault width. the maximum magnitude occurs because the conventional magnitude scale is saturated and ceases to give a meaningful measure of the earthquake size . I t should be noted that the stress in rock cannat exceed sorne maximum value.5 for the 1964 Alaska earthquake.25) + n(4.75) (5) j=o which would include all tsunamis with magnitudes from 3.using statistics for the entire trench along the Chilean coast. Ge11er lists a magnitude of 8. This agrees with historical records of tsunamis in this area. However. Applying revised information for that particular generating area.75)] ( 6) which gives a value of 0.5) = 0.5. 5) has been uti lized by Houston and Garcia (1974).5 or greater being generated in any given year is n(3. Ge11er gives a maximum earthquake magnitude. If tsunami magnitude is related to seismic moment.00) + e0. a= 0. a group of tsunamis shown with magnitude 3. For a 412year period for the Chilean coast.63.0166 or a recurrence interval of 60 years.75 actually includes all tsunamis with magnitudes from 3.5j) n(3. of 8 .22.5 for the magnitude of a design tsunami (to be used for determining potential runup in coastal areas). Because of variations in the assumed fau1t 1engthtowidth ratio. the probabili ties wou ld be summe d 2 L: n(3. Kanamori and Cipar (1974) indicate that the 1960 Chilean earthquake had the largest seismic moment ever reliably determined (2 x 1030 dynecentimeters).5 to 5. To calculate probabilities tsunamis may be placed in groups.0. etc. defined by Kanamori (1972) as a function of rigidity. To analyze the probability of an individuaZ tsunami having a magnitude greater than or equal to 3. the rock will fracture when the stress reaches that value. the probability for a tsunami with a magnitude of 3. 24 .63(3.0 where the tsunami magnitude has sorne relationship to earthquake magnitude as mentioned previously. 63(4.where the coefficients a and b are determined by a least squares analysis of the available data for the generating area.63(4. e.75) + e0.
1 07. 25 .Another major source of tsunamis in the western United States is the Aleut i an Trench. adapted from Davis and Echols.758.\l7.1 i. 1962). (1975b) and Houston and Garcia (1978). Principal fault systems and distribution of epicenters of major Alaskan earthquakes. Also. as revised in Houston.llU o 0 1. Only relatively recent records exist for the area. 1969.. of about 8. le981961 (from Wilson. the straight !ines in Figure 4 representing the occurrence of earthquakes in Ala ska and the world would intersect at an earthquake magnitude of about 2. so the plotted lines should not be extrapolated to values of earthquake magnitude less than those shawn.5 Fillodtircleo "'hypoc•nur>50 Kil Figure 3. et al. The probabi lity of tsunami occurrence is assumed to be uniform along the trench.113. gives n(m) = 0. and the mean annual number of earthquakes of any given magnitude in Figure 4. Analysis of these records by Houston and Garcia (1974). M.0 · 71 m ( 7) which is similar to the previous equation for the PeruChile Trench .5 because of the physical limits on allowable stresses in the rock forming the Earth's crust. The straight !ines in Figure 4 are not accurate above an earthquake magnitude. EXPLANATION Me pl tude • U6. The distribution of recent earthquakes along the Aleutian Trench i s shawn in Figure 3.5.
5 or greater to be generated at any particular segment of the trench in any given year.05 0. This value is based on a relatively short period of data for large tsunamis only. 1969. Using equation (7).98 '' \ '' \ ' 0.6_ _ _ __. 1964) .75) (8) which gives a value of 0 . The general equation for a part~cu l ar segment of the trench becomes n(i) = 0.1 0. Magnitude.3 0 ::Il '' ~ sœ ~ <a ~ \ \~ '' \ '' \ ::E «> '' 0. Trends of World and Japanese data are inserted for comparison.___ _.~ . Mean annual occurrence of shallowfocus earthquake shocks for the Aleutian and southeastern Alaska region (from Wilson..5 3 '' '' '' . adaptcd from Berg.___ _..00145 for a tsunami of the given magnitude of 3.0094 26 e. the probability of a tsunami with a magnitude of 3 .___ _ _ _8. 0174 for the Aleutian Trench. M Figure 4.5 0.5) = n(3.02 b= 0. '' '' \ z z 0 c c ct c 0 0.03 Log 10N=a+b(8M) a=1.01 .0 • 71 i (9) . Dividing the trench into 12 segments gives the probability of 0.7. 5 or greater is n(3.75) + n(4.25) + n(4. assum1ng an equal probability for each segment. '' \ '' \ '~ '.
The wave train of each tsunami must be superimposed on segments of the tidal cycle of an interval equal to the duration of the wave train.0. it is necessary to propagate tsunamis across the ocean by numerical ~eans frDm each segment of the trench for all tsunami magnitudes (i.0.e. THE GENERATION OF TSUNAMIS Tsunamitype waves can be generated from a number of sources. only the Aleutian Trench and the PeruChi1e Trench appear to produce significant tsunami runup. can be obtained for runup greater than or equal to a particular value. and explosions.and 500year tsunami flood elevations for the west coast of the continental United States. 2. For the west coast of the United States (excluding Hawaii). combining the tsunami with the astronomical tide. and the characteristics of the generated waves are dependent of the generating mechanism. 1980). 3. and the probability of the resulting runup determined. Determining the probability of tsunami runup at a particular coastal location for tsunamis generated in the Aleutian Trench area. would require the numerical generation of 84 tsunamis (12 segments of trench and 7 intensities of each segment). P8 (Z).0). transoceanic distances are normally generated by the tectonic activity associated with shal1owfocus earthquakes. large waves can be generated locally by the other generating mechanisms.To determine the probability of runup of a given height at a given location along the coastline.. the probability of runup to a given elevation is given by (10) Probabilities for tsunami runup can then be determined at each coastal point.0. including shallowfocus submarine earthquakes. This superposition must be made for each tidal segment of that interval for a 1year period. volcanic eruptions.5. However. By approximating the probability fs(Z) of the astronomical tide by a Gaussian distribution (Petrauskas and Borgman. although Holden (1898) indicates sorne occurrence of tsunamis from sources along the California coastline. 27 . II. 1andslides and submarine slumps. each runup value has an associated probability. Using numerical results obtained for tsunamis generated along the Aleutian and PeruChile Trenches. As shown by Houston and Garcia (1974). 1971. Each of these sources has its own generating mechanism. 4. Houston and Garcia.5. 3. Houston and Garcia (1978) have determined probable 100. The tsunami waves which travel long. i = 2. 4. and 5. and the totality of runup values at a given shoreline point defines a probability distribution from which the cumulative probability distribution. A cumulative probability can then be established for runup at a particular site.5. An analysis similar to that used for the Aleutian Trench could be applied to tsunamis generated in other source areas. Tidal variations are discussed by Harris (in preparation. 1974).
or other apparent transfer of material on the sea floor. A general expression for the lower limit of the earthquake magnitude. tsunamis are generated hy shallowfocus earthquakes of a dipslip fault type. Hammack and Segur (1974) studied the propagation of waves bath experimentally and numerically. i. However. without a net change in volume. i. of tsunamigenic earthquakes is given by !ida (1970) as M 6. Heck (1936) indicates that horizontal motion of the sea floor does not appear to generate large tsunamis. M. Van Dorn (1965) discusses the generating mechanism of the 1964 tsunami which originated in Alaska .) Waves generated from a unipolar source decay much less rapidly with distance than waves generated by a bipolar disturbance. 5). and Cox (1950) indicate that tsunamis which travel long distances across the ocean are probaoly caused by unipolar disturbances.g. i. horizontal motion along the faultline (Fig. large "local" tsunamis may be generated by horizontal motion. They indicate that where there is a positive net change in volume (e.3 + 0.e. MacDonald..1. waves of stable form (solitons) evolve. (An example of a unipolar disturbance would be the uplift of a large area of the sea floor where there is a net change in volume.e. Movement along faultlines.e. The wave record for the 1964 tsunami at Wake Island (see Fig. followed by a dispersive train of oscillatory waves.Slip Fouit StrikeSiip Fouit Figure 5. As the positive pole was the main tsunamigenerating mechanism. this was equivalent to a unipolar source. !ida (1970) shows that major tsunamis (those that cause high water levels at many different coastal locations) do not appear to occur as the result of deepfocus earthquakes or the strikeslip fault type. 6) illustrates this type of wave generation. Shepard. a combination uplifting and subsidence. vertical EOtion upward on one side of the fault and downward on the other side (Fig.005 Df (11) 28 .. a unipolar uplifting of the sea floor). As shawn by !ida (1970). having a positive pole (uplifting) under the sea and a negative pole predominantly under the land.. Oip. Submarine Earthquakes.. The ground motion was dipolar. The number and amplitude of the solitons depends on the initial generating mechanism. 5).
based on tsunamigenic earthquakes in Japan.8) 1. the configuration of the coastline. Tsunamis usually do not occur for earthquake magnitudes less than that given by equation (11).en associated wi th lesser magnitude earthquakes.5 where E is the earthquake energy in ergs. although a small number of tsunamis of lesser magnitude have be. However. the length 29 . and possible local resonance effects. the equations developed to define these limits are based on limited data and do not fully consider coastal configurations and resonant effects. i. Also. The difference results from the relationships used to determine earthquake magnitudes from surface wave magnitudes and body wave magnitudes. the definition of "disastrous tsunamis" may be more a function of the location of the origin and the population in . Richter (1958) gives higher values for earthquake magnitudes than those listed by Gutenberg and Richter (1954).e. 1964). showing arrival of tsunami (initial motion is positive and remains above normal tide curve for more than an hour) (from Van Dom. The Richter scale is given by M = (logE . the adjacent coastal zone.Figure 6 . A tsunami generated from a dipslip fault source will have the characteristics of being generated from aline source. Wave record from Wake Island. Geller and Kanamori (1977) note that care must be taken when defining earthquake magnitude. It should be noted that equation (11) does not consider the location of the earthquake with respect to the coastline. rather than an analysis of the actual waves generated.11. Attempts have been made to define lower limits for earthquake magnitudes associated with disastrous tsunamis.. where Df is the focal depth in kilometers and M the magnitude on the Richter scale.
Tsunamis with volcanic origins have the characteristics of waves generated from a small source area. in 1853. shoreline and submarine slumps. Both the 1868 and the 1975 tsunamis in Hawaii caused high waves at points on all sides of the island of Hawaii as well as waves on the other islands (PararasCarayannis. and volcanic explosions.92 cubic miles) of the island. Alaska.e. or standing edge waves may be generated along the coastline. 1975). but are often associated with earthquakes. there may be refraction effects which trap waves along the coastline. 1936. 1874. Landslides and Submarine Slumps. Also. The 1958 30 .e. 1975) indicated that trapped waves may exist with many nodes around the island. Large shoreline subsidences were associated with the eruptions and earthquakes on Hawaii. When displacement occurs along a substantial length of faultline. the spreading of wave energy along the wave crest) will be much less than for a wave generated from a small source. a small percentage have been caused by volcauic activity which includes localized earthquakes. For a "locally" generated wave. with associated earthquakes off the southeast coast of the island of Hawaii. The 1975 waves persisted for more than 4 hours at all points. and 1958. persona! communication. the main component of the wave energy will travel perpendicular to the faultline and the energy per unit length of wave crest would remain approximately constant for an unrefracted wave. leaking energy to the surrounding ocean. The waves generated by such events will spread geometrically as they propagate from their source in an open ocean. 2. Landslides and submarine slumps can occur from various causes. or if resonant or refraction effects exist. Volcanic Activity.of the generating area is much greater than the width . but may cause very large waves near the source. a wave generated near the coastline under consideration. PararasCarayannis. These trapped waves would gradually decay. Meyer (Department of Mathematics. Although most major tsunamis have been caused by shallowfocus earthquakes. and the August 1883 eruption and explosion of the island of Krakatoa near the Sunda Strait in Indonesia. Examples of these are the volcanic activity of April 1868 and November 1975 in Hawaii. 3.. but can be very high near their origin. Waves can be particularly high if they occur in a confined inlet. i. International Tsunami Information Center.. The explosion of Krakatoa destroyed an estimated 8 cubic kilometers (1. University of Wisconsin. 1969. the divergence of the wave rays of the generated wave (i. persona! communication. These waves spread geometrically and do not cause large wave runup at locations distant from the source. Examples of landslidegenerated waves have been reported by Miller (1960) for Lituya Bay.
at a smallcraft harbor breakwater. the water depth increased from 2. Alaska. and generated a 61meterhigh wave seaward in the bay.740 feet) on the opposite side of the bay. in 1845 and 1905. These later waves apparently resulted from sorne reflection or resonant effects within Prince William Sound. 1805. Data from nuc1ear explosion Baker at Bikini Atoll in 1946 show that the wave height is approximately inversely proportional ta the radial distance from the point of origin. a third wave arrived. Ambraseys (1960) indicated that the tsunami of 9 July 1956 in the Greek Archipelago was probably produced by a series of landslides on the steep banks of a submarine trench. 1755. as referenced by Wiegel (1964). A survey of the Valdez.e. The height of a wave generated by an explosion has been shawn ta be dependent on the depth of the explosion charge. destroying the dock. At a radial distance equal ta 3Sd.7 ta 15 meters (35 ta 50 feet) high. and Hwang (1968) show that two critical depths exist which will produce the highest waves for any given explosive charge. An explosion acts as an impulsivegenerating mechanism which generates dispersive waves from a point source. 4. Hr =constant where H is the height of the wave. Striem and Miloh (1975) report that tsunamis have probably been generated by slumping of the continental slope off the coast of Israel. heading toward the Valdez Narrows from the open sea. and r is the radial distance from the point source. Alaska. and only about 2 percent of the potential energy of a falling or sliding weight is converted into wave energy. and the second wave by the slump of a shoreline area sorne distance away. and will not be considered further here (see Smith. and 1934. area after the March 1964 earthquake showed that the water depth at the end of the Valdez Dock had increased from 9 ta 37 meters (30 to 120 feet). in the narrows. Extensive material is avai1able on waves generated by explosions. Explosions. Norway. The first wave ta hit Valdez was generated by the s1ump of the waterfront. reported on landslidegenerated waves in Tafjord. The critical depths are dependent on the charge yield. Also. with the wave height decreasing less rapidly. i. The owner of a fishing boat. Waves were also generated by icefalls in Yakut~t Bay. followed more than 2 hours later by a fourth wave. which dispersed after passing the narrows. Jorstad (1956). 31 . Van Dorn (1965) indicates that tsunamis generated from this type of source appear to be fairly localized and will not be large at long distances from the source. Van Dorn. The generating mechanism is extremely inefficient. destroying the breakwater. in 1718. The wave had an amplitude of 30 meters (100 feet) near its source. 1868.. (1970).7 ta 27 meters (9 ta 90 feet). reported a wave 10. Le Mehaute. et al. where d is the water depth. After about 5 ta 6 hours.wave reached an estimated maximum surge elevation of 530 meters (1. An example of a wave generated by a shoreline slump is given in Berg. the relationship changes slightly. 1967). Wilson (1963) discusses data on wave dispersion. given in equivalent pounds of TNT.
it is generally assumed in analyzing wave generation that the total uplifting occurs instantaneously.0 to 3. as reported in Berg.2 miles) per second for rupture propagation and a rupture length of from 600 to 800 kilometers (370 to 500 miles) . 1965. Crustal displacement progresses along a faultline from sorne initial source. transoceanic tsunamis results from the displacement of water above the area of uplifted sea bottom associated with a dipslip fault movement. speed.9 to 2. et al.. it may be assumed that the uplifting of the water surface equals the uplifting of the sea bottom. An uplifting of the sea bottom will produce a vertical uplifting of the overlying water. As a first approximation. Very little data are available on the size of the generating areas and the height of uplifts for various tsunamis which have been recorded at coastal points. ~ ~ ~ h. The potential energy of an incrementai area of uplifting is proportional to h 2 . The uplifted area in Prince William Sound was considered to have a limited effect on the tsunami generation because of the restricted connections between the sound and the shelf area. MECHANICS OF GENERATION The generation of large. Area and Height of Uplifting. extensive surveys were undertaken in the area of origin (Plafker. 1970). The average value of h 2 was estimated as 4. 1. Because of the high speed of rupturing.0252 grams per cubic centimeter (1.7 centimeters (32.174 feet) per second squared an incrementai area of uplifting the height of uplifting over the incrementai area Ai p g 32 . establishment of new elevations at hench marks. After the 1964 tsunami generated in Alaska. where h is the height of uplifting. and measurement of the displacement of sessile marine organisms. 2 (13) the energy in ergs (footpounds) the density of the seawater and is assumed to equal 1. Berg. The potential energy of the uplifted water is then given as n E where E i=1 L pg A. (1970) for the 1964 Alaskan tsunami give speeds from 3. The uplifted water area on the Continental Shelf was estimated as 1. new hydrographie surveys in areas previously surveyed.5 kilometers (1.h.III. These surveys included comparisons of tide levels at surviving tide gages.1 square feet). et al.1 square meters (44.184 x 10 12 square feet).1 x 10 1 1 square meters (1.989 slugs per cubic foot) gravitational acceleration and is equal to 980. and length of rupture propagating from the epicenter of a given earthquake by using recorded seismic surface waves. establishment of previous tide levels by visual observation and interviews with area residents. BenMenahem (1961) developed a method for determining the direction. Various analyses using this method.
If the incrementa! areas are equal, i.e., A 1 equation (13) can be rewritten as
E
Ayz, then
(14)
or, alternatively,
E
pg n A.
7.
n
l:
h~
7.
(15)
i=1 2n
Noting that the total area, A= n A. and that
h~
1.
A,
is given by (16)
n
h2
i= 1 2n
equation (15) becomes
l:
...J:..=
1 +
h~
+
+
h2 n
(h2) avg 2
(17)
2n
E = pg
A (h2)aVfi.
2
(18)
where (h2)
avg
is the average value of the square of the uplifted heights.
For the 1964 Alaskan earthquake the height of uplifting varied considerably over the area of uplifting, and had a maximum in excess of 15 meters (49 feet) at a point near Montague Island (Malloy, 1964). The tsunami had a calculated potential energy of 2.26 x 10 22 ergs (1.67 x 10 15 footpounds). When using equation (18) it must be remembered that the average of the height squared, (h 2 )avg. is not equal to the average height squared, (havg) 2 . This is easily illustrated by the following example problem. * * * * * * * * * * * * * EXAMPLE PROBLEM 1 * * * * * * * * * * * * * * *
~x
GIVEN:
An area of uplifting is divided into five equalsized areas of 101 2 square centimeters (2.48 x 109 square feet), with upliftings
of 30, 60, 90, 120, and 150 centimeters, respectively. FIND: (a) (b) (c) The value of (havg)2' the value of (h2)avg• and the potential energy of the uplifted seawater.
33
SOLtrriON:
(a)
h
avg
30 + 60 + 90 + 120 + 150 5
90 centimeters
(havg) 2
8,100 square centimeters (8. 72 square feet) 30 2 + 60 2 + 902 + 120 2 + 150 2 5
(b)
(h2) avg (h2)
avg
= 49 5 500 = 9,900 •
square centimeters (10.66 square feet)
(c)
From equation (16),
E
2
E E
(1.0252)(980.7)(5)(2.3 x 1012) 9 ,9oo
2
5.72 x 10 19 gramsquare centimeters per second squared 5.72 x 10 19 ergs (4.22 x 1012 footpounds)
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
The horizontal motion along a rupture line may contribute very little to tsunami generation. The maximum energy input from the horizontal motion would occur when the rupture line is normal to the continental slope. The motion along the rupture line, in that instance, would be equivalent to a wedge moving a short distance through the water (see Fig. 7). Berg, et al. (1970) show that for a motion equivalent in magnitude to that of the 1964 Alaskan earthquake, and acting normal to the continental slope, the potential energy input to the resulting tsunami would have been 3.12 x 1020 ergs (2.3 x 10 13 footpounds). This is less than 1.5 percent of the energy input from the vertical uplifting, and seems to confirm Iida's (1970) analysis which showed that major tsunamis are associated with the dipslip fault type. In fact, the rupture line of the 1964 Alaskan earthquake was almost parallel to the continental slope, and the energy input from the horizontal motion would have been negligible in this case. In other cases, the contribution of horizontal motion may be greater. 2. Initial Wave Formation.
Because of their long periods and corresponding long wavelengths, the train of waves forrning a tsunami is taken to be shallowwater waves at their origin, and propagates across the ocean as shallowwater waves. The actual form of the wave train is determined by the initial generating mechanism, i.e., the area of the uplifted sea bottom, the height of and
34
Equivalent displocement of woter surface
Figure 7.
Horizontal motion normal to continental slope (scale exaggerated).
variation of the uplift within the area of uplift, and the depth of water and coastal characteristics in the generating area. While ordinary sea waves are assumed to have a cnoidal shape as they approach a shore (i.e., high crests and shallow troughs), the waves in a tsunami may have various combinations of forms. Visual observations of tsunamis have led to reports that the initial wave was often a negative wave causing an initial drawdown of the water level at the shoreline. Shepard, MacDonald, and Cox (1950) show that s?me tide gage records indicate a small positive wave followed by a very deep trough, the amplitude of the trough being about three times the amplitude of the initial wave crest. This may have been misinterpreted by observers who reported initial negative waves. However, Striem and Miloh (1975) indicate that an initial drawdown may occur when the tsunami is generated by slumping of the continental slope. Tsunamis may sometimes produce waves with narrow, deep troughs and low, wide crests at the shore'.· line, the opposite of the cnoidal waveform. Wave records from Wake Island for the March 1964 tsunami (Van Dorn, 1964) show a positive surge with a period of 80 minutes (see Fig. 6). There was a series of positive wave crests with the elevations of the intervening troughs above the normal expected tide level. This was followed by a series of crests and troughs with the elevations of the troughs below the normal tide level. Using a shallowwater wave celerity at the source and an average depth of approximately 100 meters (325 feet) for the generating area, the period of the initial positive surge is approximately equivalent to the time ~equired for the trailÎng edge of the initial uplifted water surface to travel completely across the area
35
of generation. This indicates that the uplifted water surface at the source formed a series of solitarytype waves. The multiple crest can be accounted for by initial instabilities in the waveform caused by the generating mechanism, and the effect of the varying bathymetry of the ocean basin through which the wave passes. The lower waves following the initial series of wave crests correspond to the expected oscillations from a disturbance in the water surface as the disturbance is damped out. Wilson, Webb, and Hendrickson (1962) showed that the height of a tsunami at a coastal point near the source of generation could be given as a first approximation by the empirical equation log 10 H = 0.75 M 5.07
(19)
where H is the height in meters and M the Richter magnitude. Using the value of M = 8.3 given by Berg, et al. (1970) for the March 1964 tsunami, H = 14.29 meters (46.9 feet). However, this is an empirical relationship which does not completely account for the characteristics of the generating mechanism or the coastline. Although equation (19) might provide a first ruleofthumb estimate of wave heights, the actual heights could be above or below that estimate. Determination of actual heights would require computation by numerical or empirical means. Wilson (1969) gives the relationship of Housner (1969) for the fault length Lf in kilometers as log 10 Lf = 0.87 M 4.44
(20)
g1v1ng a fault length for the March 1964 tsunami of Lf = 604 kilometers (1.98 x 106 feet). This is within the range of estimates given in Berg, et al. (1970) and approximates the length of the generating area, i.e., the length along the initial wave crest. Wilson and T0rum (1968) give a relationship for the period minutes) of the primary tsunami (carrying maximum energy) as
0.625 M  3.31
T
(in
(21)
1
For the March 1964 tsunami, this equation gives a period of T = 75.4 minutes (using the Richter magnitude M = 8.3). This is very close to the period of the positive surge noted by Van Dorn (1964) at Wake Island, and is equal to that period if the crest of the initial oscillatory wave at the trailing edge of the surge is neglected. The initial deformation of the water surface, for any tsunami, will collapse into some system of waves which must be defined. The resulting wave system depends on the shape of the seabed deformation and the water depth above the deformation. The simplest means of analysis is to assume
36
although a computer solution can sum a relatively large number of terms. Z0 the depth of generation lnegative downward and equal to d for bottom uplifting). 8) 3/ 2 + o d)] ei'ITI4 (23) cosh(kd) n=o + L: Son (ka 2 ) . k the wave number. S) (22) where I k 11 2 cosh[k(Z (21r) and I (k. the initial wavelength (or wave period) must be known. The variables are defined as n the wave height. and the solution assumes that r is much greater than the dimension (diameter or length) of the source. 8. that the initial displacement is not timedependent. and Hendrickson (1962) and in other sources. 1975). They also indicate that only the first few terms of I(k. of varying degrees of complexity. Levy and Keller (1961) present one solution in terms of elliptic coordinates for a source region which is more elongated than circular. t) r1 Ao(k) [ (:~~)J 2rrC ll/2 /k(rct)+i(n/4) I(k. Many of the mathematical representations of waves generated from bottom uplifting are based on circular source regions. and then propagate the initial displacement outward from the generating area using longwave equations (Brandsma. Webb.l n. This is generally true for tsunamis.8) may be important in the solution.no 2 ~ (24) Ihe terms Sen a~d San are even and odd Mathieu functions. d the water depth. are described by Wilson. and a the interfocal distance of the coordinate ellipses. Levy and Keller indicate that the veloci ty of the bot tom uplifting is unimportant if the time of uplifting is small in comparison with the period of the generated waves. Other means of establishing the initial waves. t the time. and En~ne (ka/2) and Fn~no (ka/2) are even and odd Mathieu transforms. r and 8 the coordinates of a point in polar coordinates. Ca the group velocity. Divoky. however. The limitations on the solution are that the solution was derived for water of uniform depth. This solution has the form n(r.that the water surface has an initial displacement equal to the seabed displacement. cos 8 F (ka). C the wave celerity. and Hwang. 37 .
Because tsunamis are longperiod waves with long wavelengths in relation to both the water depth and the wave height. to a first approximation. longwave equations can be used. Divoky.026 grams per cubic centime ter) IV. The simplest means of analyzing the wave motion.oa (a2 6 a a 2 .. is to use the following smallamplitude solutions to the wave equations: 38 .x2 ) 5/2 dx (25) where x a y b z c p measured along the major axis of the ellipse the length of the semimajor axis measured along the minor axis of the ellipse the length of the semiminor axis the vertical direction upward from the undisturbed water surface the maximum uplifted elevation at coordinates (x = o. with an instantaneously displaced water surface. et al. is small. 1972) use a simplified monotonie displacement history to describe ground motion. and a triangular cross section parallel to the minor axis of the ellipse. as input data for a standard design tsunami in a numerical solution. having a parabolic cross section parallel to the major axis of the ellipse.Hwang and Divoky (1970. They propose that. horizontal displacement of a sloping bottom can be represented as purely vertical displacement. as discussed in Section III. H/d. They define the surface displacement as a modified elliptic paraboloid. and Hwang (1975). y= o. The numerical propagation of the wave uses the same procedure as used in Brandsma. z = c) the density of the seawater (taken as 1. TSUNAMI PROPAGATION After determining the initial disturbance of the water surface. (1975b) use an ellipticalshaped generating area. where the ratio of the wave height to water depth. the propagation of the tsunami to nearby or distant shorelines must be analyzed. 1. Houston. SmallAmplitude Waves. The potential energy of the uplifted water surface for this type of surface displacement is given by E 4 g) b 4 J (E.c.
:..e.... . (29) 39 . shallowwater equations . (27). and (28) can be reduced to more basic smallamplitude.__[_2_7T_~J=  + d)] cos [ 27T ( f . Therefore.f)J (27) L cos h [27T(Z L+ d)J a _ _.c2 ~~ tanh e~d) (26) u = 27T % h [27T(Z _c_o_s... . i.. L/d.. the ratio of wavelength to water depth..h. Sln . Letting d/L + 0.. is very large. sin . equations (26).. _ _ _ _=. Sln h [27Td] L (28) w here C L d u a T z x t ~ the wave celerity the wavelength the depth of the undisturbed water the horizontal velocity of a water particle in the direction of the wave motion = the amplitude of the wave above the undisturbed water level the wave period the vertical distance of the particle from the undisturbed water surface distance measured in the direction of wave motion time the horizontal displacement of the water particle from its undisturbed position Tsunamis are shallowwater waves.
(37).!l. it will always be true that z/L+ 0 whenever d/L + O. 1 + e~z)(2~d) T e~d) (39) Noting that z/L+ 0. and (38) into equation (27)._'..!! d T (40) 40 .. Therefore. shallowwater wave celerity as c2 = gL (30) c = lgd Because z < d. cash 2TI(z 1 + d) =1 + (2~z) (2~d) (37) Th e disturbed water surface elevation.. sinh ~+ 2TIZ L L (31) (32) (33) (34) sinh 2Tid + 2Tid L L cash 2TIZ + L cash 2Tid + L (35) Considering that cash 2TI(z L + d) = cash 2TIZ cash 2Tid L L + sinh ~ sinh 2Tid L L (36) and s ubstituting equations (32) ta (35) into equation (36)..substituting equation (29) into equation (26) gives (2Tid) 2TI L w hich gives smallamplitude.. U = 2 TI !}_ ~..:.. its undisturbed location is given by n. u=.. at any point in relation ta n = a cos [ 2TI ( ~  ~) J (38) Substituting equations (33).
n e~z) (2~d) e~d) t. and T is assurned to be constant.. (37). for a srnall. c so that = l:. and as C shallowwater wave. and (43) into equation (28)' 1 + t. nL . then /gd for a ( 49) 41 . T ( 41) u=. two points for an unrefracted wave. from the basic wave equation for all waves .2Tid ( 44) and noting again that z/L + 0.. (45) The wave energy. E. If energy is conserved between ( 4 7) ( 4 8) but as L = CT.but.!l/gd= d n gl/2 dl/2 (42) The water surface elevation can also be defined as n = a sin [2TI (f.arnp li tude wave is given by E pg H2 L 8 (46) where p is the density of the seawater.~)] ( 43) so that by substituting equations (33).
(54). (~2 y/2 = (::) (~1 y/2 = GJ'' G: r (53) (::r (54) (55) (56) Equations (49). smallamplitude. (50) aL 27Td For the unrefracted wave.280 feet) water depth into a 500meter (1.4 meter (1.31 feet) passes from a 1. * * * * * * * * * * * * * EXAMPLE PROBLEM 2 * * * * * * * * * * * * * * GIVEN: A long wave with a period of 20 minutes and a height.000meter (3. noting that (51) a2 H2 al (52) Hl jumaxl2 \umaxll lumaxl2 lumax\1 Also. of 0. shallowwater wave . The wave is assumed to be nondispersive. the ampl i tude . Noting that equations (42) and (45) can be written as agl/2 dl/2 nmax a.640 feet) water depth. and (56) provide a simple.which is the wellknown Green's Law. H. firstorder solution for the shoaling of an unrefracted. 42 .
d 1 1. From equation (54) where d2 lumax lsoo lumaxll. at d1 = 1.189) (b) From equation (50).2(9.ooo)l/ 4 500 1.ooo)34 / 3 4 / 1 000) 0.4 (I. the water particle velocity IUmaxl in each water depth.034 meter (0.189 0.FIND: (a) (b) (c) SOLUTION: The unrefracted wave height in the 500meter depth.000) 2 l/ 0.48 meter (1. 0.02 ( .4 meter. Hsoo = 0.11 foot) per second 43 .807)1/ 2 (1 .OOO 1 ( 500 .56 feet) H500 = 0. and the horizontal water particle motion l~maxl in each water depth.000 meters.000 meters.4(1.02 meter (0.2 meter. d2 500 meters From equation (49). (a) H1 = 0.' 500 0. Assuming a= H1 /2 = 0.065 foot) per second 500 meters.
8 miles) aL 2'ITd 0.(c) Firs t s olving fo r L L w here d 1 = 1.4 feet) From equation (56) where d2 = 500 meters. it is necessary to consider the curvature of the Earth.000) ( 500 1 max soo l ç.36 meters (20.max lso o 1ç. 2. or dispersion. LongWave Equations.9 feet) * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Soloviev.m \ ax lç.000 meters and T x 20 minutes. discussed later in this section. To increase the accuracy of computations.max d2 3 4 / Il . (~max )2= (~)3/ 4 ( ç. the equations may be writt en as fol l ows in rectangular coordinates: 44 . 118. diffraction. Also.78 ( 1 500 ~ / • 000 } 3 4 = 6.807 1. As given by Peregrine (1970) for twodimensional waves propagating in water of constant depth. and (56) do not account for wave refraction.800 meters (73. Equation (49) does not account for wave reflection from bottom slopes and results in calculated wave amplitudes that are tao high. the longwave equations can be expressed in various forms of partial differentiai equations which can be solved numerically. CT = /gdT = /9.OOO) 3.000 (20 x 60) L From equation (51).ooo 1 = 1 . When waves travel long distances. et al.78 meters (12.800) 27r(l. (1976) compared solutions for tsunami amplitude using equation (49) and a numerical integration method. 3. (54). equations (49). they cannat be used with any degree of accuracy when the ratio of H/d becomes large.2(118.
+ u .+ g an= au at at au ax ax .t a2 u 3 ax2 (61) ~+ Cl [(d+ n) u] ax =0 (62) In addition.(d)l/2 u + . d2 . The importance of this parameter was first noted by Stokes (1847) when he stated that the parameter must be small if his equations were to remain valid for long waves. for waves traveling in only one direction.g d3 a2 u ax 2 u2 4g .Linear equations: + g élu at an ax :: 0 au Cl x (57) an + d Clt Finiteamplitude equations:  =o (58) dU at at + u.6 (gd) l/2 (64) When considering the means of describing the propagation of longperiod waves.+ dU (59) ax ~ + a[(d + n) u] ax 0 (60) Boussinesq equations: . U.. the Boussinesq equations may be reduced to the KortewegdeVries equations which are then wri tt en as 2 au+ 2 (gd)l/2 au+ 3u au+ . the parameter.d2 at ax ax 3 (gd)l/2~= 0 ax3 (63) n .!.!. should be evaluated where U is defined as (65) and sometimes referred to as the Stokes or Ursell parameter. 45 .
the value of H/d is a measure of amplitude dispersion.807 x 1.. Zabusky and Galvin (1971) show that the KortewegdeVries equation accurately describes wave propagation for U < 800 in sorne cases. the shoreward slope of the bottom topography.e. 0.. 706 * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * In more recent investigations. and the finiteamplitude. Murty points out that the linear longwave equations are valid when U << 1.000 u = o.e.800)2 1.000 1. * * * * * * * * * * * * * * EXAMPLE PROBLEM 3 * * * * * * * * * * * * * * GIVEN: A tsunami has a period of 20 minutes and a wave height of 0.16 foot) in a 1. In this case Boussinesq or KortewegdeVries equations should be used. However.000 = 99 meters (325 feet) per second CT 99 x 20 x 60 = 118. the parameter (65) has been redefined as U* where U given by equation 46 .280 feet) water depth. The acceptable limit of the value for H/d (i.000meter (3.Murty (1977) indicates that the value of (L/d) 2 is a measure of frequency dispersion.05 meter (0. the error which results from the use of the linear equations is quite small as long as the value of H/d is small. For tsunamis with very long periods (and therefore long wavelengths).800 meters From equation (65). It should be emphasized that when U = 0(1) it is not necessary that U ~ 1. both amplitude and frequency dispersion are important. SOLUTION: C L /gd= 19. nonlinear longwave equations should be used. When U is of the order one [U = 0(1)]. the condition that U << 1 is usually never satisfied. FIND: The parameter U. i.05 (118. the point where the error in the calculations becomes significant) depends in part on the rate of shoaling of the wave. Where U >> 1 amplitude dispersion dominates the solution. In this case H/d is small and amplitude dispersion can be ignored. From equation (31).
e. The Boussinesq or the KortewegdeVries equations are used for waves approaching a shoreline where values of H/d become large. the sphericity of the Earth must be considered to determine the effects of the 47 .. When a tsunami travels a long distance across the ocean. Hammack (1973) gives the value of U* as (69) to describe a particular region of a complex waveform. This variation would indicate that using a single set of equations to describe a compZex waveform may lead to incorrect results. where (nx)~ax < (H/d).e. However.. the value of U* would be expected to vary from region to region of the waveforrn in this case. He points out that nonlinear terrns which were neglected in the linear equations (57) and (58) cause a cumulative error that may become appreciable in a numerical solution after a time given by t (68) Where rapid shoaling occurs. i. The Boussinesq equations are also applicable where U* > 1.U* and (~) (n x ~2 max (66) (67) This is discussed by Peregrine (1970). Peregrine (1970) points out that the Boussinesq approximation works well for values of H/d up to about O. The finiteamplitude equations (59) and (60) are valid as long as U* > 1 but generally become invalid after a finite time as the front face of the wave steepens. 3. where the wave passes over the same change in water depth in a relatively long period of time. i. where a wave passes over a large change in water depth in a relatively short period of time.S. Distantly Generated Tsunamis. the accumulated error will be much smaller than for slow shoaling.
As a result of the convergence of unrefraated wave rays. 39. Figure 8 illustrates the convergence of the wave rays due to the Earth's sphericity._ sin 48 (72) . Waves which diverge near their source will converge again at a point on the opposite side of the ocean.5° S. (de) .!. Honolulu. Army Engineer District...S. soo~~~~~n~~~~~. the coast of Japan suffered substantial damage and many deaths occurred (U.. 74.. 1961). (PararasCarayannis. tan 1/J Re (70) where ray separation is defined by the equation 2 d S ds2 and the coefficients p p and C + P dS ds + qS =0 (71) q ds are defined as R e a.. Chao (1970) gives the equation for wave refraation in spherical coordinates as dx = _ 1_ (de) _ cos C dw ds a. The coast of Japan lies between 30° and 45° N. An example of this was the 1960 tsunami whose source was on the Chilean coastline. and about 135° to 140° E. 1960. tan 1/J = . Convergence of wave rays.~~~~C~A~~~~~A~ 45° 30° 15° oo 15° 30° 45° p 0 Figure 8.. Hirono. 1969).5° W. a difference of 145° to 150° longitude from the source area.!.tsunami on a distant shoreline.
( + sin 2a ) (1 a2 R2 cos 1)J a1)Ja<t> e c c) (~)2 R e (! ~) c a1)J2 (2 cos a tan R e lji) (! dw dC) c (73) (74) 1 dC C dw 1 Re [ .a)(l cos 1)J c a<t> ac) + (cos a) (c ClC )~ 1 a1)J (75) where R e <1> the radius of the Earth longitude of the point on the surface the wave ray is passing through latitude of the point celerity of the wave a measure of distance in the direction the wave is traveling the angle between the wave ray and a line of equal latitude a ray separation term 1jJ C s a S Using the spherical coordinate system shawn in Figure 9...!1__ f u (J R sin 8 a<t> e (77) 49 .(sin . Hwang and Divoky (1975) give the linear longwave equations as au at and av at == + g an R ae e f (J v (76) g 2..q sin a ) ( R cos 1)J e + 2 (1 a2 ~ c c) ..
the acceleration component in the radial direction is considered to be negligible. For a wave propagating ~ay from the coastline. For instantaneous bottom displacement. in relatively shallow water. it is assumed that the 50 . the initial wave is assumed to be a water surface displacement equivalent to the bottom displacement. and aç/at =O. the parameter U may initially be quite large. For linear longwave equations. If a wave is generated near a coastline.where Re e ~ the radius of the Earth degrees latitude (measured from the pole) degrees longitude (measured eastward) the velocity in the adirection the velocity in the time the coriolis parameter ~direction u v t fe Q =2 Q cos e the rotational speed of the Earth in radians per second Figure 9. The continuity equation is È!l at = 1 L [(d R sin e { ae e + n) u sin e] + L [(d ap + n) vl} + ~ at (78) where ç is a timedependent vertical bottom displacement in the generating area and equal to zero elsewhere.. i.e. Spherical coordinate system.
v in the ydirection. the velocity u in the xdirection. An expansion is used similar to that used by Keller (1948). and the dimension Y* d 0 z ::: z* d 0 t u ::: t* (!oY/ 0 2 ' p P* d pg 0 v* (gd ) l/2 ' w 0 u* (gd ) l/2 ' v w* (gd ) l/2 0 where d0 is a length representative of water depth. and that the resulting errors are of a size that can be accepted in the calculations. Nearshore Propagation.linear longwave equations can be applied to the initial propagation. shoaling waves) which correspond to the Boussinesq equations for solitary waves in water of constant depth. The linear longwave equations may be used for the propagation of waves from a shoreline. Peregrine (1967) derived equations for threedimensional long waves in water of varying depth (i.+ w .+ 1 ::: 0 az at ax ay az ~ the =0 (81) (82) 51 . and component of Q in the ydirection. across an ocean basin. and up to an area near another shoreline. Euler' s equations of mo ti on are aq aq aq aq ap . It is also necessary to consider the propagation of a tsunami toward a shoreline from a nearby generating area. Defining q where q = (u 2 + v 2 ) 1/ Q ClQX 2 and Q =in qdz d (79) is velocity and the flow rate.+ .e. 4.x is the component of Q in the xdirection..+ lE. or into the nearshore area at a distant shoreline where the linearized longwave equations will not provide solutions with sufficient accuracy. and the other variables are defined as before.+ + ~::: 0 ay at Cl x where Q.+ u . p the pressure.+ w . The dimensional variables are defined with an less variables by the following equations: x ::: x* ' y d 0 *.+ v .+ u ..+ v . w in the zdirection. the continui ty equation is ClQy (80) .+ ap at ax ay az ax ay and aw aw aw aw .
The zero where a is the ratio of water depth to wavelength arder solution gives p 0 z The firstorder solution gives aq 1 an 1 ++ at 1 ax 1 a(du1 ) a ( dv 1 ) an 1 + .. tl) z . . and (83) Q are expanded in the form (84) (85) where E is the ratio of wave amplitude to depth w is expanded as W H/d.. z 2 2 ~ ~ + ay 1 a [au 1 (90) 52 .+ . q.) d/L.At the boundary z d....+ ax 1 ax 1 ay 1 J ëlv 1 ) z _a_ [a(du1 ) + a(dv 1 ) ay 1 ax 1 ay 1 J_.. The variable ( 86) = (J ( WO + EW l + E 2 Wz + ..!. p. .= at 1 ax 1 ayl wl 1 ax 1 él(du ) d(dvl) z 0 (87) (88) cul ax 1 + dv1) ay 1 (89) ayl The secondorder solution gives ~2 a [a(du 1 ) a(dv 1 ) (xl. u~+v~+w=O ax ay The variables n...
Chan. and Mahony (1972) show that the KortewegdeVries equations with mixed derivatives. momentum equation is an 2 aq 1 aq 1 an 2 +u +v++ 1 ay at 1 ax ax 1 1 1 and the continuity equation is . such as equation (94). where equation (63) has a thirdorder derivative with respect to x only.+ The (91) an 2 at 1 a(Q 2) x ax 1 + a(Q 2) y ay 1 = 0 (92) Peregrine (1967) points out that secondorder terms will have firstorder effects where t 1 . and Fromm (1970) expanded on the work of Peregrine. with respect to x and t. Bona. Street. Mei and Le Mehaute (1966) derived a solution for waves propagating in one direction which gives the equations as i:l + u i:l + (d + n) au at ax ax and (93) (94) where (95) and B (96) It can be seen that equation (94) has mixed derivatives. Benjamin. is not of sma11 value. d2 ~ + d ~ a u + 1 a2 d au at ax ax 3 atax2 ax atax 2 ax2 at and ( 97 ) ~~ + ~x [ ( d + n) u] 0 (98) 53 . are the preferred form for describing the behavior of long waves. t 1 ) is an arbitrary function arising from integration. + . He accounts for these effects by incorporating secondorder terms into the firstorder variables.where n2 (x 1 .!.. and for waves propagating in one direction give 2 au + u au + i:J.
4 . The momentum equations for the tidal madel are 54 . along the two distinct along the coinciding characteristics x characte"ristics (99) and Cd ( 6 + ud) dn + ~ d2 ~ dw 2 dS 12 ax dS + + (1 _ c2 .e. Butler and Durham (1976) suggest a solution using equations similar to those in a tidal hydraulic madel. Defining dn dt = w and dw dt =a constant. 2 2 d a d) crdc 2 du ax2 6 dS crCd 2 da = ( Cd 0 12 dS 2 + a~ ax 6 ) u u ( 100) where (101) and C = ±Co (x) along S = (~) = 2 S 2 (t + J ~) C (x) 0 = constant (102) The KortewegdeVries equations provide a solution for wave propagation in one direction only. Alternative methods of obtaining solutions for refracted waves in two dimensions are to use the linear longwave equations with additional terms added to account for nonlinear effects. Using the work of Mei and Le Mehaute (1966). Madsen and Mei (1969) developed characteristic equations which could be solved numerically.The numerical solution of these equations gives results comp arable to experiments for varying bottom topography where H/d s 0 . The solutions generated could be used to provide shoaling coefficients to obtain refracted wave heights. i.!. or to use solutions based on the Boussinesq equations.. for an unrefracted wave.
the computations use alternate sets of equations at alternate time steps ~t as shown below. etc.2 t. as defined in Figure 10. S. at times t + ~t/2.f v + . Divoky.S g (an) l~t 2 Cly (llO) n 55 . t + 3~t/2. Solutions of the equations for long water waves are obtained by numerical means. 5).+ v . .k + 1/2). the calculations are. . then v and n implicitly and u explicitly at time t + ~t. n at (j.+ u .TEx au au au .!.dX ()t ()t ()t c d + n =0 (103) (104) av av ~+ u .+ v+ g ()t Cl x Cl y and the continuity equation is h + f u + ___!!jj_ = 0 d + n Cl y c TB h + L (du + nu) + L (dv + nv) Clt Clx Cly where the bottom stress is given by TBx 0 (lOS) = gu /u2 + v2 c2 c2 gv /u2 + v2 (106) and TBy ( 107) Chen.. and v at (j. First u and n are calculated implicitly and v explicitly at time t + ~t/2. Leendertse (1967) gives the following method for solving the linearized longwave equations by using a space staggered scheme as shown in Figure 10. t + S~t/2. Calculating u at point (j + 1/2. and the subscript n + l to indicate the value at time t + ~t. Computer Models. (108) nn  2 1 ~ t { Cl ~s y ax [ (d x + n ) * u]n+l/2 + x ay [ (d + ii ) v]n y } (109) v n .~tfc un+l/2 . IV. . Taking the subscript n to indicate the value at time t. and Hwang (1975) give numerical equations for the twodimensional case based on the Boussinesq equations (see Sec. the subscript n + l/2 to indicate the value at time t + ~t/2.k).+ an . then u and n implicitly and v explicitly at time t + 3~t/2. k).
rectangular coordinates.1o.16.!....1 1 1 ""!' 1 1. ~t f c 2 u n+l/2 2 1 ~t ~s g (an) ay (111) n+l un+l/2 + 2 . bottom effects..!.1Q. vn+l/2 . 1 1 J1 x j+l 1 r + Woter Level ( 1J} U Velocity ( u} o Depth ( d} 1 V Velocity (v} Coordinate system. .. t + 2~t. The various terms used are defined as fo11ows: n x (114) (115) (116) (117) (118) 56 . and any forcing functions.~r1 1 y L Figure 10.1l l 'T1 I. ~t f c vn+l   2 1 ~t ~s g (an)n+l/2 ax (113) These equations omit convective inertia terms. and at times t + ~t. t + 3~t. .
. with terms such as an/ae. They use a different approach for the computation of the terms noted above which allows more direct computation. an/a~. k+l/2 + vj+l . t + 2ôt. g ôt e u n ~Mas (an) n+l/2 +Iôtfv a n 1 (120) n n . Similar equations are shawn by Houston. These are described by Hwang and Divoky (1975) show the same equations in spherical coordinates with u the velocity in the adirection.2 Resin ôt { 1 a [ (d) ~ u sin e] e Mas + n n+l/2 + 6 ~~[(d+n) 1 a e v]n } (121) v n u. . v _ g ôt n+l/2 2 R sin a· ô~ ( an) a~ _!. and v the velocity in the ~direction as defined in Figure 9. n. at times t + ôt/2._ e n+ 1  2 ôt f a un+l/2  (122) 57 . (1975b). t + 3ôt. and v computed as before. t + Sôt/2. k+l/2) {119) The additional terms and require special computational procedures for Leendertse (1967).k1/ 2 + vj . the equations given by Hwang and Divoky (with coriolis terms added) are. . t + 3ôt/2. et al. For the coordinate system in Figure 11.k1 / 2 + vj +l.= v= 4 1 (vj. At times t + ôt.
.t .X . boundary conditions for the computational area must be established.1 j1 . Terms like (~)8 are computed by averaging in the 8direction.. Hwang and Divoky (1975) use solid boundaries at coastlines and fictitious open boundaries at edges of the computational area where it is necessary to truncate the region of computation. and convective inertia terms are ignored. In addition to specifying equations of motion..X 1 l x1l j ~XtX· 1 8 j+l !11x+x k1 Water Depth (dl Water Level {1)) u Velocity in the 11 Velocity in the 9direction 1~ 0 x ~direction Figure 11.n2 nl ...2 ReM as n+l/2 + 2 ~t fe (124) Again.. the wave is assumed to travel without change in form across the final space step. { l_ nn+l/2 1 d  2 Resin 8 8 lit ~8 a8 ) ~ ( (Cf+n) <P u sin 8]n+l/ 2 (123) vn+l + ~<P ~ [(d + n) u v]n+l n+l g M (an) 1 un+l/2 . At solid boundaries. and (~)<!> by averaging in the cpdirection. complete reflection is assumed...n2 with the terms defined in Figure 12. the bottomfriction. At open boundaries.X . Coordinate system. spherical coordinates. (125) 58 .. so that nB . forcing functions.t i i ..
61 hours) after the time of generation are shawn in Figure 13. At a shoreline. The uplifting and subsidence determined from field surveys (Plafker. at a shoreline) and that equation (125) will describe an open boundary introduce errors into the computations which limit the length of realtime records which can be simulated numerically.8 1 2 (n + 1) ât nât s 1âS BOUNDARY Figure 12.. 1970) was used as the initial deformation of the water surface. Berg.e. 1965. 1975). Predicted wave heights 13. sorne amount of wave energy may be trapped so that complete reflection does not occur. et al. Alaska. 59 . The assumptions that complete reflection occurs at a solid boundary (i.. Wave trapping is discussed later in this report. Houston and Garcia (1974) and Hwang and Divoky (1975) used numerical techniques to obtain predicted wave heights for the 1964 tsunami originating in Prince William Sound. Graphical representation of the total transmission open boundary condition (from Hwang and Divoky.000 seconds (3.
t) l.. . This error will have the appearance of a wave reflected from the open boundary. .n) nr. Nearshore Computer Models. n.x.s 60 . Shaw (1974. and its normal derivatives at the boundary.n s = l. sl'. Buffalo. 1975). Peregrine (1967) developed finitedifference equations for the twodimensional case of a wave shoaling on a beach. Shaw. . those values would be used in the finitedifference solutions for the interior region. 6. 1977) suggests that an outer boundary integral equation method can be used to eliminate the error at the open boundaries.Figure 13.s u(rl'. Department of Engineering Science. personal communication. The use of equation (125) at the open boundaries results in an error in the computed wave heights at those boundaries.x.000 seconds after the 1964 Alaska earthquake (from Hwang and Divoky.. sl'. Defining u r. For waves in the nearshore region. 1975.t) (r n(rl'. State University of New York.. . the error propagates through the grid at successive time steps. Surface elevation contours 13.. As computations are carried out for the entire computational grid at each time step. The outer boundary integral equation would be used to determine the wave height.
using a MarkerandCell numerical technique to obtain solutions for waves propagating in one direction. they compute the values of u and w at time t + t:.8 [nr+1. Where values are known at time t.8+1 .k + gx t:. p is pressure.8+1 4t:.u r.Pj+1. a provisional value of nr 8+ 1 . the continuity equation is used again to give an improved value for nr38+ 1 u +u u u r+1.8 + ur1.8 r.8 r1.8 r1.j+ll2.t using the equations u . ur + is calculated from an approximation to the momentum equation 38 1 u r+1 . and Fromm (1970) extended Peregrine's work.8 .t ( Pj.8 .k. Chan.t (127) Finally.t + a2x ur+1.8+l .u + 2u .x + u r38 [ nr+ 1.he calculated the values of u and n with a timestepping procedure.x + u r+I .nr1 .2u + u .8 b.u r1 .ur1. gx 61 .u* j+l/2. by the equation r3 1 3 u r+l .ur+1. first using an approximation to the continuity equation which gives n* 8+ .u 0 (126) Then.8+1 .8 2t:.8 +a] 2 2t:.8 r1.8 t:.(ax) 2 3 1 u r+1.k. t + t:.8 2t:.8+l r1.xt:.8+1 + ur.8 t:.k ) (129) (130) where the coordinate system is shown in Figure 14.nr1.u r1 .8+1 r1. x + ur.8+1 r+1.x 2 t:.8+1 r+1. and Fromm (1970) and Chan.8 .8.8+1 .x t:. Street.x 0 (128) Street.x + a] = .
where B = !(Pj+l.k112) ~z and R. equation for u~ J+l 2) k is 1 The u* _ S .k1 + 2.k ( u* .k (134) 62 .k j11z.T1 1 1 1 +4~~~+rK=l ' Il X j= 1 j=2 j =3 j=4 j=5 Figure 14. The terms on the right side of equations (129) and (130) are taken at time t.w* j. Position of variables..k ~x + (133) The convective terms are defined as shawn by Fromm (1968).uj+lf2.k+l + Pj. and u* and w* convective terms.k+l!z j. = _ .k2 .!__ ~t J. Then. and gz the components of gravitational acceleration.1 ( ) j+l/2.uj+l/Z.k ) (131) 2(1 _1) ~x2 + ~z2 (132) w* .k1 B ~x 2 ~z 2 + RJ.k + Pj.u* j+112.uj+l/2.k + Pjl. the pressures are recomputed from the equation P. k J.1 1 1 .k .
respectively. the amplitude is (140) 63 . Chan and Street (1970a) give the equation for Pj.k uj1/2. and and vertical velocities. Chen. with the velocities and wave amplitudes calculated explicitly at alternate time steps of tJ.x tJ. A timestaggered scheme is used.k + uj+l/2.t Uj fre~ ~t (139) where nj(t) is the elevation at time t.~t/2 and velocities at ta. and Fromm (1970) at time t + ~t is nj =w. as given by Chan. = uj+l/2 . and Hwang (1975) give the equations below using dimensionless expansions similar to those proposed by Peregrine (1967).k.. The amplitudes at ta + ~t/2 will be calculated using amplitude at ta . kl :1 J (138) The freesurface position.k) (a.k as R. For refracting waves propagating in two dimensions in the plane of the water surface.2ujl/2.k ~t ~z (137) Similar expressions can be developed for other convective terms. 15). At time ta + ~t/2.k . J. Divoky. For points near the free surface (Fig.1) 2 + 2 ~t (uj3/2.where CL  1 uj+l/2.t/2.k + .k ~x 13 (136) = wj+l/2.t.njI) + nj(t) J J 2 D. the velocities at ta + ~t will be calculated using velocities at ta and amplitudes at ta + ~t/2. then. .(uj3/2. (nj+I. at the and Wj the horizontal surface at time t + D.2 .u.k) (135) a.uj+I/2. Street.
k (vj. 64 . u and v can be computed at time ta + ~t using values of u and V at time ta and amplitudes at time ta + ~t/2.  ~ Cn.(j.~vj. k 1 J 2~y J. J. where the j. Where the initial velocity field is known.k (141) (142) At time ta + ~t. and ü and v the velocities satisfying the linear longwave equations. k +1 .v. (144) where the values on the right side of the equations are at previous time steps as indicated. K1) Figure 15. Computation of pressure near the free surface. k 1 ) 2~y · J.n. K+l) (j.k subscripts refer to positions in the plane of the stillwater surface as shown in Figure 10. This gives the equations M uj.k 2~x (nj+I.k+1 . the velocities u and v are given by (143) .
k1 + d.k+1 (148) etc. J+l.Chen.dj+1. k 1 ] j1. J2. then becomes The solution lit 2lly . use a higher order solution for the amplitude when d < tu (20 llt) 1 1 3 where the variables are expressed in dimensionless form. and Hwang (1975). k 2 ] J• (147) . J1.d . k J' +1 J' 1 2(lly) [vj. and Mahony (1972).2v. Divoky.k+1 4llxlly [dj+1.lit (higher order derivative terms) The higher order derivatives are approximated by central difference equations as follows (145) 2ü. Computed surface elevations were smoothed when one of the following conditions was satisfied: (a) A crest or trough has wave amplitude less than 25 percent of the maximum wave amplitude at that instant. k a3:y + 2ü. k] (146) a0= a2d axa y 1 . (c) at a matching point where equations change from linear to higher order equations. Bona. k + 2v. using a stability criterion obtained by Benjamin. k ü. (b) the local velocity component (u) or (v) has a different sign from the average value of the surrounding four points. J1.k+2 v. Smoothing is accomplished by the average (149) 65 .
k 12k + 4 (150) where n. k + nJ.k M at where n (152) at ta . Chen.k . For the open boundary condition previously mentioned (Fig. k . 66 .k+2) (151) and k represents a weighting spline coefficient that varies from 0 to The influence from the surrounding points is controlled by the values of (k).{Cd+. the finitedifference equation becomes nB.k + nj. and (153) is taken at time to ôt/2 and n nJ. To avoid numerical instability. and Hwang (1975) for Boussinesqtype equations.5 _n_) d + n (155) Listings of typical computer programs for solutions of longwave equations can be found in Brandsma. the partial derivative with respect to time was approximated by ~ = nj. and in Chen.nB1. and T). Divoky. nj. .1 + nJ. Divcky.k2 + nj.k+l. k + nJ+l. and Hwang (1975) for linear longwave equations. k J.ôt/2.where the values on the right side are before smoothing.nj. .k(nj2.~ (nB. ( 1 + 4k) ( nJl. k J. Divoky. . k .k) CM ~~Y where [{Cd+ n) v}j.k + nj+2. For the case k = 0.kl] (154) ë = [g(d + n)] l/2 (1 + 0. the equation reduces to Laplacian interpolation.n) v}j.k .k + . and Hwang (1975) imposed the condition at matching points that nmatching = 0 · 5 Cnzinear + nhigher arder) Also. 12). k + 1) . .
. and the steepness of the beach slope may modify the wave period and wave height.e. 6 1 = e2 = O. While sorne amount of the energy in a tsunami might reflect from the ridge. and cause the waves to form bores which surge onto the shoreline. the waves are modified by the various offshore and coastal features. Ocean ridges provide very little protection to a coastline. 1. Waves may become higher and shorter. He considered only the case of a wave at a zero angle of incidence. The 1960 tsunami which originated along the coast of Chile is an example of this... '[. 1961). the major part of the energy will be transmitted across the ridge and into the coastline. TSUNAMIS APPROACHING THE SHORELINE As a tsunami approaches a coastline. headlands. and dispersion may occur. reflect wave energy. Submerged ridges and reefs. (158) or Ht where H.. r Hi Hr the incident wave height the reflected wave height the transmitted wave height the initial water depth the water depth under the transmitted wave Ht d1 d2 67 . '[. various shaped bays.. + H '[. That tsunami had high wave heights along the coast of Japan. i."= '[.= dl/2 '[. 1 + H..V. 1 H r dl/2 1 dl/2 2 + dl/2 2  (156) Ht H. Lamb (1932) gave the equations for a single wave passing over an abrupt change in water depth as shawn in Figure 16. Abrupt Depth Transitions. and Ht = 2di/2 dl/2 + dl/2 2 1 H r (157) H. cause wave resonance. continental shelves. The equations he derived are H. An ocean shelf along a coastline may cause greater modification to a tsunami than an ocean ridge. including Shikoku and Kyushu which lie behind the South Honshu Ridge (Hirono.
Wave passing onto shelf. and Et as the wave energy of the transmitted wave. E H2 L r r r == (159) (160) 68 . and also when a wave passes from shallow water to deep water. The equations predict that substantial reflection will occur when a wave passes from deep water to shallow water.Reflected Wovt Incident Wova PLAN z Seo Surfoct Bottom Bottom PROFILE Figure 16. and that a single incident wave splits into a single reflected wave and a single transmitted wave. It is assumed that no energy loss occurs. Taking Ei as the wave energy of the incident wave. Er as the wave energy of the reflected wave.
(161) which reduces to (162) Cochrane and Arthur (1948) extended Lamb's work to consider waves approaching a shelf at varying angles of incidence. z. respectively. wave dispersion on the shelf is not considered.and from equations (159) and (160) E 1' + E. For a given incident wave angle e 1 . applies to shallowwater waves. This equation also applies to a single wave with a reflected component and a transmitted component. The ratio of transmitted wave height is given by Ht to the incident wave height Ht = ___z_Œ_d_ 1 1_c_o_s_e__ __ or. The water depths d 1 and d2. (165) as before. 69 . determined using Snell's Law so that the value of e2 can be (164) Equation (163). are defined in Figure 16.. as written. alternatively. and the angles 8 1 and e 2 . They give the ratio of reflected wave height to incident wave height as Hr = ~ cos 81 ~ cos v"(Ç" cos e 2 Hi e1 + ~ cos (163) e2 for an abrupt change in water depth. The solutions to equations (163) and (164) are presented graphically in Figures 17 and 18.
6 0.4 0. Transmitted wave angle 70 82 versus incident wave angle 81 • .8 1 2 3 4 6 8 10 Figure 17. 0.4 0.0.3 0.2 0.6 0.2 0.8 2 3 4 6 8 10 20 Figure 18.1 0. Wave reflection from a shelf Cafter Cochrane and Arthur.3 0. 1948).
1 (0.* * * * * * * * * * * * * * EXAMPLE PROBLEM 4 * * * * * * * * * * * * * * GIVEN: or30 feet) at an FIND: (a) (b) (c) The angle at which the transmitted wave propagates onto the shelf.200 deep. 71 . from conservation of energy.626 meters (5. and passes onto a shelf where the water depth is 100 meters. + Hr 1.TI 81 82 e2 . z. and the height of the transmitted wave.v<Ç + cos e2 (Hi) 1' v'<Ç I<Ç cos e2 ( 1) H r h.500 meters (8. 500 h.33 feet) 1 + 0. 74° 0. angle of incidence e1 = 30° SOLUTION: (a) From equation (164) sin e2 (:~f2 sin Sl.lïOO cos 5. the height of the reflected wave.1) = 5.626 * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * When the initial angle of incidence e1 > 0.soo cos 30° .05 feet) (c) From equation (158) H. An incident wave with a height of 1 meter (3. 1( )l/2 [ 100 2.28 feet) and a period minutes approaches a coastline through water 2. 74 ° cos 30° + 1ïëi0 cos s.74° (b) From equation (163) H icÇ cos el cos 81 . As the energy equations are written for a unit length of wave crest.626 meter (2. the distance between adjacent wave rays is different for the incident and transmitted waves.500 sin 30"] sin.
cos E. 1 (168) 1 (169) Not ing that Lt/Li = Ct/Ci = v'd 2 /d 1 and that Lr/Li = Cr/Ci = v'd 1 /d 1 and substituting equations (163) and (165) into equation (169). lcÇ'" cos a y 2 = co s e2 1 (171) proving that the equations of Cochrane and Arthur conserve the energy of the incident wave. Japan.E. with an actual recorded wave height at Hanasaki. The observed 72 . it was determined that the reflected wave arriving at Hanasaki would have a height of 17 centimeters (0. 1 e2 (167) E t cos e + E r 1 where bt is the distance between adjacent wave rays for the transmitted wave . Using a rough approximation for the wave height at the top of the continental slope. which reflected from the continental slope off southern Oregon. + which gives e2 Ei cos e1 Et cos E r E. = 1. 1 (166) or.lcÇ ~cos el + lcÇ'" (~cos a1 ~cos e1 + + cos cos 2 02 ) 1 (170) e2 which redu ces ta ra. + (~cos el . alternatively. and bi the distance between adjacent wave rays for the incident wave. Rewriting equation (167). Cochrane and Arthur (1948) compared a calcu1ated value for a wave from the 1946 tsunami.56 foot).
Defining a parameter. * * * * * * * * * * * * * * EXAMPLE PROBLEM 5 * * * * * * * * * * * * * * GIVEN: An incident wave. as well as the difference in water depth.925 feet) deep.5 meter (1. when d 1 /d 2 < 1.0 the value of the refle ction coefficient Kr is negative. were reflected waves from the continental slopes of Japan and the Bonin Islands. Dean (1964) considered the case of a wave normally incident on a linear change in water depth shown in Figure 19. Cochrane and Arthur note that reflected waves are normally of secondary . 2. recedes from the coastline through water 100 meters deep and passes from the shal low water over a shelf into water 3. convergence may cause reflected waves to be of primary magnitude. For these same values of d 1 /d 2 . At given stations. (174) 1 Dean found the results shown in Figure 20. but not negligible. MacDonald. for the zero angle of incidence assumed by Dean. The transition between the two water 73 . Alaska. as (172) the transmission coefficient (173) and the reflection coefficient H K r r H. As before.0 the value of Kr is positive. Shepard.wave height. for an arrivai time equal to the calculated time fo r the reflected wave. on the island of Hawaii. zl.025 meters (9. should be considered in the calculations. the predicted transmitted wave height would equal the incident wave height. When d 1 /d 2 > 1. Cochrane and Arthur (1948) indicate that the length of a continental slope. from the 1 April 1946 tsunami originating in the Aleutian Islands. but this occurs only in relatively few cases. and Cox (1950) note that the highest and most damaging waves at Napoopoo and Hokeena. magnitude according to theory. Figure 17 shows that for waves arriving with a higher angle of incidence there will be sorne value of d 1 /d 2 for which equation (163) predicts no reflected wave. Linear Depth Transitions.64 feet) high and has a peT"iod of 40 minutes. was in good agreement with the calculated wave height. which is 0.
.5 ~~:t::r::. .:.F·<1. ~1.01 0...=:... .::... Ref1ection and transmission coefficients (from Dean.• 0 f=..05 0.= ~~=:.. 74 .' .2 0.02 0. ~~·ft_.:::...... Il/ l' i/ li J j.. 1964). / l/.1 0._...Reflected Wove Tronsmitted Wove Incident Wove dz Figure 19. r: ~. 10 20 50 100 200 " 500 1000 Figure 20.. Linear s1ope and she1f._[:~ 0.1 1/ //.
5) = 0.62 (the negative sign indicates that the ref1ected wave is TI radians out of phase with incident wave) 0.e.62(0.52 foot) (c) y= pg = (1. and (c) show that energy is conserved. el= o.025 = 0.8 x 100 (40 60) 75.02 feet) (b) From Figure 20.second squared 75 .055 ki1ograms per square meter . SOLUTION: /g"ëÇ L1 T x = /9.. 167 it is found th at K r 0.100 100 x 0.31 meter (1.167 From Figure 20. where ~= d2 _!QQ_ 3.16 meter (0. FIND: (a) (b) The height of the reflected wave. that the total energy in the reflected and transmitted waves equa1s the incident wave energy.100 meters From equation (172)..depths is a linear slope S = 0 .e.033 and z1 0 .5) = 0.62 Hi H r = 0.8 meters per second squared) y = 10. the height of the transmitted wave. 47T x z1 = 75.32 Hi = 0. Ht = 0.32(0. i. i. 1 and the wave is at a zero angle of incidence with the slope transition.1 0.026 ki1ograms per cubic meter) (9.
L2 is the wavelength at depth d 2 . y H~ Li . Nonlinear Depth Transitions. f7T(L~ L~ )~ ~ sinh 1 ["(~ 2 + ~~)J u 76 .The energy per meter length of wave crest is E.36 x 10 7 joul~s y H2 L 8 Er Lt E = 1' 1' 10.055(0. The difference results from the minor errors which occur using Figure 20.8 x 3.e.9. Lt 8 9.055(0.36 x 10 2.36 x 10 7 . the reflected wave is out of phase.100 ~ ~ 8 7 kilogrammeters per second squared 2.5) 2 75.07 x 106 = 1. i.33 x 10 7 joules.025 (40 x 60) 10.33 10 7 joules As Kr is negative.055(0. ~ Er E 1' Et 2. and n is an arbitrary small number 1n equation (175) which fits the equation to the actual slope and determines the length of the slope in equation (176).16) 2 413. Kajiura (1963) investigated waves passing from deep water to shallow water over the nonlinear slope profile shawn in Figure 21.07 x 106 joules 413..100 8 8 c2 T y H. then for energy to be conserved Ei E.31) 2 75. 3.000 8 t 1.10. The profile is defined by the equation (175) where the effective slope length ~ ~ is given by n = 27T (176) L1 is the wavelength at depth d 1 . The reflection coefficient obtained by Kajiura is given by the equation sinh Hl' = L\ 2 l (177) H.000 meters x /g(Ç T 19.45 x 107 joules compared to the computed value of Et= 1.
3 0. the total of the reflected and trans mi tted wave energy equals the incident wave energy. .1 0. Ref1ection coefficients (from Kajiura. the reflection coefficient approaches zero as the slope length ~ approaches the length of the incident wave L1 . 7 0.000 meters (14. It is assumed that the slope is defined by equation (175) and that energy is conserved. o. J. 7_7 . The solution of equation (177) is plotted in Figure 22.4 0.8 0.500 meters deep and passes onto a shelf where the water depth is 100 meters. at a zero angle of incidence (8 1 = 0).2 0. The effective length of the slope between the two water depths is ~ = 24. 1963). i.5 0.1 0. S1ope and shelf. Figure 22.9 miles). As shawn in the figure.8 * * * * * * * * * * * * * * EXAMPLE PROBLEM 6 * * * * * * * * * * * * * *' GIVEN: An incident wave which is 0.Figure 21.97 feet) high and has a përiod of 20 minutes. 0. approaches the coastline through water 2.3 0.e.6 meter (1.6 L.
46(0.3 miles) 78 .6) 2 187.28) 2 187.6) = 0. 1. 1" = 6. 1" 0.FIND: (a) (b) The height of the reflected wave .rg. .ç T = 0.E 1.85 x 10 7 joules E. Ll 1.800 8 10.28 meter (0. SOLUTION: C1 T = . y H~ E.46 Hl" (b) = 0.46 Hi = 0.800 8 8.055(0. and (c) the height of the reflected wave for a linear slope of the same length.600 meters (23.800 meters (116. the height of the transmitted wave. = 187.91 foot) From conservation of energy.65 x 10 7 joules Lt = 37.7 miles) 24.128 19. 8 y H2 10.000 r. 187.50 x 10 7 joules E1" = 1" Ll 8 1.8(2.055(0.800 From Figure 22 H H:"= 1.500) (20 x 60) L1 Q.
Gagnon and Bocco (1962) obtained measurements for waves passing from shallow water to deep water at an abrupt transition in depth.08 feet) which indicates that a linear slope gives a higher reflected wave and 1ower transmitted wave than a slope defined by equation (176). their results indicated a fairly constant value of IKrl ~ 0. 1.67 From Figure 20. where no transition occurs).e. 41fdl L1 S 0.55 H.33 meter (1.89 feet) (c) The slope is defined by (2. Their experimental curves for Kt and Kr as functions of d 1 /d 2 are given in Figure 23.lOO) 24..6) = 0.600 )1/ 2 Ht 1. 79 . H 0.55 H.000 s From equation (172).65 x 10 7 ( 10.1) 1. However. 65 x 10 7 joules Ht 8 x 6. = 0.800(0. including the case whçre d 1 = d2 (i. where d/d2 = 25 2 : 0.SOO) 187.055 x 37.2. * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * 4.500 . H r 1. Bourodimos and Ippen (1968) obtained experimental results for waves passing from deep water to shallow water over a slope where S = 0.Et = y H~ Lt 8 6. Experimental Measurements.1 = 41f(2.55(0.125.19 meters (3.
So 1itons and ShoalingInduced Dispersion. Madsen and ei M (1969).80 0.6 Kr 0. the equations wi l l be hyperbolic and the waves will be stable. Byrne (1969).1. Street. Chan.60 0 0. For certain conditions.8 0. This train of waves will consist of an initial wave having the highest amplitude. Wave decomposition has been investigated by Mason and Keu1egan (1944). where k is the wave numb er 2n/L. and hyp erbolic if kd < 1. For tsunamis. and Fromm (1970). and ~bitford (1968). where d/L << 1. Street. <.9). and indicate t hat t he waves will only be unstable if kd > 1. Benj amin and Feir (1967) discuss the stability of waves. Ben j amin and Feir (1967).363. Zabusky and Galvin (1971).4 0.2 0. Burgess. Galvin (1970).45 0. S.90 Figure 23. Ref1ection and transmission coefficients (modified from Bourodimos and Ippen.0 v + 0.363.2 1. Exampl es of this are shown in Figures 24 to 28. Whitham (1967) showed that equations governing extremely graduai variations in wave properties are elliptic if kd > 1.70 0. Horikawa and Wiege! (195. a wave will decompose into a train of waves. followed by a fini t e number of waves of decreasing amplitude.4 Kt 1. 80 .363. 1968). and Hammack and Segur (1974). at !east in a constant water depth.50 (:~ r 0.
c . 1969). 81 .2 c 3 feet from Gene rotor 0 êi CD > ~ 6 5 4 0 T. Separation of so1itons (from Ga1vin. ~ 1 / ~ Figure 25..25 0.ime (s) 0. 2 ia i / Induced wave generation over a submerged bar (from Byrne.2 39 feet from Generator 0 êi CD > ~ 0 Time (s) Figure 24. 1970).125::.
Slope 5=0." 10 "7 20 0.2 (c) 0.05 d.1 ~ "7 Figure 26..(o) 0. 82 .1 "7 ~ 0 20 30 . dl 40 0.1 d1+f+. Solitary wave propagating over a slope onto a shelf (from Madsen and Mei.2 0.!. 1969).
__I t•: t(g/dl) 1/2 "'10...5 x 15 x 60 154.J.15 t•:t(g/dl )1/2 d.50 0... 70 75 . 1970). and Fromm...o. 0.. 0.840 feet) deep.______.. r•=45 T*=90 '1 0.75 1. The wave ce1erity is given by x 3.9 miles) 83 .. Figure 28.___.50 0.==r===0=:::::=="=r==~45.~T*:~~90:.02 · :~J. and Fromm...5 meters (563 feet) per second The wavelength is L = CT = 171...350 meters (95. Chan.!.~ did·= 0. Chan. 1970).00 LL"'. SOLUTION: C = lgëi = v'9..25 1 '0. x * * * * * * * * * * * * * * EXAMPLE PROBLEM 7 * * * * * * * * * * * * * * GIVEN: FIND: A tsunami with a wave period of 15 minutes travels through water If the wave is stable.25 So1itary wave propagating onto a she1f (from Street.75 eosine Slope d2/dl =0...'"""''.25 d.... Figure 27.15 eosine Slope 0 5 10 15 20 25 .__.807 = 171. Wave train propagating onto a she1f (from Street._ 30 35 ~0 45 50 55 60 65 _.000 ~00 meters (9. d..
where U is defined as (H/d)(L/d) 2 . He noted that a wave passing over a bar would sornetirnes produce a second. for cases where 22 ~ U ~ 777. from uplifting of the sea bottorn). If a generated tsunami had the characteristics of a group of solitons. They found that soliton generation is dependent on the net volume change in the body of water.350 and kd = Therefore. but rather a combination of ~everal solitons. depending on the distance from the generating area.2 seconds. As these additional waves developed near the shoreline.15 which would indicate that the waves are stable.15 rneter (0. using the KortewegdeVries equations. * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Galvin (1970) investigated waves propagating through water of uniforrn depth in a laboratory wave tank. separate again into solitons. Taking these waves as shallowwater waves. They found good cornparisons for slightly dissipative waves. If the initial generating rnechanisrn was negative everywhere (sea bottorn subsidence). etc. kd < 0. but as many as seven could exist in sorne instances. no solitons evolved. He found that the initial generated waves broke clown into several waves which are called solitons.g. each of the initial waves broke clown into five solitons. it may be assumed that the generated waves were not actua~ty single waves.000 0. it could appear differently at various coastal points. However. trailing wave as shown schernatically in Figure 25.0194 = 0. Figure 24 illustrates an example where. When the net volume of the initial wave system was positive (e.122 1. Zabusky and Galvin (1971) cornpared nurnerical and experimental results for solitons. the solitons would recombine into single waves.. 84 .From this I = 154.5 foot) and a generator period of 5. and kd ~ 0. There are cornrnonly two or thre~ solitons.0194(2~) d 3. Galvin noted that if a group of such waves traveled over a sufficiently long distance. solitons evolved followed by a dispersive train of oscillatory waves. Byrne (1969) made field observations of waves passing over a nearshore bar. for a water depth of 0.363 and the wave is stable.4 rneters. he was unable to determine if such waves would recombine with the waves in the initial wave train. Harnrnack and Segur (1974) also studied numerical and experimental results. the wavelength is approximately 6.
807 1 x 3.5 meters per second 85 . SOLUTION: The wave celerity in deep water is c1 = L rgcç = 19.000meter water depth into a 200meter water depth. FIND: The maximum wave amplitude for a stable wave which will not decompûse into a train of waves.807 x 3. The wave period is 60 minutes.Masan and Keulegan (1944) investigated waves passing into a shallower water depth.000) l/2 = 171. * * * * * * * * * * * * * * EXAMPLE PROBLEM 9 * * * * * * * * * * * * * * GIVEN: A tsunami travels from a 3. Thcir results were confirmed by Horika~a and Wiegel (1959).9 miles) The condition for wave instability is given by equation (178) as (alLI) 1/2 (a 1 x 154.350) 11 2 > 2d2 > 2 x 200 a 1 > 1. * * * * * * * * * * * * * * EXAMPLE PROBLEM 8 * * * * * * * * * * * * * * GIVEN: A tsunami with a period of 15 minutes passes from water 3. The condition for instability obtained from their experiments was (178) where a 1 is the wave amplitude in the deeper water.000 = 171.40 feet) Thus. with an abrupt change in depth.2. the right side of equation (178) has been multiplied by 1.5 meters per second and the wavelength is 1 ~ C T = 171.5 x 15 x 60 = 154. SOLUTION: C1 The deepwater wave celerity is given as = (gd 1) 112 = (9.04 meters (3. L1 the wavclength in the deeper water. FIND: The maximum wave amplitude for a stable wave which will not aëcompose into a train of waves.350 meters (95. although in the latter report thcrc is an apparent discrepancy in the presentation of the results.000 niëters deep onto a shelf \vhcrc the water depth is 200 meters (656 feet). and d 2 the depth in the shallmver \vater.04 meters would not decompose. waves with a deepwater amplitude less than 1.
and the wave height in the shallower water depth as Ht. Burgess.5 x 60 x 60 = 617. and the results shown in Figure 28 for a train of waves. and Whitford (1968) investigated solitary waves passing from an initial water depth. Chan. the shallower depth as d2. Defining the initial water depth as d1. relative wave height Ht/Hi reached a maximum value for any initial wave height ~ and then decreased. They obtained results similar to those of other investigators. and into a shallower water depth. over a steep slope. His results are similar to those of Street. showing that each wave changed from a single wave into a train of several waves.85 foot) Waves with a deepwater amplitude less than 0.400 meters (384 miles) From equation (178). the first wave crest of the series was the highest. and Fromm (1970) and Madsen and Mei (1969). The numerical results of Street. Goring (1978) has also recently carried out experiments on solitary waves propagating onto a shelf. Street. Ht/~.400) 1/ 2 a1 > > 2 x 200 0. as the ratio d1/d2 increased. 86 . and Fromm (1970) for wave trains is inconclusive in this regard as Figure 28 shows the additional wave crests. The locus of the maximum values of wave enhancement. * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Using the results of Mason and Keulegan (1944). each having a high initial wave followed by smaller waves. In sorne instances. The numerical work of Street. Madsen and Mei's (1969) numerical results for the propagation of long waves give the results shown in Figure 26 for a solitary wave passing over a slope and onto a shelf. but does not separate the waves into groups associated with the initial crests. Chan. wave instability is given by (alLl)l/2 > 2d2 (a 1 x 617. the above examples illustrate that the longer period tsunamis are much more likely to decompose where the waves have the same height in the deep ocean. the initial wave height as Hi. the maximum value of Htl~ became greater and occurred at a higher value of d 1/d 2 .26 meter (0. It may be presumed that a number of initial wave crests will produce the same number of groups of wave crests. Chan. and Fromm (1970) give the results shown in Figure 27 for a solitary wave. As ~/d 1 decreased. In all cases where a single wave produced a series of wave crests. are shown in Figure 29 with the results for the solitary wave experiments. there was also a significant increase in wave height.and the wavelength is L1 = c1T = 171.26 meter will not decompose.
_  .... . and Whitford....3 ~Locus of Maximum Wave Enhancement . ...... ....... 87 ....050 Figure 29......... ..025  . Wave enhancement (from Street.. .........0.. 1968)... '.  ...0.......... .......... ' ' ....... Burgess. ...
Line Normal ta Tangent Wave~ Figure 30. 1. a tsunami may interact with a shoreline in a number of different ways. All of the above interactions depend on wave reflection at the shoreline. TSUNAMISHORELINE INTERACTION In addition to the shoaling of waves on the nearshore slope. and. 88 . the trapping of reflected incident waves by refraction. The angle. the generation of edge waves by the impulse of the incident waves. the reflection of wave energy from an abrupt change in water'depth at the seaward edge of a shelf. including standing wave resonance at the shoreline.VI. Wave Reflection. the reflected wave will be in phase with the incident wave. Wave reflection from a shoreline. a 1 . The reflection of an incident wave ray from a shoreline is illustrated in Figure 30. as the reflected wave from the shoreline propagates seaward. Also. For a steep nearshore slope. LeBlond and Mysak (1977) provide a general discussion of edge waves and wave trapping. a wave arriving at an oblique angle to the shoreline may produce a Machstem along the shoreline. Tsunamis entering inlets and harbors may also produce resonant conditions within the inlets and harbors. between the wave ray and aline normal to a tangent to the shoreline will have the same value for the incident and the reflected wave rays.
997)1/ 2 sin2(~.sy/ 2 si~ 2 s = (2 x ~. the wave is completely refle cted at the shoreline. * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Wiege! (1964) indicates that where Ï > (Ï)c 89 (181) .5 meter and a period of 20 minutes 1n a 1. 8oo L I = H 4. (H/L)c. which is given by (179) where S is the angle of the beach slope in radians.800 meters 0.0997 radians).0997) ~ < (~) c thus .1 (S = 0.(!!) L L < (180) c * * * * * * * * * * * * * * EXAMPLE PROBLEM 10 * * * * * * * * * * * * * * GIVEN: A tsunami has a height of 0.5 ll8.6 From equation (179) (~) c (.Miche (1944) defined the wave reflection at a shoreline in terms of a critical wave steepness. the wave celerity. C C. In the deeper water. Complete reflection will occur if the wave steepness. The nearshore slope s 3 = 0. H/L.000 = 99 meters per second CT= 99 x 20 x 60 = 118. FIND: If the wave is completely reflected at the shoreline.21 x 10. in deeper water is given by !! . is SOLUTION: lgd = 19.807 x 1.000meter water depth .
5 Ï = 2 ~:~oo = 2.6 5 2.o.54 x 106 !L. the wave celerity. * * * * * * * * * * * * * * * * * * ~ * * * * * * * * * * * * * * * * * * 90 . > (!:!.10 From equation (179).000meter water depth. (~) = (.that the reflection will be defined by (182) where cR is a coefficient of roughness and permeability which has a value of cR= 0. The nearshore slope s 3 = 0.)c :! L 0.5 meter and a period of 4 minutes 1n a 1.OI)l/2 sin2.000 = 99 meters per second CT= 99 x 4 x 60 = 23.097 which indicates a lowreflected wave height where the shoreline has a very graduai slope. * * * * * * * * * * ·~< * * EXAMPLE PROBLEM Il * * * * * * * * * * * * * * GIVEN: A tsunami has a height of 0.800 meters (14.OI) (![) c = 2. However.8 for a smooth impervious beach.807 x 1. where c S is given in radians. is SOLUTION: In the deeper water.8 miles) x Io.01 (S = 0. and the slope is smooth and impervious.01 radians).10 x 10· 0.8 2. C.sy/2 si~2s = (2 x'ITO. FIND: The coefficient of reflection ~/Hi at the shoreline. the effect of the slope roughness on longer period waves has not been adequately determined. C L !gd= 19. Various values of cR were defined for rough slopes for shortperiod waves.54 x 10.
s2. and the sea bottom sloped seaward. s 2 . 1964). and a is the distance from the shoreline to the depth d = 1:2 d 8 • The depth variation defined by equation (183) can be compared to a linear (constant) bottom slope. from the shoreline 12 d s a . For the linear bottom slope. Hidaka obtained a theoretical solution for wave resonance on the sloping shelf defined by equation (183) using Mathieu functions (see Blanch. d .2. The depth d at any arbitrary distance x from the shoreline is given by x2)l/2 a2 d d 1 s ( + (183) where the horizontal distance.d s (185) from which a ( /2 = 1) d s ( 186) The variables are shawn in Figure 31. a. between the toe qf the nearshore slope (taken to be a vertical wall) and a point at the distance x = a from the. and t time. x. shoreline. where the water depth at the wall was d8 . is positive measured seaward from the shoreline. The primary mode of oscillation 91 .7) Hidaka defined the surface elevation n n above the undisturbed water as = u cos 27rt) (T (188) and U a dimensionless amplitude obtained by dividing the amplitude at any point by the amplitude at the shoreline (U = 1 at the shoreline). T the wave period. Shelf Resonance . Defining the wave by the equation ( 18.d s (184) x or at a distance. x = 0 at the shoreline. Hidaka (1935a. 1935b) carried out a theoretical investigation of a vertical wall at the shoreline.
Ce shelf is defined by the Mathieu function 2a el = 7. 92 . ~. The values obtained by Hidaka for resonant periods are for a shelf extending a long distance offshore.: Ill ~]  ~~~~~~~~~r~ 0 ~~ . as shown in example problem 14.51361.29863.lll "'E ~~>. but Hidaka's results of the variation of wave amplitude agree very well with those obtained by Wilson (1972) for a constant (linear) shelf slope.: Zo 1. These results have not been verified by other investigators..s::. e1 ) where (189) The second mode of oscillation is proportional to the Mathieu function Ce (~. for th~ Shelf resonance.2417 _a_ ~~ s 1 (~. and the period T 2 is 2 2 2 1.. The period is given by 3.i.O ul <.. the shelf width i 8 >> L.9254 _a_ (190) % The first and second modes are shown in Figure 31. i. e ) where e = 21.. where L is the wavelength of the incident wave.0 xFigure 31.e.
d8 .0810 0. The wave profile is defined in Table 1.7827 0.0 u 0.7985 0.6665 0.5 0.0716 o.9 u l.5149 0.4 0.7432 0.1239 0. and it can be assumed to behave as a vertical slope.0 0.7392 0.2241 0.6 3.4 3.9265 0.0 3. 2572 0.2 0.8 2. 3975 0.3 1. 5 889 0.7 3.3 3.3428 0.6998 0.4313 0.4520 0. 7773 0.1 3.3018 0.4473 x/a u l.5 3.5668 o.5940 0.2 3.7346 0.8 0.3110 0.7 3.0176 0.0 1.0 3.7818 0.2 2.2972 0.0 0. 0072 0.0688 o.3443 0. First mode Distribution of amplitude U (from Hidaka.8 0.4 2.8887 0.3 2. 2868 xia 2. at the toe of a nearshore slope is 30 meters.4392 0.0 u 0.2 1.3 0.8566 0.5 3.8076 0.9 1.3 1.8096 0. Table 1.9813 0.2 0. 7244 x/a 2.7 0.5200 0.1261 0.0115 0.6995 0.3781 0.3 2. the distance a= 12.5357 0.6233 0.0 2.1110 0.6 1.2 2.7 1.72 miles).7548 0.4 3.8 3.9 3.5 2.2 3.2327 0.7 0.8028 0.430 meters (7.5224 0.6754 o.5 0.1 1.7 2.4 0.9 4.9 3. 1953).9 4. 1 0.8 2.To determine the variation of wave amplitude with respect to distance from the shoreline.6 0.1521 0. This equation was solved by Hidaka using Stormer's method (see Milne.1 2. 0000 0.4 1. 1 2.2800 0. Second mode x/a 0.6 1. 4221 * * * * * * * * * * * * * EXAMPLE PROBLEM 12 * * * * * * * * * * * * * * GIVEN: Water depth. 2822 0. 1493 0.3256 0.6476 0.6 2.8 1.6 2.5600 0.0 2.8 3.8 1.1 0.6 3.3 3.1766 0.4430 0. 0000 0.7466 0.9474 0.1 1.9070 0.5 2.6436 0. 1935b).1 3.7766 0.5445 0.6 0.7 2.2834 0.2 1.3 0. the equation for U is put in the form d2u [ dp2 + (1 + p2) l/2 + _(_l_+_p_2_) + t t (191) where p = x/a. 93 .7 1.2197 0.5880 0.6197 0.8391 o. Complete reflection occurs at the nearshore slope.6807 0.7964 0.0305 0.1462 0.9 1.5 1.5267 0.9 0.0 1.4 1.5 1.4 2.6936 0.4052 0.
Taking h8 as the wave height at the shoreline. for the second mode. 807 ( 30) From equation (190) T 2 = 1. a height equal to 0._43~019. d8 . FIND: The primary and secondary periods of oscillation.5 miles). Table 1 gives. Complete reflection occurs at the nearshore slope.gers 2. for the second mode.3 minutes) (b) Both the first and second modes of oscillation are in the range of tsunami periods which are likely to occur.807(30) 1. SOLliTION: (a) d s = 30 meters and a 12.430 meters From equation (189) T1 = 3. 94 .430) = 29.7818 hg where x/a= 2.395 seconds (23.350 seconds (39.2417 _a_ . the distance a 621 meters (2.800 meters (18.9254 _1_2~. 9254 _a_ licÇ 1.2417 _1_2~. and (b) the relative wave height of the wave at a distance one wavelength from the shoreline in relation to the wave height at the shoreline.FIND: (a) The primary and secondary periods of oscillation.2 minutes) 3.4 or where x= 2. * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * ~ EXAMPLE PROBLEM 13 * * * * * * * * * * * * * * GIVEN: The water depth. at the toe of a nearshore slope is 15 meters._43_0_ 19.039 feet). The values in Table 1 and Figure 32 show that this is approximately the distance between secondmode wave crests (one wavelength).4 (12.
= 15 meters and a = 621 From equation (189) T = 3.gers 1.1.77 minutes) = 1.001 0.::::.5 u 0 Exomple 4 5=0. 9254 . These equations are N.64 minutes) 19. meters.5 1.6 seconds (1.C) Nj .9254 ___ a_ .e. 807 (15) * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * A different means of calculating the amplitude u.. is suggested by Wilson (1972).2417 From equation (190) T 2 621 19.:. (BjDj .2417 _a_ 1 . SOLUTION: d8 Resonan~ amplification on a shelf.D.:~~~ 6 21 98.Uj C + B. the effect of a nonuniform offshore bathymetry).0 0.807(15) 166 seconds (2.5 X/a Figure 32.. which will also account for refraction effects (i.gd" s T1 = 3.0 0 0. J J+ 1 J+l (192) 95 .
uj+l B.) J (193) 2 (194) [fl(b. l + b j)] J+ llw 2 c D. (197) U = 1 at the shoreline (as in the case of Hidaka) and N0 0 shoreline. from equation ( 186) .!!.= dsSz 52 a ( 12  1) ds rz (200) 1 96 .9254 as previously defined in equations 1 (189) and (190). From this 21T . 2. 3.d. + 2C(N . and il the horizontal distance between 1. where bj and bj+l rays at stations j placement of a water stations j and j + represent the distance between refracted wave and j_+ 1. J u . J J+ 1 + N . J (4g) b.2417 and K2 = 1. N the horizontal disparticle. respectively.. T n K a s n . J J (195) (196) j = 1.gd s (199) and for the constant slope. d _. For an unrefracted wave.gd (198) where K = 3. =0 at the Looking at unrefracted waves on a constant shelf slope s 2 for the same wave periods previously defined by equations (189) and (190)..
2417 ___ a_= 2.001. sa sa BjDj BjDj (201) (202) s[ 1 + s(/2.1) 30 0.430 meters s v'gd s 3... and using values of t:. the slope of the shelf is s 2 = 0... The equations of Wilson (1972) can then be expressed as Nj+l .1) j] at the jth increment.350 seconds 1. Complete reflection occurs at the nearshore slope. cr9 72) • SOLUTION: FIND: The wave profiles using the methods of Hidaka (1935b) and Wilson From example problem 12.s (li  1) d ( 12 ..::.1) + 1) S 12 j Kn 2 ( j Kn 2( (203) j + u. dj = dj = ds + jS sa /::. 9254 ___ a_ v'gd s = 1. d8 .1) N t(<c//.001 12. sa where s is an arbitrary increment.1) + s(121.. 395 seconds Exploring the second mode of oscillation as before. J + (204) * * * * * * * * * * * * * EXAMPLE PROBLEM 14 * * * * * * * * * * * * * * GIVEN: The water depth...1l • ( •2< .243 meters (4. = 0.078 feet) 97 . a= _ _ _ _. at the toe of a nearshore s1ope is 30 meters...1 a= 1.2E ) uj] 12 .Defining t:.
23 107. 3. * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * In comparing the work of Hidaka (193Sa.41 85.2 2.6 0 38.1 2.9002 0.4001 0.6035 x/a 0.8391 0.90 23. U. however.9501 0. The wave profiles are plotted in Figure 32.4 2.8090 0.6729 0. Values of horizontal water particle displacement.95 94.6 1.9 1.001. s2 = 0. 8117 0. 7292 0. It was shawn that when a wave propagates seaward from 98 .5928 0.15 32.48 110.0000 0.0 1.43 28.5 0.9131 0. 0571 0. Section V discussed the reflection of waves from an abrupt transition in water depth.31 5.6941 For the given conditions. and wave amplitude.47 53.56 104.2 0. The values obtained by Wilson (1972) are given in Table 2.7507 x/a 1.5 2. Reflection from Seaward Edge of Shelf.50 80.14 70.49 80.8194 0.8 1.0 0.3 N u 1.7 1. Figure 32 shows that Wilson's method produces almost the same results as those obtained by Hidaka.9 2.71 42.29 69.1865 0. w2 = 0.7 0.6 0.the values of U obtained by Hidaka (1935b) are given in Table 1.8484 0.71 85.1 0.20 89.4 1.8609 0.79 70. Wilson's results would predict slightly less amplification over the same distance.3 2.001 u 0. used in example problem 14 is typical of slopes found on many continental shelves.7862 0.20 12.5 1. 1935b) has the advantage of being a general solution. The value of shelf slope.92 40.3374 0.4 0. Table 2.2190 0.4684 0.27 44.17 8.3 0.66 18. 1935b) with that of Wilson (1972).2465 0. Figure 32 shows that the two methods produce comparable results.0045 ds T2 = 1. the numerical method proposed by Wilson for determining the wave profile·is much more readily used for a particular shelf slope.33 56. Wilson's method has the added feature of considering wave refraction.4399 0.0 2.90 o. 82 67.8 0.395 s = 30 rn N s = 0.0336 0.1 1. N.2 1. Table 1 (Hidaka.5967 0.
2. . This was illustrated in example problem S. 3. t . The time.gd where d8 ~ d ~ d2 .J the shoreline sorne of the wave energy is reflected shoreward from the transition in water depth at the seaward limit of the shelf. d2 the water depth at the seaward limit of the shelf. for a wave wi th a normal angle of incidence. However. . This is further illustrated in Figure 33 where d8 is the water depth at the toe of the nearshore slope. s2 the slope of the shelf. and the water depth. . the wave reflected from the shoreline will be in phase with the initial wave incident on the shoreline. sl the slope of the steep transition. for the wave to travel the distance. Reflected waves on a shelf. For perfect reflection. where the wave reflected from the transition is ~ radians out of phase with the incident wave. ~s• from the steep transitio~ to the nearshore slope will be the same as the time required for the reflected wave from the nearshore slope to travel back to the steep transition in depth. the actual phase difference will depend on the geometry of the shelf and transition. d 1 the water depth at the seaward limit of the steep transition in water depth. Wove Reflected Seoword from Shoreline wove Reflected Shoreword from Oepth Transition Figure 33. Therefore. dx c 99 (206) . resonance will occur if 2 where nT (205) T is the incident wave period. and s 3 the nearshore slope. and n = 1. The wave reflected shoreward from the steep transition may be ~ radians out of phase with the wave transmitted seaward across the transition. No ting th at C = .
dx This gives (209) (208) = s2 (207) and. By definition dd dx or defining dx. FIND: The resonant wave periods for the shelf. 2.9 feet). for a constant shelf slope. .000 meters (18. d8 . at the the seaward edge of the shelf is 60 meters (196. * * * * * * * * * * * * * * EXAMPLE PROBLEM 15 * * * * * * * * * * * * * * GIVEN: The water depth. (210) Substituting equation (210) into equation (205) gives 8 Cdi/2 . The slope of the shelf..where x is measured seaward from the toe of the nearshore slope . . d8 . at the toe of a nearshore slope is 30 meters. is 30. ~8 . . and n = 1. 3.000 0. The width of the shelf.6 miles) and the water depth.30 30. for a constant slope is given by 60 . SOLUTION: s2 .001 100 . Equation (211) provides a first approximation for the resonant wave periods.dJ12) T n s2 gl/2 (211) where T is a resonant wave period where the reflected wave and incident wave are rr radians out of phase.
933 seconds (32. Edge Waves.301/2) n 0. have wave periods which will be longer than the incident wave periods.dy2) ' n 1. n = 3 1. He speculated that the currents would act as quasielastic boundaries. Tl T2 T3 etc. . Guza and Bowen (1975) indicate that experimental results confirm the work of Galvin (1965) and Bowen and Inman (1971) which show that incident waves that are not strongly reflected will not excite edge waves visible at the shoreline.2 minutes). standing edge waves will have peaks and nodes at points along the shoreline. . 4.2 minutes). n = 2 1.450 seconds (24. x the distance measured from the shoreline in the seaward direction. creating a resonant condition between two boundaries. 7 minutes).001 (9.900 seconds (48. . The impulse of incident waves reflecting from the shoreline may generate edge waves in the longshore direction. . . These edge waves. In this case waves generated near a shoreline could be trapped between the shoreline and an offshore current. Guza and Inman (1975) have defined the water surface profile of edge waves in the seaward direction using the dimensionless wave amplitude. U.807) 11 2 5. T4 5. . and a dimensionless distance.800 n ' n 1' 2' 3. although edge waves may be either standing or progressive waves. x. 2' 3. s2 gll2 T 8 (601/2 .From equation (211). the trapped mode of longshore wave motion. n = 4 * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Nagaoka (1901) considered the possibility of currents parallel to the coast acting as boundaries which would reflect waves. The water surface profile is given in Figure 34 which shows that higher modes of standing edge waves 101 .3 minutes). in the seaward direction given as x w2 x g tan S (212) where w is the radian frequency (2~/T) of the edge wave.800 seconds (96. T 8 n (d~/2 . n 1 2. and S the angle of the nearshore slope (tanS= S).
1.6 Figure 34. shallowwater They define the longshore wavelength wave nurnber. Therefore.0 0.2 0 0. Offshore profiles of edge waves (from Guza and Inrnan.will have peaks and nades in the seaward direction in addition to the peaks and nades in tne longshore direction. Defining the edge wave with the lower frequency by 102 .6 0.e.Reflected normolly incident wove u 0. and S the angle of the nearshore slope 1n radians. ~y the period of the edge wave.4 0. but propagate in opposite directions along the shoreline. given by 2 Guza and Davis (1974) carried mechanism of edge wave generation waves on a constant beach slope. as shawn in Table 3.. Where the frequencies of the two progressive edge waves forrning the pair are different. the edge wave pair forrns a standing wave. This standing wave will always have a frequency equal to onehalf the incident wave frequency (a period twice the incident wave period) even though the frequencies of the edge wave pairs may vary. Guza and Davis attribute the generating mechanism to a nonlinear interaction between the incident wave and a pair of progressive edge waves with frequencies defined by w1 and w2 where w = 2TI/T and (215) i. The two edge waves have the same wavelength. 1975). ky.2 0. the nades and antinodes of the standing wave will move in the direction of the edge wave with the higher frequency (shorter period). the incident wave frequency is equal to the surn of the two edge wave frequencies.4 Edge woves .8 0. of the edge wave by the longshore k y ~: =(~:) (2n + 1) g(2n + Î) tanS (213) where S « 1 (214) Ly is the wavelength of the edge wave. out a theoretical investigation of the by norrnally incident.
2n/Ly. Therefore.Table 3 .5 0. In the Southern Hemisphere the opposite would be true.56 0. and values of the parameters are given in Table 3 for various modes of resonance.5.134 0.5 < p) (l) < where p is a variable given as 0 higher frequency by (1)2 p + 0.56 0. 035 1. the frequency of edge waves moving left (looking seaward) exceeds the frequency of waves moving to the right in the Northern Hemisphere. Guza and Davis (1974) obtained values for resonant edge wave parameters. Corrected values were presented by Guza and Bowen (1975). in general.5 0. Snodgrass.5 0.366 0.083 0.05 0. is given by ca.3 x 103 x 103 x 103 = (0. of the nodes and antinodes of the standing wave ca = p ~ ky ky (218) the wave where w is the radian frequency of the incident wave and number.095 0.063 0.5 (l)l o.88 0.458 0.274 0.634 0. The parameters shown in Table 3 are in dimensionless form.40 2.5 p) (217) the drift speed.542 0.5 0. Munk.68 4. for a uniform straight coastline.44 1. and the edge wave with the (l) = (0.052 0. (1)2 ky c K 0 0 0 0 1 1 1 2 2 3 0 1 2 3 1 2 3 2 3 3 0.36 0.5 0.28 1.604 0.691 0.075 0.563 0. the higher frequency edge waves would display a preference for moving in a particular direction.041 o. of the edge wave.28 x 102 x 103 x 103 x 103 x 103 x 103 x 13 51 100 160 180 330 520 610 810 1. 396 0.25 0.52 0.5 0. and Gilbert (1964) note that because of coriolis splitting.5 10. Nl N2 (l)l Resonant edge wave parameters.427 0. The parameter K defines a critical value of incident wave amplitude by a2 w K _c_= \} k2(w w )1/2 Y 12 (219) 103 . 726 0.200 (216) o.309 0.
i. the primary pair of edge waves would experience the greatest growth.001 (B = 0.5 w ~ T1 ~ T = 0. The dimensionless value of ky used in Table 3 and equation (219) is given by k* . respectively. v the dimensional kinematic viscosity.where c is a coupling coefficient reevaluated by Guza and Bowen and given in Table 3 in dimensionless form. SOLUTION: (a) From Table 3 0) w1 = 0.5 2 TI 20 x 60 104 .y (w*) 2 gB (220) where ky is the dimensional wave number of the edge wave defined by equation (213) and w* the dimensional radian frequency. * * * * * * * * * * * * * * EXAMPLE PROBLEM 16 * * * * * * * * * * * * * * GIVEN: A tsunami with a period of 20 minutes approaches the shoreline on a constant shelf slope s 2 = 0. where wr and w2 are the dimensional values. The dimensionless values of w1 and w2 used in Table 3 and equation (219) are given by w* 1 w* 2. of the incident wave. the primary pair (N 1 = N2 = O) would be excited while the wave is still sorne distance from the shoreline and the other pairs would be excited closer to the shoreline as the wave amplitude increases. Therefore. and (N 1 0.. 2TI/T. and (b) the wave amplitudes necessary to excite the first three edge wave pairs. ac the dimensional critical incident wave amplitude.5 = 0. It is assumed that the nearshore slope is steep enough for the wave to reflect strongly from the shoreline. (N 1 = 0. N2 = 0).. N 2 = 1). N2 = 2). and the values N1 and N2 define n for w1 and w2 .e.001). Table 3 shows that the number of edge wave pairs would increase as the incident wave amplitude increases. FIND: (a) The wave periods and wavelengths of the first three edge wave pai rs (N 1 = 0 1 .
7 minutes) T2 = ___T = 20 ___ 0..( 2rr 3.001 = 6._0_0_1 1 10 5 3.880 seconds (64.5 0.634 x 60 0. 366 Tl w= o.737 seconds (28.5 w 2 .400 seconds o..280 seconds (54.6 miles) for both edge waves.( 2 2 7T ) . 2 ky=~ =(~rry k g(2n +~)tan f3 y. 74 x 10'+ . .8_0_7_(1)_0_. N = 0).893 seconds (31.366 0.9 minutes) The wavelength of the first edge wave pair is given by equation (213) where n = 0 éN 1 0.6 minutes) = ______ = 20 T T2 0. .309 =r___ w x 60 0. 400 seconds (40 minutes) T1 = 2. 366 2rr r = ______ = 20 x 60 T 0.691 T2 = 1.5 minutes) 2) w1 Tl = . 1 9..366 T1 = 3.2»4_0_0 . For the second edge wave pair.807(1) 0.(20 x 60) 0. 280 N 1 =0 for w1 and N 2 1 fo r k y  )2 9.309 T = 20 ___ 0.990 meters (5..99 x 10'+ and Ly = 8.634 T2 = 1.0.309 T1 = 3..691 x 60 0..
5 x 10.and Ly 16.4 miles) for the first edge wave k y   (~) 1.200 ro6) 0.500 meters (14.200 a c = (13 112 x 1.4 2 for 21T ) 2 _ __ _ 1 ( 3. in Table 3. N1 = 0 for w1 2.12 meter (0.17 meter (0. and w = 2TI/1.893 2 1 .061 meter (0.500 meters for the second edge wave.20 foot) For the second edge wave pair ac= el \ x 1.737 2 1 9.66 x 10.6 x ros square feet per second) = 1.3.5 x 21T 1. Note that the two edge waves of a pair have the same length as indicated by the single value for the dimensionless wave nurnber.6 square meters per second.807(5) 0.200 0.880 ky = 9. For the third edge wave pair.780 meters for the second edge wave.66 x 10 4 and Ly = 23.001 and Ly 23.200 106)1/2 112 x 0. (b) The incident wave amplitude needed to generate the first edge wave pair is given by equation (219) as where K is given in Table 3 and v = 1.5 x ro. ky.001 2.2 stokes (square centimeters per second) (1.4 and Ly = 16.5 x 106) 21T 1.74 9.6 miles) for the first edge wave 21T ) ky= ( 1.56 foot) * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * 106 .807(3) 0.001  x 10.39 foot) For the third edge wave pair ac = (JOO x ~~5 1.807(1) 0.780 meters (10.
Guza and Bowen (1975) investigated edge waves generated by incident waves at sorne arbitrary angle of incidence. and where w is the radian frequency of the incident wave. is small.. ca. equations (222) and (223) reduce to equation (214). and the drift speed. 107 .e. A progressive edge wave moving along a coastline may be reflected from an obstacle such as a large headland. it can be seen that leaky modes (i. 3. N = 0) are given as 2 1 kl and k2 and of the primary edge wave pair w2 4g tan w2 s (l + 2y) (222) = 4g tan ( 1 . The edge waves associated with the tsunami are assumed to occur over the wider and flatter shelf slope shown in Figure 33. i. will produce greater edge wave energy at higher frequencies (shorter periods). They show that the maximum edge wave amplitude at the shoreline is theoretically three times the amplitude of the incident wave for a straight coastline.. a 1 . and are not of concern here. edge waves radiating energy to deep water) will only occur on steep nearshore slopes . with the shoreline (see Fig. Guza and Bowen (1976) discuss the height of the edge waves occurring along a coastline. a 1 . Where a 1 = 0.e.2y) s (223) when the angle of incidence. as the edge waves increase in height their natural frequency increases and no longer matches the forcing frequency. 30). The standing edge wave where a 1 > 0 will progress along the shoreline. Snodgrass. k 1 (N = 0. Guza and Bowen (1975) demonstrate that this could produce a standing edge wave with higher amplitudes near the obstacle. These nearshore slopes are very short in comparison to the tsunami wavelength. (221) the longshore wave numbers. Gallagher (1971) indicates that energy would be lost because of bottom friction and the disperison caused by irregularities in the coastline. Guza and Bowen (1976) indicate that edge wave growth is limited by radiation of energy to deep water and by finiteamplitude demodulation. Defining a parameter y = sin a 1 tan S k2 . and Gilbert (1964) relating trapped modes to leaky modes. From equation (213) and the work of Munk. a 1 . Reflection could also occur from a depth discontinuity such as a submarine canyon in the manner described in Section VI. of a node or antinode of the primary edge wave pair is now given as ca=2ky yw (224) Gallagher (1971) shows that an increase in the angle of incidence.
shallowwater wave on a straight section of shoreline. 108 . Hawaii. Schematic of caustic (uniform bottom slope). 1970. and the wave energy tends to be trapped. Pierson. Refraction diagrams of these waves show a tendency for the waves to turn parallel to the shoreline as they move into deeper water. It is assumed that the wave reflects from the shoreline slope and refracts on the shelf. the waves may be turned back shoreward (see Fig. and extends a sufficient distance seaward. with sorne water depth. a substantial amount of wave energy will be reflected seaward from the shoreline. although sorne wave energy will leak across the caustic (Chao. and Funasaki (1965) illustrated the effects of wave trapping at Hilo.5. parts of a wave crest reflected from one section of coastline may be refracted and trapped so that they coincide with an incident wave on another section of coastline. Refracted Waves and Caustics. As the wave rays are not normal to the shoreline. only the lower frequency waves are significant. where the reflected wave rays were turned by refraction so that they arrived simultaneously at a point inside Hilo Harbor (see Fig. Chao and Pierson. Chao (1970) and Chao and Pierson (1970) investigated higher frequency waves trapped by a caustic. Mulvihill. Where a coastline is irregular. The case of wave energy being trapped by refraction can be most easily illustrated for a long. Figure 35. and with a constant shelf slope extending seaward. When a shelf slopes away from the shoreline. and will refract as they travel seaward. For tsunamis. Palmer. and that waves with frequencies above sorne maximum value will propagate seaward into deep water. When very long waves such as tsunamis arrive at a shoreline. They demonstrate that lower frequency (longer period) waves will form caustics closer to the coastline. at the toe of the shoreline slope. The line tangent to the wave rays where they turn shoreward is a caustic. 1972). d8 . different parts of the wave crest would arrive at the shoreline at different times. The wave rays will not cross the caustic. 35). 1970. These ref!ected waves will interact with the bottom topography. 36).
Hi1o. 1960 tsunami refraction. Appiying Sneii's Law where the wave maves through sorne incrementai distance. (d + dd) d (227) 109 . and assuming that there is a shaiiowwater wave so that the ceierity C is given by c= (gd) l/2 (225) then the incrementai refraction of a wave ray is defined by sin(a. + da. 1965). = (d +d dd) 112 (226) or.Figure 36. Hawaii (after Palmer. and Funasaki. with an incrementai change in water depth d. sin 2 (a.) sin a. Mu1vihi11. + da. squaring both sides of equation (226).) sin 2 a.
(cos 2a.sin 2a)(l) + 4 sin a cos a da] l [2 sin2a + 4 sin a cos a da] (229) 2 and equation (227) becomes 2 sin 2 a + 4 sin a cos a da 2 sin 2 a which reduces to 1 + 2 cot a da 1 + dd (d + dd) d (230) d d (231) (232) 2 cot a da = dd Now. the term can be written sin 2 (a + da) = 1 2 [1 . if the numerator of the left side of equation (227) is expanded. and noting that a= n/2 radians for a straight.Now. 35).cos 2(a + da)] 1 [1 . integrating along the wave ray from the shoreline to the point where it turns parallel to the shoreline. uniform coastline at the point where the wave ray turns parallel to the bottom contour. cos da sin da ~ ~ 1 da Then equation (228) can be written as sin 2 (a +da) = = l 2 [1. where da ~ (228) 0. a 1 as the initial direction of the wave ray at the shoreline (Fig. 110 . taking ~ as the water depth where the ray is parallel to the shoreline and ~ as the distance from the shoreline at that point.cos(2a) cos(2da) + sin(2a) sin(2da)] 2 + 2 sin a cos a(2 sin da cos da)] But.
d sin 2 a s S sin 2 a 1 d s S a. (239) or.sin a cos a tan a da 2 2 d (241) S sin a1 111 .. so d (236) cp ds + Sxp where s x p =2 S sin s a1 s (237) To compute the coordinate parallel to the shoreline of a point on the wave ray.. equation (238) becomes dy 8 = ..(ddp ). note that dy (238) tan a = dx The x coordinate is given by x=.l ddd lrr/2 al 2 cot a da (233) s which when integrated gives 2 fu(sin (234) which reduces to fu d p fu d s 2 fu sin al (235) Taking the antilogs d d But p 8. sin 2 al is the bottom slope. differentiating equation (240) with respect to dx = 2 d s sin a cos a da (240) substituting equation (240) into equation (238).
The water depth at the toe of the shoreline slope 30 meters and the shelf at the toe of the shoreline slope has a uniform seaward slope S2 = 0.= 2 y 2 0._21 sin 2a 1 ] + 4 _____s.Collecting terms and integrating l This gives y/2 dy 0 = 2 S sin a1 2ds f r /2 sin a 1 2 a da (242) .003 sin 2 t [ 30 0.=. IV.07)2 [~~+si: T] 112 25.00~~~~.s_ [~4 2 y 2 S sin a1 sin 2a 1] 4 (243) 1hese equations are limited to the particular case of a long. Refraction diagrams would be required to obtain exact solutions for irregular coastlines.___ S sin 2 a 1 r. it may be desirable to use wave refraction equations in spherical coordinates such as the equations given by Chao (1970) (see Sec.97 miles) . 3).= __2____.. FIND: The distance the wave ray will travel away from the shoreline. If the waves travel for long distances over a shelf. * * * * * * * * * * * * * * EXAMPLE PROBLEM 17 * * * * * * * * * * * * * * A wave ray reflects from a straight shoreline at an initial angle rr/4 radians.:_s___ d d s 30 0.::.700 meters (15. straight coastline.000 meters (6.d..214 miles) 2 d a _4rr .003 Xp From equation (243) y 2 10.003. but may provide a first approximation for solutions on sorne sections of continental shelves. and ~e distance along the shoreline to the point where the reflected wave ray will impinge upon the shoreline. SOLUTION: From equation (237) ~= d~ = GIVEN : ______.
Trapping of generated tsunami. * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * For the particular case of a shallowwater wave on a straight section of coastline and uniform shelf slope.94 miles) along the coast. The wave ray which was reflected from the shoreline will impinge upon the shoreline again at a point 51.000 meters from the toe of the shoreline slope to have the wave ray turn parallel to the bottom contours. For a wave ray originating within the coastal area. as the angle a 1 decreases the distance of the caustic from the shoreline increases and the distance y between the point of reflection and the point where the wave ray impinges again on the coastline also increases. This will excite longshore edge waves along the coastline. the energy will tend to propagate along the coastline. the shelf needs to extend 10. as waves radiating away from a source area may become trapped within a caustic in the same manner as reflected waves.For this example. Figure 37. 113 . When the tsunami energy becomes trapped between a caustic and a coastline. An investigation of the wave rays using the usual wave refraction techniques will define the caustic locations. Tsunamis generated in coastal areas may have part of their energy trapped along the coastline. When the coastline is irregular. the trapped waves may concentrate their energy at particular coastal points. This is illustrated in the following example problem and in Figure 37. and a 1 the angle between the wave ray and the orthogonal to the bottom contours as before. and the locations of any coastal points where energy concentrates. provided the wave is trapped. and may substantially increase observed wave heights.400 meters (31. Xp the distance seaward from the point of origin. given by equations (237) and (243) and illustrated in Figure 35. d 8 is the water depth at the point of origin.
The depth of water at the outer edge of the shelf is d d 8 SOLUTION: + Sx = 30 + 0.14 miles) from the toe of the shoreline slope.000) 330 meters (1. at the inner end of the faultline. Noting that Xp varies 1inear1y with d3 . 37).08 miles) which is beyond the 1imits of the she1f. 114 . The faultline extends to the outer edge of the shelf which is 100 kilometers (62.14.355 or 35.2) (100 ~ + 156..003(100.* * * * * * * * * * * * * * EXAMPLE PROBLEM 18 * * * * * * * * * * * * * * GIVEN: A tsunami is generated by uplifting along a faultline located on a coastal shelf (Fig.23 .23 kilometers (97. FIND: The percentage of the wave energy trapped on the shelf. The faultline is oriented so that the angle between the coastline and the faultline is 8 = 40°. ~ s d s sin 2 al d s s 30 . i.200 meters (8.2) Lf = 0.003 sin 2 40° Xp 330 0.83 miles) At the outer end of the faultline.14.5 percent of the energy generated seaward in the examp1e is trapped. The uplifting is uniform along the faultline so that it acts as a line source.083 feet) Looking first at wave rays traveling seaward from the faultline. using equation (237).003 30 0. from proportionality the 1ength ~ of the part of the fau1tline (of 1ength Lf) contributing trapped energy is Lf(100 . where the wave ray originates 330 0. The shelf has a uniform slope seaward of S = 0.e.003 sin 2 40° xp = 14. taking d8 as the depth of water at the faultline.003 = 156.0.003 and a 30meter water depth at the toe of the shoreline slope.
where severa! wave crests exist between the point of reflection and the caustic. a caustic is formed after reflection as shown by the dashline in Figure 37. A caustic. uniform section of shoreline (Fig. Chao and Pierson (1970). producing high waves at the points of convergence. the wave ray will reflect from the shoreline and be directed at the angle a 1 at the position shown in the figure. For the waves generated seaward. c = /gd= 115 lg(d 8 + Sx) (24 7) . the traveltime. The wave rays follow similar paths to those discussed above. by definition. For the wave rays generated landward from the faultline... 35. parallel to the bottom contours) at the edge of the shelf.5 percent of the energy generated shoreward in the example will also be trapped. For the straight coastline shown in Figure 35. If the wave ray generated from the point at distance 2 along the faultline is considered for the straight. but the wave crests propagating shoreward from the caustic will interact with the wave crests propagating seaward f r om the coastline. These wave rays may reconverge at various points along the coastline. c but d so that ds = dt ~ cos a (244) s (245) J rOt dt = fOXp l _dx _ _ (246) c cos a Also.Consider now the wave rays generated shoreward . wave rays trapped on a shelf may diverge apart.e. from shallowwater assumptions. after they have reflected from the shoreline. the trapped wave rays). 37). along the wave ray between the shoreline and the caustic can be easily determined. Chao (1970). a caustic will be formed by those wave rays refracted back to the shoreline (i. * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * As illustrated in Figure 37. Therefore. This wave ray will turn parallel to the shoreline (i. t. Note that the energy generated shoreward would have a tendency to form a Machstem along the shoreline if S is greater than 45°. and Pierson (1972) discuss the case of shortperiod waves reflected from a shoreline. is a line tangent to a family of wave rays. producing many peaks and nodes between the coastline and the caustic. Where the distance s is measured along the wave ray.e.
8 .= 2 dx 2 d sin a da a1 cos a s sin (248) Using equation (239) to define x and substituting equations (247) and (248) into equation (246). 3 . wave trapping can be defined using the equations tan a and 2 n tanh [n(: ) d] where r e c n d the radius of curvature of a contour line (taking circular bottom contours to define the shelf around the coastline) the coordinate angle in polar coordinates of a point along the wave ray at radius r a constant a variable along the wave ray the water depth at radius r 1.s) (250) which gives t (251) Shen and Meyer (1967) and Shen (1972) give a solution for curved coastlines. . t + s (d (249) s S sin 2 a 1 . 2. defines an integer number of wave crests around a circular island rde = ±(n2r2 _ c2)l/2 dr 1 (252) (253) 116 .and from equation (240). For a circular arc. . it becomes t = lrr/2 a l 2ds S sin 2 al lg sin çx dx sin2 a \j s Collecting terms.
Taking the radius at the shoreline as R8 . Camfield (1979) gives the following development to express the solution in terms of several dimensionless parameters. where the wave ray turns parallel to the bottom contours) is an appropriate length scale.e. (259) where d is the water depth at radius rp. Shen and Meyer (1967) indicate that a number of caustics may exist. Equations (252) and (253) were derived for dimensionless variables where the dimensional values of length had been divided by sorne horizontal length scale. resonance will occur between the shoreline and the caustics for wave periods defined by integer values of n 2 for which solutions exist. For a circular island.. Shen and Meyer (1967) infer that the radius rp of the caustic (i. and the radius wher e the wave ray turns parallel to the bottom contours as rp• at the shoreline (254) (255) 1 (256) Where the wave ray turns parallel to the bottom contours 0 n "' r (257) (258) c p Substituting into equation (253). the radius of curvature rp is normalized so that its dimensionless value is r p = 1 (260) The derivation of the equations also assumes that vertical length scale horizontal length scale = E (261) 117 . and equation (259) can then be solved to obtain a value for rp corresponding to each value of c which will provide a solution. Equation (256) can be solved to determine c for any integer value n 2 satisfying the equation. Using an asterisk (*) to define dimensional values.
Shen (1972) taKes EM = 1. (264) tanh . < R. Also the dimensional form of equation (259) becomes c tanh ct. so that M must be large. Consider first a concave coastline.] in its dimensional form. 1 (2 63) 1/M = c/n 2 . all wave rays could obviously be trapped by this type of coastline. a concave coastline would forma closeo circular basin with radius R. 118 . tsunamis) where the period T is large. Shen defines M as M (w*)2(horizontal length scale) g which gives (262) For longperiod waves (e.so the vertical dimensions in equations (2~2) to (259) are assume~ to represent dimensional values divided by rp E. In the limiting case. R. It is assumed E + o. where is the dimensional depth at the toe of the shoreline slope.g.. the caustic radius rp must be large in arder for M to be large. and rP. the solution is for cases where the shoreline radius (and therefore the caustic radius rp) js much greater than the wavelength. where 0 < o << 1. Equation (263) further Shen has defined M = n 2/c so that E reduces to 2 2 c2 tan2 a 1 ~(1 + c2 tan2 a 1 n~ R.. [n~ p d*] r* c 1 (265) where d* is the water depth at radius rp.=tan 2 a l tan 2 a l c 2 R* r* s p )l/ )l/ ct. such as a large bay. where the term rp E is the vertical length'scale. ~ the dimensional·radius of curvature of the shoreline. equation (236) now reduces to R. In general.. Therefore. where the radii r:p and are measured from the center of curvature offshore. Using Shen's work. and rp the dimensional radius to the point where the wave ray turns paralle1 to the bottom contours.
rp becomes smaller so that the caustics are farther from the shoreline . rp ~ 0. it will be trapped closer to the shoreline. where a 1 is the angle between the chard and a radius drawn to the end 119 .Where a 1 ~ 90° the wave rays become trapped very close to the shoreline so that rp ~ R.e. It is of interest to note that equation (270) provides a solution independent of water depth or shelf slope. To determine the limiting values of n 2 which will provide solutions. defining the angle between the reflected wave ray at the shoreline and the normal to the shoreline (see Fig. circular bay. 35). as c ~ 1. r. Finding the caustic location. meaning that a wave ray reflected along an orthogonal to the shoreline will pass through the center of curvature.. it may be noted that. ~ 1 (267) (268) and as c ~ 1. Equations (264) and 265) are used to investigate wave rays at any angle a 1 . From equation (265). R* s > 1 r* p (266) From equation (265). R.. defining the shoreline if the concave shoreline extends a sufficient distance. a large. e. when n ~ oo. As a 1 ~ 0. (269) p ~n = R* s sin a 1 (270) Where the angle a 1 is known. the wave energy will always be trapped between the radius. which is a restatement of the fact that all waves would be trapped where the coastline is concave. 262). ~~ R* p r* which reduces to (r*) .. Where the wave period becomes longer.g. in equation (264). Equation (270) defines the distance from the center of curvature to a chard across a circular arc. From equation (264). From the definition that 1/M œ T2 (eq... tanh which then gives ~ 1 + c ( 2 tan 2 a ) _/ 1 tan2 al l/2 n 2 d* ] c~ R. c ~ 1 as tanh [n~ d*/Crp c)] ~ 1. i. As a 1 becomes smaller. nz ~ oo. the radius. when n ~ oo and T ~ 0 is of interest. and that 1/M = c/n 2 . defining the caustic and the radius. and a 1 > 0. 1t can be seen that waves would be trapped where T ~ 0. rp.
d$. the wave rays which would probably be trapped are those where a 1 is large.c R* s ( n~ d. in relation to the water depth at the shoreline.) c R* s = R. Where a circular island has a small radius.) = R. R.e. a necessary condition of wave trapping. The minimum and maximum values of can be found as follows: From equation (264). (274) as a condition of wave trapping on a convex coastline. c R. r* p (272) Substituting equation (265) in equation (272) above. r* > R* p s tanh (n~ d*) c r:p < (273) as rp > R. and therefore verifies equations (264) and (265) for a concave coastline such as a large bay. .. Therefore. Letting a 1 + n/2. This is necessary in order to have the rate of curvature of the wave ray exceed the rate of curvature of the bottom contours. tanh (n~ d. the wave rays most nearly parallel to the shoreline. then R.e. This defines the path of an unrefracted wave ray (the expected result when T + 0). a circular island). tan a 1 + ~ so that the term (271) From equation (264) c tanh . i.of the chord.g./rp 1.. for a convex coastline. r* p tan a 1 > 0 + o (275) 120. r:p d* d.. where Ici c tanh + ~ n2 which will produce solutions (n~ d.) = R. there is a greater probability of the wave rays spiraling off into deep water than there would be for an island with a large radius.. i. For a convex coastline (e. This means that the slope of the shelf must be greater than sorne minimum value defined by d~/R~ in order to have a caustic. to have waves trapped on the shelf.
c R* s and from equation (275) ~ n2 d* 2 s c R* s (276) (277) (R*) 2 s d* r* s p From equation (265)..r* c p it is found that [ ~ n~ d* r* c p (279) . n~ d*] tanh .(d*) 2] . R* s R* s n~ + (282) which provides minimum rmn s2  d* s (283) (n2) ~ax (R*) 2 S2 s . and noting that for + (r* . n2 d*) 2 s tanh ( .R* d* s s R* s .= 1 p n~ d* (280) Substituting the value of a uniform slope d* equation (280) is then r* p d* s from equation (278). that for c ~ (278) oo.R*) S p s 2 (R*) 2 = 0 s (281) 4 n 2 (d* R* S2 .. But as c ~ oo.d* s 2 s (284) 121 .where o is sorne small value.n 2 (R*) 2 S2 s s s 2 s Equation (282) is a quadratic equation for and maximum values given by (n2) 2. noting..
S. equation (265) is plotted on lines of constant d~/(R~ S2 ). s = R*S s 2 d* s (285) and at the maximum value of n2 .At the minimum value of n2 . From equation (262) note that 21T M gT2 = 21T rp (289) 122 .determines wave trapping. which define resonant periods for a circular island.5 (288) s The parameter d. the maximum value of n~ d~/R~ defined by equation (285) must be less than or equal to 1. Equation (264) is plotted on lines of constant n~ d~/~./CS 2 R. so the minimum value of s 2 for wave trapping to occur is given by ~ d* 1. n2 2 R* d* 2. Solutions are not limited to integer values of n 2 . for values of n 2 between the minimum and maximum values defined by equations (283) and (284).__ s_ d* s 1 (287) R./CR. Solutions for trapped waves are obtained where the two families of curves intersect. Solutions for equations (264) and (265) are plotted in Figure 38.) is a shelf parameter which .. d* S R* 0. 2 s2 < which reduces to 8 . From equation (288) it is seen that solutions will only exist where d. S2 ) ~O. A continuous band of solutions exists for equations (264) and (265). n2 s 2 R* d* (286) s For a solution to exist for given values of d~ and ~.
1.t :::.8 R* s rr o.s t' N 0.2 ï:__:= 0. a.. =5o Figure 38.1 ~· __ l  . Solution to equations (264) and (265).4 ·. ··::::: + oc.0 0 .. ~ . ..~ b w 0 .""'" ]" ooo~i "0" ~ 01 2 3 4 5 6 7 8 9 10 Il c 12 13 14 15 16 17 18 19 20 a.·::=t:·:::_:::..
a 1 =10° Figure 38..Continued .3 . 0.. Solution to equations· (264) and (265).ë.· .. ~VJL. Ill .. ..:J ..1 ..J..t' N ~ IJ!. . ··1~}~' 1 .o.(Jii·~.~~ r"·' ~~:.... 1 .. '"''"'4:'~''"j. L. t~ · 'S ·U'. i .·=•·~~i ~~2 !) · ... ·.. ~ .: . ·.... 20 c b... ''"'.
al c = 15o Solution to equations (264) and (265).Continued .3 0 . 1 c.9 0.5 t' 1'J Vl 0.0 0.8 0.7 0 . Figure 38.6 R* rj 0 .2 0.4 0 .1.
 1+++·H~ 11 Figure 38 . Solution to equations (264) and (265).9 .! 1  0.Continued 126 .
..~ ·~ t . l. ~.li'! 1 • ' L.. 1 1 •.ri.~ .t~ 1. :'   1': 1 R . 0.. . f" + T j ._ C! H 2 fo' ~ . " .. ~ j ~' J H ~ ~l ~ 1 r:·i +.l i 1 1 3 ~ L .6  1 ~~+l . _f t . Solution to equations (264) and (265)..9 ~· . ~r 1   ...2 ~. .~o~ ~t  l~  H  .C'{~ : ·~~ ~ ·~ ... Ir' k r1 .r:· · 1 H  ... ~. ' l'+ i _ ~. _. . tt1 !tt ~ ~2 •t . .·:\ t ! • : . ...r.0  0.1. ._ t'. ' 't ~ 1 J" . l ' :~   1~ .< L ._ N i.t :tt wr IJ ~ 4 :: .1 r'~ 'L'=·. 1~ \ Il 1 jl~ 1. "' ._ c . +  .1 Il''' iLt~ 0 1 1 1 ff ~ ~ lb" l±r ~~ jl ~ J...4 r. i~ 1 1t "· i :i)\ rr.1±llJ.8 1 t~ . ' ._ .Continued 127 .~...l.5 0.0~ ~~ ~ Ir..i1 11 1 :~ 1~  .il . .'it ..7 0..~ 1 .. '"' ·~ ..i· ·'· ' 9 5 7 8 10 Figure 38 .3 ~··"' ·· ~t'li ~ 1J '' 1 ri.Ill ' ....' T j · _ . t __.~  11 U lW ~7" ... ' >11~~~~ +. "!4..~ rr.  :: ' '"() H· [1 . ~._". ~Il If. ..ri '~i 0.:P< ~~. ' 1(1t 1 { _  0. t· ~ ..[{J\tr ~1. t 1 "' ! ··.~i Il !il ·jHI i j ~~ IJ~·jj!fjl___ ·f:f ~ i....~ yV. i1 II.. .l   ·  0.A.6 R* s 110.  . · f. l 0.. 1" .P< .. 1 .._.a '  ~ . ~.._ · tl~ j1' ~ 1O./ v~1 •t .' ·ttt. :r~r.
000 meters (62. These solutions will give the maximum trapped wave periods. A tsunami reflects from the shoreline slope and refracts over a shelf where the bottom slope of the shelf 128 . for varying values of a 1 .and as M n 2/c. Solutions for equations (264) and (265) at greater values of R. Figure 38 shows that./rp (at longer wave periods) define trapped waves. the minimum trapped wave period and the outer limit of the trapped wave zone can be approximated using equations (264) and (265). Carver. as previously defined.. and R~ . where rp ~ R~). the m1n1mum trapped wave period will increase as a 1 decreases. defined by equation (286). at lower values of a 1 . a * * * * * * * * * * * * * * EXAMPLE PROBLEM 19 * * * * * * * * * ·k * * * * GIVEN: A curved section of coastline is convex. U. The theoretical solutions given by equations (264) and (265) are for the case of a coastline approximated by circular arc. Solutions of equations (264) and (265) for values of n 2 . define caustics at the inner limit of the trapped wave zone near the shoreline (i.14 miles). The theoretical solutions presented here can be used to verify numerical methods used for more complex solutions for irregular coastlines. with a radius of curvature R~ = 100. This is expected since shorter period waves. and Markle (1977) using a finiteelement numerical model developed by Chen and Mei (1974). and the depth at the toe of the shoreline slope d~ = 30 meters. Solutions for irregular coastlines must be obtained by numerical methods. s 2 . this will approach a straight coastline. In the limiting case. The minimum period of the trapped waves is defined where a line of constant 21T d~ n 2/(gT 2) is tangent to the line where d~/(S 2 R~) is constant for the given values of d~. tend to pass into deep water and are not trapped. becomes very large (U >> 1). However.e. An example of a numerical solution is given by Houston. Tmax• but the solutions tend to break down at this point as the parameter. Solutions for equations (264) and (265) for smaller values of R~/rp (at longer wave periods) define the damping zone discussed by Lozano and Meyer (1976). defined by equation (66) as U = (H/d)(L/d) 2 . this can be rewritten as (290) which reduces to 21T d* n 2 s gT 2 1 = 21TC (R*)\n2 d*) s 2 s r* ~ p s (291) Lines of constant 21T d~ n 2/(gT 2 ) are also plotted in Figure 38.
 d* s2 R.is s 2 = 0.068 where n 2 d* ~= 0. 71 2.000) 30 300 0. rp. for wave trapping to exist 8 .000 0. (a) From equation (288).5 s2 (b) > 2 _!!_ d* R. FIND: (a) The minimum shelf slope required for trapped waves to exist.000 30 0.003(100. < 0. where a 1 45° . and (c) SOLUTION: the minimum period of waves trapped by refraction.93 and R* ~ r* p 129 . for a 1 45°. 71 n2 = 0.71 R* s R* d! 100.367 s = 48. The angle between the reflected wave ray and the orthogonal to the shoreline a 1 = 45°. the minimum wave period is at 0.65 = 0.1 From Figure 38. defining the outer limit of the trapped wave zone where a 1 = 45 0 . = 2(30) 100. (b) the radius.0006 The shelf parameter is s = S R* 2 d* s 30 0.003.
giving the Mach stem the appearance of a large wave moving along the shoreline.068) = 13.93 R* 100. Perroud (1957) showed that a 1 = 45° defines a critical angle for wave reflection.530 0. i. Experimental measurements by Perroud (1957) show that the Mach stem has a profile at the shoreline similar to the profile of the incident wave.R* p 8 2~ 107.8 miles) The width of the trapped wave zone from the shoreline to the outer limit is given by r* . C~. When a 1 = 45° the end of the wave crest at the shoreline turns perpendicular to the shoreline (see Fig. * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * 6.530 . between the reflected wave ray and the normal to the shoreline defined by a 2 < a 1 • The second component is a Mach stem which moves along the shoreline in the direction of the longshore component of the incident wave. 1 and S. growing in size as it progresses along the shoreline. The first is a reflected wave.95 minutes) for the minimum trapped wave period.e. Regular reflection no longer occurs when a 1 > 45°.068 or T2 = 7.2 seconds (1. the incident wave produces two components. Figure 38 illustrates solutions for trapped waves for angles a 1 ~ 45°. MachStem Formation. lower than the incident wave. n2 meters (4.000 0. Perroud showed that. When a 1 < 45° regular reflection occurs.which gives r* p = ~ = . along the shoreline is given as c c~ = sin where a1 (292) C is the celerity of the incident wave near the shoreline.= 107. and with the angle.068 g 2~(30)(48.000 meters (66.100. 130 .81(0.93 0. the wave reflects in a manner described in Section VI. 39). a 2 .65) 9.68 miles) (c) d* n2 s gT2 T2 T = 2~ 0.747 = 117. so its speed of propagation.. The Mach stem remains attached to the shoreline end of the incident wave crest.530 d. for a 1 > 45°. Figure 39 shows the initial growth of a Mach stem along a vertical wall for the critical angle a 1 = 45°.
g. lt was generally found by Perroud (1957) and Chen (1961) that the incident waves neither reflected from the shoreline nor formed Mach stems when a 1 > 70°. 1957).Refleeted Wove Cres t Figure 39. solitary wave. Machstem formation. the Mach stem formed a breaking wave along the shoreline. e. Wigen. Chen (1961) studied Machstem development using a range of values for the nearshore slope. This would be the case for an incident wave traveling nearly parallel to the section of shoreline. He showed that where the angle of the nearshore slope S < 60°. and a 1 > 55°. Murty. and Chawla (1975) have tabulated the approximate periods of inlets on the Pacifie coast of North and South America based on the formula (293) 131 . the tsunami may cause resonance if a natural mode of oscillation of a bay or harbor corresponds to the period of the tsunami. When a bay or harbor is very long in relation to the tsunami wavelength. 7. Lines of equal surface elevation above still water normalized to unit incident wave amplitude (after Perroud. a wave entering an inlet with a gradually varying cross section. The relationship between the wavelength and the slope length was not considered in this case. and was not varied during the experiments.. Bay and Harbor Resonance.
and for Puget Sound are shawn in Table 4. and da the average depth of the inlet. This is not consistent with the work of other investigators. B. of the secondary undulations could be given as I (294) Values of I for inlets of Alaska. They point out that equations (293) and (294) are based on a onedimensional theory which is not valid for law ratios of Lb/B. and the effects of branched inlets are not considered. Where the larger inlet has a width B1 . and Sullivan (1962) carried out a hydraulic madel investigation of an inlet connected to an "infinite ocean. they give the amplitude a 2 at the head of the smaller inlet as (295) where a is the incident tsunami amplitude in the larger inlet. T the per1od o} the tsunami. As indicated by Murty. Wigen. 6 ." The ocean was simulated in a wave basin. and T1 the period of the smaller inlet as given by equation (293). . Wigen. and to d~/ 2 . and that transverse motion is important in these cases. width. while the minimum amplitude would be at T1 /T = 2. using wave absorbers to minimize reflected 132 . Lb the length of the inlet. British Columbia. and relative intensity of secondary oscillations of the water level. for inlets on the coast of Alaska and British Columbia. Referring to the work of Nakano (1932) which showed secondary undulations to be proportional to the length of an inlet. Ippen. 40). I. . and Chawla (1975) proposed that the relative intensity. Raichlen. restricted entrances. and for Puget Sound. etc. and Chawla. . and Chawla. These values are only approximate because variations in inlet cross section. Values of length. would be expected to excite larger amplitudes of oscillation. and inversely proportional to the width. Equation (295) would predict very high values of a 2/a1 where B1/B 2 is large. are given in Table 4. depth. Lb. The maximum amplitude a 2 will occur when T1/T = 1. 4. sorne hays which have small ratios of Lb/B have large secondary oscillations.where T 1 is the primary period. and the width narrows to a width B2 in the smaller inlet (see Fig. Fukuuchi and Ito (1966) consider a tsunami passing from a larger bay or inlet into a smaller inlet. 3. period. Wigen. as given by Murty. I. Inlets with higher relative intensities. 5 . Murty. .
and of Puget Sound (from Murty.135 0.9 0.0 5.0 1.220 0.9 ll7 32 48 42 61 79 31 33 46 43 12 26 35 ll5 28 51 Ill 24 Kwatna Inlet South Bentinck Arm Rivers Inlet loklses Inlet 37 46 26 57 39 133 .3 3.0 24.5 1.175 0.1 0.0 2.3 1.140 0.6 3.0 22.8 3.5 1.5 14.0 0.3 12.240 0.295 0.7 192 100 102 1.725 64 170 100 54 122 382 185 81 50 119 348 355 587 147 20. mean width (km) ~/8 Alaska Tarr InletGlacier Bay Muir Inlet Lynn Canal Gastineau Canal Taku InletStephens Passage Tracy Arm Endicott Arm 'Frederick Sound Thomas Bay Tenakee Inlet Peril Strai t Bradfi e ld CanalErnest Sound Behm Canal WestBell Arm Burroughs BayBehm Canal East Rudyerd Bay Boca de Quadra carroll Inlet Georse Inlet British Coli.5 6.5 1.255 0.8 3.9 1.210 0.240 0.3 24.2 1.0 2.3 22.3 13.4 43.0.2 3.360 0. Wigen.225 159 51 164 61 165 56 58 133 35 5.8 14.0 13.0 22.2 39.295 .345 0.8 32.8 1.634 125 173 332 236 162 153 188 788 810 189 170 59 143 95 323 3.5 0.1 15.275 0.6 7.1 27.170 0.045 0.270 0.255 0.Uilbia Portland Canal Cllservatory InletHastings Arm Alice Arm IChutzeymateen Inlet Work Channel Prince Rupert Inlet Douglas Channel Kildala Arm Gardner Canal Surf Inlet Laredo Inlet 5heep PassageMusse! lnlet Spiller Channel Roscoe Inlet Cousins Inlet Cascade Inlet Dean Channel Ill 35 146 18 133 43 44 80 0. Tl (min) B.9 24.7 47. and intensity of secondary undulations of inlets of Alaska and British Columbia.9 1.7 5.425 0.8 15.3 34.220 0.6 1.4 26.200 153 82 26 49 74 60 97 31 2.5 1.4 2.2 3.215 0. and Chawla.2 2. In let length (km) ~· da· mean depth (km) Period.6 1.1 20.295 0.385 0.130 0.120 0.260 0.070 0.240 0. periods of fundamental mode.9 406 144 124 601 230 1.8 10.3 1.4 19.6 25.6 46.1 2.0 23.310 0.250 0.275 0.0 1.2 52.245 0.330 0.9 1.0 27.2 4.9 13. Dimensions.420 0.165 0. 64 71 80 ll5 104 97 74 174 36 72 ll3 22 56 44 22 76 82 31 ll5 76 19 25 54 19 83 19 91 22 39 0.2 23.5 15.0 17. 1975).Table 4.0 16.040 0.7 12.150 0.1 13.6 23.8 10.420 0.0 15.3 28.2 2.
5 47.1 181 1.9 276 1. and of Puget Sound (from Murty.> Inlet length (km) Smith Inlet Mereworth Sound Belize In let Nugent Sound Seymour Inlet Drury Inlet Knight Inlet Ca11 Inlet Loughborough In! et Bute Inlet Toba Inlet Jervis Inlet Howe Sound Vancouver Island British Cohunbia Holberg.295 0.1 9.5 1.090 0. Dimensions.045 0.4 16.7 10.135 0.1 d3/2 a 1.8 11.390 0.0 2.120 0.035 0.2 16.050 0.7 1. 7 1.5 1.Table 4 .909 775 281 310 681 374 364 1. 7 2.8 10.2 469 461 1.8 18.3 1.925 2.118 700 134 .7 1.165 0.7 3.0 o.100 0.255 0.9 14. 7 1.150 0.040 0.110 0.190 0.5 14.030 0.0 15.5 3.2 27. Wigen.351 485 962 120 o.3 39.1 0.4 2. 1975) .6 11.8 6.669 145 2.Rupert In let Quatsino SoundNeroutsos Inlet Forward Inlet Klaskino Inlet Ououkinsh Inlet Port Eliza Espinosa Inlet Nuchali tz Inlet Tahsis Inlet Cook ChannelTlupana Inlet Zuciarte ChannelMechalat Inlet Sydney Inlet Shel ter In let Herbert Inlet Pipestem Inlet Effingham Inlet Alberni Inlet Saanich Inlet Puget Sound Puget Sound Hood Canal Possession SoundSaratoga Passage 111 102 70 44 59 11 11 14 11 14 15 29 31 48 20 19 23 9 17 69 23 33 19 52 24 67 22 130 28 35 76 37 89 43 da.150 0.112 270 377 249 56 58 80 57 o.2 1.215 0.9 43.0 0.075 0.270 0.5 0.398 472 1.6 3.4 1.3 3.9 1.Continued Lz.3 18.095 0.7 15. periods of fundamental mode . mean width (km) Io/B io/B 25.0 1.3 0. and Chawla.4 14.2 7.3 34. and intensi ty of secondary undulations of in lets of Alaska and British Columbia.5 40.7 11.7 18.220 0.4 26.180 73 103 43 40 32 33 20 64 56 54 69 48 38 49 29 37 122 37 1.9 1. mean depth (km) 0.495 0.3 0. 759 367 1.5 12.3 32.6 20.3 2.145 0.225 Period.510 0. 7 1.420 0.3 2.085 0.025 0.404 108 2.080 0.2 1.5 32.3 1.4 47.115 0.0 15. Tl (min) 43 43 69 59 70 74 161 51 54 72 40 85 61 8.090 184 207 157 6.7 20.165 0.2 53.2 31.
. The maximum amplification occurring where T1 /T < 1 is equivalent to resonance for a longer inlet. since a node does not exist at the entrance. T1 /T ~ 3. It is assumed in this case that B1 + oo. (296) Therefore. The results are for a fixed inlet width and varying wavelength. etc. Figure 41 shows that the maximum amplification occurs where T1 /T < 1. Ippen. 7 . that the inlet has an effective length Le extending into the open sea. i. Tle. is (297) The length Le. waves.e. therefore. the·variation in the curves illustrating the dependence of the results on the ratio of wavelength to inlet width. Ippen. where k is the wave number 2~/L. and 135 . . i. Horikawa. The experimental results of Ippen.e. Raichlen. and Shuto (1971) carried out similar experiments for an inlet with the entrance partially closed by a breakwater. ' Figure 40. Raichlen. wide inlets. Each curve in Figure 41 was obtained by varying the inlet length for a fixed wavelength. is defined by equation (297) if it is assumed that T1e/T = 1 where maximum amplification occurs.Lb Bz f 1 1 Plan view of inlet. The results were dependent upon the efficiency of the wave filters and wave absorbers used in the experiments.. and Sullivan (1962) and Nishimura. and Sullivan are shown in Figure 41.. 5. Using equation (293) to define T1 . and Sullivan did not explore the amplification for shorter period waves. particularly for short. Raichlen. It can be assumed. Horikawa. Nishimura. Le > ~ and the effective primary period. They also found that the inlet had an effective length greater than the actual length.
Nishimura.4 02 a.427 1.0 0. and Shuto reported that variations in opening width at the mouth of an inlet did not affect the effective length.2 1.907 e k=4.72 cm Harbor length varies œ k=4.5 00 0. and Sullivan.157 0 k=4. 40) . The resonant amplification would occur where TLe/T = 1 as before. However. 1. Ippen and Goda (1963) indicated this would not be true because the half harbor has a asymmetric entrance onehalf the width of the centered entrance of the full harbor. Raichlen.119 e k=3. 1 Depth=25.5 1. Horikawa.0 klb Figure 41.2. Ippen and Goda defined r esonant amplification (the ratio of an amplitude in the harbor to the amplitude at the closed harbor entrance) as 1 (298) where ljJ1 and ljJ 2 are wave radiation functions given in Figure 42. The funct ions shawn in Figure 42 apply to all harbor openings. 136 .6 Absorber No. For a fully open inlet or harbor (see Fig. Amplification factor versus rel ative harbor length (from Ippen. 1962).71 cm Harbor width fixcd b=5. 1 Filter No. Shuto all indicate that the effective length may be determined by the r atio of inlet width to inlet length. Ippen and Goda showed that the harbor entrance width determined the value of wave radiation functions whi ch are used to dete rmine water surface elevations. they investigated a halfharbor width and assumed symmetry would produce the same resonant motion in the half harbor as it would in a full harbor. where b is the width of th e op ening.290 () k=4.8 1.
The incident wave1ength L = 258.0 1.. 04L) 2L 0. Wave radiation functions (from Ippen and Goda.1257 From Figure 42.. where is the length of the in1et. 0.4 1.2 v / '2 . ~ O.B ~ ~ 0. * * * * * * * * * * * * * * EXAMPLE PROBLEM 20 * * * * * * * * * * * * * * GIVEN: ~ A ful1y open inlet has a width.6 0...8 1..:. a2 1 al [ (cos kLb  lji 2 sin kLb) 2 + ljl 1 sin 2 kLb] l/2 137 .6 '!!> 2 Figure 42. "'2  ~~ 0. FIND: SOLUTION: kB 2 27T(O..6 0 / /_. 1963).4 0. The resonant amplification in the inlet.2 1./ v / / /  ..4 / 1(. rr.2 1. B. where b = B for a fully open inlet ljll 0..12 and From equation (298)..194 Lb. given by B = 0.2 1v / / 0.0 ~1 . ../ 0.1.
Their experimental results for partially closed harbors generally showed that amplification factors were less than those predicted theoretically. 44).~02.54~.5~5.~53~. 43).24 sin{21T 0 · }) 0.~o~5.16 + (0.o~1.f5 0 • .12)2 sin2(21r 0 04 · 0.. 0 .1 2 04 0. 9 Wove Period (s) '1!e: 8 N C! 0 "' ID 0 .194 4 3 2 0o~o~. ci 10 0 .194 ~ )~1/ 2 * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Ippen and Goda (1963) compared theoretical and experimental results of a fully open harbor (Fig.. .5 klb Figure 43..... 0 7 6 Theory Harbor .5~2.·. 1963).. Frequency response of a fully open harbor (from Ippen and Goda.5. 138 . They also obtained theoretical results for partially closed harbors with both symmetric and asymmetric entrances (Fig.o3•.o4~.. comparisons between experiments and theory were only obtained for higher modes of oscillation. j ~8 1jJ ~ 8/Lb= 0. However.194 8.
..1 .20 18 16 14 . 1963). Theoretical frequency response curves of harbors (from Ippen and Goda.1. Wi de Harbor 3 4 6 1 8 9 10 Figure 44.01 0..01 0. 139 ..1 .0 6 4 2 0 0 2 3 4 6 9 10 o.1..... Norrow Harbor 8 7 x Point of predicted amplification . 5 N 0 0  6 b/8 ·0.0 4 3 2 2 b. 10 N 0 0  12 8 b/8 ·0.
..0 b/8 :0.E: 0. Theoretical frequency response curves of harbors (from Ippen and Goda.5 ( symmetrical ) . 1963).9 8/Lb= 1.5 ( symmetrical) .Squore Harbor with Asymmetric Entronce 9~~~~~~~~~~ 8/Lb: 2.1 x Point of predicted amplification .E : 0..... N 0 2 3 klb 4 5 6 d.... N  5 4 0 bl!t ~€8 x Point of predicted amplification .Continued 140 ....95 (most asymmetric) 8~ 0 3 2 oL~L~~~~==~L~~==~==~==~zd 0 2 3 4 5 6 klb c...1 6 0 ..95 (most osymmetric l 0 .75 .0 b/8 : 0..E : 0.E: 0.75 .· ..E: 0..· .E:: 0. Wide Harbor with Asymmetric Entronce Figure 44.
and Miyata (1971) show that the wave number.e. 45). and Miyata. Shaw. Carrier. represented as kinetic energy.. for k a (299) La is the length of the entrance channel (Fig. Shaw. 1971). 1945. the water in the harbor has a much lower velocity. (b/~) where 1 Lb~ f~Lcb 1 B . Harbor with an entrance channel. Equation (299) is restricted to very limited cases. and the rise and fall of the water level in the harbor is represented as potential energy. Figure 46 shows that equation (299) will generally predict resonant wavelengths that are tao short (and therefore predicted resonant periods with values lower than the actual resonant periods).Miles (1972) indicates that for waves passing from a Continental Shelf into a harbor. The terrn ln (ka b/2) in the denominator represents the effect of energy radiation from the seaward end of the entrance channel (Rayleigh. Miles. Helmholtz resonance is represented ka. The harbor undergoes a pumping motion where the water leve! in the harbor is assumed to rise and fall uniformly across the total area of the harbor (Carrier. 1948). The water passing through the entry channel is assurned to have a high velocity. 141 . i.L~~1 Figure 45. where the dimensions of the harbor and the entry channel are small compared to the local wavelength of the tsunami. and Figure 46 shows a comparison of equation (299) and the results of Miles (1971) for a harbor with a zerolength entrance channel (La= 0). the response of the harbor is essentially restricted to the Helmholtz mode. the lowest mode of resonance.
l iù1l!t111i=iJ 1.j..08 0. s lllftUMJ1M+1il'tr·il~l_J rn '.il.fl1l i+U+TI1Il.: J!!:l:!l": ll+l++l il+II+I++Jl+l+"' ..~J:ttt~W i:Jl jo.5 ~~ki!i i:i IUIJn .~ 1 lil~ j:::~ r• '.06 i'i! 0.6 0.lw+ .02 0.+·: '· ·TI'· I ++ 1Hi+H1++ i: :~I l+l+++lHl+Hl+Hi li 1W11 +1++Hl1l S:T' I~. ...0 1 ·  CD 1.04 0. i . entrance length La= 0) . j ·i! •l : : 1:!i! l ittmttH~Jt!t1Hlmtttttlt*11U 1 11 :::: ' t' N 01 ' 1~ .01 0...0 Figure 46...L...j.... : .4 0. l .l 2.5 ~+H+rrHTI##~M4~~~~~~~~~ 'fr+""9 ._ ... j:::j::: . ..'1 !' 1 +llrrr:m:m:nmm1=11+1 :~' t i_ iHI·JJI·JHilil lttttj tH 1 :[j! !!i: :~: !::! 0 H±HjH !+Htttt !ffi 1t1 tHtHI IIHII!·IH!IIH!!!J 0..w.! 11 11l! ll lllillll 1 m til i 2.3.0 1 .3 0.0 ::l.2 0.011u:rn1... Wavelength for Helmholtz resonance (centered harbor entrance. b J.t. O..1 b/8 0. : :li J .03 0.8 1.ffii i'I W! i:l i~T :~ ri.. +:N ~ 1...! :..
Q the flow rate through the entrance channel. and Herchenroder (1977) present a numerical means for analyzing harbors responding to the Helmholtz mode of resonance. and Miyata (1971) suggest an approximate method for determining resonant wavelengths.1 2 ) Q2 . 46). 1971). L0 (L0 = 0). The governing differentiai equation is ~_ dt where 1 .Carrier.h ) .1 2 .g I (h . Correcting an error which appears in Carrier. Shaw. the resonant wavelength for the harbor with an entrance channel is then given by the equation (300) The resonant wavelength where L0 results (see Fig. Miles (1971) indicates that narrowing the entrance width or increasing the length of the entrance channel will significantly increase the response of the harbor to the Helmholtz mode.e. Shaw.9' ( . can be obtained. i. and Ab the area of the harbor (Ab= Lb B). Shaw. which may dominate tsunami response. The method uses a RungeKuttaGill technique where (301) hb is the surface elevation of the water in the harbor above sorne arbitrary fixed datum. Their method assumes that the resonant wavelength. Carrier.I F 2 A A g b s g be sc 1 (302) g (303) 143 .. This narrowing or lengthening also has the effect of decreasing the resonant frequency (Carrier. Harris. 45). Shaw.. and Miyata. which will be more generally applicable than equation (299). and Miyata. for an equivalent harbor of the same dimensions but having no entrance channel (L0 = 0). for harbors with entrance channels. 0 can be obtained using Miles' (1971) For a harbor with an entrance channel (Fig . Seelig. and Miyata point out that lengthening the entrance channel to a harbor also increases the frictional resistance so amplification factors for a very long entrance channel may be significantly reduced (although the resonant frequencies would still be less than for a harbor without an entrance channel. where L0 = 0).
5 Time (hr) Figure 47.y) ax 2 + a2 f(x. 47). Also. the peak water levels were slightly lower in this case. and F defined as the total bottom friction in the entrance channel. 1977). and Herchenroder (1977) (Fig. A80 the crosssectional area at the sea end.A is the crosssectional area of flow through the entrance channel at 0 any point X between the seaward end at X and the harbor end Xb 3 (A0 . 1971). 0.y) ay 2 + k2 f(x.y) 0 (305) 144 .Woter levet in boy 0. Harris.75 Inlet length =122 m Jnlet depth =7.25 ~ ~ 0 0. Tsunami water levels in a bay (tide excluded) (after Seelig.6 x J0Sm2 0. therefore.25 0. A sample computation for a tsunami entering a bay is given in Seelig. hs the height of the sea level above the arbitrary fixed datum. and Herchenroder. being a function of X). Aba the crosssectional area at the bay end. Harris. Miles (1971) found that he could transform his equations for waveinduced oscillations in a harbor to an integral equation equivalent to the equation formulated by Lee (1969.3 m Inlet width =24 m Bay oreo=4. It can be seen that the peak water levels in the bay occur slightly after the peak water levels just seaward from the entrance channel. Lee expresses the governing equations for wave oscillations in an arbitraryshaped harbor as (304) and a2 f(x.50 0.50 Tsunami at boy entronce .75 0 0.
!_ J 2so H(l) (kr) a [f 2 (o.y) is the known incident wave function.y) at the entrance.a(o.y0 ).y) in the open sea and the function f 2 (x.y0 ) and the interior point (x.y) by deterrnining the function f 1 (x. fp(x.e. The reflected wave function is deterrnined from the incident wave function for total reflection. The following boundary conditions are assurned: (a) f(x.y) within the harbor is defined by a line integral ! 8 taken around the harbor boundary in a counterclockwise direction giving f 2 (x.L = . Equation (305) is the Helmholtz equation. and fra(x.y).. n is (b) The harbor does not affect the wave system where (x2 + y2)l/2 + oo.y) the wave function for the wave radiating seaward from the harbor entrance.e.y) is a wave function to be deterrnined.y) the reflected wave function. y) (307) where HJ 1 ) is a zeroorder Hankel function of the first kind.y) at sorne position (x.y) and f 2 (x. The wave function in the open sea is represented by the surn of three functions (308) ) where fi(x. i.where z ag cosh [k(z + d)] cosh(kd) (306) and f(x.y)/an = 0 along all fixed boundaries where in the normal direction to the boundary. the wave amplitude and the slope of the water surface must be the sarne for f 1 (x. then rnatching the functions at the harbor entrance. Lee determines the value of the unknown wave function f(x.' at large distances from the harbor entrance.y ) ] d 0 an s (309) 145 . and r the distance between the boundary point (x0 . The radiated wave function is deterrnined as f.y) in the harbor. i. f 2 (x0 ...y0 the function at a boundary point (x 0 .y) . The function f2(x.
. by the 146 . The large increase in water · leve!. C of the Long Beach harbor Chen and Mei (1974) have developed a finiteelement numerical mode! which can be used to study water leve! oscillations in a harbor.. VII. 1977) applied Chen and Mei's mode! to studies of Los Angeles and Long Beach harbors.. Damage may be caused by strong currents produced by waves overtopping the structures.8 feet) are not uncommon. by the direct force of the surge produced by a wave. 7rrr~~~~~~ 6 5 . y is is Lee (1969) expressed equations (307) and (309) in matrix form and solved them numerically.Arbitrory shoped harbor theory • Experiment 4 2 6 Figure 48. Structures may be seriously damaged or destroyed by the tsunami. can impose powerful forces on shore protection structures and on structures located near the shoreline. TSUNAMI RUNUP AND INTERACTION WITH STRUCTURES The arriva! of a tsunami at a shoreline may cause an increase in water leve! as much as 30 meters or greater in an extreme case. Houston (1976. 1969). and x measured normal to the coastline.where the line integral ! 8 is taken across the harbor entrance. Figure 48 is an example of his experimental results for a small laboratory mode!_ of an arbitraryshaped harbor.. combined with the surge of the tsunami. measured along the coastline (across the harbor entrance). Response curve at point mode! (from Lee. Increases of 10 meters (32.
1. the vertical runup height at that point can be estirnated. the higher heights should norrnally be used for conservative design. must be estirnated.. but less than 1 mile to the east. Where tsunamis of a known height have produced variations in runup at a particular section of coastline. Hawaii. waves rushed onshore. the predicted runup heights may contain substantial errors. and by erosion at the base of the structure. MacDonald. 147 . Later waves were so violent that they destroyed houses and left a line of debris against trees 150 meters (500 feet) inland. It should be noted that the characteristics of the waves may vary from one wave to another at the sarne coastal point. i. After the height of the tsunami at a point along the shoreline has been deterrnined. the probable increase in water level caused by the tsunami. and Cox (1950) for Haena. An example of how extrerne this variation can be is given by Shepard. Variations in tsunami height and shoreline topography will actually cause sorne variation in runup characteristics along any section of coastline. and Cox (1950) cite a case in Hawaii where the first waves came in so gently that a man was able to wade through chesthigh water ahead of the rising water. To determine the potential damage to structures located along a shoreline. Alaska (Fig. Estimates of tsunami runup are also needed for flood zone planning along the shoreline. Shepard. the runup height may be assurned to be constant along that section of coastline as a first approximation. calculations for tsunami heights should be carried out at additional shoreline points between those points.e. must be applied to a sufficient nurnber of points along the shoreline to determine this variation. The nurnerical rnodels for prediction of tsunami height at the shoreline. i. the elevation of water at the shoreline due to the tsunami. with variations from about 5 ta 8 rneters (17 to 27 feet). 49).hydrostatic pressure created by flooding behind a structure combined with the loss of equalizing forces at the front of a structure due to extreme drawdown of the water level when the waves recede. When the variation is large between adjacent points.e. Tsunami Runup on a Shoreline. MacDonald. An exarnple of the variation in runup height is given by Wilson and T~rurn (1968) for Kodiak City. on the Island of Kauai. When the tsunami height along a section of coastline is relatively constant. and the variations in onshore topography are relatively rninor. the runup height. The mean runup height at Kodiak City was a little more than 6 rneters (20 feet) above mean lower law water (MLLW).. Major damage may also be caused by debris carried forward by the tsunami in the nearshore area. flattening graves of trees and destroying houses. The height of a tsunami will vary from point to point along a coastline. and for operation of the tsunami warning system to evacuate people from endangered areas. where there was a gentle rise of water level on the western side of the bay. Because these variations are difficult to predict.
The prediction of maximum runup heights would require the consideration of joint probabilities of tsunamis and storm waves. such an event did occur in Newfoundland in 1929 (Hodgson and Doxsee. tsunami.g. Storm waves riding on top the tsunami will have a wave celerity corresponding t o the depth (including tsunami height) at any particular point. the same period and wave 148 . and storm waves occurring simultaneously may appear to be small. 1968). Kodiak City. An . If two storm waves are otherwise equivalent (e. it causes an apparent variation in water depth over a long distance. which is an important consideration in computing runup heights..added complication. 1930). however. heavy line is maximum flood level (from Wilson and T0rum.Figure 49. 1964 tsunami runup. is the possibility of storm waves occurring simultaneously with the tsunami. Because a tsunami has a very long period relative to storm waves. as well as the probability of a high tidal stage. Alaska (contours in feet). The probability of a high tide.
at the head of a bay. California. and reached a maximum runup height equal to about 10. 149 . Hawaii. Magoon (1965) indicates that the 1964 tsunami at Crescent City had an elevation of about 6 meters above MLLW along a substantial length of shoreline. the tsunami intruded into a small inlet opening onto the main bay. and Cox (1950) refer to the Kalaupapa Peninsula. Wiegel (1965) reports that maximum runup elevations above MLLW at Crescent City. this assumption cannot always be used with accuracy. the tsunami can cause one storm wave to overtake and superimpose itself on another storm wave. the predicted value of runup equal to wave height at the shoreline compared well with the maximum runup measured in the area. and the possible convergence or divergence of the runup must be considered. and one is at the crest of the tsunami while the other is at the leading edge. the storm wave at the tsunami crest will have a higher celerity (U. equal to slightly more than 3 meters. on the Island of Molokai. 1960). MacDonald. At Ryoisi (Fig. the tsunami flowed directly up a canyon along a streambed. Therefore. The assumption was compared with a few cases where both height and runup data were available. 52).5 meters (34 feet). producing higher waves at the shoreline. but winter storms had left driftwood 6 meters above the same datum plane. A combination of a winter storm and a large tsunami could be very destructive.S. At Hongo (Fig. the runup height was actually somewhat less than the wave height at the shoreline.height). The results of Nasu (1934) indicate that the tsunami height at the shoreline and the runup height are dependent on the configuration of the coastline. Nasu's (1934) data for a tsunami occurring in Japan also indicate that the total runup was about equal to the wave height at the shoreline at many locations. 50). and 1964 tsunamis. and the maximum runup height was about 11 meters (36 feet). which included the 1960 tsunami at Hilo. where the 1946 tsunami left driftwood at elevations slightly more than 2 meters (7 or 8 feet) above normal sea level. 51). that produced a borefronted wave. Honolulu. Japan (Fig. and that the line of maximum tsunam. flowed up a canyon along a streambed and highway. inundation generally followed a contour at that elevation. on the north side of a bay. For those cases. were equal to or greater than the maximum wave height (cresttotrough) at the Crescent City tide gage for the 1952. The effects of ground slope. This assumption is based on the idea that a tsunami will act like a rapidly rising tide. While the assumption that maximum runup heights will equal the tsunami height at the shoreline provides an initial estimate. Shepard. wave period. Houston and Garcia (1974) assume that tsunami runup on a shoreline will have a runup height (vertical rise) equal to the wave height at the shoreline. Hawaii. At Kamaisi. Army Engineer District. 1960. The wave at Hongo and Ryoisi was probably traveling in a direction oriented directly along the axis of the canyons as the surge came onshore. Storm waves alone may be more severe than a tsunami at sorne exposed coastal points. The wave at Kamaisi was probably traveling parallel to the shoreline as it flooded into the bay.
Contour intervol in meters. Tsunami runup at Kamaisi. while the 1933 tsunami reported on by Nasu (1934) had wave periods of about 12 minutes. Iwasaki and Horikawa (1960) show that the period of the waves will be a major factor in determining maximum inundation levels. Smoller numerols ore ground elevations in meters (meosured from the sa me dotum) SCALE 0 50 100 200 MET ERS Figure 50. the orientation of the bays appeared to have no influence on the runup heights. Japan (after Nasu. The waves from the tsunami which struck the coast of Japan on 24 May 1960 had periods of about 60 minutes. the height of the 150 . 1934). For the longer period waves of the 1960 tsunami. heovier numerols ore woter heights in centimeters. and the water level gradually gained height over the entire surface of the bays where it was observed. The 1960 tsunami did not forma bore or a spilling front.
1934) o runup was equal to (or sometimes less than) the height of the wave at the shoreline. For the nonbreaking wave. Tsunami runup at Hongo in Toni..g. there are sorne cases where reflected waves become bores after reflecting from a shoreline. heovier numerols ore woter heights in centimeters 0 50 100 200 c=w+ IIETERS Figure 51. waves which break far from the shoreline and become fully develop ed bores before reaching the shoreline.e. the assumption that the runup height equals the wave height at the shoreline may be reasonable and possibly even conservative. 1934) indicate that the runup 151 . and waves which break near the shoreline and act as partially developed bores which are not uniform in height.SCAl€ Contour intervol in meters. Field observations (e. a tsunami which acts as a rapidly rising tide).. Tsunamis at a shoreline could be categorized into three types of waves: nonbreaking waves (i. In addition. Japan (after Nasu. Nasu.
01 indicated that the runup takes the initial form of a horizontal water surface at an elevation equal to the wave height at the shoreline (Fig. where maximum runup heights have been observed to be much higher than the wave or bore height at the shoreline..01. ...03. 53). The higher.. heovler numerols ore woter heights in centimeters Mi 50 ... . Results for plunging breakers on 4° and 8° slopes fronted by a slope S2 = 0. Tsunami runup at Ryoisi. . and that the higher runup on the slope washes up the slope at a shallow depth. indicated similar runup characteristics. 8°. and on 4°. height is sometimes less than that value. ICAU 100 200 Figure 52. 1934). Japan (after Nasu. shallow runup may cause sorne flooding. To analyze the runup of breaking waves and fully developed bores. but would not be expected to otherwise cause damage 152 . it is necessary to consider the actual form of the runup. Camfield and Street's (1967) experimental results for an 8° nearshore slope fronted by a slope S2 = 0. Using solitary waves. and 12° slopes fronted by a slope s 2 = 0. Contour intervol in meters.
Also. Freeman and Le Mehaute (1964) give a formula for surge runup as u 2 (1 + A) (1 + 2A) R s (310) 153' . Miller (1968). Sorne attempts have been made to develop theoretical solutions.5 5. Miller (1968) gives results for borelike waves which act as surge runup on a shoreline. Kononkova and Reihrudel (1976) studied the runup of solitary waves on uniform slopes which were apparently fronted by a horizontal tank bottom. For nearshore slopes greater than 8°. In general. the runup height increases as the slope increases.~w'"0 0.5 3.5 2." The experimental work of Camfield and Street (1967).5 1.0 Distance (ft) 3. fastmoving greatly elongated wedge. the higher runup on the steeper slopes appears to have a relatively shallow depth. This percolation tends to partially dissipate the shallow part of the runup observed on the impermeable model beach. and Kononkova and Reihrudel (1976) was for flat. their results were comparable to those of Camfield and Street (1967). and that the higher runup flows up the slope as a thin sheet. However. the form of the wave was of a thin.0 2. uniform slopes with no convergence of the wave crest. "ln the later stages of runup. the experiments show that for flatter slopes (less than 8°) the runup height appears equal to or less than the wave height at the shoreline.0 4. permeable beach.S=MI~~~~. For steeper slopes.0 Figure 53.5 4. and the ratio of runup height to wave height at the shoreline appears to reach a maximum value for vertical walls. they found runup values higher than the wave height at the shoreline. Solitary wave runup (from Camfield and Street.0 1. He shows that the runup in this case also takes the initial form of a horizontal water surface at an elevation equal to the wave height at the shoreline. Miller comments that. 1967). For nearshore slopes less than 8°. O'Brien (1977) points out that a fraction of the uprush percolates into a natural. shallow runup may not be representative of prototype runup. this higher. because of the shallow depth.
Freeman and Le Mehaute (1964) noted that coefficient A in equation (310) should be somewhat less than O. and using the maximum value for equation (310) reduces to R A. f.S.somewhat higher than experimentally obtained values. the relative runup would decrease as the roughness increases. This value is . as slope increases. hs = 1 + 6 C2s h 32g (313) As Ch varies with depth. 1968) obtained values of relative runup between 4.O from solitary wave experiments for breaking waves running up on a vertical wall. the relative runup R/hs = 6. Kishi and Saeki (1966) inèicate that the value of A decreases as the slope decreases. this equation would predict that the relative runup R/hs varies between a prototype and madel unless proper roughness scaling is used. as f 8g c~ (312) where Ch is the Chezy coefficient. Equation (313) does not consider the effects of wave period. Because Ch decreases with increasing roughness.S and S. the relative runup increases. Camfield and Street (1967. they obtain a maximum value of A= O. taking the friction factor. Using a value for us given by u s 1 . Also.where R the vertical height of runup above the stillwater level the current velocity of the surge at the shoreline the friction factor the ground slope gravitational acceleration a coefficient us f S g A Adapting the work of Keulegan (19SO). 154 . which is consistent with Freeman and Le Mehaute that the value of A depends on the form of the wave at the shoreline.S.(gh ) 1/2 A s (311) where hs is the surge height at the shoreline. As the slope approaches infinity (a vertical wall).
91 A2 s h 1 / s (316) 3 in metric units (the coefficient 0.02 in the footpoundsecond system of units). Freeman and Le Mehaute have carried out numerical calculations for slopes S ~ 0. R/h8 would have a calculated value of 2.01.2 foot) on a smooth aluminum slope. Camfield and Street's (1967) laboratory results for borelike solitary waves running up a 4° slope (S = 0.25 when S = 0. of the surge traveling up the onshore slope. If only the great~r water depths were coitsidered.3 for a value of h8 = 0. give a total relative runup R/H8 of 3. h.01 in equation (316). It should be noted that the above equations assume a uniform slope. For runup on a shoreline where the slope varies. 314 is multiplied by 2.67. Kishi and Saeki give a loglog plot for A versus S. the effect of wave period on the results was apparently not investigated.07.91 hl/ 3 s (315) in metric units.0699). and using a value of A= 0.22 for the footpoundsecond system of units). It is proposed equation (314) can then be written 0.03. and A= 0. the runup values of Camfield and Street (1967) include a shallow flooding which may not be an accurate prediction of prototype conditions.4 suggested by Kishi and Saeki (1966). This allows equation (310) to be rewritten as ~ = _1_ 2A 2 ( 1 + A) ( 1 + 2A) h 8 1 8gn2 + o.75 h 8 . and that a plot of h 1/ 3 versus h for 0 ~ h ~ h 8 will give an average value of h ~ 0. with A= 0. fronted by a slope S = 0. 155 . Very little data exist to verify such equations or to determine their full range of application.1.061 meter (0. but present no results for very flat slopes. Using a value of n = 0.04 when S = 0. then the measured value of R/h 8 "' 1.91 on the right side is equal to 2. Ch can be related to the Manning roughness coefficient n by (314) in metric units (the right side of eq. Noting that. Also. such as shown in Figure 53. Kishi and Saeki obtain similar results for rough slopes. which is close to the measured value. it would be necessary to use a numerical solution to determine the limits of the runup. As previously mentioned.It is also necessary to account for the dependence of Ch on the varying height. for uniform flow. O. Values of A were only obtained for that range of slopes.
It was shown that decreasing the Manning roughn. n. invalid case wguld be a nearshore bore or breaking wave. and the wave height. eut grass and pavement) could increase the distance required for dissipation of the surge by 160 percent (from 670 to 1.. However. Bretschnei der and Wybro also demonstrated that a bore would be dissipated faster than a tsunami acting as a rapidly rising tide. ~as limited application where 2. e. Their results compared well with the experimental resu lts of Street and Camfield (1966). and is invalid where h8 /H < 1. where h8 = 10 meters or 33 feet). h8 . Heitn~r (1969) developed a numerical method based on finite elements. 156 .ess coefficient. from n = 0.. Onl y very limited data are presently available for estimating values of the roughness coefficient n. Street. the "roughness" may consist of groves of trees or subdivisions of houses. it provides a simple means of investigating the effects of roughness on the limits of inundation. and Strelkoff (1969) and Chan and Street (1970a.76.9~4:.76.8 H Equation (317) indicates thatAthe higher values of relative runup.200 to 5. trees and houses. occur w hen the values of h8 /H are the 1owest. Although this is not entirely correct (the Manning relationship was developed for uniform flow). it is necessary to consider the drawdo wn of the water when the wave trough arrives at the shoreline.12. Replotting Spielvogel's results into a more usable form gives the equation ~ = _2~. and effectively includes the influence of the bottom slope and the wave period.015 (short. and the results depend on the choice of a bottomfriction factor and an artificial viscosity. Bretschneider and Wybro (1976) investigated the effect of bottom friction on tsunami inundation by using the Manning n to describe the roughness of th e onshore slope. 8  h (317) s 0. at the point where the leading edge of the wave is at the shoreline. he provides only limited results for simulating waves in laboratory channels. This relates the runup to the rate of shoaling just before the wave reaches the shoreline. In add i tion to considering wave runup. This latter.74 > h 8 /H > 2. R/h8 . Spielvogel (1975) developed a theoretical solution for tsunami runup based on !he wave or surge height at the shoreline.. may be moved by the waves.1 2 > h /H > 1. the roughness elements...g.770 meters or 2. Spielvogel indicates that equation (317) i~ correct for 3. 1970b) use a mod i f ied Marker and Cell (SUMMAC) numerical finitedifference technique fo r ca l culating the wave runup of solitary waves on a 45° slope and on a ver tical wal l . For prototype conditions.025 (long grass with brush) ton = 0. but their numerical method was not applied t o wave runup on the shoreline for flatter slopes. Chan .Th e solut ion of equation (316) is very dependent on a correct choice of the roughness coefficient .800 f ee t in the example used. Also. H.___ h .
. indicated sorne very narrow. or expose seawaterintak e pipelines.Not as much attention has been given to wave rundown . with very low velocity currents. and friction effects must be considered. 7 meters at a breakwater and wharf. the drawdown of the water level may result in the seaward collapse of seawalls. Most of the damage was caused by the violent withdrawal of the water. Houston and Garcia (1974) estimated that small tsunamis in southern California acting as rapidly rising tides would have maximum current velocities of about 0. drawdown will start at the shoreline.6 meters below normal sea l evel between waves (Shepard.5 meter per second . reversing the direction of flow). Hawaii. The quantity of water overflowing the dunes was insufficient in sorne instances to fill the low areas to landward. It should also be noted that a graduai increase in water level. waves at Hanamaulu Bay rose 2. destroyed sorne buildings when the water depth reached a height of 2 meters (6. The current velocity for the 1933 tsunami. As previously noted. rather than acting like a rapidly rising tide. large drain openings must be provided to prevent water levels from building up behind the barrier if it is overtopped by successive waves. and estimated current velocities with a maximum value of 1 meter per second . during the 1964 tsunami where water flowed over narrow coastal dunes. ho wever. may be followed by a sudd en withdrawal of water producing very strong currents. Consideration must also be given to the current velocities of the runup. and Cox. the currents associated with the rundown might have much higher velocities than the currents associated with the runup. result in damage to ships in a harbor. deep wave troughs with the initial troughs having greater amplitudes than the initial crests. Magoon (1965) cites one example south of Crescent City.15 feet). which was about doubl e the velocity estimated by Houston and Garcia for small tsunamis. however. California. The rundown elevations will depend on the wave train generated at the tsunami source. the tide gage record at Honolulu. 1950). but the water receded to a level 5 .e. During the 1946 tsunami in Hawaii. No estimates are available for the rundown currents. The barrier will also limit the height of the runup. as well as the quantity of water overtopping the barrier. reducing the resulting runup height. Japan. For the 1946 tsunami. Where the slope is very long compared to the wavelength. Ishimoto and Hagiwara (1934) investigated the large 1933 tsunami at Kamaisi. MacDona ld. it can be seen that for low velocities the retarding effect of the slope roughness (deceleration) may prevent the water from rising to a runup height equal to the wave height at the shoreline (i. 157 . Water overflowing a coastal barrier will have a current velocity determined by the difference in height between the top of the barrier and the ground level behind the barrier.
The receding water gradually widened the gap. Japan. et al. Keulegan (1950) gives u = 2(gh) 1/ 2 (318) where h is the surge height at any point and u the water velocity at the same point. However.83(gh) 1/ 2 (319) The higher value would be conservative. and that a quay wall constructed of reinforced concrete sheet piles at Hachinohe collapsed due to a lack of interlocking strength after backfilling was washed away. Fukui. Breakwaters and seawalls may provide coastal areas protection from tsunamis.000 feet) (U.S. When a tsunami occurs. which had been constructed to protect a section of coastline. A high seawall along a coastline may prevent flooding of the backshore areas. and it was swept away. Numerous instances of tsunamis damaging or destroying protective structures have been recorded. therefore. They indicated that a sea dike at Kesennuma Bay failed during the 1960 tsunami because the water from the incident waves. flush with the original ground level. Honolulu. care must be exercised in the design of the structures. and seawalls may reflect waves within a harbor. Proper placement of breakwaters may also decrease wave heights by changing the natural period of an inlet discussed in Section VI. Interaction with Shore Protection Structures. 1960). Iwasaki and Horikawa also indicated that receding water may seriously scour the seaward base of a revetment or seawall. The 1946 tsunami in Hawaii overtopped and breached the breakwater at Hilo. Matuo (1934) refers to the case of an earthern embankment at Yosihama on the northeast coast of Honshu. (1963) give a lower value of velocity as u = 1. The 1933 Sanriku tsunami overtopped the embankment. removing 7. breakwaters may decrease the volume of water flowing into a harbor and onto the coastline. breakwaters may also affect the resonant period of a harbor so that wave heights are increased. The combination of this 158 . 2.800 meters (6.25metric ton (8 tons) stones to a depth 0. They also noted that a quay wall at Ofunato failed because of scouring of the backfilling.~urge runup on a dry bed will have a much higher velocity than the values given by Houston and Garcia (1974) for a tsunami which acts like a rapidly rising tide. A tsunami may damage shore protection structures. which had overtopped the dike. Army Engineer District. 7. caused extensive erosion receding at a gap in the dike. Iwasaki and Horikawa (1960) investigated areas along the northeast coast of Honshu after the 1960 tsunami.9 meter (3 feet) below the water surface along nine sections of the breakwater crest with a total length of over 1.
and b t he top width of the breakwater in meters. where u is the current velocity in meters per second. Also.5 feet) below the level of the water surface at the time of arrivai of the tsunami. and Cox (1950) mention a case where water overtopping sand dunes eut a channel about 30 mèters wide and about 5 meters (15 feet) deep. Large stones from a seawall. Nasu (1948) developed sorne empirical criteria for the stability of breakwaters based on the geometrie shape of the breakwater.5 feet) and a period of 6 minutes. Japan. and Cox. Kaplan (1955) gives an empirical equation for the volume of overtopping of a seawall at ·t he shorel ine.2 meters (10. during the 1933 Sanriku tsunami.9 meter high were washed out and a gully about 3 meters deep and 27 meters wide was washed into a highway along the shoreline at Hilo. Richter.89b u 2 < __. Matuo (1934) reports on a dynamometer located on a breakwater at Hatinohe harbor. were carried inland (Eaton. 1961).scouring and the increased hydrostatic pressure from initial overtopping may cause failure. The dynamometer was located 0. The seawall was undamaged by the tsunami. Tsunamis will not always produce the maximum forces on a structure. Similar failures occurred along a highway on Onagawa Bay and along a quay wall at Kamaishi. weighing up to 20 metric tons (22 tons). The recorded maximum pressure was 38. a woodpilemooring dolphin was destroyed as a result of the loss of sand at its base. 54) collapsed seaward. Japan. extending inland about 18 meters.65(Kh s h) ü) 3 v (321) 159 ._V_ _ __ 0. he gives h + 0. For a breakwater with a seaward slope of 1:2. The damage from the 1960 tsunami in Hawaii is evidence of the erosive force of a tsunami. The concrete seawall along a highway between Hadenya and Mitobe on Shizukawa Bay (Fig.5 and a landward slope of 1:2. This equation can be rewri tt en as 21.0358 (320) for the condition of geometrie stability.76 meter (2. and Ault. California. MacDonald. hv the height in meters of the vertical segment of the face of the breakwater against which the current acts.. Magoon (1962) noted that approximately 2 meters of sand was scoured at the seaward toe of a steel sheetpile retaining wall at Crescent City. but a few months later storm waves destroyed parts of the wall and damaged the lower floor of the hospital (Shepard. 1950). in 1960 which contributed to its partial failure. Matuo (1934) mentions a concrete retaining wall which was overturned seaward by the 1933 Sanriku tsunami. MacDonald. A concrete seawall protected the buildings at the Puu Maile Hospital at Hilo during the 1946 tsunami. Shepard. Concrete seawalls 0.300 newtons per square meter (800 pounds per square foot) for a wave with a height of 3.
1'
(j\
0
Figure 54.
Concrete seawall destroyed by 1960 tsunami, Shizukawa Bay, Japan (from Iwasaki and Horikawa, 1960).
where V is the quantity of water overtopping the wall in cubic meters per meter or cubic feet per foot length of wall, h8 the wave height at the shoreline in meters or feet, hw the wall height in meters or feet, and
K
R
hs
(322)
where R would be the vertical height of wave runup on a similar wall high enough to prevent overtopping. Wiegel (1970) gives the following empirical equation for overtopping volume in cubic meters per meter length of wall V=0.287
h
t t2(1 . h
1 2
s
c o21Tt  h,_,)3/2 dt sw
T
(323)
where h 8 is the total wave height in meters (cresttotrough) of the wave at the shoreline, T the wave period, t 1 the point in time where overtopping begins, and t 2 the time when overtopping ends. As the wall height, hw, is measured in meters from the sea level at the time the tsunami occurs, it varies but its lowest value (i.e., the greatest overtopping) would occur when the sea level is at the highest tidal stage. Values for overtopping are shown in Figure 55 .
10,000 8,000 6,000
=
c;
~
J:
4,000 3,000 2,000
en
'0
~
e
c
~
...
~
0
1,000 800 600 400 300 200
~
'0
§ ~
.
100 0.1
0.2
0.3 0.4
0.6 0.8 1.0
3
4
6
e •o
Elevation of Tsunami Crest obove Seowoll, (ml
Figure 55.
Overtopping volumes Cafter Wiegel, 1970).
161
Based on stability tests carried out in a hydraulic madel, Karnel (1967) developed suggested breakwater sections for Hilo, Hawaii (Figs. 56 and 57). Allowable overtopping heights are given in Table S.
Table S.
Allowable overtopping heights (after Karnel, 1967). Allowable height of overtopping (rn)
Slope of harborside of barrier 1:2 1:2.5 1:3 1:3.5 1:4 1:4.5 1:5 1:6 1:7
9.1rnetric ton Armor stones 0.3 0.8 1.3 1.6 2.0 2.3 2.4 2.7 2.9
18.2rnetric ton Armor stones 1.0 1.2 1.5
Iwasaki and Horikawa (1960) show typical cross sections of seawalls at locations on the northeast coast of Honshu (see Fig. 58). In sorne instances, such as in fishing ports or harbor areas, it is undesirable to have high seawalls directly on the waterfront. The seawall at Yarnada (Fig. 59) is in two stages. A law seawall along the waterfront allows access to the water; a higher seawall, set back from the shoreline, protects the town from higher waves. Figures 58 and 59 show that the seaward toe of a wall is protected by rubble to deter scouring. Also, the area behind the top of the shoreline wall, such as at Yarnada, is paved to prevent saturation and erosion of the backfill rnaterial. The protection provided by a breakwater depends on its location and the width of the navigation channel through the breakwater. Iwasaki, Miura, and Terada (1961) ran madel tests for breakwaters in Kesennurna Bay. They discovered that a breakwater at the rnouth of the bay would substantially reduce wave heights in the bay for all wave periods tested. As expected, the greatest reduction in wave height occurred when the area of the breakwater opening was the least. When the ratio of the breakwater opening area to the crosssectional area of the bay was equal to about 0.1, the wave height was reduced to about 0.25 times the height which would occur without the breakwater. Surprisingly, when the breakwater was rnoved to the rnouth of Kesennurna harbor in the madel, at the inner end of the bay, the breakwater had alrnost no effect in reducing wave heights. The location of the breakwater would be expected to affect the resonant periods of the bay and the harbor. Therefore, care should be exercised in placing a breakwater in any bay or harbor.
162
SWOE
0, !XImHG AHO RfHAIIIUTATr:D SARi fR
fL tif FOR JO. 'l'OH STOHfS fL t:IO FOR »TOH STOHfS
HARBORSIDE
•
..J ..J
=
;:
~
IIJ IIJ
:::1!
0
z
..J
70
60
20
60
70
50
DISTANCE FROM CENTERLINE 1 fil
(Dimensions
ahown in feet)
HOT'E: CROSSHATCHEO AREA REPAESENTI
THE EXISTIHG BARRIER.
Figure 56.
Suggested design for rehabilitated breakwater, Hilo, Hawaii (from Kamel, 1967).
S1osid6
Horborsid6
i
_.s20lb tl2ton Stonu/
1
(Dimension shawn in feet)
Figure 57.
Suggested design, typical nonovertopping barrier section, Hilo, Hawaii (from Kamel, 1967).
16 3
16 4 . Dimensions in meters. 1960). Seowoll ot Toro Seoside b.Seaside o. Seawall cross sections (from Iwasaki and Horikawa. Seowoll at Yoshihomo Figure 58.
Hawaii. 1960). Other types of trees may be easily uprooted and flattened. Honolulu. 60) may withstand a tsunami surge but may be sheared off by debris carried forward by the tsunami.80 Figure 59. trees may offer sorne protection against a tsunami surge. Cross sections of seawall. Groves of coconut palms (Fig.WL:!:O ~ Section B +5. Yamada. Figure 61 shows a grove of pandanus 165 .50(1896) El. A seawall may cause waves to reflect into the harbor. Groves of trees alone or as supplements to shore protection structures may dissipate tsunami energy and reduce surge heights. It was determined at Hilo.S. that a seawall might aggravate surge conditions within the harbor (U.+ 4.Section A El +2 80 1 9 90 · H""all :. Dimensions in meters. Caution is also necessary when placing seawalls in a harbor area.L W · R•2$0 L. Army Engineer District. Japan (from Iwasaki and Horikawa. 1960). In sorne instances.
Coconut palms near shoreline. Hawaii (from Matlock. Hilo.t' 0'\ 0'\ ' Figure 60. and Matlock. 1962). Reese. .
and Cox. Damage from a tsunami may occur to structures located at the shoreline or along river channels near the shoreline. Hawaii (from Shepard. and Cox (1950) 167 . 62). a dock at Crescent City. MacDonald. 1968). destroyed a bridge over the Necanicum River and a railraad trestle over Neawanna Creek. MacDonald. Grove of pandanus trees knocked down by 1946 tsunami on the Island of Kauai. 1950). California (Fig. Oregon.Figure 61. MacDonald. uplifting a large lumber barge moored to the dock (Wilson and T~rum. Matuo (1934) calculated that trees could be broken by water velocities of 2 meters per second or greater. MacDonald. was damaged when the water elevation increased to 2 meters above the deck elevation. 3. Reid and Taber (1919) noted that palm trees were uprooted by the 1918 Puerto Rico tsunami. and Cox. He indicated that trees broken off by higher velocities may add debris to the surge and increase the damages resulting from the surge. Shepard. and may cause the entire structure to be swept away by the tsunami. Magoon (1965) indicates that a buildup of debris in front of a structure may increase its effective area. This would result in an increased drag force. Shepard. 1950). and Cox (1950) indicated that dense thickets of hau trees provided effective shields in many places during the 1946 tsunami in Hawaii. In 1964. The tsunami surge at Seaside. Other Shoreline Structures. but did not analyze specifie types of trees. trees which were knocked over in parallel rows by the 1946 tsunami in Hawaii (Shephard.
.v~ .. Coast Guard).S... Dock darnaged by 1964 tsunami at Crescent City.~"'"i~~~~~~ Figure 62.. California (photo by U.....l' 0"\ (X) .. .t " ·. .
When the tsunami forms a borelike wave. These velocities would be on the order of those normally associated with an incident surge. The current velocities associated with the surge are proportional to the square root of the surge height. and approximations of the current velocities can be obtained from equations (318) and (319).illustrate the damage to the railroad bridge on the Wailuku River (Fig . 4. 1.ight of a tsunami is discussed in Section VII. For a surge height approaching 5 meters. When the tsunami runup acts as a high velocity surge 169 . damage may result from the higher current velocities associated with the withdrawal. After the runup height of a tsunami has been established. and that this is followed by a flattening of the face of the surge on the dry slope. MacDonald. Iwasaki and Horikawa (1960) show a case of Mangoku. Miller (1968) noted from laboratory observations that a bore approaching a shoreline exhibits a relative steepening of the bore face just before reaching the shoreline. the estimated current velocity would be about 14 meters (46 feet) per second. the effects of this runup on structures and other objects located near the shoreline must be determined. and most initial damage will result from buoyant and hydrostatic ·forces and the effects of flooding. Shepard. Japan. 64) caused by the 1946 tsunami in Hawaii. the resulting incident current velocities are relatively low. The determination of the runup he. with equation (318) providing the more conservative result. More concern is therefore given to a tsunami which approaches the shoreline as a bore. 65) slumped almost ·1 meter due to the heavy scouring of the channel bottom. as different equations govern the motion and profile of the surge. In sorne instances. Tsunami Surge on the Shoreline. This surge shouZd not be confused with the bore approaching the shoreZine. and Cox (1950) noted that in many instances the withdrawal of the water occurred much more rapidly than the runup and flooding. When the tsunami acts as a rapidly rising tide. where a bridge support (Fig. the runup on the shoreline has the form of a surge on dry ground. 63) and the railroad trestle on Kilekole Stream (Fig.
Island of Hawaii (from Shepard. 1950). Hawaii (photo by Shigeru Ushijima. Tsunami damage t o r ailroad bri dge on Wailuku River.Figure 63. Figure 64. fr om Shepa:rd. 17 0 . 1950) . Tsunami damage to r ail r oàd trestle on Kolekole Stream. and Cox. Hi lo. MacDonald. ~lacDona l d. and Cox.
1960) .. Bridge damaged by 1960 tsunami at Mangoku.. Japan (from Iwasaki and Horikawa.J t' Figure 65. ..t' .
of water across the ground. This can result in cracking or collapse of a structure or wall. This leading edge has the appearance of an elongated wedge. (e) Hydrostatic forces caused by partial or total submergence of structures by the tsunami. For saltwater.99 poundseconds squared per foot 4). when added to other forces. (b) Surge forces caused by the leading edge of the surge impinging on a structure.026 grams per cubic centimeter (1. When water or water pressure intrudes under a structure. 172 . a. and the high velocity flow may cause severe erosion of the ground and damage waterfront structures by scouring material at the base of the structure. Vehicles and other large items may also be lifted up into the surging water. (c) Drag forces caused by the high velocity of the surging water. Buoyant forces are defined by the weight of the displaced water when abjects are partially or totally submerged. The buoyant force also increases as a function of surge height. so that a structure may be carried forward by the leading edge of the surge. or other material carried forward by the surging water. or may be destroyed in place if the surge force is high enough and the buoyant force is not sufficient to lift the structure from its foundations. These forces will displace buildings or other items in the direction of the current. will move a structure in the direction of the current. taking the density p = 1. five types of forces may result from the surging water: (a) Buoyant forces caused by partial or total submergence in the surging water. Calculations to predict tsunami runup have indicated a probable surge depth of 2 meters at that location. and the force of the surge on a structure gradually increases as a function of the increase in surge height.422 square feet). * * * * * * * * * * * * * * EXAMPLE PROBLEM 21 * * * * * * * * * * * * * * GIVEN: A house occupies a floor area of 225 square meters (2. Buoyant Forces. It is assumed that the flow of water will be at a constant depth around the house. (d) Impact forces caused by buildings. where the water level is relatively constant. These forces may either destroy other structures on impact or create momentum which. the buoyant force is FB = pg V (324) This assumes water intrudes where V is the displaced volume of water: under the structure. boats. the buoyant force tends to lift the structure from its foundations.
026 kilograms per cubic meter (9. 5.5) x F = 6.81)(~)(6. and the force holding the tank in place. so that F = 3.000 poundmass).53 x 10 6 kilogrammeters per second squared 4. M.49 105 pounds) 173 . and that it is filled to a depth of 2.2 feet) with oil having a specifie gravity of 0.An empty oil storage tank is 3 meters high and 6. The buoyant force is given by pg SOLUTION: v 1.88 (density p = 880 kilograms per cubic meter).1 meters in diameter.180(9.81 meters per second squared) (225 square meters) (2 meters) 4.62 105 newtons (1. The tsunami water depth is 1.29 x 10 5 newtons (1.FIND: The buoyant force on the house.53 x 106 newtons (1.180 kilograms (7.81) x + 880(9.19 x 105 pounds) (b) The force holding the tank in place is the mass.8 meters.02 x 106 pounds) * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * EXAMPLE PROBLEM 22 * * * * * * * * * * * * * GIVEN: .5 meters (8. Assume that the tank has a mass of 3. FIND: (a) (b) SOLUTION: (a) The buoyant force is given by The buoyant force on the tank.1) 2 (2. of the empty tank multiplied by g plus the volume of oil multiplied by its density and g.29 x 105 kilogrammeters per second squared FB 5.
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * In Shepard. and minor damage from the breaking of water pipes and electric lines. the current velocity associated with the waves is very low. and Cox's (1950) discussion of the 1946 tsunami in Hawaii. the fact that sorne buildings had weak structures and broke apart when lifted from their foundations. so that the major damages are similar to those discussed above. where tsunamis act like rapidly rising tides. In many instances. and sorne houses could be moved back to their original foundations with very little repair work required. leading edge.. from the leading edge of a surge impinging normally to the wall could be given as F (325) where F is the force in newtons per meter of width. MacDonald. Shepard. and Cox (1950) mentioned instances of people wading through chesthigh water to escape from the tsunami. MacDonald. a house at Kawela Bay on Oahu was floated off its foundation and deposited in a canefield 61 meters inland. Damage caused by buoyant forces was the result of buildings being deposited on uneven ground. Surge Forces. leaving breakfast cooking on the stove and dishes intact on shelves. the buoyant force could overcome the mass of the tank and the oil plus the strength of any structural anchorages. Cross (1967) showed that the force per unit length of vertical wall. u the surge velocity in meters per second. With a lower level of oil in the tank. and Cp a force coefficient defined by CF = (tan 8) 1 • 2 + 1 (326) where 8 is the inclination of the water surface of the surge shown in Figure 66. b. and b given by b Z the distance from the . Many other houses were also gently floated from their foundations.!_ du _ S g dt (328) 17 4 . tan 8 is given by the equation tan e (327) where Ch is the Chezy roughness coefficient. h the surge height in meters.It can be seen that very little reserve force remains to resist drag forces from the surge.
and du/dt the acceleration term for flow under the tip of the surge.. where S is the bottom slope (negative upward).= gS dt (330) i.. and (328) into equation (325) F = .e.!_ 2 pgh2 + [(~ + . (327). the acceleration is equal to the influence of gravity acting along the slope (see Fig.!.. defining u from equation (318) and assuming that the value is relatively constant under the surge (332) where Ch varies with depth and is given by (333) 17 5 .9 cos 13 S=gS ( for smoll 13 ) Surge Direction ~ g on {3 Figure 66. Definition sketch of surge on a dry bed (slope exaggerated). ·z + 1] (329) If it is assumed that du . then b ~ 0 and (331) or. Substituting equations (326). du C~h g dt _ s\ ) 1. 66).
and (334) into equation (329) for metric units and collecting terms (335) It must be remembered that many approximations have been used in this solution. n is Manning's rough Substituting equations (330). or (334) in the meterkilogramsecond system of units.)l/ + 1] 100.5 meters impacts normally against the vertical side of a building.033 h2 + [(O~~.81(2." so that u = 2~ = 219. The Manning roughness coefficien~ n = 0.l) 2 ) • h1/3 1 2 + (1 026)(9. It may be noted that when e = 0.in the footpoundsecond system of units. SOLUTION: The surge velocity is determined from the maximum surge height. u = 2~. The coefficient Cp accounts for both inertial forces and drag forces.. and the velocity remains constant. remains relatively constant and that the surge profile remains constant. * * * * * * * * * * * * * * EXAMPLE PROBLEM 23 * * * * * * * * * * * * * GIVEN: A surge with a maximum height of 2.1.5) = 9. (332).560 h 176 . and ness coefficient. the force F is simply the hydrostatic force plus a drag force (per meter width) where Cv = 2.9 meters per second The surge force is given by equation (335) as F F =! 2 =! 2 pgh 2 + [(~) ' h1/3 1 2 + 1] pu 2 h 4 ( 9 · 81 ) (O. + [( 2 1] 1.9) 2 h F= 5. and it is assumed that the surge velocity.026(9.8l)h 2 . FIND: The surge force per meter of building width as a function of sürge height.
Wilson and T~rum (1968) report on the case at Seward..e. c. This indicates that the calculated value is conservative for design purposes. F.0 270. The water high.026 grams per cubic centimeter 1.900 2. it is possible that the surge force structure before the buoyant force lifts it into the flow. and the force of the surge will then carry these abjects forward. Alaska. acceleration is negligible and its effects can be ignored. * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * As indicated by example problem 23 and shawn in Figure 66. Therefore.5 203. and the buoyant force may destroy a velocity near the leading edge of a surge is relatively height of the leading edge is relatively law (i. of a tsunami surge overtaking a pickup truck being driven from the shoreline. The buoyant force of the leading edge of the surge tends to lift abjects into the surging water. Drag Forces.. CM. the force. although this change in water level appears to occur rapidly with respect to time because of the forward velocity of the surge. 1. depending on the body (Table 6) the projected area of the body normal to the direction of flow in square meters the velocity of the water in meters per second CD A u 177 .0 138. the is law). The truck was swept up by the surge and carried forward like a surfboard into nearby woods. is tabulated below: 1.300 F. then it can be assumed that the inertia or mass coefficient.e. If the velocity is assumed to remain relatively constant under the surge. A surge on a dry bed has a much flatter front than a bore approaching a shoreline.5 73.For various values of h. meters newtons per meter 0. approaches zero so that the drag force in newtons is (336) where p the density of seawater = 1.800 2.100 NOTE:Calculations will show that Cp> 1 at the maximum surg~ height (where the rate of change of surge height ~ 0). i. This is seen in laboratory tests.500 h. It can be seen that the hydrostatic pressure component of the force is a relatively small part of the total force.026 kilograms per cubic meter a coefficient of drag.5 339. The velocity of the water in the surge produced by the tsunami runup creates a drag force which tends to move a structure in the direction of the surge. there is a graduai rise in water level at the front of the surge.
Object Drag coe ff' . 1c1ents.0 1 NOTE.1 1.5 2. The projected dimension shown.Table 6. or the length of the flat plate.2 1. L/d Reynolds number cv Circuler cylinder 1 5 10 5 10 5 10 5 >5 x 10 5 0. 20 0.33  OJ 00 00 Square cylinder 0] 00 3. or the width of the flat plate.6 Rectangular flat plate ( totally submerged) 1 5 20 00 >10 3 >10 3 >10 3 >10 3 1.5 x 10 4 2.74 1.63 0. 17 8 .0 <>à t 00 10 4 to 105 1.L d d _l_ The height of a submerged cylinder.
For a situation in which there is either overflow or underflow~ the coefficient of drag can be determined by using an approximation. the ratio of lengthtowidth for the structure should be determined. The coefficient of drag. the coefficients of drag. p is in units of poundseconds per foot 4 . The tsunami surge has a depth of 2. is dimensionless and retains the same value as in the kilogrammetersecond system. 5 meter·s (11. u 2 /(gd). 17 9 . giving a wetted height on the structure equal to 1. Assume that the depth of flow around the structure is twice the actual depth.3 feet). the area in square feet. where d is the projected horizontal dimension of the structure transverse to the direction of flow. existing tables must be used to establish maximum coefficients to Table 6 gives examples of drag coefficients. FIND: (a) The coefficient of drag of the structure. 5 feet) high and supported on columns so there is a 1.velocity in feet per second. In cases where there are overtopping and underflow. Therefore.9 feet) open spate under the base of the structure. Where this parameter approaches unity there are strong unidirectional freesurface flow effects. Individual model tests would be required to determine a more exact interaction between the tsunami and the structure. In that case. An example of this type of calculation follows: * * * * * * * * * * * * * * EXAMPLE PROBLEM 24 * * * * * * * * * * * * * * GIVEN: A flatsided structure is 14 meters wide and normal to the direction of flow. CD. given in Table 6 may be too low. 67). This ratio should then be used for determining the coefficient of drag. Then obtain a coefficient of drag as if there were both underflow and overflow (see Fig. and the. flow around a vertical cylindrical storage tank would be treated as flow around an infinitely long cylinder in order to obtain a drag coefficient. the flow may be treated as flow around an "infini tel y long" structure where the ground and the free surface define the boundaries of a layer of fluid. and (b) the coefficient of drag of a similar structure located at ground level with no underflow. Tabulated values freesurface flow at of drag coefficients ensure safe design.5 meters. of drag coefficients are generally not available for high Reynolds numbers.5meterhigh (4. Hallermeier (1976) discusses the importance of the parameter. and that the height of the structure is equal to twice the wetted height. The structure is 3.To determine the drag force in pounds.0 meter (3. and where there is no underflow. For cases where flow does not overtop a structure. CD. For example.
. CD 2..0.. v J Flow Direction h' w ___1 ! l Tsunami surge flowing post elevated structure h hw L = surge height = wetted height on structure = length meosured perpendicular sections shawn above and below to the L/d = L/(2hwl ~11 Flow Direction 1 .2hw.25.. the coefficient of drag CD~ 1. From Table 6. 1 Equivalent body used for determining coefficient of Drag Co Flow is ossumed symmetricol about the doshline Figure 67... that the structuré is equivalent to a structure 14 meters wide and 2. From Table 6. * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * 180 . This corresponds to an infinitely high s tructure where L/d = oo. SOLUTION : (a) It can be assurned. (b) The structure is higher than the depth of flow so there would be neither underflow or overflow. for purposes of deterrnining the coefficient of drag.. 67).0 rneters high with both underflow and overflow (Fig.. for a flat plate normal to the flow direction where L/d = 7.....2h j Determination of CD when flow passes under a structure.
91 Z + (1. the coefficient of drag can be determined by doubling the wetted height and assuming both underflow and overflow for a flat surface 1.5) FB 3.5 feet) 181 .24 1.83 .026(9.17 x 105 pounds) which will act against the side of the locomotive at a distance.83.83) = 8.* * * * * * * * * * * * * EXAMPLE PROBLEM 25 * * * * * * * * * * * * * (This example is taken from an actual situation which occurred at Seward.0.83 meters The velocity can be obtained from equation (318).5) ( 8 · 47 2 ) 2 F x 105 newtons (1.52 meters (5 feet) and the width of the locomotive body was 3.700 pounds) As indicated previously. Interpolating in Table 6 for a flat plate for L/d = 6.91) 2 1. The clear space under the locomotive was approximately 0.8 gives CD = 1. see Wilson and T~rum.81(1.83= 5.0.5metric ton (230.83 meters.53 x 10 5 newtons (78.5 meters (41 feet). z = 0. The surge was assumed to have a depth of 1.24)(1. given as Z.47 meters per second From equation (336).24 0.05 meters (10 feet). above the ground. 1968.) GIVEN: A 104.05) (12.91)(12.000 pounds) railroad locomotive was overturned by a tsunami surge. so for h u = 2~ = 219. FIND: The overturning force on the locomotive.91) (3. the drag force is F D D = 1.5 meters long. The surge was assumed to act normal to the side of the locomotive. in 1964. The buoyant force is given by equation (324) as pg SOLUTION: v 1.37 meters (4.026(1.81) (1.91 meter (3 feet) and the length of the locomotive body was 12. Alaska. The width between the rails was 1.83 meters high and 12.
44) (0 .F(O. Cv = 2. From equation (318) u = 2~ = 219. 76) 3.07 x 10 5 newtonmeters (1.53 x 10 5 (0.44 meters (8 feet) under the platform. 10 6 newtons (2. The surge acts normal to the sides of the columns. is supported by square columns with 14.24 x 105(1. A tsunami creates a surge with a depth of 2.36 x 10 4 newtons 182 .026(2) (2 . the locomotive will be overturned. Therefore.025 x 106(0. times gravitational acceleration.37) . 3 meters above ground level.The downward force from the mass of the locomotive is the mass. or F F mg= 104.5 inches) cross sections.500 kilograms (9. * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * EXAMPLE PROBLEM 26 * * * * * * * * * * * * * GIVEN: A platform. the center of mass of the locomotive is equidistant from the two rails.5 feet) from the rail . which are rigidly fixed at ground level.3 x 10 5 pounds) Taking overturning moments about a rail.44) = 9.48 x 10 5 footpounds) indicating that the overturning moments are greater than the restraining moment.76 meter (2.1.76) 2.5 by 5. FIND: The moment of the surge force about the base of a column. 76) + FD Z .81 meters per second squared) 1. and from Table 6.by 14centimeter (5. The buoyancy and drag forces produce overturning moments (+) and the mass of the locomotive a restraining force(). the columns may be considered as infinitely long columns.79 meters (32. Summing moments M M M FB(O.0. or 0.76) + 5.1 feet) per second The drag force on a column is given by equation (336) as Au P CD Z = 1. SOLUTION: To determine the coefficient of drag.025 x m.81(2.14) 2 ( 9 · 79 ) 2 2 3. g.
d. trees.36 x 10 4 (1. and Cox (1950) noted numerous instances of severe erosion caused by the 1946 tsunami in Hawaii. Other instances of erosion were also noted. Impact Forces.22 meters (4 feet) above ground l evel . The drag forces can be lessened by constructing a building on an elevated platform sorne distance above the ground. and Cox noted that dense stands of grass prevented or greatly diminished ground erosion during the 1946 tsunami in Hawaii. and Cox (1950) in Hawaii.1 x 10 4 newtonmeters (2. and by Reid and Taber (1919) in Puerto Rico. the backwash from the tsunami undermined a raad. Conversely. this may be an expensive solution and has the undesirable feature of adding debris to the water. The buoyant forces can lift buildings from their foundations. Imamura (1942) gives an example from 1707. At Kalaupapa. the first floor of a building may be designed to be carried away by the tsunami. The high velocity of a tsunami surge can also damage structures by scouring material near the structures' foundations.44meter depth so that the resultant drag force acts 1. The moment is then M 3. In higher latitudes.93 x 104 footpounds) on each of the columns. However. This material may include automobiles. Shepard. was eut back about 21 meters (70 feet). and the surge or drag forces can slam them into such things as trees or other structures. A large boulder moved by the 1960 tsunami at Hawaii is shawn in Figure 68. The high velocity of a tsunami surge will sweep large quantities of material forward with the surge. the material may include large quantities of broken ice. thereby reducing the forces on the building and protecting the higher floors. or other materials in the path of the surge. MacDonald. Buildings that are firmly attached to their foundations to resist the buoyant forces must also have sufficient structural strength to withstand the drag forces acting against them. MacDonald. when a tsunami washed away layered sediments which had covered an old ricefield.22) 4. Erosion and deposition of surface material are quite common when severe tsunamis occur. A section of shoreline at Moloaa.2 meters deep across a highway. the Earthquake Research Institute (1934) reported instances of ricefields being covered with sand by the 1933 Sanriku tsunami. * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * As indicated in example problem 25 drag forces and surge forces can act in conjunction with buoyant forces. At Haena Bay. In sorne instances. Instances of deposition of sand are also indicated by Shepard. buildings.The velocity is assumed to be equal over the 2 . Shepard. petroleum tanks. a sand beach eroded and sand was deposited 1. 183 . debris from buildings. MacDonald. when tsunamis occur during the winter.
Hawaii (from Matlock. Reese. .r~ ~ ·.!~ ~1 ... Hilo. 1962) . and Matlock.. ! t' 00 ~ ··~ Figure 68.~ . Large boulder moved by 1960 tsunami.
with respect to the ground is assumed to be constant. the velocity of the abject approaches sorne terminal velocity).. u.Impact forces from material carried forward by the current are not as easily analyzed as other forces. However. the velocity of the fluid. and is given by the equation F (337) where the coefficient of drag p the density of water the crosssectional area of the abject transverse to the direction of motion the velocity of the water with respect to the abject at any instant in time the inertia or mass coefficient the volume of water displaced by the abject A t ti me 185 . the impact force is spread over a wide area. of the fluid after the abject has moved sorne distance (i. the water within the structure becomes a part of the mass of the structure. the ratio of the drag force to the actual mass of the abject in motion. if a corner of the first structure impacts with the flat side of the second structure. Considering an abject being swept forward from a stationary position by a moving fluid such as a tsunami surge. The veloci ty. u. ub. If the flat side of one structure impacts with the flat side of a second structure. with respect to the ground varies as the abject is accelerated. u'b. The force accelerating the body is a combination of drag forces and inertia forces. of the obj ect approaches the veloci ty.e. The velocity of forward motion of such material depends on the distance the material has moved. the force is concentrated and there will be a greater tendency to crush the impacting structures. and the resistance created by the abject dragging against the ground or impacting and grinding against other abjects. and the velocity of the abject. It should be remembered that if a structure is partially flooded. Analyzing the effects of a structure impacting with another structure also requires knowledge of the rigidity of the structures and the angle of impact. The drag force will initially accelerate material which is swept up into the current.
For a structure or any other large object floating in the water, the mass, rn, of the object is equal to the displaced mass, pV, of the water. This mass may vary as wa~er gradually floods the interior of a structure, but for the analysis presented here the mass will be assumed constant. From Newton's second law
F
(338)
At any instant in time the magnitude of the deceleration of the fluid wi th respect to the object is equal to the magnitude of the acceleration of the ground with respect to the object (which is equal to the acceleration of the object with respect to the ground), i.e., where u is assumed constant,
d~
dt so equation (337) becomes
F
2
(339)
=
dub
pV
dt
= CD
(u  ub) pA      C 2
dub
pV 
M
dt
(340)
or, rearranging terms,
d~ =
dt
(341)
For an object moving a short distance, the coefficients CD and CM will be assumed constant. This is not entirely correct (e.g., the value of Cv will vary as a function of velocity), but will be assumed as approximately correct for a short distance. A constant, a, can then be defined by
a=2V(l + CM)
Cif
(342)
Substituting equation (342) into equation (341) and rearranging terms give
a dt
=
d~
(u  ~) 2
(343)
Integrat ing equation (343),
1 u~
1 u
(344)
186
whi ch reduces to
ub
=u
u
 cmt + 1
(345)
t.
defining the velocity of the object at any time
The distance, x, traveled by the object as a function of time can be determined by noting (346) Substituting equation (345) into equation (346) and integrating give (34 7) which gives x
= ut  lCl fu (aut
+
1)
(348)
Typical drag coefficients are given in Table 6. The coefficient of added mass, CM, can be estimated for a rectangular structure by using the results of Riabouchinski (1920) as given by Brater, McNown, and Stair (1958) (Fig. 69). The values in Figure 69 are for irrotational flow without separation, and the formation of a wake behind the structure would be expected to modify these values. Individual madel tests would be required to obtain exact values. Example solutions of equation (348) are shown in Figure 70.
0.2
Figure 69.
0.4 0.6 0.8 1.0
o/b
2
4
6
8 10
CM for twodimensional flow past rectangular
bodies (irrotational flow with no separation) (from Riabouchinski, 1920).
187
Ois tance,
1
(ml
Figure 70.
Example plots of x versus
t
for abjects moved by tsunami surge.
* * * * * * * * * * * * * EXAMPLE PROBLEM 27 * * * * * * * * * * * * * *
GIVEN: A tsunami surge is 5 meters high .at the shoreline. A building rüeated at the shoreline is swept forward a distance of 6.1 meters and impacts with another building. The building is rectangular, 12 meters (39.4 feet) wide and 6 meters (19.7 feet) deep in the direction of flow, and is submerged to a 3meter depth as it is carried forward (see Fig. 71). The velocity of the surge is approximated as u = 14 meters per second. FIND: (a) The time required for the building to impact with the other building, (b) (c) SOLUTION: (a) The submerged crosssectional area of the building, transverse to the direction of the surge, is given as A= width
x
the force accelerating the building at the moment of impact, and the momentum of the building at the moment of impact.
submerged depth = 12.0
x
3.0
= 36
square meters
and the submerged volume (the displaced water) is V = width x length x depth = 12.0 x 6.0 x 3.0
= 216
cubic meters
188
.,.,.,.,
Building set into motion Building moving Building impocts with r"' \ second build,..in...;g;.__ _i\
\
...
\
\
\
F
F
F
\
\
\
\
t = 0 x = 0 F = 9.12xlos newtons Mo=O t = x = F = Mo= 1. 0 second 1. 7 meters 5,45xl0 5 newtons 7.05x10 5 newtonseconds t = x = F = Mo=
\
\
2. 02 seconds 6.1 meters 3.59xl0 5 newtons 1.16xl0 6 newton·seconds
u
= 14
meters/second
Cv= 1.13 CM= 3.5
A "' 36 meters 2 V = 216 meters 3
Figure 71.
Building moved by tsunami surge.
The coefficient of drag can be approximated by assuming the side of the building is a flat plate. To determine an equivalent flat plate using Table 6, assume that the submerged depth for underflow and overflow (a totally submerged plate) is twice the depth of the building, or
.!: =
d
12.0 2 x 3.0
2.0
and from Table 6
From Figure 69, where
b = 12.0
th en
a
~=
°' 5
and equation (342) gives
1.13 x 36
0.021
2 x 216(1 + 3.5)
189
Buildings and larger tanks were able to withstand the force of the tsunami. The impact forces either destroyed the loadcarrying capacity of walls. which gives. however.021 x 14 x 2. automobiles.The relationship between distance and time is shawn in Figure 70.16 x 106 kilogrammeters per second (2.22) 2 3. 105 newtons (8. =~ x mass taking the mass of the building equal to the mass of the displaced water for a partially submerged building which is floating (the mass includes water within the building).59 M 0 F (c) x .02 seconds (b) From equation (345). for x= 6.021(14 .5. ~ ~ = u  u aut + 1 14  14 (0. and baled lumber.1 at impact is M 0 x 10 4 pounds) Momentum.1 meters. F F pVa (u . or caused bending or breaking of light columns.1 feet) per second Substituting equations (342) and (341) into equation (338). mass = pV = 1.026 x 216 x 0. Wilson and t~rum (1968) discussed some instances of impact damage resulting from the 1964 Alaskan tsunami. 02) + 1 5. Figure 72 illustrates the damage at the Union Oil Company tank farm at Whittier.59 x 10 5 kilogrammeters per second squared 3.~) 2 1.56 105 poundseconds) * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Magoon (1965) indicates that substantial damage occurred at Crescent City during the 1964 tsunami as a result of debris impacting on structures. This debris included logs. smaller 190 .22 meters (17.026 kilograms per cubic meter x 216 cubic meters M ub mass x and the momentum is 0 x = 5.026 x 216 1.22 x 1. t = 2.
..... "1 / \ / \ \ \ / 1 1 ...IIIIIl....s..IUlMQUSf COMP\. Alaska (from Wilson and data from Union Oil Company of California).4BOBBL.... \ \ '... '.. SWEPT IWAYCOIIIPL...(f( IU1t'0UT ..· · . Damage to oil tank farm at Whittier.. ~\ \ \ \ 1 · 1 ... ._... •< " ·~.LA(f u GAA&G[ 1 1 1 1 : •u•OE~~t':f ~g....:······r='··~ · ··~······ . .(lt H()~$[ • CC..····. NOult~ P\.. 1 \ : // '.. 1968..... . \ \ 1 1 UI\III.ETf LOSS \ 1 : 53.. / .___j CALir(..~8~i. .Y 1 COOII[P&Gf & lllllllii[MC:JSf COMPLET[ IUANOUl lllf IlL AC[ 1 0 '• UfoiO<.. ... ~ \ 1 1l' . ____ . ..~~::~ •..Atf 1 t' \.... }9}6 ~· '\ 1 D Pu..A~IA 1 Figure 72.0 t' D D 801\..· ·. .'·n ~110P(It1't lUSC!I ....&MAG(00.ALASI(~... / /... .CAJ(8[A[PAIA(0 \ 8\IR~EO ·~PLACE 4106 ! / 1 1 \ \ Buo•l5 ~...11'1 .•.. T~rum. .....LACE 53 t 480 BBLS • SCAAP \ 1 \ \ \ \ 1 / 1 1 \ \170 FEU WHAA. ~:A... CCY•A•n OF !< . 01\.... V'M~lt !itL CG...
destroying the end of the dock. per foot width of wall. There was an unusual occurrence at the abandoned Kahuku Airfield on Oahu during the 1946 tsunami (Shepard. Van Dorn (1965) notes that harbor regulations could be instituted requiring ships large enough to damage harbor structures to stand clear of a harbor in the event of a tsunami warning. assuming that water levels are equal on all sides of the structure. the hydrostatic force would probably not exceed 10 to 20 percent of the drag force at higher water levels. adding substantially to the damage. T~rum e. for a water depth h is F (349) As seen in example problem 4. Sorne of the larger tanks were apparently set into motion by the impact. Wilson and T~rum also mentioned the problem of a smallcraft harbor located immediately in front of a developed shoreline at Kodiak City. 1950). large hydrostatic forces on the wall may result. An interesting example of impact forces is reported by Wilson and (1968). Hydrostatic Forces. and most of the tanks were ruptured. and Cox. However. A resulting fire destroyed the tank farm. and would appear tq be relatively insignificant at lower water levels. The hydrostatic force on a wall. the hydrostatic force will not contribute to the motion or potential motion of the structure. and rammed the dock. (Spaeth and Berkman. During the 1964 tsunami a house was washed out to sea near Point Whiteshed. Magoon (1962) indicated that the flooding caused by a tsunami can saturate the fil! behind a retaining wall. The house was swept more than 12 miles along the shoreline. California. Hydrostatic forces are normally relatively small compared to surge and drag forces. Once the initial surge has passed a structure. this force can cause cracking of exterior walls and interior flooding of the structure. However. Blacks 192 . Combined with the large drawdowri of the water leve! which may occur at the seaward toe of a wall during the withdrawal of a tsunami wave.tanks were carried forward by the surge and impacted with other tanks. 1972). It is believed that this contributed to the partial failure of a retaining wall at Crescent City. MacDonald. In the case at Kodiak City there was only about 30 minutes between the tsunami warning and the arriva! of the first large wave crest of the tsunami. when tsunamis are generated from distant sources there may be enough time to clear the harbors. The boat harbor contained a large number of fishing boats and yachts which were carried into the adjacent waterfront business area by the tsunami. carried into the harbor at Cordova.
NOAA). Tsunami detectors were designed and installed at tide stations to alert personnel of forthcoming tsunamis. apparently as a result of hydraulic pressure from water penetrat ing into the sand under the pavement when the tsunami flooded the area. Coast and Geodetic Survey (now the National Ocean Survey. The Japan Meteorological Agency organized the present Japanese system in 1941. 19 3 . Meetings to discuss implementation of the warning system throughout the Pacifie coastal areas were held in 1948. (b) effects of flooding. 5 meters (3 to 5 feet) across. informai warning systems operated sporadically for centuries. (c) fire and explosion from the impact and rupturing of petroleum tanks or containers of chemicals (see Fig. Initially. and contamination of water supplies with saltwater. the warning system supplied information to civil authorities in Hawaii and to military bases throughout the Pacifie. Civil Defense Agencies of California. shorting of electrical lines and transformers. The ne ed for a warning system in the United States was recognized following the 1 April 1946 tsunami generated in the Aleutian Islands. VIII.S. Oregon. Spaeth and Berkman (1972) discuss the early history of the seismic sea wave (tsunami) warning system in the United States. additional hazards should be considered in addition to the actual forces of the surge. National Oceanic and Atmospheric Administration. 72). particularly at Hilo. and (d) release of poisonous gas or toxic materials from ruptured containers. That tsunami caused heavy damage and resulted in the loss of many lives in Hawaii. In 1953. Sorne of these are listed below: (a) Contamination from debris carried in the surging water. The higher pressure under the pavement has not been explained.of pavement were tilted in circular areas 1 to 1 . and a formai tsunami warning system was recommended as early as the late 19th century. The present tsunami warning system was organized by the former U. TSUNAMI WARNING SYSTEM AND INSTRUMENTATION Cox (1964) discusses the development of the tsunami warning system in Japan. The first detector was installed at Honolulu in 1947. and the system has expanded since that tirne. The tentative communication plan for the warning system was approved in 1948. including spoilage of goods and materials. and Washington were added to the agencies receiving tsunami warning information. When considering the total effects of a tsunami surge. but could have resulted from water trapped in the sand during a rapid withdrawal of the tsunami. Local. Other Hazards. f.
The 1964 Alas ka earthquake began at 0336 G. tsunamis that would create a hazard on t he coast line s of the United States do not arrive from those direct ions .Wei gel (1 974) notes that management of the tsunami warning system was transferred to the Environrnenta l Research Laboratori es. digital tide gages whi ch pr ovide a tide height every several minute s will not provide suffic i ent data for recording tsunamis. Korea. Guatemala. in 1971. Char t s showing tsunami tyaveltimes between variou s coa stal points and t sunami generating areas have been prepared . and the United States. The Tsunami Warning System. France..) After the alarm. cont i nuou sly recording tide gages are required . as well as the remaining 14 seismograph and 35 tide stations of the Tsunami Warning System in the Paci f ie (including Pacifie Ocean territories). China. Ch i le.t. U. However. Stat ion s i n the Pacifie Ocean territories of t he United States and t hose in t he rema inder of the Pacifie provide additional informati on on the t sunami s .m. NOAA .t. 1. an advisory bullet in i s sent to dis semination agencies i n the warning system.t.R. Japan . but generally. The tsunami warning system functi on s best for dis t antly generated tsunamis.e . I f there i s a possibility of tsunami generation. Phil ippine s . t sunamis where the a r r i va i t ime i s sev era l hours after the initial wave generation. and the alarm sounded at Honolulu Observatory at 0344 G. The tsunami warning system i n the United States is based at t he Honolulu Observatory and receives dat a from 18 seismograph s t ations and 16 tide states in the United States (including Alaska and Hawaii . i . The s e charts are used to pr edict the arrivai time of a poten t ial tsunami at the various coastal points after the epic enter of the earthquake has been determined. times are used to provide a uniform time at all points in the system. a deci sion is mad e as to whether or not an adv i s ory bulletin should be i ssu ed. i n quir ies are sent to various seismic obs ervatories in the system to obt ain sei smic r eadings. After receiving and evaluating init1al dat a .S. 1972). initiat ing activity in the Tsunami Warning System. (Spaeth and Berkman. An alarm at tached to the seismogr aph at Hono lulu Observatory is triggered by the arrivai of seismic waves. then to the National Weather Serv ice in 1973.S. The warning system is based on 51 tide stations and 32 seismograph stat ions (1976 lOC dat a ). Peru . The Intergovernmental Oceanograph i e Commission (l OC) maintains an International Coordination Group for t he Tsunami Warning System in the Pacifie.m. New Zealand. Membe r countries are Canada. Tsunami warnings fo r the U ted St ates are based ni primarily on the stations in the United State s and on thos e stations located in both North and South America. As the period of the tsunami varie s from several minutes to approximately 30 to 40 minut es . 194 .m. the warning s ys t em can also alert the population to the possibility of t sunami generat ion from nearby seismic activi ty. but exc luding Pacifie Ocean territories) . Thailand. Ecuador. Modern. (G..
Maps of potential tsunami hazard areas are included in the county telephone directories in Hawaii to define areas which should be evacuated. It is the responsibility of agencies receiving tsunami warnings to disseminate information to the civilian population. an advisory bulletin is issued giving estimated times of arrivai (ETA's) of the tsunami at various coastal points. by police and civil defense personnel. (b) No reply is received from a tide station in a critical recording position in response to a query from the Honolulu Observatory after the occurrence of an earthquake large enough to trigger the seismograph alarm. Various agencies are kept advised of the status of the tsunami until an allclear bulletin is issued.e. tsunami alerts when no significant tsunami occurs." If it is considered probable that a tsunami will cause property damage and loss of life. Cox (1978) considered the cost of false alarms. If there is reasonable cause to assume that a tsunami may have been generated (based on seismic records or observations at tide stations). i. as distinct from advisory bulletins. various reporting stations continue to report water levels to the Honolulu Observatory until all information has been obtained from the tide stations of the Tsunami Warning System in the Pacifie. The Honolulu Observatory continues to receive reports and issue bulletins to Civil Defense agencies as information becomes available. The alert is repeated at 1 hour and then a half hour before the estimated ETA. fire. forestry. The first alert is given 2 hours before the estimated ETA by sounding an alert on all of the sirens in the State. g1v1ng the cost of false alarms at $264. a tsunami alert is given. This may be clone through broadcasting news media (radio and television). are issued only under the following circumstances: (a) Unusual sea level disturbances having tsunami characteristics are recorded at one or more of the warning system tide stations scattered about the Pacifie. If a tsunami was generated. Cox (1964) indicates that tsunami warnings. and Civil Air Patrol personnel. (c) An earthquake occurs whose epicenter is in or on the borders of the Pacifie Ocean in such a location that a tsunami generated there would not arrive at any tide stations sufficiently in advance of its arrivai at a particular shoreline to allow warning that shoreline. The alert is extended by police. Schank (1978) discusses the public warning system in the State of Hawaii. CivAlert broadcasts information and instructions.. The cost of false alarms must 195 .Inquiries are sent to tide stations near the earthquake epicenter to determine if tsunami waves have been generated. or by sounding signais on sirens. Broadcasters in the State are linked together in a system called "CivAlert.000 per year (1977) in Hawaii alone.
1972). and had the Armed Forces Radio Station broadcast a tsunami warning. so a prompt response to the warning was essential.be balanced against the high cast of casualties to the local population if a tsunami occurs without sufficient warning. the U. Military and government personnel promptly evacuated the endangered areas. at a distant source. the evacuation of the city of Kodiak was not as well carried out (there were eight deaths at Kodiak). A decision is then made at the local level as to whether or not the local population should be alerted and evacuated. Approximately 30 minutes after the Alaska earthquake of 1964. (a) Type I. Severe Earth temblors not present. Although reasonably prompt. Cox and Stewart (1972) discuss the particular problems of providing tsunami warnings to areas near a tsunami source. Shoreline slumping. 2. Haas (1978) separates tsunamis into the following four types (summarized in Table 7). Human Response. Seismic detection instruments have been placed in police and fire stations and similar locations. These include the unusual feeding habits of fish before a tsunami due to the presence of large quantities of bottomadherent diatoms in the upper layer of the sea (Suyehiro. (b) Type II. Adams (1978) discusses the local tsunami warning system used in Hawaii. Alaska. Fleet Weather Central at Kodiak Naval Station received ward of a large tsunami at Cape Chiniak.S. The tsunami arrives within 10 minutes. Tsunami warnings must be timely to be effective. (d) Type IV. Various investigators have called attention to phenomena occurring just before a tsunami. Very heavy Earth temblors can be felt by the local population for a period up to severa! minutes. To reduce the hazards on particular coastlines. Large waves generated onto the shoreline almost immediately. Tsunami arrives within 30 minutes. The tsunami is generated 196 . and large rock and ice falls coïncident with the earthquake. The study of such phenomena has not been developed enough to be included in a forma! tsunami warning system. (c) Type III. Noticeable Earth shocks felt by the local population for a period up to severa! minutes. 1934) and increases in the Earth's magnetic field preceding an earthquake (Moore. The first large tsunami wave crest arrived at Kodiak about 30 minutes after the warning. earthslides. No local Earth shocks. Spaeth and Berkman (1972) note that the response of a local population to a tsunami warning may be slow unless the population is well trained to respond. a policy of regional evaluation was adopted in 1966. Seismic activity with a magnitude which may generate a tsunami triggers an alarm.
such as at Seward. the possibility of warnings for types II. Many lives are lost either because sorne residents do not respond to visible signs of a possible tsunami. residents moved to high ground when they observed the initial signs of a tsunami. At Ouzinkie. such as the Tsunami Warning System in the Pacifie. or residents 197 . a large warning system.Table 7. and most people evacuated the low areas. For a type IV tsunami. the initial slumping of the waterfront gave warning to the residents of the town. and little time is available for an organized warning system to operate. but 11 people were killed by a wave 9 to 12 meters (30 to 40 feet) in height. At the village of Kaguyak on Kodiak Island. 1978) . is required. there is almost total reliance on prior education. and the population must be educated to respond to the alert. Therefore. Physical elues Visible slumping or sliding Severe Earth temblors Noticeable Earth shocks None Approximate time for evacuation Less th an 1 minute 5 to 10 minutes Maximum credible preventive action Almost none II Ambulatory persans can be evacuated Sorne persans can be evacuated Most persans can be evacuated and up to 75 percent of all movable property III 15 to 30 minutes IV 45 minutes to 12 hours Haas notes that while no effective warning can be given for a type tsunami. Weller (1972) cites a number of instances of human response during the 1964 earthquake and tsunami in Alaska. A reliable local warning system is needed to alert the public. For a type III tsunami. on Spruce ~sland. Evacuation for a type II tsunami requires prompt response by the population based on their individual sensing of strong earthquake shocks. Tsunami type I Typology of tsunami events Cafter Haas . III. and IV will depend on the education of the public and the effectiveness of warning systems. public education alone is insufficient because the physical evidence of a possible tsunami is not as strong. At Seward. and there was no loss of life. the residents evacuated the town when they observed the initial development of wave action offshore. but three people were killed when they returned to low areas before the arrival of the largest wave.
California. Row (1972) discusses the atmospheric waves associated with the Alaska earthquake in greater detail. DeepOcean Tsunami Gages. Spaeth and Berkman (1972) note that severa! people were killed at Crescent City. would arrive weil in advance of gravity waves traveling through the ocean.e. the tide stations provide records of tsunami heights and periods. For example. 3. there is a high probability that a tsunami may have been generated. 1970). the local topography distorts the tsunami recorded on tide gages near the coastline.return tao saon to law areas. The use of atmospheric waves to estimate the ground motion of a tsunami source requires further investigation. i. In addition to supplying information for the Tsunami Warning System. If the earthquake is a dipslip type. a tsunami traveling across the ocean at a speed of 200 meters per second (447 miles per hour) will arrive approximately 2 hours 20 minutes after the atmospheric wave. whether the earthquake is a dipslip type or a strikeslip type. If the earthquake is a strikeslip type. and these records do not provide information on the 198 . The Doppler shift associated with the ionospheric waves can be monitored at relatively law cast (Furumoto. A large amount of damage at Crescent City resulted from the failure to remove vehicles. Ionospheric Waves. Row (1972) discussed the possible association between ground motion and ionospheric waves. and additional information must be obtained from tide stations. at a distance of 5. and the latearriving disturbance (propagating at about 300 meters per second) previously mentioned by Van Dorn. To evaluate the possibility of a tsunami being generated by an earthquake. Row indicates that there is bath an earlyarriving pressure disturbance. such as at Kaguyak. associated with seismic waves in the Earth. this method may be useful for the Tsunami Warning System. it is desirable to have information about the source mechanism of the earthquake. He notes that this provides a rapid approach to source mechanism estimation. because they returned to a law area before the arrivai of the largest wave. from the endangered area. Murty (1977) provides further discussion on ionospheric effects. Van Dorn (1965) indicated that a dipolar barometric wave in the atmosphere was associated with the dipolar ground motion of the 1964 Alaska earthquake . and tsunami alerts can be canceled at ali locations except those near the epicenter. Furumoto (1970) reports on the use of a lümegahertz Doppler recording of Rayleigh waves to estimate the initial phase of the source of the 11 August 1969 Kuril Islands earthquake. However. it may be assumed that a large transoceanic tsunami will not be generated. Unfortunately. at distances far from the source. Pressure disturbances also propagate through the ionosphere. The latearriving disturbance is associated with the tectonic deformation and. and that this raised a possibility for early tsunami prediction..100 miles) from the source. including a gasoline tank truck. 4.000 kilometers (3.
in or near seismologically active regions. and may have substantial effects on flood insurance premiums and permits for utilization of property. and port facilities are necessarily located near the shoreline. Vitousek and Miller discuss four possible methods of measuring a tsunami in the open ocean: (a) Freedrop recoverable instrument package. so that welldesigned protection is required for highcost facilities. and only at specifie coastal areas. Numerical procedures have been developed that allow the simulation of a tsunami source. which are compared to the more limited field measurements for verification. Procedures have also been developed to simulate tsunami flooding shoreward of the coastline. A system of this type would provide deepocean data. a means of recording tsunamis Vitousek (1961) proposed placing permanent instrument packages in the ocean.deepocean form of the tsunami. While sorne deployment of openocean gages has been carried out. and to obtain openocean tsunami data. and (d) the cableconnected instrument previously discussed by Vitousek. so it is necessary to have as high a degree of accuracy as possible in defining flood zones. the 100year flood level. Also. by tsunamis. Numerical data are used to supplement the available field data on tsunami flood levels. usually by untrained observers. e. Sorne data can be obtained from historical accounts. Openocean data on tsunamis. Shinmoto and Vitousek (1978) give details of an airdeployable freedrop tsunami gage which can be emplaced quickly after a tsunami occurs. Available data on tsunami inundation come from visual observations (including posttsunami surveys) and from tide gage records. 199 . and that the cost of laying special cables would be unrealistic. Vitousek and Miller (1970) indicate that cableconnected systems 'would be expensive. Therefore. The numerical results. Small variations in predicted flood levels may affect property worth millions of dollars. Data are generally only available for a few occurrences. and would also provide additional useful information for the Tsunami Warning System because of the direct connection with the gage. experience in operating such gages is limited. Future use of such gages is required to determine their practicality and reliability as part of the Tsunami Warning System. However. in the open ocean is needed. IX. (c) an underbuoy instrument. large powerplants are typically located at low elevations because of pumping requirements for cooling water. the generation and propagation of the tsunami waves across the ocean. connected to abandoned transpacific telegraph cables. and the interaction of the tsunami and coastal topography. are almost nonexistent. but such data are dependent on incomplete persona! observations..g. (b) an undership instrument. needed for verification of numerical investigations. SUMMARY AND CONCLUSIONS The potentially high value of property in the coastal zone and the intensive development of such land for both private and public use require that careful consideration be given to the possibility of catastrophic flooding of areas of the coastal zone.
Theoretical solutions exist for wave r efraction at coastlines with uniform topography. and openocean tsunami data so that errors in simulating tsunami generation can be separated from errors in the simulati on of nearshore propagation. and the maintenance of a system for placing the instrument packages quickly and on short notice. It is also desirable to maintain a s t andby capability for dropping instrument packages into the open ocean immediately after a tsunami occurs. It becomes necessary to use large time steps to r educe computational errors. This latter capability requires the maintenance of gages and associated instrument packages in operating condition over long periods of time. continuing improvements are needed in the numerical procedures for simulating tsunamis. A particular area of possible improvement is t he treatment of boundaries of the computational grid. waves passing over mathematically defined transitions from deep water to shallow water. Deviations between the numerical results and the theoretical solutions indicate the degree of accuracy where the numerical procedures are applied to more complex topography.provide the additional data needed to construct tsunami flood level maps for various probabilities of recurrence. etc. Errors in the wave reflection from solid boundaries. These errors grow with increasing time so that the solution is not accurate for long per iods of real time. It is necessary to maintain tide gages with the capability of recording tsunamis. 200 .. Because of the long periods of time between the occurrence of tsunamis. An airdropped system is probably the most practical for this purpose. including the periodic testing of the system by placing and recovering the instruments. this requires both accurate data on real tsunamigenerating mechanisms. However. to us e long real distances between grid points. the numerical solutions cannat always be verified and adjusted to match field data. and errors at openocean boundaries where the waves must pass completely through the boundary. At the present time. and also to have standby plans with designated personnel to obtain field observations immediately after tsunamis occur. This smooths out the topographical variations so that wave scattering caused by small topographical features is not properly accounted for. Because of the limited field data available. A continuing program of gathering field data on tsunamis in the open ocean and coastal inundation by tsunamis is needed. and consequently to use a course grid. theoretical solutions are needed for simple topography to provide verification for numerical procedures where field data do not exist.e. Improvements in the numerical simulation of tsunami generation are a l so desirable. Also. Continued research should be carried out in areas such as shelf resonance. the accumulation of data for particular coastal points is very slow. propagate through the computational grid at each succeeding time step. the various effects on tsunami propagation cannat be adequately separated in the computational procedure as t he available data are mainly from tide gages and visual observations of maximum inundation levels. i. Numerical procedures can also be verified by comparing with theoretical results for idealized cases. In particular.
A. BENMENAHEM.E. "Statistical Methods in Hydrology. AMBRASEYS. "Reliable Estimation of the Seismic Moment of Large Earthquakes. N. 1978. No. pp. Part 3. BERG." U. 1970.. T. No. " Journal of Physics of the Earth~ Vol. pp. Jan. 1962.. Hawaii Institute of Geophysics. Army Engineer District. BENJAMIN." Journal of Geophysical Research~ Vol. 13. pp. Apr. T. National Bureau of Standards.N. France. Its Location and Seismic Setting.. "On Model Equations for Long Waves in Nonlinear Dispersive Systems. Sacramento." Proceedings of the Symposium on Tsunamis~ Department of Fisheries and the Environment. 417430. Hawaii..LITERATURE CITED ABE." Philosophical Transactions of the Royal Society~ Series A. E.B.L. European Seismological Commission. Vol.. and IPPEN. 1968... 201 . 1964. 65. 1972. 205217. 4778. "Data for the Investigation of Seismic SeaWaves in Europe. July 1961. 401435.T. J . "Relationship of Instruments and Policy in the Hawaii Warning System. Paris. E. "Wave Reflection and Transmission in Channels of Variable Section. W." Bulletin of the Seismological Society of America~ Vol. 29. BONA. BENJAMIN. 1964. pp. Ottawa. B.M. and Conclusions as to the Origin of the Major Tsunami. BERG. 1965." Report No.." Proceedings of the 15th Alaska Science Conference~ American Association for the Advancement of Science. Apr. AMBRASEYS. Ch.N. 1967.. pp... No.R. 272.S. N.. 51. in the Greek Archipelago. 15611568. E. BEARD. "Radiation of Seismic SurfaceWaves from Finite Moving Sources. and FEIR. 27. Canada. J . "Size of Great Earthquakes of 18371974 Inferred from Tsunami Data. and MAHONY..S.J. G.. K. L. pp. BLANCH. 1960. 381390. "Mathieu Functions. 218232." Joùrnal of Geophysical Research~ Vol. A. B4. 23. 1956.. Nov. 4. ADAMS. J. 20.. Honolulu. U. "The Disintegration of Wave Trains on Deep Water. Jan.. Sacramento. pp. Washington. 1975." Handbook of Mathematical Functions~ Ch. K. Calif. "Field Survey of the Tsunamis of 28 March 1964 in Alaska.. 12571265. pp. BOURODIMOS." Proceedings of the llth Conference on Coastal Engineering~ American Society of Civil Engineers. ABE. "The Alaska Earthquake. L. International Union of Geodesy and Geophysics. Government Printing Office. 195212.C. et al.. pp. 3." Journal of Fluid Mechanics~ Vol. HIG702. "The Seismic Sea Wave of July 9." Monograph No. 84.. 1979. D.
Breaking and Runup of Solitary Waves.. 1970a. L.. CAMFIELD.L. G. 1971. BRANDSMA." Proceedings of the Symposium on Long Waves in the Ocean 3 Department of Fisheries and the Environment.E." Port Townsend Leader3 Summer Supplement.. TC486." Proceedings of the 12th Coastal Engineering Conference3 American Society of Civil Engineers. BYRNE.K. 25902596 (copyrighted by American Geophysical Union). R." Proceedings of the 19?6 A~y Numerical Analysis and Computers Conference 3 Sept.L. "The Effects of Bottom Configuration of the Deformation. pp. T." Proceedings of the 15th Conference on Coastal Engineering 3 American Society of Civil Engineers. BRETSCHNEIDER. Washington. Palo Alto. Nuclear Regulatory Commission. 1975. Dec. Vol. National Science Foundation.BOWEN. McNOWN. R. R. 6894. 1979.. F.J. 1976. 1971. Calif. No..E... C. 1979.E. M.. BRANDSMA.. Sept. and WYBRO. 1967. No. LS.." Journal of the Engineering Mechanics Division. Wash. "Applications of Numerical Modeling to Coastal Engineering Problems. and STREET. R. CAMFIELD." Journal ofComputational Physics 3 Vol. 10.. Aug. M.C. H." Proceedings of the llth Conference on Coastal Engineering 3 American Society of Civil Engineers. and STREET. and HWANG. Stanford University. D. DIVOKY.. Nov.L. and DURHAM.L. pp.. F. U. D. Canada.S.P. SHAW. BUTLER. pp." Journal of Geophysical Research 3 Vol. 10061024. No.. D.F.A Revised Madel for Tsunami Applications. D.S. 6.. 86628671. 17031716. pp.W. 36.L. 1976. 1958. unpublished data. pp. DIVOKY. pp. J. D.. 173189. LS. 74.." Journal of the HydrauUcs Division 3 Vol. HY6. No. 183187. "Channel Effects in Harbor Resonance.. 1970b." Journal of Geophysical Research 3 Vol. Ottawa. "Wave Forces on Submerged Structures. pp.F. F. R.G. "Tsunami Atlas for the Coasts of the United States. CAMFIELD. Apr.L. 97." Report No. "The 'Great Flood' of 1866.. 76. 1968.. Nov.L.. CARRIER. "A Computer Study of FiniteAmplitude Water Waves. EM6.C. 84. and STAIR.. BRATER. and HWANG... M. R.J. A.. Department of Civil Engineering.L. Washington. 60. Ch. "Seawave. 11. 471508. CHAN. pp. "Wave Trapping of a Coastline Approximated by a Circular Arc. Dec. P." Final Report. and STREET.C. D.... May 1969. R. Port Townsend. 202 . and MIYATA. CHAN..D. E. 1975. and STREET... Ch. and INMAN. "Shoaling of FiniteAmplitude Waves on Plane Beaches..K.. "Tsunami Inundation Prediction. "Field Occurrences of Induced Multiple Gravity Waves. CAMFIELD. "Edge Waves and Crescentic Bars. R.C.
203 .. and HWANG. and STEWART. U. Fairbanks. STREET.C. 7. and ECHOLS. June 1970. Washington. Wôrld Data Center A. E. COCHRANE. Washington.N. Palo Alto. and STRELKOFF. Rockville. R.." Report No.. 1976." Journal of Marine Research~ Vol. Dec. 1962. C. 239251.C." Report No.. Cambridge. N. T..P. c. "Tsunami Surge Forces.. R. W. Mar. "An Experimental Study of Gravity Wave Behavior Near a Straight Caustic.. Stanford University. Geophysical Sciences Laboratory. "A Table of Alaska Earthquakes. 1967. National Oceanic and Atmospheric Administration. 1964. Calif." TM 124." Report No. YY. CROSS. D. "The Digital Simulation of Water WavesAn Evaluation of SUMMAC.. Massachusetts Institute of Technology. Department of Civil Engineering. D. Hawaii. Corps of Engineers. R. R." The Great Alaska Earthquake of 1964. T. 93.B. C. D.. Vol. STREET. YY.. Solid Earth Geophysics. T.. 1948. "Computer Studies of FiniteAmplitude Water Waves. DIVOKY.." Technical Report No. and FROMM. COX.. 1961.J. "Nearfield Tsunami Behavior. pp. 3.C. CÀEN. National Research Council.Y..L.. New York. TR707. Palo Alto. and PIERSON.. Aug. "Tsunami Forecasting.K. "Reflection of Tsunami. 201231.C.. Oceanography and Coastal Engineering~ National Academy of Sciences. G. D. 104. COX." Proceedings of the Second International Conference on Numerical Methods in Fluid Dynamics~ Department of Civil Engineering. "Catalog of Tsunamis in Alaska." Report SEl. CHEN. L.. pp. 17881961. New York. Calif. N. Stanford University. D. CHAO.. June 1969. No.. No.. D. M. pp.... and MEl. Including the Effects of Caustics and the Spherical Earth. "The Theory of Wave Refraction in Shoaling Water.C. PARARASCARAYANNIS. New York University. and ARTHUR. 190.Y. University of Alaska Geophysical Institute.." Final Report. 229245. 1970.. New York University. and Coastal Engineering Division. CHEN. Army. R.. Mar.CHAN.. "Oscillations and Wave Forces in an Offshore ' Harbor (Applications of the Hybrid Finite Element Method to WaterWave Scattering).. H. Nov.S.C." Report No. 1972. LS. COX. 1970.. Mass.. 1974. WW4.C.O. Apr. Hawaii Institute of Geophysics. and CALEBAUGH. Geophysical Sciences Laboratory. "Experimental Study on the Solitary Wave Reflection Along a Straight Sloped Wall at Oblique Angle of Incidence.." Report HIG6415.S.H... J. DAVIS. Md. National Science Foundation. 8 (UAGRl31)." Journal of the Wateruays~ Harbors. CHAN.S. J.. C. Beach Erosion Board. TR7017. "Technical Evaluation of the Seismic Sea Wave Warning System. H.e.. 1975. Washington. K. CHAO. J. Honolulu. Alaska. R.
D. 6782. Part II (7).. 1960. 47. "Ionospheric Recordings of Rayleigh Waves for Estimating Source Mechanisms. 120.. pp. "Wave Breakers on a Beach and Surges on a Dry Bed.. pp. No. Japan. W.. M. 1. IBM Research Laboratory. A. and AULT. 2. Part 1." Journal of Fluid Mechanics~ Vol. 1961. Honolulu. 187216. FREEMAN. et al..P. San Uose. Thesis.. .C.M.. Shallow WaterWaves of Periodically Recurring Form. 9.." Proceedings of the lOth Conference on Coastal Engineering~ American Society of Civil Engineers..J. 1971.." Journal of the HyéJ:r>auZics Division~ Vol. "Resonant Edge Waves on Laboratory Beaches. Cal if. C. pp. Oct. GALVIN. Feb." Tsunamis in the Pacifie Ocean~ Ch." Transactions of the American Geophysical Union~ Vol. Mar. No.S. 46.L. 132 (also Reprint 572.." Proceedings of the Symposium on Long Waves~ University of Delaware. pp. Coastal Engineering Research Center. Y. "The Tsunami of May 23. 1963. and ITO. No. 1934. 51. J. 112. pp. FUKUUCHI. U. RICHTER.." Bulletin of the Seismological Society of America~ Vol. "Reports on the Tsunami of 1933 on the Sanriku Coast. pp.S.U.. 1968. London.. 1966. Army. GALVIN. 129. 1962. p.. J.J. "FiniteAmplitude." Journal of the Waterways and Harbors Division~ Vol. GALLAGHER.H. 135157. HY2." Bulletin of the Seismological Society of America~ Vol.J.. on the Island of Hawaii. 119134.DAVISON. and LE MEHAUTE.. FUKUI.." Bulletin of the Earthquake Research Institute~ Tokyo Imperial University. "Long Wave Modification by Linear Transitions. 5. C. 204 . Corps of Engineers. Del. DEAN. GELLER. Vol. M.. 1970. Fort Belvoir. J. and BOCCO.E. Newark. C. R. 66. pp. FURUMOTO. B. No. H. 1965. "Hydraulic Study on Tsunami.." S. Japan. 1970. NTIS AD 754 868). Y. pp. GAGNON. "Generation of Surf Beat by NonLinear Wave Interactions. Great Earthquakes~ Thomas Murby &Co. 1964.V.. 15011523.. Tokyo. Cambridge. Hawaii.G. Ch. 1976. B. Massachusetts Institute of Technology.. "On the Effect of Breakwaters Against Tsunami. IBM Research Report RJ531. 49. 821839. 90. EastWest Center Press. Va. 6. FROMM." Coastal Engineering in Japan~ Vol. Department of Civil Engineering. "The Effect of a Gradual Change of Depth on a Train of Surface Waves. Mass. EARTHQUAKE RESEARCH INSTITUTE. WWl. "Scaling Relations for Earthquake Source Parameters and Magnitudes. R. Apr. 1936. 90. pp. 1964. EATON.
Part 4. GUZA. "Edge Waves and Beach Cusps. 269293. A. C." Journal of Geophysical Research~ Vol. 1980) . and RICHTER... (in preparation. pp. New York... GUZA. WW4. GORING. 1976. Part 2. 33. Nov. 21.S. 1974. 1954. Calif. R. R. 289314. B.L. Nov. Corps of Engineers.L. 45294534.M.. pp. "A Note on Tsunamis: Their Generation and Propagation in an Ocean of Uniform Depth. Pasadena. pp .. pp. 9.E. and DAVIS. J. "Tides and Tidal Datums." Proceedings of the Symposium on Tsunamis~ Department of Fisheries and the Environment. 224235. Princeton University Press.. 34. No. "FiniteAmplitude Edge Waves. "Human Response to the Tsunami Warning System.J. "Magnitudes of Great Shallow Earthquakes from . 205 .T. Seismicity of the Earth and Associated Phenomena~ Hafner.GELLER.. R.L.. D. 12 8 5 12 91. 80. Part 2. 365377. 60.E.G. Nov.J. J. HAAS. Keck Laboratory Report No. Princeton. Fort Belvoir." Journal of Fluid Mechanics~ Vol. D..." Journal of Fluid Mechanics~ Vol.. June 1977. and RICHTER. 1978.T. Army. R. KHR38. 102. R. pp. GUTENBERG." Journal of the Waterways~ Harbors and Coastal Engineering Division~ Vol. "The KortewegdeVries Equation and Water Waves." U. 1969. 3. HALLERMEIER. HASKELL. H.F. No. New Jersey.J. 1973. B. 59. and KANA.. pp. H. 769799. and BOWEN.T. "The Resonant Instabilities of Long Waves Obliquely Incident on a Beach. GUZA... Va. and BOWEN. 1965. 67. pp. A. 1975." W. 587598. Canada." Journal of Geophysical Research~ Vol. 80.." Journal of Marine Research~ Vol.MORI. D. 65. Mar. HAMMACK. 865908." Bulletin of the Seismological Society of America~ Vol. C. 1976.A. GUZA.. and INMAN. HARRIS. pp. Seismicity of the Earth and Associated Phenomena~ 2d ed. No.F. Coastal Engineering Research Center.J. and SEGUR.T.1904 to 1952. Comparison with Experiments. "Tsunamis The Propagation of Long Waves Onto a Shelf. J. N. pp. Ottawa. 1. July 1975.L. R. California Institute of Technology.. No.. 1978.." Bulletin of the Seismological Society of America~ Vol. 79.. "Elastic Displacements in the NearField of a Propagating Faul t. "Nonlinear Flow of Wave Crests Past a Thin Pile.. HAMMACK.. "Excitation of Edge Waves by Waves Incident on a Beach.. GUTENBERG. pp." Journal of Geophysical Research~ Vol. No. No. R. 29973012 (copyrighted by American Geophysical Union). 1974.
"The Chilean Earthquake of 1960.S." Technical Report H743. HOUSTON. "The Grand Banks Earthquake.S. 6. "TsunamiWave Elevation Frequency of Occurrence for the Hawaiian Islands. and DOXSEE. Tokyo. 3.L. Smithsonian Insti tutian.D. Princeton. K. 1." Miscellaneous Paper H7620. Sept." Earthquake Notes~ Vol. Vol. 911. No. 1935a. Dec.. Aug.W.H. 2031 (in Japanese with English abstract). 1930.. D. K. Army Engineer Waterways Experiment Station... E. November 18. ." BuUetin of the Seismo logical Society of America~ Vol. Miss. 1947. Washington.. University of California. 18. Existing Conditions and Proposed Improvements. N. 1977. 1976. Oct. Vicksburg.R. R.G. N. 2... pp.. HOUSNER. Miss. Pasadena. 1 and 2. "List of Seismic Sea Waves." Smi thsonian Miscellaneous Collections 1087.. K. HECK. J. T.HECK. U..S. Vol. Vicksburg. Vicksburg. HIRONO. "Los Angeles Harbor Numerical Analysis of Harbor Oscillations. HOUSTON. Japan..J. U. Earthquake~ Princeton University Press. K. J.." Dissertation. D. 37. No.. 1.. HIDAKA.. California Institute of Technology.S. HOUSTON. HODGSON. Institute of Engineering Research. Miss.Japan." Series 89. Wave Research Laboratory. HEITNER.. CARVER. Vicksburg. A. May 1974. 1769 to 1897. pp.C. HIDAKA. U. HOLDEN. and WIEGEL. HOUSTON. H." The Memoirs of the Imperial Marine Observatory~ Kobe. "Seiches Due to a Submarine Bank.. .R. 70. pp..L. Issue 4." Proceedings of the Fourth World Conference on Earthquake Engineering~ 1969... Vol. J. 4. 1935b. Japan." Journal of Geography~ Tokyo Geographical Society. Feb. Army Engineer Waterways Experiment Station. 6. Berkeley. pp. G.A. R. Report 1. U. 1977. "Long Beach Harbor Numerical Analysis of Harbor Oscillations.S.. 1936." The Memoirs of the Imperial Marine Observatory~ Kobe. 206 .R. W. Miss. "Secondary Wave Crest Formation. Dec. "Engineering Estimates of Ground Shaking and Maximum Earthquake Magnitude. 1959. 1969. Nos. 1898.. 1929.. 1961. N.. 7281.W. Feb. J. Army Engineer Waterways Experiment Station. HORIKAWA. Army Engineer Waterways Experiment Station. "A Catalogue of Earthquakes on the Pacifie Coast. No. "Type 16 Flood Insurance Study: Tsunami Predictions for Pacifie Coastal Communities.. E. and MARKLE. "A Mathematical Model for Calculation of the Runup of Tsunamis." Miscellaneous Paper H772. "A Theory of Shelf Seiches.R. S. No." Technical Report H7716. Calif. Calif. Earthquake Engineering Research Laboratory.. and GARCIA.
Monograph No.. D. D. J. ISHIMOTO.. Army Engineer Waterways Experiment Station. l. and DIVOKY. "Numerical Investigations of Tsunami Behaviour.T. K.. Oceanography and Coastal Engineering.. 1961. 1934." Tsunamis in the Pacifie Ocean. Hydrodynamics Laboratory.. NORCONP1 PSAR. Puerto Rico. National Science Foundation." Final Report . U. pp. Mar. D. "The Generation of Tsunam is and the Focal Mechanism of Earthquakes. 1978. Vicksburg.R. Cambridge. "Tsunami Generation. 1942. A. Vol." The Great Alaska Earthquake of 1964." Report No. 207 . 1723.4B. HWANG. 1970.. and GODA. 6. HWANG. U. and Generation Mechanisms of Tsunamis and a Catalogue of Earthquakes Associated with Tsunamis. LS." Journal of Geophysical Research~ Vol. HOUSTON.. Washington... ..HOUSTON. "Wave Induced OscilEffect of Energy Dissipators in Coupled Basin 52. Jr. Army Engineer Waterways Experiment Station. 14. 191210. l.. Tokyo Imperial University. National Academy of Sciences . National Research Counc il." Bulletin of the Earthquake Research Institute . "Effect of Source Orientation and Location in the Aleutian Trench on Tsunami Amplitude Along the Pacifie Coast of the Continental United States. A. "Wave Induced Oscillations _in Harbors: The Solution for a Rectangular Harbor Connected to the OpenSea. " Proceedings of the lOth Pacifie Science Congress Symposium.. "Magnitude. et al. HWANG.. Miss. Dec. pp .. LS. Nov." Technical Report H7826..R. 75. 33.. 318. J .C. pp. No." Report No. Suppl ementary Vol. 68026817. LS. IIDA. pp. Miss. and SULLIVAN. July 1963. A. "The Phenomenon of Sea Water Overflowing the Land.. W.. D. 1975. Mass. "Probable Maximum Tsunami Runup for Distant Seismic Events. Ch.'' Research Report H754." Zisin. HOUSTON. "An Interpretation of the Reexposure of an Old Rice Field Buried in th e Ground. K. "Type 16 Flood Insurance Study: Tsunami Predictions for the West Coast of the Continental United States. Honolulu. Massachusetts Institute of Technology. May l975a. 59. IIDA.S. No.. Tokyo. Hawaii. International Union of Geodesy and Geophysics.. lations in Harbors: Systems. and DIVOKY. R. IPPEN..." Appendix 2. IMAMURA. T. Amendment 23.S.C. Mass. D. "A Numerical Model of the Major Tsunami.T. pp. 1970.. pp. 149153 (in Japanese). dynamics Laboratory . Y. et al. and DIVOKY. and GARCIA. Puerto Rico Water Resources Authority. 24... A. RAICHLEN. IPPEN. Washington. Massachusetts Institute of Technology.. Japan. July 1962. and HAGIWARA.R. Vicksburg.. July l975b. EastWest Center Press. K. J. 718. F.. M. HydroCambridge. Energy. 1972.
MIURA.. "A Big Atlantic Earthquake. Royal Society of New Zealand. 123135." Proceedings of the Tsunami Research Symposium~ Bulletin 15. "Focal Processes of the Great Chilean Earthquake. JORSTAD. 128136.. J. JAGGAR. pp.. H. Army Engineer Waterways Experiment Station . John Wiley and Sons." 80.H." Proceedings of the lOth Pacifie Science Congress on Tsunami Hydrodynamics~ International Union of Geodesy and Geophysics. Hawaii. KAPLAN. 1967. "Experimental Study of Solitary Tsunami Waves. KAMEL. New York.. 1932. III. 1974. .. Part 1 (The Model Experiment of Tsunami in Kesennuma Bay) . T. Miss.". 1948. "On the Effect of the Breakwater in Case of Tsunami. 1960. and CIPAR. No.S. T. 21. and TERADA. No. KANAMORI. KISHI. 6." Technical Report No. H.A. H. 1955 . . Inc. Washington. Army. Vol. 1956. 1966. pp. Monograph No. pp... 208 . "Mechanism of Tsunami Earthquakes. 711768.. "Tsunami Caused by Chile Earthquake in May. 3348. H. G. 2792. and HORIKAWA. D.. K. Vicksburg. pp. 1972. "The Shoal ing." CoastaZ Engi neering in Japan~ Vol. 11. pp. Naturen~ KAJIURA. 1960 and Outline of Disasters in Northeastern Coasts of Japan. Vol. "Generalized Laboratory Study of Tsunami Runup. pp. 1960. 1961. 24. Dover Publications. et 200ars minne. U. "Wave Motion. T. pp.. A..S.B.. T. "The Solitary Wave and Periodic Waves in Shallow Water. KONONKOVA." Physics of the Earth and PZanetary Interiors~ Vol. IWASAKI .A. Corps of Engineers.. 1950. F. 6.J." The VoZcano Letter~ Dec.C. KEULEGAN. 1963." Communications i n Pure and AppZied Mathematics~ Vol.. 322348. Hy drodynamics~ LAMB . 1929. "Stability of RubbleMound Tsunami Barrier Hilo Harbor. J. 346359. 206230.IWASAKI. 1.. A. 9.E. and SAEKI. KELLER. pp. K.. Ch. Beach Erosion Board. U. "On the Partial Reflection of Water Waves Passing Over a Bottom of Variable Depth.." Engineering HydrauZics~ Ch. May 22. Breaking and Runup of the Solitary Wave on Impermeable Rough Slopes.Proceedings of the lOth Conference on CoastaZ Engineering~ American Society of Civil Engineers. "Fjellskredet ved Tjelle. Aug.. 2.E. " The TechnoZogy Reports of the Tohoku University~ Sendai... New York. A. 1976. and REIHRUDEL. 323333. 323339.. 25.. Inc. S. Japan. pp." Physics of the Earth and PZanetary Interio1•s~ Vol. 6th ed. Jan.. KANAMORI. K." TM60. G.
. 1962.. Tex. Washington. 1976.. Beach Erosion Board. 3568. LEVY. LOZANO. J. May 1944. 1962 (unpublished). LINSLEY.. "Leakage and Response of Waves Trapped by Round Islands. "Crustal Uplift Southwest of Montague Island. M.. L.. O. "Wave Induced Oscillations in Harbors of Arbitrary Shape. Rand Corporation..H.J. Aug. Structural Mechanics Research Laboratory. Santa Monica." Memorandum RM5294PR. LEE.A. 459495. pp. E . U.. 1971. California Institute of Technology. LEENDERTSE.J. pp.. . 6.. pp. "A Wave Method for Determining Depths Over Bottom Discontinuities..S. and PAULHUS. Ch. "The Transformation of a Solitary Wave Over an Uneven Bottom. KOHLER. C." TM5. Pasadena. and KELLER. Aug. pp. REESE. M.e. Corps of Engineers. L.H. MASON.C. presented at the Hydraulics Division Conference. 10751088.. 10. 209 . Dec. and MATLOCK.N. 19. MAGOON. J. and MEYER. "The Tsunami of May 1960 as it Affected Northern California. "Water Wave Production by Point Sources and Extended Sources." Report No.C. 375393. MATLOCK. and KEULEGAN. . and MYSAK. 1961. Hydrology for Engineers 3 McGrawHill Book Company. 45. pp. c. and MEI. "Trapped Coastal Waves and Their Role in Shelf Dynamics. 1969. P. 146." Journal of Fluid Mechanics 3 Vol. Alaska. R.. Jr.M. No. MALLOY. MAGOON. O.J. D. "Wave Induced Oscillations in Harbors of Arbitrary Geometry. MADSEN. pp. G... 1964." Mcœine Modeling 3 The Sea 3 Ch." Proceedings of the Coastal Engineering Specialty Conference 3 Arnerican Society of Civil Engineers. KHR20." The Physics of Fluids 3 Vol.. 39.." Science 3 Vol. Arrny. ." Proceedings of the lOth Pacifie Science Congress on Tsunami Hydrodynamics 3 International Union of Geodesy and Geophysics. 4. "Aspects of a Computational Model for LongPeriod WaterWave Propagation. Vol. D. B. 1965.H. Part 4." Report DASA1268." Journal of Fluid Mechanics 3 Vol.. Austin. 10481049. New York.A. R. J . 24. pp. 162166. May 1967. J. Keck Laboratory of Hydraulics and Water Resources. O. "Analysis of Structural Damage from the 1960 Tsunami at Hilo." Arnerican Society of Civil Engineers.LEBLOND. University of Texas.B...S.L.. 1977.T. Hawaii. Calif. 1958.J. W .A. 8. R..K. 781791. R. Calif... 1969. H. Mar... Monograph No.T. "Structural Damage by Tsunamis. J. LEE.
MILLER. Ottawa. "Sorne Features of Tsunamis on the Pacifie Coast of ·south and North America." The Great Alaska Earthquake of 1964~ Seismology and Geodesy~ National Academy of Sciences. John Wiley Sons. Vol. New York.... C. 652664. Inc. 1948.J. "Experimental Determination of Runup of Undular and Fully Developed Bores and an Examination of Transition Modes and Internai Structure. 1966. WIGEN. 5564. Chicago.L.W.." Journal of Fluid Mechanics." Journal of the Acoustical Society of America~ Vol. F. 114. 8. "Long Waves on the Continental Shelf: An Experiment to Separate Trapped and Leaky Modes. MURTY. "Coupling of a Cylindrical Tube to a Half Space. 1960. T." Technical Report No. F. Tokyo.W. Numerical Solution of Differential Equations. Mar." Bulletin 198. pp. 1944. 1972. 1977. W. 1953. Part 1. Bulletin 15. pp. C.W.. National Research Council. and WIGEN.S. SNODGRASS. 210 . and CHAWLA.E. Japan." Series No.. J. "Wave Propagation Across the Continental Shelf. 393400. "Tsunami Behavior on the Atlantic Coast of Canada and Sorne Similarities to the Peru Coast... 36. Part 2.. "Resonant Response of Harbors: An EquivalentCircuit Analysis.S. 518525. T. Department of the Environment. 529554. W. R. 20. R. June 1968. pp. No. H. ''Movement Ondulatores de Mers en Profondour Cons.. MILES. 5160. R.. 20. Department of Fisheries and the Environment. 54. "Estimation of Energy of Tsunami and Protection of Coasts. MURTY. "Giant Waves in Lituya Bay.C. U. "Note on the Equations of Long Waves Over an Uneven Bottom. Ill. Fisheries and Marine Service. S... 241265. Canada. 71.S. "Seismic Sea Wavès. 1975." Journal of Fluid Mechanics~ Vol.S. 46. and GILBERT. MILLER." Bulletin of the Earthquake Research Institute~ Tokyo Imperial University. Alaska. MICHE. pp. 1971. Ponts Chaussees~ Vol. MILNE. "Magnetic Disturbances Preceding the Earthquake..Tsunamis. G. W." Proceedings of the Tsunami Research Symposium~ Royal Society of New Zealand. Department of Geophysical Sciences. T.. MURTY. Ottawa.. Part 4. MEl. pp. Canada." Journal of Geophysical Research~ Vol.C. D. 6380. pp." Professional Paper 354c.MATUO." Journal of Fluid Mechanics~ Vol. Washington. 1972." Ann. and LE MEHAUTE. pp.O. D..O. S. No. J. D. & MOORE... University of Chicago. J.. Geological Survey.. MILES.. Marine Sciences Directorate. B. Jan. 2. MILES. MUNK. Washington. 5. pp. 1934. 1964... 1976.
NAGAOKA. pp. 14. K. G. 1967." Proceedings of the PhysicoMathematical Society of Japan~ Vol. H. 126136. "Frequencies of Crest Heights for Random Combinations of Astronomical Tides and Tsunamis Recorded at Crescent City. 1972. NASU. "Wave Behavior Near Caustics in Models and in Nature. 2131. PEREGRINE. pp. Journal of the Waterway~ Port~ Coastal~ and Ocean Division~ Vol. pp. Discussion of "Similitude in Coastal Engineering" by B. New York. 1970. PARARASCARAYANNIS. 1. NISHIMURA. Calif. "Heights of Tsunamis and Damage to Structures. "Local Phenomena of Tsunami. NASU. PIERSON.H. PERROUD." Technical Report HEL 168.. O'BRIEN.. and FUNASAKI. 1971. D. M. pp. E. Vol." Proceedings of the Tokyo MathematicoPhysical Society~ Vol." Bulletin of the Earthquake Research Institute~ Tokyo Imperial University. PETRAUSKAS. Cola. 218226. Part 2. 1971. 2735. Series 99. N. 2. Berkeley. "The Solitary Wave Reflection Along a Straight Vertical Wall at Oblique Incidence.. 14. N.... . 1965. C. G. Sept. 1932. R.. Tokyo. Japan. Vol.Q. "Catalog of Tsunamis in the Hawaiian Islands. NAKANO. D." University of California. N. M." Proceedings of the Coastal Engineering Santa Barbara Specialty Conference~ American Society of Civil Engineers. 1977." Bulletin of the Earthquake Research Institute~ Tokyo Imperial University. 1948. Mar.. T. Berkeley. PALMER." Journal of Fluid Vol. Mechanics~ PEREGRINE. 14." Coastal Engineering in Japan~ Tokyo.H. M. pp. 26. ESSA.. "On Destructive Sea Waves (Tsunami). 4456.. pp.J." Waves on Beaches~ Academie Press. Hydraulic Engineering Laboratory. Nos..Reflected Waves Superimposed..E. W." Report WDCAT 692. .Coast and Geodetic Survey. WW3. Ch. 4. Mar.. Institute of Engineering Research. "Preliminary Note on the Accumulation and Dissipation of Energy of the Secondary Undulations in a Bay. and BORGMAN.P. 1901... Japan. "Long Waves in Two and Three Dimensions. pp. 1957. Calif." Proceedings of the Symposium on Long Waves~ University of Delaware. 6390. 163180. Japan. 103. "Hilo Harbor Tsunami Madel . Aug. 1934. 27. 6372. HORIKAWA. 2. pp. Boulder. H. "Long Waves on a Beach. P. No.. Vol. University of California. No. 211 . pp. L.H.. 815827. 393400. No. California. pp. Issue 3. 2. and SHUTO. "On the Function of Tsunami Breakwaters. Le Mehaute. May 1969.. Tokyo.. MULVIHILL.
Report of the Earthquake Investigation Commission... H. REID. SCHANK. "Sur la Resistance des Fluides. M. Army Mathematics Rese arch Center. .. Va. 48. C. 4. 42. M . 1919." The Great Alaska Earthquake of 1964~ W. The Theory of Sound~ Vol. 568585. New York. 1972. H. B... pp." H.. 252254. 1866 (newspaper) . Fort Belvoir. .. Strasbourg. Locally Inhomogeneous Medium. SHA R.." Waves on . Washington. pp. RICHTER. C. 212 .L. No. 147152.. University of Hawaii. 1. June 1914. RO R. pp. 1920. Dover Publications. G. "An Outer Boundary Integral Equation Applied to Transient W.. 494507. lst Sess.D. New York. II." International Congress of Mathematics. 5380. France. U." BuUetin of the Seismological Society of America~ Vol. "Surface Wave Resonance on Continental and I s land Slopes. S. Seismology~ W. 1974.P.. REID. Elementary Cali f .. San Francisco. U." Science~ Vol. pp. D. 1975.C. Ch.H. Series No.S. 66th Cong.. pp. D. V. "Tectonic Deformation Associated with the 1964 Alaska Earthquake. "Hazard Reduction and the Mitigation of Tsunami Effects Through Effective Public Warning in Hawaii. June 1965. University of Wisconsin. and MEYER.F. Miss. Honolulu. 781.. Nov.N. and U." GITI Report 14. E. "Atmospheric Waves.. D.. 1945.PLAF KER. pp. .E.P . PUGET SOUND WEEKEY~ Seattle. Army Coastal Engineering Research Center. LORD. "The Puerto Rico Earthquake of 1918. SEELIG. SHEN . 123161. Army Engineer Waterways Experiment Station . Government Printing Office. 31 Dec." Proceedings of the Symposium on Tsunamis~ Department of Fisheries and the Environment.E. . W. 1755. 269. 16751687. Hawaii. HIG746 (NOAAJTRE126). Seismo logy and Geodesy~ National Academy of Sciences. RAYLEIGH. 1978. No. SHAW . "The Lis bon Earthquake of November 1. XVI." Journal of Applied Mechanics~ Vol. 148. and HERCHENRODER. . RIABOUCHINSKI. 1958. R. . Sept. Wise.. Reprint of 1896 ed.C.." Technical Summary Report No. Beaches~ Academie SHEN. Vicksburg. 2. Madison. Washington. and TABER. Mar." Report No. "Wave Resonance Near Shores. 1977.. Sept. Hawaii Institute of Geophysics and Joint Tsunami Research Effort. 1967. HARRIS. Wave Scattering in an Inhomogeneous Medium.S.S. Inc. Freeman & Sons.. R. pp.F.C Press. "The Outer Boundary Integral Equation Method Applied to Wave Scattering in an Infinite. "A Spatially Integrated Numerical Madel of Inlet Hydraulics.F. Wash. D.. R.
.e. SPAETH.. 5. 11. "An Inexpensive AirDeployable Midocean Tsunami Gauge." Transactions of the Cambridge Philosophical Society.L.Q. and VITOUSEK... C. A. P." The Great Alaska Earthquake of 1964. SPIELVOGEL. Series No. Aug. pp. "Recurrence of Tsunamis in the Pacifie. pp. STOKES. 2938. M. "SingleWave Runup on Sloping Beaches. "Summary of Da ta on Tsunamis in the USSR. 6. 1976.. Monograph No.S. 27. Academy of Sciences of the U. 213 . 149164. Ca1if. 196i. L. S...." Proceedings of the Symposium on Tsunamis.. "Annotated Bibliography on Water Waves Caused by Exp1osions1946 to 1966.J. SMITH. BURGESS. F. and FER CHEV. Stanford University. SOLOVIEV. P.J. D. S. Apr. G. 48. "Materials for Preliminary Tsunami Zonation of KurileKamchatka Coast Based on Hydrodynamica1 Calculations.L.Y.SHEPARD. Palo Alto.C. EastWest Center Press. S. Paris. . 1968. SOLOVIEV. s. Hawaii Institute of Geophysics). pp. L. July 1964. DASA Information and Analysis Center. et al. MacDONALD." Journal of Fluid Mechanics. pp. 391528.. 1980). D. "On the Theory of Oscillatory Waves. S.. . ." DASIAC Special Report 58.. STREET. pp... Department of Civil Engineering.S. SPAETH. National Academy of Sciences. 1970." Technical Report No. 242244. D. Part 4. M. 1950. 9. and COX. "The Tsunamis as Recorded at Tide Stations and the Seismic Sea Wave Warning System. Ch. 38110. and BERKMAN. and WHITFORD.L.G. S. 74... No.. 1967. No.G. Royal Society of New Zealand.G. STEIN. pp. 1972. International Union of Geodesy and Geophysics. 1975.W." Bulletin of the Council for Seismology." Proceedings of the Tsunami Research Symposium.. SOLOVIEV. "The Behavior of Solitary Waves on a Stepped Slope. M. pp. 1978. D. pp. Van Campen. "Earthquakes Along the Passive Margin of Canada Induced by Déglaciation.. R. Vol. 1847." Bulletin of the Scripps Institution of Oceanography. "The Tsunami of Apri 1 1. 93." Geophysical Research Letters (in preparation. 685694. et al. . 441455. Hawaii. Bulletin 15. Washington. Oceanography and Coastal Engineering. 8. Department of Fisheries and the Environment. . Vol. Honolulu. 1946. France.G. R. SHINMOTO." Tsunamis in the Pacifie Ocean. Vol.A. Annotated Bibliography on Tsunamis.R. 2355 (translated by W. G. M.
"The Numerical Simulation of Long Water Waves: Progress on Two Fronts.W. W.M. R. Calif. "Tsunamis Induced by Submarine Slumpings off the Coast of Israel. G. 1960. STRIEM. 239252. LS. VITOUSEK. and· FROMM." Tsunamis in the Pacifie Ocean~ Ch.. pp. "Sorne Observations on the Unusual Behaviour of Fishes Prior to an Earthquake.K.. pp. Tetra Tech." Bulletin of the Earthquake Research Institute~ Tokyo Imperial University. Jan. "Handbook of ExplosionGenerated Water Waves. W.L. L. W. and MILOH. "An Instrumentation System for Measuring Tsunamis in the the Deep Ocean. University of Hawaii." Proceedings of the lOth Conference on Coastal Engineering~ American Society of Civil Engineers. 1961. 1. VAN DORN.. Y. 19. "Descriptive Catalog of Earthquakes of the Pacifie Coast of the United States 1769 to 1928." Advances in Hydroscience~ Vol. 1964. Vol." Report No.. pp.D. Hawaii. AD457729. 1970. and HWANG. T. Hawaii. J. 3.StateoftheArt. 1965 (also published as Scripps Institution of Oceanography Report. H." Proceedings of the Ninth Conference on Coastal Engineering~ American Society of Civil Engineers. ARMY ENGINEER DISTRICT. New York. Academie Press. 453473. TC130.G.C. CHAN." Report IALD1102. VITOUSEK. No.. Volume I . S. "Source Mechanism of the Tsunami of March 28...E. pp.L. TOWNLEY. ARMY ENGINEER DISTRICT.. VAN DORN.Final Post Flood Report... TABER. 4.STREET." Nov. Israel Atomic Energy Commission. 30. 1939. M.G. 166190. M. and CAMFIELD. Federal Clearinghouse No. Pasadena. Report on Survey for Tidal Wave Protection and Navigation. Hawaii. STREET. 1965). EastWest Center Press.. No. Honolulu." Bulletin of the Seismological Society of America~ Vol.E. Adams. and MILLER.. LE MEHAUTE. 1297. "The Tsunami of 23 May 1960 in Hawaii . 16. pp. U. 29. 1964 in Alaska. "Tsunamis.J. ed. F. HONOLULU. Honolulu. "Seismic Activity in the Atlantic Coastal Plain Near Charleston. B. 1934.J. 1966. 1..G... Sept. R. 284301. Oct. VAN DORN. 131133.. Ch. Israel. R. Inc." 1960. 214 . and ALLEN. HONOLULU. 108160.. U. 1914.. 1970... South Carolina. EastWest Center Press. pp. SUYEHIRO.S." Proceedings of the Tsunami Meetings Associated With the lOth Pacifie Science Conference~ International Union of Geodesy and Geophysics. M. 10. "Hilo Harbor. 1968. "Observations and Experiments on Solitary Wave Deformation. July 1975. W. Japan. Ch. 2.. S." Tsunamis in the Pacifie Ocean~ Ch.S. pp. Jan. Tokyo.. "Proposed MidOcean Tsunami Gage and Oceanography Instrument System." Bulletin of the Seismological Society of America~ Vol.
WHITMAN. Calif.L. B." Report for the Redevelopment Agency of the City of Crescent City. Part 4. 1." Technical Report No. .. 1974. DASA1355.. B.C. 47. 1964." Earthquake Engineering 3 Ch. and GALVIN. pp.W. Army Coastal Engineering Research Center." Technical Report CR. 1963. National Engineering Science Co. Tsunami! National Oceanic and Atmospheric Administration .... Vol.C. WILSON. Port Hueneme. B.. California from Tsunami Waves.J.. 4. 11. N.P. "ShallowWater Waves. 1971. 215 . D.J. National Engineering Science Co. WILSON.B. National Research Council. Jan. Engineering Evaluation. Mar. and T~RUM. N. PrenticeHall.. ed. Calif. "NonLinear Dispersion of Water Waves. Crescent City.A.M. 69. R..WEIGEL ." TM25. J .. and HENDRICKSON.. pp." The Great Alaska Earthquake of 1964 3 Oaeanography and Coastal Engineering 3 National Academy of Sciences. Feb.S. O. WEBB." Southern California Edison Co. WIEGEL. Englewood Cliffs. May 1968. "Human Response to Tsunami Warnings. Investigation of Long W aves and Their Effects on the Coastal and Harbor Environment of the Lower Chesapeake Bay. Report No.. 811824." Journal of Fluid Meahanias 3 Vol.027. . N. E. 222228. G.. Part 2. "Long Waves Generated by Nuclear Explosions. R..S. U.W. Oaeanographiaal Engineering 3 PrenticeHall. "Estimate of Tsunami Effect at San Onofre Nuclear Generating Station Units 2 and 3. Wiegel. J. .. "The Nature of Tsunamis. No . II. 1965.J. 1967 .. Dec.M. Englewood Cliffs. WIEGEL.W.W. WIEGEL. WELLER.L. 339412. "Earthquake Occurrence and Effects in Ocean Areas. 1964. California. "Tsunamis. 1962. 1970..L. 27. 1972. L.. Their Generation and Dispersion in Water of Finite Depth. Apr. "The Tsunami of the Al as kan Earthquake. B. Los Angeles. pp.. R. Washington.. WILSON. pp.L.W. 1969. C. WILSON. Calif.." Appendix I. 1972. 253306. ZABUSKY. "Protection of Crescent City. the KortewegdeVries Equation and Solitons.. Aug." Journal of Fluid Meahanias 3 Vol. B. WILSON. Vol.. A.. Naval Civil Engineering Laboratory. SN 572.J. R. U. Washington.
.
1897 29 Nov. Large tsunami on south coast of Ceram. Seattle. Tunisia. Indonesia. Hawaii. minor damage. and Keauhou.APPENDIX TSUNAMIS OCCURRING BETWEEN 1891 AND 1961 Date 30 July 29 Nov. 6 meters observed at Hilo. Philippines (western shores of Basilan Island). 1899 7 Oct. l'laves observed in Hawaii.Columbia Possible tsunami. Much damage and loss of li fe. Tsunami. 1894 1894 1894 1896 1896 Apr. Nemuro. Tunisia. Severe damage at Tacloban. Alaska Banda Sea Kyushu Rikuchu. Maximum height of 30 feet at Napoopoo. no damage. 1897 Tohoku District. Japan Lokris. Tsunami in Tagonoura. 10 June 9 Jan. Tsunami at Hilo. Waves generated did little damage. Japan El Salvador Thessaloniki. Wave at Avachinskaya Bay. Gree ce Kamchatka 21 Sopt. maximum height 3. Kai1ua. 1897 1898 22 Jan. Tsunami in El Sai v ad or and Guatemala. 15 June 5 Aug. Washington 16 May 1892 1893 1894 Marian as Islands Gree ce Wave 4 feet high at Bizerte. 1891 1891 Lerdo. 26 Feb. Runup 80 feet high at Shirahama. Sept. Saloniki Harbor flooded by waves. Source ~lexico Remarks Tidal wave of considerable height at head of the Gulf of California. 1906 Calabria. Hawaii. Variation in terres trial magnetism observed at Sendai preceding the tsunami. Japan. 1899 30 Sept. Ital y Ecuador . Tsunamis associated with these earthquakes were generally small. Slight tsunami in Japan. Min or tsunami at Guam. believed min or. 1905 31 Jan. 9 Aug. devastated ports along the northeast coast of Japan. Greece Yakutat Bay. Japan Large tsunamis induced by earthquakes at Constantinople (Istanbul). Large tsunami at Montserrat. Japan Sulu Sea West Indies Gree ce Southwest Peloponessus. Water in Lake Washington surged on to the beach 2 feet above the mark of the highest water and 8 feet above the lake stage on that date. sorne damage. 22 Mar. Gree ce • Turkey Kashima Sea Sanriku. 217 . Tsunami at Bizerte. 1899 Tsunami reported at Marathos. 5 Ju1y 25 June 1899 1901 1902 1902 1904 1905 8 Sept.
26 May 7 Aug. Tonga New Britain New Hebrides Islands Tsunami he igh t 5 met ers at Port of Cal der a where i t caused damage. 24 men killed on Sirnusirijimi. Mexico. Ch ile North of Vava 'u. Wave swept coast from Lebak ta Glan. causeway washed out. Hawaii. 1906 Source Valparaiso. Ital y Puna. Italy Philippine Islands Remarks Recorded in Hawaii and Japan. Hawaii Messina. Tsunami in New Guinea. 1913 1913 1913 1914 1914 1915 1916 1917 1917 1918 South Island. Bonin Islands. Tsunami generated. Small tsunami. Sept.5 ta 3 meters high on shoreline. 1920 218 . 1908 28 Dec. 23 Oct. height estimated at 24 feet at sorne points. Nov. also damage at Mayaguez. Tsunami caused fata1i t ies and damage at Point Borinqucn and Aguadilla. Weak tsunami in Hawaii. 4 Dec. breakwater destroyed at Messina. Weak tsunami. Wave with 10foot amplitude caused serious damage ta small boats in Kagoshima Harbor. Observed in Hawaii. Philippine Islands Kuril Islands Puerto Rico 12 Jan. Seismic waves in Lake Bombon washed away severa! villages. Wave similar ta 1 January 1916. 1 May. 15 Sept. 26 June 15 Aug. main damage at Kingston. 1908 1911 26 Feb. New Zealand Sangi. Heavy damage along shoreline. Water level in Rabaul Harbor fell 15 feet and rose ag ain rapidly. Tsunami recorded at Messina and Ca tania. 14 Apr. Slight tsunami at Samoa. Ch ile Dampier Strait Jamaica Mexico Calabria. minor damage at Hilo. Thirtyfoot wave at Acapulco. 11 Oct. Weak tsunami. East Indies Near east end of New Guinca Sakuraj ima. waves 2. Japan North coast of New Guinea Ionian Sea New Britain Kermadec Islands Tonga Islands Southern Mindanao. Tsunami generated. 1918 11 Oct. 1 Jan. 1907 1907 1907 20 Sept. Probable tsunami. sorne loss of life.TSUNAMIS OCCURRING BETWEEN 1891 AND 1961Continued Date 19 Aug . Tsunami observed at Futami Harbor. S May 1918 1918 1918 1919 1919 20 Sept. Fortyfoot wave in Samoa. 30 Apr. wave recorded at Honolulu. negligible along Chilean coast. Wav'e height about 5 feet on Greek coast. Tsunami at Uruppu Island. 14 Mar. 30 Jan. 1906 14 Jan. Copiapo. some damage in Hawaii.
all buildings swept away except the chur ch. Mexico J ali seo. Height of tsunami about 5 feet on Sea of Japan coastline.S. Japan Remarks Wave height 10 feet on Al banian coast. Many boats wrecked. Ne ar Piraeus. U. 1927 1927 1928 1928 1929 1929 1929 1929 H June 17 Jan. Pangasman. Tsunami destructive in many places along the coast of Chile. 3 June 18 June 1931 1932 1932 Solomon Islands Jalisco. 14 Apr. damage and loss of life on Burin Pi)ninsula. Railroad track swept away between Cayutlan and Manzanillo. possibly a storm surge. 6 Mar. 16 Mar. caused min or damage. 1924 1924 1926 16 Sept. maximum amplitude 15 feët at Hilo.. Wave swept Palmerston Island. Me xi co 219 . Tsunami at Guadalcanal. Mexico. one pers on killed. Tsunami measured in Hawaii (maximum ampli tude 1 foot at Hilo). Tsunami hit Newfoundland. Tsunami observed in Hawaii. maximum height of 26 feet at Atami. 1923 9 Jan. 7 Mar. 26 May 18 Nov. Possible minor tsunami. Small tsunami recorded in Crime a. Tsunami in Hawaii. Fourfoot wave at Que en Charlotte City. La~Tolla. Sixfoot wave at Surf. Waterfront damaged at Puerto Angel. Tsunami on French coast (Atlantic coast). Japan Black Sea Black Sea California Kamcha t ka. Venezuela Aleutiari Islands Queen Charlotte Islands. Canada Grand Banks 11 and 12 Sept. Eighteen native villages destroyed on San Christobal Island. Tsunami hi t towns on shore of Sagami Bay. mu ch damage. 300 miles northwest of Raratonga. Mexico.TSUNAMIS OCCURRING BETWEEN 1891 AND 1961Continued Date 18 Dec. 0. 1927 Nov. approximately 50 persons killed.R.24 inch high at Minor tsunami recorded at Hilo. tsunami on coast of Kamchatka. 11 Nov. minor damage in Hawaii. South China Sea Tonga Island Small tsunami at Agno. 1920 1922 1923 1923 Source Strait of Otranto At a cuma. 3 Oct. 26 June 1926 1927 1927 Solomon Islands Saint Pierre and Miquelon Tango. Ch ile East Kamchatka Kamchatka Kwanto. one person killed. 13 Apr. Waves 7 fe et high on north coast of Crete. Gree ce Mexico Cumana. 25 Apr. 1926 18 Nov. Small tsunami recorded in Crime a. Small tsunami .S. 1 Sept. 28 Dec. 3 Feb.
Kurile Islands 220 . Mi nor tsunami. Philippine Islands. British Columbia. Japan Blanche Bay. waves along Pacifie coast of Japan 2. Costa Rica. 5 to 5 meters high ne ar the Kii Peninsula. sorne damage. Tsunami at San Esteban. 5 Dec. 1945 Arabian Sea 1 Apr. Town of Matanzas badly damaged and abandoned. Alaska Off Kinkazan. 1932 3 J. 23 June 1946 1946 1946 Aleutian Islands. waves generated in Georgia Strai t flooded fields and highways. Tsunami observed in nort hern Japan. Tsunami on coast of Hokkaido. Alaska. 22 Mar. 1946 Island. Bot tom of Deep Bay sank from 9 to 84 fe et. 28 and 29 May 1937 6 Mar. 24 Aug. Volcanic eruption. Immense damage. mi nor damage on coast of Hai ti. Japan China Se a Li tuya Bay. 1938 1938 19 May 23 t. 10 Nov. Peru Chile Japan (Kumanonoda) Remarks Small tsunami. 1941 1942 1943 1944 27 Nov. Tsunami observed along coast of Sea of Japan. Damage and loss of life in Alaska. Japan Off south coast of Alaska Near Ogasima.Costa Rica Near Lima. Canada Dominican Republic ~1atua Aug. 1933 1934 1936 1936 3 Nov. Gree ce Sanriku coast. 8. Mexico Hierissos.TSUNAMIS OCCURRING BETWEEN 1819 AND 1961Continued Date 22 June 1932 Source Jalisco. Canada Macassar Strai t Ibaraki. Ampli tude less than 2 feet in Aleutians.022 people killed. 75 foot.180 vessels wrecked. Highest waves 14 to 16 feet from 26 Sept. Tsunami generated. Volcanic explosion. Japan Off Iwaki. Japan Northern part of Sea of Japan Panama . 8. Small tsunami.lar. more than 100 persons ki lled. Mi nor tsunami. 27 Oct. 14 Feb. height of O. 6 Apr. 12 and 13 Nov. Minor tsunami. tsunami at Karachi and Bombay where there was damage and loss of l i fe. Heavy damage on coast of Pakistan and India. Slight tsunami recorded in Peru. Maximum runup of 400 feet.lay 5 Nov. ampli tu des more th an 3 fe et. 3. New Britain Solomon Islands Queen Charlotte Islands. two new islands appeared in the Arabi an Se a. crest to trough.831 houses destroyed. heavy damage and many li v es !ost in Hawaï i. Small tsunami noted in Gulf of Orphano. Small tsunami at Valparaiso. Mi nor tsunami on Japan coast. 1938 1938 1938 1938 1939 1940 Mi nor tsunami. 7 Dec. sorne damage. 1 May 2 Aug. Slight tsunami at Punta Arenas.
1952 in Reef 4 Nov. 1946 1947 1948 Source Hon s hu.. Japan Sept.TSUNAMIS OCCURRING BETWEEN 1891 AND 1961Continued Date 21 Dec. 1952 17 Mar. maximum height of 12 feet at Hilo. Kurile Islands Near Dominican Republic lonian Islands Cyprus Suva Earthquake BosoOki. Cyprus (no major damage ).rthqu3. 14 Dec. Mexico Kona. 19 Apr.2 feet at Talara. Tsunami caused damage at LaVela. largest wave 24 September. Hawaii.ke 10 Sept. Hawaii Tokachi Oki Earthquake. Slight tsunami at Kalapana. 1. Tsunami.ot wave at Ketchikan. Tsunami in Fiji Islands. Japan). Feb. 23 Oct. Submarine volcano eruption at Myojin Reef (400 kilo meters south of Tokyo.500 people killed on Shikoku Island. 1953 14 Sept. 6 Oct. Philippines Near New llebrides ~lyoj 13 July Sept. Large wave on coast of Methoni. Large tsunami. Small tsunami. Sea of Crete Tonga Island Queen Charlotte Islands. 1949 1949 1949 1950 1950 1950 1951 4 Mar. believed caused by a submarine lands 1 ide. Very slight tsunami in Hawaii. 20 Oct. Venezuela. 1948 22 Aug. Alaska. 5 Oct. ampli tude of 0. East Kamchatka Earthquake. Tsunami gen erat ed. Twofo. amplitude of 0. Eightfoot wave. 1952 1953 1953 Paramushir Island.7 foot. Mi nor tsunami in Hawaii. Tsunami on Pacifie coast of Hokkaido from Nemuro to Hidaka (main damage in Ki ri ttapu and Tokotan ). 221 . Minor (very slight) tsunami in Hawaii. 19 ~!ar. mass of ice accompanied the tsunami. 12 Dec. Series of Waves at Paphus. Japan Peruvian E. Tsunami had maximum height of 3 met ers at Choshi. Tsunami kil led one pers on near Mercedes. Gree ce ~le s senia. 5 Mar. Wave amplitude of 3. Small tsunami. Very slight tsunami in Hawaii. 1952 1952 1952 South of Hawaii East of Mindanao. Japan Slight Tsunami in Hawaii. Remarks Damage on south coast of Japan (1\'akayama Prefec t ure) . Peru. Very slight tsunami in Hawaii. Japan (Nanka i Earthquake) Kononi. 1953 26 Nov. Canada Solomon Sea Philippine Islands Costa Rica Guatemala Guerrero. Very slight tsunami. 29 Dec. 2 foot at Puerto Plata. Cliff collapsed near Napoopoo creating a 12foot wave which destroyed a boat dock. 1953 1953 1955 1955 1956 Central Chi le North coast of Hokkaido. presumed to have destroyed hydrographie research vessel. 18 Jan. 21 Aug. 31 May 11 Aug.
damage at Hilo. height small. Greece Aleu tian Islands Me. Tsunami on coast of Miyazaki prefecture occurred at low tide. 4 May 1959 18 Aug. 1956 9 . wave 100 feet high near source. Ofunato. 1958 9 July 1958 6 Nov. and Kiritappu heavily damaged in Japan.xico Ecuador Lituya Bay. 1959 7 Feb.607 222 . Shizukawa. wave runup estimated at 1. Tsunami damaged Esmeraldas and Quayaquil. Small tsunami. Sma11 tsunami struck Ancon. 1960 29 Feb. Waves in lake from earthquake. GOVERNMENT PRINTING OFF I CE : 1980 . Giant wave from rockfall. heavy damage.S. Tsunami recorded at Talara.Mar. U.622 .740 feet at· one point. Slight tsunami in Hawaii and the Aleutian Islands. Waves 3 to 5 meters high in South Kurile Islands (referred to as EtorofuOki Earthquake by Japanese). Hawaii. 1960 16 Jan. 1960 20 Mar. sorne deaths.S. Damage and deaths along Peruvian coast. East coast of Honshu Northern Peru Near Kamchatka Hebgen Lake. Tsunami waves in Hawaii. 1961 tr U . 1961 27 Feb.R. Fourfoot wave. S. Morocco. Tsunami at Agadir. 1959 13 Jan. Tsunami recorded at Acapulco and Salina Cruz. Montana Southern Peru Morocco Coast of Japan Chile Coast of Peru Ibaraki Oki Earthquake HiugaNada. 1960 22 May 1960 20 Nov. 1958 Wave believed to be caused by submarine landslides. 1957 28 July 1957 19 Jan. Pacifie coast. Fivefoot wave locally. Source 1956 Ne ar Kamchatka Greek Archipelago Volos. Alaska Iturup Earthquake. 1959 at Miyako. Japan Remarks 9 Ju1y 1956 2 Nov.TSUNAMIS OCCURRING BETWEEN 1891 AND 1961Continued Date 30 ~lar. Very faint tsun~~i 22 Jan.
.
.
Edge waves. 1980. Coast al engineering. Water waves. and discusses methods of predicting tsunami flooding. Coast al Engineering Research Center .Fort Belvoir. Title. I. Special report no. Coast al Engineering Research Center.he report summarizes available information. 1. Camfield. Mathematical models. Coastal engineering. 201. T.US81sr no. Camfield.US81sr no. Coastal Engineering Research Center.(Special report .S. TC203 . TC203 . 222 p. 1980. S. 2. Series: U. 3. II. Ti tle. no. . Mathematical models.Camfield. I. Math. This report provides a source of stateofthe·art information on tsunami engineering. Coastal Engineering Research Center . Tsunami engineering 1 by Frederick E. Tsunami engineering 1 by Frederick E.U. Fort Belvoir.S. : U. TC203 . Water waves. Bibliography : p. 4. S. 27 cm. I. Tsunami engineering 1 by Frederick E. no. identifies gaps in existing knowledge. II. Coastal Engineering Research Center. Tsunamis. This report provides a source of stateof theart information on tsunami engineering. and discusses methods of predicting tsunami flooding. 27 cm. S. S. Coast al Engineering Research Center. 6) Cover title.U. S. Bibliography : p. 4.U581sr no. . : ill .US81sr no. Camfield. The report summarizes available information. Coastal Engineering Research Center. 2. Va . 1. 3. The generating mechanisms of tsunamis and the method of determining the prob. Edge waves. : Hi . 201. Edge waves. II. 1980. Series: U. Series: U. Water waves. The report summarizes available information. 3. : U.Fort Belvoir. Frederick E. identifies gaps in existing knowledge. . Special report no .(Special report . 6 627 . 201. : U. Tsunami engineering 1 by Frederick E.(Special report . Camfield. 6. Special report no.U.S. 6 627 Camfield. Coast al Engineering Research Center. Coastal engineering. : U. Series: U. Fort Belvoir. identifies gaps in existing knowledge. Va. Va. 6) Cover title. 222 p. Coast al Engineering Research Center. . : ill .S. 6. Coastal Engineering Research Center . 2. II. 27 cm. 201 • This report provides a source of stateoftheart information on tsunami engineering. Title. Tsunamis. Coastal engineering. Title. Tsunamis. Bibliography : p. and discusses methods of predicting tsunami flooding. 4. 4.U. 6. 222 p. Coast al Engineering Research Center. Va. no. 6 (27 Camfield. Bibliography : p.bility of occurrence are given . Offshore structures. 3. Frederick E. S. . 2. I. S. The report summarizes available information. 1980. 6 627 Camfield. identifies gaps in existing knowledge. . S. 27 cm. Offshore structures. TC203 . 6) Cover title. Special report no. and discusses methods of predicting tsunami flooding. 1. 6) Cover title. Offshore structures. Mathematical models. The generating mechanisms of tsunamis and the method of determining the probability of occurrence are given. 222 p. The generating mechanisms of tsunamis and the method of determining the probability of occurrence are given. This report provides a source of state·oftheart information on tsunami engineering. Offshore structures. S..S. S. Tsunamis. no.ematical models.(Special report . Edge waves. S. Water waves. Frederick E. : ill . Coastal Engineering Research Center . 6. 1. The generating mechanisms of tsunamis and the method of determining the probability of occurrence are given. Frederick E.
.
Camfield. Frederick E.U. : U. Mathematical models. 5. and discusses methods of predicting tsunami flooding. 3. 2. Coastal Engineering Research Center. Coast al Engineering Research Center . I. 222 p. TC203 . Special report no. Coast al Engineering Research Center.U. The generating mechanisms of tsunamis and the method of determining the probability of occurrence are given. identifies gaps in existing knowledge. no. 4. Coast al engineering. Coastal engineering. Coast al engineering. Series: U. This report provides a source of stateoftheart information on tsunami engineering. Tsunami engineering 1 by Frederick E. 6 627 6. and discusses methods of predicting tsunami flooding. : U. 6 (27 Camf ield. Coast al Engineering Research Center. S. .S. Title. Tsunamis. Water waves. Tsunamis. Series: U. II. Frederick E. Coastal Engineering Research Center.U581sr no. Coastal Engineering Research Center. 5. . Tsunami· engineering 1 by Frederick E. S. Tsunami engineering 1 by Frederick E. : ill .U. Coastal Engineering Research Center. 1980. . TC203 .(Special report . Special report no. 3. 201 • This report provides a source of stateoftheart information on tsunami engineering.S. Special report no. 2. The report summarizes available information. 201. Bibliography : p. 6. Frederick E. II. 27 cm. Water waves. S. 2. 201 . Camfield.U581sr no. 2. Tsunami engineering 1 by Frederick E. : ill . 3. S. 1. Special report no. 1980.Camfield. identifies gaps in existing knowledge. Mathematical models. II. 1. Edge waves. Series: U. 6 627 Camfield. Edge waves. S. 222 p. Offshore structures. no. Water waves.(Special report . Offshore structures. Coast al Engineering Research Center . Title. 4. 6) Cover title. I. 6) Cover title. Series: U. S. . TC203 . Coastal engineering. Camfield. . .Fort Belvoir. : ill . and discusses methods of predicting tsunami flooding. Va. Tsunamis. Va. Water waves.Fort Belvoir. : ill . and discusses methods of predicting tsunami flooding. 6) Cover title. 222 p. I. This report provides a source of stateoftheart information on tsunami engineering. Bibliography : p. identifies gaps in existing knowledge. Coast al Engineering Research Center. Bibliography : p. 27 cm.Fort Belvoir. Camfield. S. 27 cm. Offshore structures. : U. 1.tr581sr no. Title. . The generating mechanisms of tsunamis and the method of determining the probability of occurrence are given. . This report provides a source of stateoftheart information on tsunami engineering. no. Coast al Engineering Research Center. 3. The report summarizes available information. Frederick E. S. Va. Edge waves. Mathematical models. The generating mechanisms of tsunamis and the method of determining the probability of occurrence are given. Title. : U.(Special report . TC203 . Edge waves. 6. 222 p. Va. I. 1980.(Special report . 201. 5. Tsunamis. The report summarizes available information.Fort Belvoir. Coastal Engineering Research Center . II. Coastal Engineering Research Center . The report summarizes available information. S. 5. 1. 6 627 . The generating mechanisms of tsunamis and the method of determining the probability of occurrence are given. no. identifies gaps in existing knowledge. 6. Camfield. Bibliography : p.U. S.U581sr no. 4. 6) Cover title. Offshore structures. 4. 27 cm. Mathematical models. 1980.
l ..