Development
of a Model for the Real Time Evaluation of the Tsunami Risk
George Pararas-Carayannis
(Excerpts from Paper
presented at the The International Association for the Physical
Sciences of the Oceans (IAPSO) Conference in Honolulu, Hawaii,
August 1995)
Abstract
The real time assessment of the tsunami risk from large Pacific-rim
earthquakes is not always possible because the dynamic interactions
of major and minor crustal plates and source mechanisms for tsunami
generation are not well established or understood. The tectonic
interactions of crustal subplates that have caused recent destructive
earthquakes and local tsunamis have distinctively different seismic
source characteristics than those of major crustal plates which
involve extensive subduction and are responsible for destructive
Pacific-wide tsunamis. Determination of earthquake parameters
of Richter and moment magnitude, as well as epicenter and focal
depth is not adequate seismic information to differentiate and
evaluate, in real-time, the tsunami risk potential from all large
earthquakes.
Using the Japan/Kuril islands region as an example, this paper
reviews, compares and evaluates the seismic and hydrodynamic
source parameters of recent and historical earthquakes, the tectonics
of the region, the complex dynamic source interactions responsible
for destructive tsunami generation and the conventional tsunami
risk assessment methods. Finally, the present study proposes
a model for a more accurate, real-time, evaluation of tsunami
generation and tsunami risk for the source area as well as for
distant terminal points.
The proposed model would provide a real-time analysis of periods
and patterns of unaltered, long seismic waves and trends of first
motions of such waves, to deduce earthquake characteristics at
the source region through the development of a pattern recognition
methodology. The latter can be developed from comparison of long
period waves of an actual earthquake with mathematically equivalent
theoretical radiation patterns of simulated sources and analytical
solutions of all possible hypothetical torque couples embedded
in an unruptured elastic medium. These analytical torque couples,
simulating numerous theoretical deformation models, rotating
in opposite directions, would deform the medium thus radiating
elastic waves in patterns identical or similar with those in
which an earthquake source radiates seismic waves. By using a
number of such pre-solved, theoretical-analytical models, the
pattern of the actual seismic radiation pattern from an earthquake
source could be compared to arrive at a fairly reasonable source
mechanism for the actual event. This could include dimensions
of the source area, orientation and length of faulting, degree
and percentage of vertical subduction, energy transfer to the
overlying elastic medium, energy radiation and other pertinent
source data.
Thus deduced source parameters could serve independently to assess
qualitatively, in real-time, the relative tsunami risk from the
earthquake or they could be used as input to existing hydrodynamic
models to provide quantitative estimates of the distribution
of tsunami energy flux and expected tsunami run-up at terminal
points. An integrated computer model with algorithms can be developed
which can provide, in real time, fairly good evaluation of the
tsunami potential and possible tsunami risk for the source area
as well as for the rest of the Pacific Ocean. In addition to
the benefits in the real time assessment of the tsunami risk
for warning purposes, the proposed model would help in the understanding
of earthquake-tsunami focal mechanisms. Furthermore, the proposed
model could become a useful research tool which could be applied
to evaluate the plate and subplate tectonic interactions of different
regions of the world.
Introduction
The real time assessment
of the tsunami risk from large Pacific-rim earthquakes is not
always possible because the dynamic interactions of major and
minor crustal plates and source mechanisms for tsunami generation
are not well established or understood. The tectonic interactions
of crustal subplates that have caused recent destructive earthquakes
and local tsunamis have distinctively different seismic source
characteristics than those of major crustal plates which involve
extensive subduction and are responsible for destructive Pacific-wide
tsunamis. Determination of earthquake parameters of Richter and
moment magnitude, as well as epicenter and focal depth is not
adequate seismic information to differentiate and evaluate, in
real-time, the tsunami risk potential from all large earthquakes.
In assessing the tsunami risk for the region the present study
reviews historical tsunamis and makes comparisons of seismological
and hydrodynamic parameters and reviews the dynamics of tectonic
interactions that are responsible for tsunami generation. Finally
the study provides a quantitative estimate of risk for each subsection
of the region.
Historical
Earthquakes and Tsunamis in the Japan/Kurile Island Region
Subduction along the
Japanese Trench is responsible for many large historical earthquakes
and catastrophic tsunamis in the immediate region of Japan/Kuril
islands with minor damaging effects elsewhere in the Pacific.
Tectonic movements in the East Sea/Sea of Japan, Sea of Okhtsok
and inland seas have been responsible for large earthquakes and
tsunamis but their effects have been confined to the region.
The Earthquakes
of October 4, 1994 and January 17, 1995
Two major and very
destructive earthquakes ocurred in the region in October 4, 1994
and on January 17, 1995. The October 4, 1994 earthquake generated
a small tsunami which was damaging in Japan and the Kuril islands
but did not pose a Pacific-wide threat. The January 17, 1995
earthquake produced no noticeable tsunami. A Pacific-wide Tsunami
Warning was issued by PTWC at great cost to the Pacific community
on October 4, 1994.
Comparison
of Similarities of the 1963 and 1994 Earthquakes and Tsunamis
The historical record shows that this event was very similar
in magnitude and epicenter to two quakes which occurred on October
13, and October 19, 1963, from the same exact region (Northern
Hokkaido, Kuril islands). for which two Pacific-wide tsunami
warnings were issued then. Neither of these events generated
a Pacific-wide damaging tsunami in Hawaii or elsewhere.
A simple reference to the historical Tsunami data Tsunamis in
the Hawaiian Islands, would have revealed that there has never
been a Pacific-wide tsunami threat, even from the largest earthquakes
in the Northern Hokkaido, Kuril Islands region; only local damaging
tsunamis have been generated. Table 1 provides a comparison between
the earthquakes and tsunamis of October 4, 1994 and October 13,
1963.
Table 1. Comparison
of Similarities of the 1963 and 1994 Earthquake and Tsunami Events (Re: Catalog of Tsunamis in
the Hawaiian Islands/Pacific Catalog)
Oct 12,1963
:
Magnitude 8.2 , Epicenter
44.8 N, 149.5 E
Depth: 33 km
Time 05:18 UTC
(Area Affected by local damaging tsunamis: South Kuril Islands
(Urup, Shikotan,Kunashir, Urup, Iturup, Paramushir, etc )
Japan: Hanasaki, Kushiro, Hachinohe,Chichijima
Highest Recorded at Midway 1.03 meters adjusted to 0.6 meters
peak to trough
Highest in Hawaii: Kahului 0.4 meters adjusted
to 0.8 meters peak to trough
PTWC issued Pacific-wide Tsunami Warning
Oct 4,
1994 :
Magnitude 8.2 Epicenter
43.668 N, 147.333 E
Depth: 33 km
Time 0:323 UTC
(Area Affected by local damaging tsunamis: South Kuril Islands
(Shikotan,Kunashir,Iturup, etc.)
Japan: Hanasaki, Kushiro,Hachinohe,Chichijima
Highest Recorded at Midway I. 0.54 meters (peak to trough)
Highest in Hawaii: Kahului 0.8 meters (peak to trough)
PTWC issued Pacific-wide Tsunami Warning
There are similarities in magnitude, and epicenter location.
There is overlap in tsunami generation area. Tsunami heights
and damage in immediate area are very similar. Both tsunamis
recorded similarly at the Midway Island. The Kahului tide gauge
recordings were the highest for both events in the Hawaiian islands.
Although it is only a coincidence, both events occurred in the
same month in October and even at approximately the same UTC
time.
Table 2 , provides a similar comparison for the major aftershocks
of the two events.
Table 2. Comparison
of the two major aftershocks of the October 4,1994 and October
19, 1963 earthquakes and tsunamis.
Oct 19,1963:
Magnitude 6,75-7 Epicenter
44.7 N , 150.7 E
Depth: 33 km
Time 00:53 UTC
(Area Affected by small observe or recorded local tsunamis: South
Kuril Islands (Urup, Shikotan,Kunashir, Urup, Iturup, Paramushir)
Japan: Recorded or observed at Hanasaki, Kushiro, Hachinohe,Chichijima
Highest Recorded at Midway I. 0.2 meters adjusted to 0.4 meters
peak to trough
Highest in Hawaii: Kahului 0.4 meters adjusted to 0.8 meters
peak to trough
PTWC issued Pacific-wide Tsunami Warning
Oct 9, 1994:
Magnitude 7.2 Epicenter
Same approximate location as major quake
Depth: Shallow
Time 0756 UTC
(Area probably affected by local tsunamis: South Kuril Islands,
unknown height but expected to be small
Japan: Hanasaki (9cm), Kushiro (3cm),
Highest Recorded at Midway I.: unknown
Highest in Hawaii: unknown
PTWC did not issue a Watch or a Warning.
The Great Sanriku
Earthquake and Tsunami of 1933
Even the great Sanriku earthquake of 1933 did not generate a
tsunami that was of any consequence elsewhere in the Pacific.
The 1933 tsunami had originated from the Sanriku area, but was
considerably south and westward of the October 4, 1994 quake's
generating area. The epicenter of the 1933 earthquake had been
at 39.1 N, 144.7 E,, in a region of much greater vertical subduction
along the Japanese Trench, near the island of Honshu and approximately
300 nautical miles south and 120 nautical miles westward of the
October 4, 1994 event. Furthermore, the orientation of the fault
zone and the tsunami generating area for the 1933 earthquake
were different; greater amount of energy radiated unobstructed
towards Hawaii. Crescent City, California, recorded the largest
tsunami oscillation (1.1 meters), anywhere in the Pacific from
that event even with the more optimum tsunami energy orientation
effect from that source.
Plate
Tectonics of the Japan/Kuril Island Region 
The plate tectonics of the Southern Kuril islands-Northern Hokkaido
region are quite different than those near Honshu or the islands
near the southern portion of the Japanese Trench. Specifically,
the October 4, 1994 earthquake (and the 1963 earthquakes) ocurred
at the Pacific side boundary of a smaller tectonic subplate which
includes the Sea of Okhotsk and perhaps a portion of the northern
part of the East Sea/Sea of Japan. This subplate is characterized
with large earthquakes such as the 1963 and 1994 events but with
lesser vertical subduction and rotational movement (possibly
counterclockwise) as the North Pacific Plate grinds against it.
The whole area appears to be highly fractured in an east-west
direction and the crustal displacements appear to be ocurring
along the boundaries of subplates that may not be longer than
200-250 miles. The fractured smaller plates along the northern
part of the Japanese Trench limit the extent of crustal displacements
and therefore the size of the resulting tsunami. The historical
record supports this as well. This is the reason why very large
magnitude earthquakes from that region produce only locally catastrophic
tsunamis.
The East Sea/Sea of Japan represents another subplate with some
subduction and possible clockwise or counterclockwise rotational
movement as it interacts against the Okhotsk plate, along the
inland sea boundary. The 1983 earthquake and tsunami in the East
Sea/Sea of Japan was along the boundary of these interacting
subplates. Thus, the mechanisms and tectonic interactions of
subplates that caused both the 1983 earthquake and Sea of Japan
tsunami and the October 4, 1994 earthquake and tsunami, are different
than those of other typical subductions caused by the Pacific
plate movement along the northern part of the Japanese Trench.
This appears to hold true for other regions of the Pacific.
Development
of a Model for the Real Time Evaluation of the Tsunami Risk
With proper applied
type of research and analysis, the source mechanisms of large
earthquakes can be analyzed in real time to assess the relative
tsunami risk at distant terminal points. Similarly, a methodology
can be developed which may provide in real time even quantitative
estimates of tsunami run-up for warning purposes. Unecessary
tsunami warnings can be eliminated through such real-time assessment
Presently, the only data available within the time constraints
of a potential warning decision is some basic seismic data. Such
data is simply not sufficient. Tsunami risk evaluation for warning
decisions, must concentrate in the real time understanding of
the earthquake-tsunami focal mechanisms and evaluation of the
seismic parameters. For example we now use seismic moment (Mo)
as the more reliable method for measuring the energy and magnitude
of the larger earthquakes which saturate the Richter scale. Why
not extend these measurements for further assessment of earthquake
source parameters and mechanism, in real time?
The same measurements that provide an estimate of the moment
magnitude can be used to study other earthquake source parameters.
For example, since the seismic moment is related, not only to
the size of the earthquake but also to the overall deformation
at the source and to the fundamental parameters of the faulting
process, why not use these measurements to determine indirectly
earthquake source mechanism that can help with the real time
tsunami evaluation?
Proposed
Methodology and Model Development for the Real Time Assessment
of Earthquake Source Parameters.