Earthquake-Generated Tsunami - Tsunami - page 86

tsunami towards the center of the ocean. Thus places like
Hawaii and French Polynesia are particularly vulnerable. An
exception to this occurs with tsunami originating in Alaska.
Here most of the energy in a tsunami is beamed towards
California and Chile. This certainly was the case with the
Alaskan earthquake of 1964 (Ben-Menahem and Rosenman
1972 ). Alaskan earthquakes do not affect French Polynesia
because of topographic modification across the intervening
ocean. As one goes westwards along the Aleutian Island
chain, the directivity of wave propagation sweeps towards
Hawaii. For these reasons, once the epicenter of an earth-
quake has been located around the Pacific Rim, it is a simple
task to plot the resulting tsunami's direction of travel and
likely area of impact. Outside the Pacific Ocean, the east
coasts of Indian and Sri Lanka received the full brunt of the
Indian Ocean Tsunami of December 2004 because they were
orientated parallel to the rupture zone extending northwards
from Indonesia (Ammon et al. 2005 ; Subarya et al. 2006 ).
Very little energy for this tsunami propagated into the
southeastern Indian Ocean because this region lay in a sha-
dow zone were beaming was minimal.
It is assumed that faulting occurs in massive rock units.
Faulting is not this simple. In subduction zones, sedimentary
rocks are often being buried (Geist 1997 ). Accretional
wedges can built up on the seabed as surface sediments are
scraped off as one plate dips below another. This is espe-
cially prominent where low-angle trusting is occurring.
Marine sediments are also water saturated. Close to conti-
nents these sediments can contain organic material that
decomposes into methane, leading to gas-rich layers. Both of
these types of sedimentary layers have low densities. Thrust
rupturing into these types of sedimentary layers can increase
the excitation of tsunami waves by a factor of ten. Only 10 %
of the force of the rupture needs to occur in the overlying
sedimentary layer for this to happen. Tsunami earthquakes
may also be generated by aftershocks associated with rup-
turing of an active fault through softer sediments near the
seabed. In summary, the most prone area in the ocean for the
generation of large tsunami is along a subduction zone where
one plate moves upwards over another at a low angle, and
where this movement propagates through less consolidated
sediments near the seabed. In these circumstances, while the
moment magnitude, M w , of the tsunamigenic earthquake
may be an order of magnitude smaller than expected, the
resulting tsunami can be very large.

the Pacific Rim, plates are moving at consistent rates. For
instance, in the Alaskan region, the Pacific and North
American Plates have generated continual earthquake
activity over the past 150 years—as stresses build up to
crucial limits and are periodically released at various points
along the plate margin. However, stresses may not be
released at some points. These appear in the historical
record as abnormally aseismic zones surrounded by seis-
mically active regions (Bryant 2005 ). The former locations
are called seismic gaps and are believed to be prime sites
for future earthquake activity. The Alaskan earthquake of
1964 filled in one of these gaps, and a major gap now exists
in the Los Angeles area.
The seismic gap concept is flawed. Many earthquakes
occur in swarms, with the leading earthquake not necessarily
being the largest one. Tsunami generation also tends to occur
in the area encompassing aftershocks. More significantly,
when many tsunami events are examined, the arrival times
of the first wave along different coastlines tend not to orig-
inate from a single point source. Also it is know that many
segments along a subduction zone can fail in a single
earthquake event (Stein and Okal 2011 ). The more segments
that rupture, the greater the earthquake and the resulting
tsunami. Hence sudden adjustment of plate contact along
subduction zones can potentially occur along hundreds of
kilometers of plate boundaries. The 2004 Indian Ocean and
2011 Japanese T ¯ hoku events were generated by this pro-
cess. Additionally, earthquakes and the tsunami they gen-
erate are chaotic geophysical phenomena (Bak 1997 ). They
should thus be generated by a spectrum of seismic waves
with varying amplitudes and periods. If this is the case, then
the system of tsunamigenic earthquakes behaves as white
noise. One of the aspects of such systems is that earthquakes
can recur at the same location rather than in areas that are
more quiescent. This behavior is characteristic of subduction
zone earthquakes (Satake 1996 ). For example, the October 4,
1994 Kuril Islands Tsunami occurred at the same location as
a previous event in 1969. A casual glance at the source
location of tsunami over time in the Pacific will show that
tsunami originate repetitively within a 100 km radius of the
same location in many regions (Fig. 1.2 ). Finally, seismic
research is limited by the length of the seismological
record—just over a century (Stein and Okal 2011 ). This
record is no indicator of the frequency of tsunamigenic
earthquakes at a particular location. Just because a large
earthquake hasn't happened, doesn't mean it won't, espe-
cially along subduction zones. This point was brought home
by the T ¯hoku Tsunami of March 11, 2011. Paleo-studies
indicated a magnitude 9 earthquake was possible in the
region; historical records did not (Minoura et al. 2001 ).
There is thus a pressing need for more paleo-studies in
regions adjacent to any subducting plate.

5.3.2 Seismic Gaps and Tsunami Occurrence

The concept of earthquake cycles depends upon crustal
movement occurring at constant rates over geological time
and the build-up of frictional drag along fault lines. Around