Náttúrufræðingurinn - 1985, Síða 34

C. Scholz, R. Turner & F. Wu. 1977.
Prediction of the Haicheng earthquake.
— EOS, Trans. Am. Geophys. Union
58: 236-272.
Scholz, C.H., L.R. Sykes & Y.P. Aggar-
wal. 1973. Earthquake prediction: A
physical basis. - Science 181: 803-
809.
Semenov, A.N. 1969. Variation in the
travel time of transverse and
longitudina waves before violent earth-
quakes. - Izv. Acad. Sci. USSR Phys.
Solid Earth 4: 72-77.
Sveinbjörn Björnsson & Páll Einarsson.
1981. Jarðskjálftar. — í: Náttúra ís-
lands: 121-155. Alnrenna Bókafélagið,
Reykjavík.
Þorvaldur Thoroddsen. 1899. Jarðskjálftar
á Suðurlandi. - Hið íslenska Bók-
menntafélag, Kaupmannahöfn: 199 bls.
SUMMARY
Earthquake prediction
by
Páll Einarsson
Science Institute
University of Iceland
Reykjavík
Considerable progress has been made in the
last 15 years or so towards the goal of seismolog-
ists to predict earthquakes. This progress is
intimately linked with new theories on global
tectonics, improvements in the seismograph
coverage of the world and better understanding
of physical processes taking place at the earth-
quake source. The theory of plate tectonics,
c.g., led to the concept of the „seismicity gap",
which has been successfully applied to delineate
areas along plate boundaries where the prob-
ability for large earthquakes is highest. The
discovery and documentation of changes in the
velocity of seismic waves prior to earthquakes in
thc Garm region of the USSR, and in New York
demonstrated that measureable changes prc-
cede at least some earthquakes, and provided
the impetus for further research. Physical mod-
els were developed to explain these changes.
The models are based on experimental results
from rock mechanics, especially the observed
phenomenon of dilatancy, i.e. the volume
increase that occurs in a rock specimen shortly
before it breaks under loading. The models
predict thc changes in velocities, seismicity,
resistivity, ground water, radon cxhalation and
surface uplift that have been observed prior to
some earthquakes.
The success of the models to explain the
observations led to optimism, that culminated in
1975, when Chinese seismologists successfully
predicted an earthquake of magnitude 7.3 near
Haicheng in NE-China, but the prediction was
refined and narrowed as time went on. Premoni-
tory changes were observed in tilt, earth cur-
rents, the geomagnetic field, seismicity and the
flow of ground water. Anomalous animal
behavior also secms to have played a large role
in the short term prediction of the event. A
pronounced foreshock sequence was considered
to be the final warning, resulting in evacuation
of buildings few hours bcfore the earthquake
occurred.
Most destructive, historic earthquakes in lce-
land have originated in the South Iceland Seis-
mic Zone, a transform fault zone connecting the
submarine Reykjanes Ridge and the Eastern
Volcanic Zone. At least 33 earthquakes have
caused considerable damage in this zone in the
last 1000 years. Sequences of earthquakes affec-
ting most of the 70 knr long zone recur at
intervals ranging between 45 and 112 years. The
sequences characteristically begin with a large
event (M 7) in the eastern part of the zonc
followed by similar or slightly smaller events
farther west. A sequence may last from a few
days to three years. The last sequence of this
type occurred in 1896, and one must conclude
that there is more than 80% probability that a
major sequence recurs in thc next 25 years. A
modest effort to monitor changes in physical
parameters preceding the next sequence has
been initiated, including the operation of a
seismograph network, installation of volumetric
strainmeters and sampling of fluid from gcoth-
ermal wells for radon analysis.
28