Advances in Conception about Earthquakes - Earthquake Engineering for Structural Design - page 135

to be expected. Algorithm is designed to determine the TIP, which can be
considered as the alarm time. Using these algorithms, some earthquakes in the
USA and Japan were announced with reliable prediction (Rotwain and Novikova,
1999). The Keilis-Borok team in June 2003 predicted the 6.5 magnitude
earthquake which struck Central California on 22 December and also in July 2003
predicted the 8.1 magnitude earthquake which occurred in September on
Hokkaido. Previously, this team predicted the Northridge earthquake (Science
Daily, 1994). But the scientific community is fairly skeptical about these successes
in prediction, considering that a larger number of successful predictions is
necessary to guarantee that the developed methodology is correct.
Another methodology for earthquake predictions has been developed in China
under the leadership of professor Yin (Yin et al, 1994, 1995, 2000, 2002), using the
so-call Load-Unload Response Ratio (LURR). This parameter can be defined from
the constitutive curve of the focal zone for the controlling-state parameters. In case
of the earthquake prediction, this constitutive curve can be drawn for one factor
defining precursor parameters (see the next details concerning the precursors of an
earthquake). The LURR parameter is the ratio between the slopes of curve for
loading and unloading behavior. If the nonlinear system is stable, this parameter is
1. When the system is close to instability, the LURR value significantly differs
from 1. For this reason, it can quantitatively indicate the degree of the imminence
of instability. The temporal variation of LURR in many seismic areas in China,
Japan and the USA has been calculated and analyzed, showing in all cases that this
ratio is close to 1 in the periods of stability and significantly higher than 1 before
the occurrence of earthquakes with magnitude higher than 6.0.
For short-time prediction, it is very important to determine the geophysical
precursors of the earthquake (Gray, 1996). The monitoring of these geophysical
precursors enables scientists to make short-term predictions of an earthquake.
There are a variety of geophysical precursors. However, only five of them are the
most important; they are: the P-wave changes, ground uplift and tilt, random
emission, electrical resistivity of rocks and the water level fluctuations. In order to
monitor the changes in the Earth's magnetic field, the Demeter (Detection of
Electro-Magnetic Emissions Transmitted from Earthquake Regions) micro-satellite
was launched in 2004 (Wikipedia, nd).
Another prediction for very short-time earthquakes is given by animal behavior
(Kirschvink, 2000, Wikipedia, nd, Carayannis, 2007). Animals can detect the P-
wave or ultrasonic wave generated by an earthquake, even if the waves are too
small for humans' senses, or when there are some changes in low-frequency
electromagnetic signals. When this happens, animals start behaving agitatedly and
nervously. The Chinese began a systematic study of this unusual animal behavior
and these studies had an important role in the prediction of the 1975 Haicheng
earthquake. Scientists said snakes could sense a quake from 120 km away, up to
five days before it happens, the reptiles being perhaps the most sensitive to
earthquakes. Thus, the crate of some reptile farms, where a camera monitors the
snake behavior, is one solution for short-term prediction.
In case of early warning , the effectiveness depends on the distance from source
to site, because it is based on the difference between the velocity of P and S waves,