Geoscience Reference
In-Depth Information
of conditional probability and acceptance of this fact by risk assessors that the
longer it has been since the last event the sooner it is until the next one. This is
because time is required for pressure and stress to build for volcanic eruptions
and earthquakes, respectively, after the last event. The same is true to a certain
extent with landslides, where these events may be caused by the accumulation
of weathered bedrock or saprolite on a slope. Once this material fails and moves
downslope, it takes some time before the fresh bedrock is able to weather to a suf-
ficient degree so that a substantial saprolitic mantle can accumulate. This then is
really a function of the sensitivity of the landscape, particularly in non-seismic
tropical regions where many landslides can be generated by intense rainfall
events saturating a thick mantle of weathered material on a steep slope (Thomas,
2004).
The palaeoseismic record allows us to gain an appreciation of the magnitude
and frequency of events likely in a region, especially when such events have
not occurred for some considerable time. Meghraoui et al .(2003), for example,
recognised that the absence of major earthquakes along sections of the Dead Sea
Fault in Syria over the past 830 years is not a reasonable guide to the likely future
occurrence of earthquakes. They found that three major events of magnitude
M w 7. 0and 7.5 occurred over the last 2000 years. The average recurrence interval
of 550 years for large earthquakes at this location shows that the fault zone
is not inactive and also that the next large earthquake may be imminent, or
possibly overdue. Such detailed conclusions would not have been possible from
the historical record.
Palaeoseismic records from the San Andreas fault, California do suggest that
recurrence intervals for large earthquakes here may not have been consistent
over time. Fumal et al .(2002a)andBiasi et al .(2002)interpret these records to
show that there may have been a relative acceleration of earthquake activity
between the early AD 600s and the early 800s followed by a lower rate around
AD 1500 and then by a higher rate after this. Biasi et al .(2002)suggest that strain
transfer due to earthquakes along other sections of the same fault or other faults
may influence the timing of earthquakes. Recurrence intervals can be affected
by the accumulation and release of fault normal stresses at restraining bends
in a fault causing the recurrence intervals of earthquakes along nearby faults
to be different. The stress release and accumulation may be influenced by the
occurrence of earthquakes on these nearby faults. The two systems therefore
interact. Strain conditioning of the recurrence interval may also be another
possibility. Biasi et al .(2002)suggest that earthquakes might occur at regular
intervals for some period of time but stress may not be completely released and
it gradually accumulates. They also suggest that the physical properties of the
fault may vary with time such as the characteristic size and spacing of asperities
Search WWH ::




Custom Search