Landslides - Extreme Events: A Physical Reconstruction and Risk Assessment - page 200

Statistical--modelling analysis
Modelling and statistical analyses can also be utilised to investigate the
occurrence of past landslides. Jibson and Keefer (1994) used modelling techniques
to determine the cause of landslides in the New Madrid seismic zone. Like the
Crozier et al .(1995)study in New Zealand, Jibson and Keefer (1994)aimed to
ascertain whether extreme hydrological conditions or seismic activity was the
likely cause. The landslides are assumed to have occurred during the 1811--1812
earthquake based on historical notes and past literature. To test this, Jibson and
Keefer (1994)analysed the slides under static or aseismic conditions and dynamic
(seismic) conditions. The static conditions assume failure caused by changes in
ground-water conditions in the absence of earthquake shaking. A broad range of
physical or geotechnical factors are investigated in the field including the shear
strength of the materials likely to fail, their plasticity, grain size distribution
and unit weight. These factors are then incorporated into a numerical model.
The dynamic analysis uses the Newmark displacement method which models a
landslide as a rigid friction block of known critical acceleration, which is the
acceleration required to overcome frictional resistance and initiate sliding on
an inclined plane. Estimates of the minimum intensity of an earthquake event
can also be made using this technique.
Jibson and Keefer (1994) concluded that groundwater conditions, even during
extreme events, could not have triggered the investigated landslides. Based on
prior knowledge of the earthquake in 1811--1812 and the described analysis it was
determined that seismic shaking triggered the landslides. While this approach
wasundertaken in a known seismically active area, the method can also be
applied in areas where the seismic conditions are unknown.
Conclusion
Landslides, like other geological extreme events, leave a distinct signa-
ture in the landscape. Block falls probably leave the most prominent imprint,
simply because the blocks can be of substantial size. Of course they will eventu-
ally break down or weather and erode and/or be buried by other sediments but
this may take considerable time. Until then lichenometry is a useful method,
especially using Bull and Brandon's 'FALL' technique to establish the age of
theevent. Dammed lakes are more ephemeral but still a useful guide to past
landslide events. Debris flows, mud flows and slumps are also obvious features
but could be reasonably expected to last a shorter time in the landscape, at
least as far as being an obvious event. Of course flows and slumps will remain
in a stratigraphic sequence for longer but will require a trained eye for their
detection.