The challenge in rock instabilities detection and hazard assessment
is to develop methods that are both modular (i.e. that can be adapted
to the funding available for the study) and evolutive (i.e. that
can be upgraded with the development of new tools and new digital
documents).
Basically, instability factors can be divided in two categories: intrinsic
parameters (IP) that change with time, and external variables (EV),
that cause these changes. Hazard assessment is based on two different
ways of combining IP and EV: (1) by summing and weighting them together
without introducing a hierarchy; (2) by forming them into a hierarchy,
leading to a physical-like modeling.
A specific hazard scale for every case study can thus be defined,
taking into account that some instability factors are lacking, or
on the contrary that they are more detailed, i.e. depending on the
scale of the study, the knowledge of IP and EV can be more or less
accurate. The assessment methods are usually performed as a first
step by combining geometrically IP and EV; and in further steps by
introducing physical modeling and by field observations. Combinations
as well as iterations can be performed along with theses steps.
A certain number of IP descriptors can be easily computed by analyzing
Digital Elevation Models (DEM): for instance, 1) slope angles, indicating
whether slopes are near their equilibrium angle or not; 2) Sloping
Local Base Level (SLBL) to calculate the volumes that are potentially
erodible by landslide activity; 3) the main structural features shaping
the topography that can be extracted using 3D topographic orientation
histograms; 4) faults or discontinuities whose traces can be built
with the help of known points; 5) hydrographic networks; 6) kinematics
tests and factors of safety for different rockfall mechanism types;
7) rockfall activity that can be assessed by examining aerial photographs
or vectorial topographic map to map cliffs and fresh rockfall deposits;
and 7) geology; EV descriptors can be computed either by DEM analysis
or with the help of external data such as 1) precipitation contribution
to the watersheds; 2) hydraulic head index; 3) water table level
that estimated by a smoothed DEM or by data; 4) earthquake activity;
5) recent tectonic movements; 6) freeze and thaw cycles; 7) human
activities and protection works; etc.
The above approaches were applied to linear elements at risk in Quebec
and Switzerland: The rockfall hazard assessment along the Quebec
City Promontory (Quebec, Canada) shows a good agreement with the
observed data. Based on five instability factors, the rockfall hazard
assessment along mountain roads in the canton of Valais (Switzerland)
was highly predictive.
Applied to planar elements at risk, the hazard assessment of large
rock instabilities was efficient in the area surrounding the 1991
Randa rockfall (30 M m3). The main instabilities were detected using
instability factors extracted from a DEM, or deduced from aerial
photographs and field surveys. A simple mechanical modeling including
pore water pressure improves the results. Moreover, such methods
were tested on soil slopes, giving promising results.