Analysis
Analysis of instability data
To determine the degree of landslide hazard within a given area
requires that the information used to make such a clarification
is given some qualitative or semi-quantitative fashion. The product
of land instability investigations is usually compiled as a map,
spot localities or data banks of attributes of localities or areas
on the ground. Landslide hazard zonation maps often superimpose
and integrate information on slope angles, landslide deposits, geomorphology,
geology, hydrology, rainfall and climate, earthquake activity and
expected seismic response (the differing response of rock types
to earthquakes is explained in study guide 4). The methods can be
quite detailed, or simplified. The simpler methods allow more rapid
evaluation. Here a (1) rapid evaluation and simplified presentation
and (2) thorough evaluation are described (Varnes, 21-28).
Rapid evaluation of slope stability of the San Francisco Bay
region at 1:25000 scale
Procedure:
Existing maps were used to determine geologic formations and slope
ranges and air photos interpretation determined the extent of landslide
deposits. Maps of slope ranges were then joined with maps of landslide
deposits creating 4 map units of relative slope stability; these
were then combined with the map of susceptible bedrock and surficial
geology to define 6 zones of relative stability (Table 6).
Table 6. A rapid method of establishing landslide zones
of relative safety (from Varnes 1984)
| |
Slope
|
|
<
5 %
|
5
- 15 %
|
>
15 %
|
No
Landslide
Deposits
|
1
Stable
|
2
Generally stable
|
3
Moderately
stable
|
Susceptible
Bedrock
|
4
Moderately
stable
|
Susceptible
Surficial
Deposits
|
1A
Subject to
liquefaction
|
None
|
Landslide
Deposits
|
5
Unstable
|
Thorough evaluation of slope instability: San Mateo County, California.
Procedure:
A. The area of outcrop was estimated for each rock unit of the
geologic map using a grid overlay with resolution of 2.6 hectares
(0.01 m2).
B. The landslide inventory map was superimposed on the geologic
map in order to identify the units in which failures had occurred,
and the areas that had failed within each unit were estimated using
the grid.
C. The geologic map units were then listed in order of percentage
of their outcrop areas that have failed by landsliding.
D. The highest class of susceptibility was assigned to landslide
deposits, which contain many more failure surfaces than the rock
unit from which they are derived.
E. Other class limits were selected at convenient intervals on
the list, and a roman numeral class number was assigned to the map
units. That numeral represents the relative susceptibility of any
particular map unit.
F. The slope map was then superimposed on the combined geologic
map and landslide inventory and systematically examined to determine
the slope intervals that displayed maximum landslide frequency for
each map unit. Those slope intervals having the highest maxima were
then labelled with the highest roman numeral class. Slope intervals
showing significantly fewer slides were reduced in rank and labelled
with numerals of lower classes. Thus a unit having a maximum susceptibility
of III may be labelled with that number only where slopes exceed
30 %, and because lower slopes may be expected to have significantly
fewer slides, the final map labels for the same geologic units on
lower slopes may be II or even I, depending on the rate of change
of relative susceptibility with slope.
These methods do not distinguish between the types of land instability
that might occur.
In New Zealand Crozier (1984) proposes a landslide classification
based on the likelihood and frequency of landslides occurring (Table
7).
Table 7. A New Zealand slope stability classification
(Crozier 1984).
| Class
I |
Slopes
with active landslides. Material is continuously moving, and
landslide forms are fresh and well-defined. Movement may be
continuous or seasonal |
| Class
II |
Slopes
frequently subject to new or renewed landslide activity. Movement
is not a regular, seasonal phenomenon. Triggering of landslides
results from events with recurrence intervals of up to 5 years.
|
| Class
III |
Slopes
infrequently subject to new or renewed landslide activity. Triggering
of landslides results from events with recurrence intervals
greater than 5 years |
| Class
IV |
Slopes
with evidence of previous landslide activity but which have
not undergone movement in the preceding 100 years
IVa Erosional
forms still evident
IVb Erosional
forms no longer present - previous activity indicated by landslide
deposits |
| Class
V |
Slopes
which show no evidence of previous landslide activity but which
are considered likely to develop landslides in the future. Landslide
potential indicated by stress analysis or analogy with other
slopes. |
| Class
VI |
Slopes
which show no evidence of previous landslide activity and which
by stress analysis or analogy with other slopes are considered
stable. |
The preparation and presentation of landslide hazard zonation maps
is not simple. Different end users will require different types
of information. Regional and local administrators will probably
only need uncomplicated maps which delineate hazard zones and on
which they can make planning decisions. These maps may not include
geological or geomorphological detail, and may not show any of the
past landslide activity in the region of concern. However, engineers
may require to know the distribution of these factors and any others
that were used to delineate the hazard zones. Generally this requires
that two or more maps are prepared, e.g. an analytical map showing
geology, geomorphology and hydrology and an interpretative map which
synthesises this information.
Numerical rating systems
Numerical rating systems can be useful in assessing the role of
factors which contribute to slope instability. In Table 8 an empirical
approach for mapping landslide hazard and risk in clay slopes in
Northern Tasmania is used. The scoring reflects the extent to which
the slope is affected by the different variables which contribute
to instability, e.g. the susceptibility of the subsurface material
to sliding, the ground water level, steepness and complexity of
the slope, whether or nor the slope had previously failed and what
use is made of the land. Clay behaviour is affected by water level
and slope angle and slope complexity are also closely associated
and in this system their combined effects are summed and effectively
determine landslide hazard.
Since including a landuse variable introduces a measure of vulnerability
the product of the factors determines risk as well as hazard. According
to this method Risk, R = (P+2W) (S+2C)
(U), where R > 60 is associated with failure, R
> 50 is associated with possible failure.
Table 8. An empirical approach to numerical rating of
landslide hazard (from Varnes 1984).
|
Contributing
factor
|
Score
|
|
P
|
Clay
factor: Use range of available values of plasticity index
(PI) for the geologic unit involved in sliding |
|
| |
PI
in lower third of range
PI in mid third of range
PI in upper third of range |
1
2
3 |
|
W
|
Water
factor: Highest position annually of piezometric surface
relative to typical failure plane |
|
| |
Below
plane
Between plane and mid-depth of slide
Above mid-depth of slide |
1
2
3 |
|
S
|
Slope
angle: Use range of values appropriate to local geology
|
|
| |
Within
lower third of range
Within mid-third of range
Within upper third of range |
1
2
3 |
|
C
|
Slope
complexity |
|
|
|
Simple
slope
Old failure, now partly obliterated by erosion
New failure, stable but not eroded |
1
2
3 |
|
U
|
Land
use |
|
|
|
Woodland
Cleared or built on with special precautions
Built on without special precautions
|
1
2
3 |
Numerical and numerical-cartographic methods
In many examples of hazard assessments factors such as slope angle,
percentage of the outcrop area of geological formations occupied
by landslides etc. are given numerical ratings. It is often an advantage
to generalise and quantify the areal distribution of landslides
deposits in contour form so that it can be more easily combined
with other information. The most commonly used method of assessing
slope stability is to determine soil properties, accurate slope
profiles, and related hydrologic data and calculate a factor of
safety. If this factor is determined for many sites within a region
the resulting spread of values can also be zoned.
This method has its limitations though and it depends on the accuracy
of the initial measurements and tests, and on whether the locations
selected are representative of the region being assessed. One way
to improve the applicability of the method is to make an infinite
number of measurements to improve coverage, or to use statistical
methods to analyse the interrelationship between the quantitative
and qualitative measurements which may be obtained over large areas,
rather than specific localities.
One statistical method (discriminant analysis) involves isolating
the most significant factors influencing slope stability. These
can then be used to determine expressions using small numbers of
measurable variables which can be combined using formulas to yield
a specific number. By correlating the magnitude of that number with
slopes that have failed and unfailed slopes it then becomes a predictor
of whether or not a slope is likely to fail, with an estimated probability
of being accurate.
Computers have made these sorts of calculations very fast meaning
that large numbers of calculations can be made improving the accuracy
of the maps produced. Furthermore, automated plotting of hazard
zones, risk maps and/or slope stability variables means that producing
maps is also very fast. Correcting and updating these maps using
new information and results is now much easier. Although the basic
information required in computer-assisted statistical analysis and
mapping still needs to be acquired by human hand, which is time
consuming, computer-assisted compilation is no slower than producing
the maps by hand, has fewer errors and provides an easily accessible
data bank for future use. A large data bank of different types of
information which can then be used in different combinations to
produce different maps.
|
