Professional Documents
Culture Documents
net/publication/258497129
CITATIONS READS
92 1,317
4 authors:
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by E. Aharonov on 12 June 2014.
F02001 1 of 10
F02001 KLAR ET AL.: LANDSLIDES AND VOLUME‐TO‐AREA RELATIONS F02001
2 of 10
F02001 KLAR ET AL.: LANDSLIDES AND VOLUME‐TO‐AREA RELATIONS F02001
3 of 10
F02001 KLAR ET AL.: LANDSLIDES AND VOLUME‐TO‐AREA RELATIONS F02001
of the slope lies: (1) a deep failure mode which does not
end at the toe of the slope (i.e., xe < 0); (2) an intermediate
failure mode where xe = 0, but Yf(x) (the value of the height
of the failure surface) has values smaller than 0; and (3) a
shallow failure mode where xe = 0 and Yf(x) > 0 everywhere.
These three failure modes are illustrated along the g line of
the example slope in Figure 5.
[20] It can be shown that the g line is unique also for a
normalized cohesion, defined as c/gH, for a given slope
angle b. Thus it is shown that the failure type changes as
and c are changed, resulting in the known characteristic
that as c tends to zero, shallow failure is expected, and as
is reduced (or as H increases, which is the same thing on
the g line) the failure plane becomes deeper.
Figure 4. (a) Method of calculating the factor of safety [21] Figure 6 shows a family of g lines for different slope
using g line. (b) Method of calculating probability of failure angles from 15 to 60 degrees. The dashed lines distinguish
using the g line. between the three different failure modes (first defined by
Taylor [1937] in his seminal work on slope stability using
the friction circle method). Figure 6 illustrates that, regard-
involved in other methods such as the point estimate method less of the material properties and the slope height, if the
(PEM) or Monte Carlo [e.g., Harr, 1987]. slope is steep (e.g., b ≥ 60°) a shallow failure mode is
[16] Figure 4 illustrates this link between the factor of safety to be observed; while if the slope is gentle (e.g., b ≤
and probability of failure, and that for a given material, with
given standard deviations in strength properties, the probability
of failure P(f) decreases with increasing factor of safety.
[17] A link between the probability of failure and the
slope material properties is needed when relating statisti-
cal field observations to probability models of slope
failure. The above approach can be used for this purpose,
given enough information about topography and soil or
rock properties. Clearly, other approaches can be used for
statistical slope failure modeling, such as that suggested
by Stark and Guzzetti [2009]. The current work focuses
on the geometrical consideration of slopes at failure,
rather than on the probability density of observed failures.
3. Geometry of Landslides
3.1. Two‐Dimensional Landslides
[18] Since observed failures are, by definition, associated
with FS = 1, they lie on the g line. Each point on the g line is
associated with distinct geometrical characteristics.
[19] Slope failures occur in three different failure modes Figure 6. Normalized g lines for different slope angles b. The
depending on where on the g line the combination of c and dashed lines distinguish between the different failure modes.
4 of 10
F02001 KLAR ET AL.: LANDSLIDES AND VOLUME‐TO‐AREA RELATIONS F02001
t / l ð7Þ
5 of 10
F02001 KLAR ET AL.: LANDSLIDES AND VOLUME‐TO‐AREA RELATIONS F02001
6 of 10
F02001 KLAR ET AL.: LANDSLIDES AND VOLUME‐TO‐AREA RELATIONS F02001
7 of 10
F02001 KLAR ET AL.: LANDSLIDES AND VOLUME‐TO‐AREA RELATIONS F02001
Figure 14. Analytical model (based on w = t) superimposed on the observational field data.
8 of 10
F02001 KLAR ET AL.: LANDSLIDES AND VOLUME‐TO‐AREA RELATIONS F02001
Figure 15. Analytical model (based on 2w = l) superimposed on the observational field data.
[33] It might be useful to avoid fractional units such as soil and the bedrock, and fewer large ones that fail on deep
arise in the above expressions. For this purpose an alterna- weak layers [Stark and Guzzetti, 2009]. When we turn to
tive scaling, with a respect to a reference area Ar = (c/g)2, look at the geometry of deep versus shallow slides we find
can be suggested. Using this new scaling, the volume to that, while deep slides often fail the slope from head to toe
surface area relation becomes: (as modeled here, and as is usual for homogenous slopes;
see Katz and Aharonov [2006]), shallow natural slides often
V ¼ kAr3=2
As
ð13Þ
“bite off” only a portion of the slope. The small length l of
Ar shallow slides (often smaller than the slope length) is a
result of the scaling between t and l (equation (7)), since
where k is a dimensionless constant (equal to 0.45 in the their thickness is dictated to be small by the shallow depth
first model, and 0.53 in the second model) and Ar is a of the unconsolidated soil [Katz and Aharonov, 2006].
reference surface area. Such a nondimensional form is Because both deep and shallow slides obey the same
useful when the scaling exponents are uncertain and are mechanics, and the same scaling relations between t and l,
functionally dependent on other model parameters. The they also fall on the same trend in field measurements of
values of Ar associated with the comparison are given in V‐As distributions. An accurate study of this question is
Figures 14 and 15. still needed, but since it requires analysis of failure in
[34] It should be noted that three‐dimensional slopes are heterogeneous environments, preferably in three dimen-
expected to possess somewhat higher factors of safety than sions, it is left for future work.
those calculated assuming a two‐dimensional failure mode.
This, however, should not hinder the validity of the pro- 5. Conclusions
posed relation, as the decrease in the required cohesion for
stability is only about 20% in the worst case [e.g., [36] It has been established, on the basis of the analytical
Leshchinsky and Baker, 1985]. solution of limit equilibrium slope stability formulation, that
[35] Finally, an important question that arises is the a universal relation between volume and area of landslides
applicability of our homogeneous model results to a natural exists. This relation follows a power law (V / Ads ) with d
environment that is often highly heterogeneous in terms of value (1.32–1.38) that is nearly independent of slope
its mechanical properties. To explain why a homogeneous geometry (slope height and angle), and is also not very
model is able to predict and fit the scaling behavior of sensitive to the width to length ratio of the slides. The power
natural slopes we summarize a conceptual suggestion made law has an exponent similar to that found for the distribu-
by Katz and Aharonov [2006]: heterogeneity (e.g., a weak tions of natural landslides, despite differences in scale and
layer) is expected to constrain one of the landslide dimen- processes in nature. Comparison between the observational
sions (e.g., the slide thickness), but instability and failure relations and the analytical model indicated c/g ranging
will still follow the same log‐spiral curve. Because of this, between 5 × 10−5 and 10 m. These values may be roughly
landslides in heterogeneous slopes are expected to still associated with a cohesion range of 0.001 to 200 kPa. The
retain the basic scaling of V versus As obtained here. Het- very large range (more than five orders of magnitude) of
erogeneity does play a crucial role in constraining the shape admissible cohesions, viewed in conjunction with the
of the probability distribution function of slide sizes insensitivity of the power exponent (d confined between 1.32
[Malamud et al., 2004], causing for example many shallow and 1.38), indicates that landslide geometry is remarkably
slides that fail on the boundary between the unconsolidated insensitive to the specific mechanical properties of the slope
9 of 10
F02001 KLAR ET AL.: LANDSLIDES AND VOLUME‐TO‐AREA RELATIONS F02001
material. This clearly supports and partially explains the Hovius, N., C. P. Stark, and P. A. Allen (1997), Sediment flux from a
observations that landslides from diverse conditions follow mountain belt derived by landslide mapping, Geology, 25, 231–234,
doi:10.1130/0091-7613(1997)025<0231:SFFAMB>2.3.CO;2.
the same power law. The model presented in the work was Imaizumi, F., and R. C. Sidle (2007), Linkage of sediment supply and
developed for a homogenous slope, even though natural transport processes in Miyagawa Dam catchment, Japan, J. Geophys.
slopes are never homogenous. As discussed in the work of Res., 112, F03012, doi:10.1029/2006JF000495.
Katz, O., and E. Aharonov (2006), Landslides in vibrating sand box: What
Katz and Aharonov [2006], heterogeneity in the form of controls types of slope failure and frequency magnitude relations?, Earth
weak planes (like faults or clay layers) might have the effect Planet. Sci. Lett., 247, 280–294, doi:10.1016/j.epsl.2006.05.009.
of fixing one of the length scales (e.g., depth of slide deter- Korup, O. (2005), Geomorphic imprint of landslides on alpine river sys-
mined by depth of clay layer). Once one dimension is pinned, tems, southwest New Zealand, Earth Surf. Processes Landforms, 30,
783–800, doi:10.1002/esp.1171.
the other dimensions are expected to follow suit and correlate Korup, O. (2006), Effects of large deep‐seated landslides on hillslope mor-
with the pinned dimension following the power laws pre- phology, western Southern Alps, New Zealand, J. Geophys. Res., 111,
sented above. That is, if the depth is dictated, the length of F01018, doi:10.1029/2004JF000242.
Larsen, I. J., D. R. Montgomery, and O. Korup (2010), Landslide ero-
the slide is not randomly chosen but is constrained to a sion controlled by hillslope material, Nat. Geosci., 3(4), 247–251,
specific relation with the depth. An important prediction that doi:10.1038/ngeo776.
arises from the fact that the power law d is smaller than 1.5 is Leshchinsky, D., and R. Baker (1985), Three dimensional analysis of the
that landslide’s geometry is not self‐similar, instead slides stability of slopes, Int. J. Numer. Anal. Methods Geomech., 9, 199–223,
doi:10.1002/nag.1610090302.
tend to be shallower (relative to their surface area) with Malamud, B. D., D. L. Turcotte, F. Guzzetti, and P. Reichenbach (2004),
increasing dimensions of the slope and the slide. This Landslide inventories and their statistical properties, Earth Surf. Processes
behavior is self‐affine, in that the length and the thickness Landforms, 29, 687–711, doi:10.1002/esp.1064.
Malkawi, A. I. H., W. F. Hassan, and S. K. Sarma (2001), Global search
increase by a different factor as the slide size increases. In method for locating general slip surface using Monte Carlo techniques,
addition to the above analysis, the paper suggests a new J. Geotech. Geoenviron. Eng., 127, 688–698, doi:10.1061/(ASCE)1090-
graphical interpretation of the factor of safety and its relation 0241(2001)127:8(688).
Rendulic, L. (1935), Ein Beitrag zur Bestimmung der Gleitsicherheit,
to the slope probability of failure. Bauingenieur, 16, 230–233.
Rice, R. M., and G. T. Foggin III (1971), Effects of high‐intensity storms
References on soil slippage on mountainous watersheds in Southern California,
Water Resour. Res., 7, 1485–1496, doi:10.1029/WR007i006p01485.
Abele, G. (1974), Bergsturze in den Alpen—Ihre Verbreitung, Morphologie Rice, R. M., E. S. Corbett, and R. G. Bailey (1969), Soil slips related to
und Folgeerscheinungen, Wiss. Alpenvereinshefte, vol. 25, 247 pp., vegetation, topography, and soil in Southern California, Water Resour.
Wiss. Alpenvereinshefte, Munich, Germany. Res., 5, 647–659, doi:10.1029/WR005i003p00647.
Baker, R. (2003), Sufficient conditions for existence of physically signifi- Simonett, D. S. (1967), Landslide distribution and earthquakes in the
cant solutions in limiting equilibrium slope stability analysis, Int. J. Bewani and Torricelli mountains, New Guinea, in Landform Studies
Solids Struct., 40, 3717–3735, doi:10.1016/S0020-7683(03)00075-1. From Australia and New Guinea, edited by J. N. Jennings and J. A.
Baker, R., and M. Garber (1978), Theoretical analysis of stability of slopes, Mabbutt, pp. 64–84, Cambridge Univ. Press, Cambridge, U. K.
Geotechnique, 28, 395–411, doi:10.1680/geot.1978.28.4.395. Stark, C. P., and F. Guzzetti (2009), Landslide rupture and the probability
Castillo, E., and A. Luceno (1982), A critical analysis of some variational distribution of mobilized debris volumes, J. Geophys. Res., 114, F00A02,
methods in slope stability analysis, Int. J. Numer. Anal. Methods Geo- doi:10.1029/2008JF001008.
mech., 6, 195–209, doi:10.1002/nag.1610060206. Storn, R., and K. Price (1997), Differential evolution—A simple and effi-
Chen, W. F. (1975), Limit Analysis and Soil Plasticity, Elsevier Sci., cient heuristic for global optimization over continuous spaces, J. Global
New York. Optim., 11, 341–359, doi:10.1023/A:1008202821328.
Cheng, Y. M., L. Liang, S. C. Chi, and W. B. Wei (2008), Determination of Sun, J., J. Li, and Q. Liu (2008), Search for critical slip surface in slope
the critical slip surface using artificial fish swarms algorithm, J. Geotech. stability analysis by spline‐based GA method, J. Geotech. Geoenviron.
Geoenviron. Eng., 134, 244–251, doi:10.1061/(ASCE)1090-0241(2008) Eng., 134, 252–256, doi:10.1061/(ASCE)1090-0241(2008)134:2(252).
134:2(244). Taylor, D. W. (1937), Stability of Earth slopes, J. Boston Soc. Civ. Eng.,
Craig, R. F. (1997), Soil Mechanics, 6th ed., 104 pp., Spon Press, London. 24, 197–247.
De Joesselin De Jong, G. (1981), Variational fallacy, Geotechnique, 31, ten Brink, U. S., E. L. Geist, and B. D. Andrews (2006), Size distribution of
289–290, doi:10.1680/geot.1981.31.2.289. submarine landslides and its implication to tsunami hazard in Puerto
Fellenious, W. (1936), Calculations of the stability of earth dams, in Trans- Rico, Geophys. Res. Lett., 33, L11307, doi:10.1029/2006GL026125.
actions of the 2nd International Congress on Large Dams, vol. 4, Waltham, A. C. (1994), Foundations of Engineering Geology, Blackie
pp. 445–459, U.S. Gov. Print. Off., Washington, D. C. Acad. and Prof., Glasgow, U. K.
Guzzetti, F., F. Ardizzone, M. Cardinali, M. Galli, P. Reichenbach, Whitehouse, I. E. (1983), Distribution of large rock avalanche deposits in
and M. Rossi (2008), Distribution of landslides in the upper Tiber central Southern Alps, New Zealand, N. Z. J. Geol. Geophys., 26,
River basin, central Italy, Geomorphology, 96, 105–122, doi:10.1016/j. 272–279.
geomorph.2007.07.015.
Guzzetti, F., F. Ardizzone, M. Cardinali, M. Rossi, and D. Valigi (2009),
Landslide volumes and landslide mobilization rates in Umbria, central E. Aharonov, Institute of Earth Sciences, Hebrew University of Jerusalem,
Italy, Earth Planet. Sci. Lett., 279, 222–229, doi:10.1016/j.epsl.2009. Givat Ram Campus, Jerusalem 91904, Israel. (einatah@cc.huji.ac.il)
01.005. B. Kalderon‐Asael, Department of Geophysics and Planetary Sciences,
Harr, M. E. (1987), Reliability‐Based Design in Civil Engineering, Dover, Tel‐Aviv University, Tel Aviv 69978, Israel.
Mineola, N. Y. O. Katz, Geological Survey of Israel, 30 Malkhe Israel St., Jerusalem
Hewitt, K. (2002), Styles of rock avalanche depositional complex in very 95501, Israel.
rugged terrain, Karakoram Himalaya, Pakistan, in Catastrophic Land- A. Klar, Faculty of Civil and Environmental Engineering, Technion‐
slides: Effects, Occurrence and Mechanisms, Rev. Eng. Geol., vol. 15, Israel Institute of Technology, Technion City, Haifa 32000, Israel.
edited by S. G. Evans and J. V. DeGraff, pp. 345–378, Geol. Soc. of
Am., Boulder, Colo.
10 of 10