Document Actions

IJEGE-11_BS-Usuki-&-Mizuyama

background image
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
787
DOI: 10.4408/IJEGE.2011-03.B-086
SOIL PROPERTIES AND FLUIDITY OF LONG-TRAVELING LANDSLIDES
n. USUKI
(*)
& t. MIZUYAMA
(**)
(*)
Asia Air Survey Co., Ltd. - Japan
(**)
Graduate School of Agriculture, Kyoto University, Japan
The July 1985 Jizuki-yama landslide in Nagano Pre-
fecture, for example, caused severe damage because
it traveled over a long distance. s
assa
et alii (2000)
studied the travel speed of these kinds of highly flu-
idized long-traveling landslides, but did not focus
on their travel distance. Other studies have included
estimates of the travel distance of landslide masses
(m
oRiwaki
, 1987), case analyses of the travel distance
of landslides (k
usumoto
et alii, 2003), and an exami-
nation of the topographical conditions of large-dis-
placement landslides (m
oRisHita
et alii, 2003), but did
not evaluate the movement mode of these landslides.
Our study therefore focused on the movement mode
of long-traveling landslides in Japan by examining
their features, soil properties, and distribution charac-
teristics in relation to travel distance.
ACTUAL STATE OF LONG-TRAVELING
LANDSLIDES
LANDSLIDE DATA FOR STUDY
We collected data on landslides caused by rainfall
and snowmelt, and whose masses moved continuously
after sliding began. Of the 376 major landslide disas-
ters in Japan from 1947 to 2000, we selected land-
slides that have a complete dataset including a plan
chart; data on site geology, landslide scale, soil vol-
ume, and travel distance; and photos showing the con-
ditions of landslide masses. We excluded 1) landslides
caused by earthquakes and artificial causes, 2) land-
slides that blocked a mountain stream first and then
ABSTRACT
Better knowledge of the movement mode of fluid-
ized landslide masses, which cause severe damage over
wide areas, is very important for preventing sediment-
related disasters. We analyzed a population of long-
traveling landslides in Japan in terms of their travel
coefficient and the conditions of the landslide mass,
defining the ratio of travel distance (L2) to the original
length of a landslide mass (L1) as the travel coefficient
(Tr = L2/L1). After classifying long-traveling landslides
by movement mode into fully fluidized and partly fluid-
ized landslides, we found that the travel coefficient of
fully fluidized landslides is roughly Tr ≥ 0.5. The grain
size distributions of landslide masses suggest that the
proportion of clay and silt is one factor behind a large
travel coefficient. To evaluate the effect of grain size
distribution on landslide travel coefficient and move-
ment mode, we conducted soil mechanics tests on
experimental soil specimens. These tests showed that
soils having an intermediate grain size distribution,
containing roughly equal proportions of gravel and
sand, clay, and silt, have the smallest shear stress in the
depth range of 10 - 15 m below the ground surface.
K
ey
words
: long-traveling landslide, movement characteri-
stic, soil properties of landslides
INTRODUCTION
Some landslide masses travel long distances in a
fluidized state without developing into a debris flow.
background image
N. USUkI & T. MIZUYAMA
788
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
of landslide masses (Fig. 3). The travel coefficient is
distributed between 1 and 3.4. We defined landslides
with a travel coefficient of Tr ≥ 0.5, which is equal to
those with a cumulative relative frequency of 20% or
more, as long-traveling landslides. The average travel
distance of these long-traveling landslides is about 350
m, and their maximum travel distance is about 2000 m.
Of these landslides, 22 have a travel coefficient of 1.0
or greater (which means the travel distance is longer
than the original length of the landslide mass).
DEFINITION OF A LONG-TRAVELING LANDSLIDE
Figure 4 shows the distribution of long-traveling
landslides in Japan. Their distribution corresponds
roughly to areas of Neogene formations and meta-
morphic rocks, which are known to be landslide-
prone areas. The 109 landslide locations selected
for our study have a similar distribution. According
to the landslide classification by k
oide
(1955), these
are areas where the so-called Tertiary formation
landslide and fracture zone landslide tend to occur.
Long-traveling landslides have occurred mainly in
the Hokuriku and Shikoku regions but occur in other
areas as well, in particular in areas of Neogene rocks
and along the geological tectonic lines.
underwent secondary movement, as identified from
interviews and field surveys, and 3) landslides that
formed a landslide dam that subsequently burst. We
measured the travel distance and related dimensions
of the remaining 109 landslides as shown in Fig. 1.
DEFINITION OF A LONG-TRAVELING LANDSLIDE
Figure 2 shows the distribution of the travel dis-
tances (L2) of our set of landslides. The travel dis-
tance of landslides varies widely, ranging from a slight
movement at the landslide tip to as long as 2300 m. We
defined the ratio of travel distance (L2) to the original
length of a landslide mass (L1) as the travel coefficient
(Tr = L2/L1) and used it to evaluate the movement scale
Fig. 1 - Measurements related to the travel distance of a
landslide mass
Fig. 2 - Histogram of the travel dis-
tance (L2) of landslides
Fig. 3 - Histogram of the travel coefficient
(Tr) of landslides
background image
SOIL PROPERTIES AND FLUIDITY OF LONG-TRAVELING LANDSLIDES
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
789
CHARACTERISTICS
OF
LONG-TRAVELING
LAN-
DSLIDES
CLASSIFICATION OF MOVE-
MENT MODES OF LANDSLIDE
MASSES
Landslides are often classified by
sliding area, movement mode, moving
speed, and various landslide scales. The
movement modes of landslides are top-
ple, fall, spread, flow, slide, and creep.
Longtraveling landslides are mainly of
the flow mode. We classified the land-
slide masses in our study in one of the
following three classes (see Fig. 5):
1. The landslide mass and its surfaces
are hardly disturbed, and it slides as an
intact mass.
2. The surfaces of the landslide mass
are slightly disturbed. Most trees in the
landslide site, if they exist, have not fallen.
3. The landslide mass does not retain its original shape, and its surfaces are
severely disturbed. Most trees in the landslide site, if they exist, have fallen.
In the first case, sediment moves along the slide plane as a mass that re-
tains its original shape with little or no fracturing of the internal structure. In the
second case, the landslide mass is deformed but flows without fully fracturing
the internal structure (partly fluidized). In the third case, the internal structure is
completely fractured after sliding begins and the flow is fully fluidized.
EQUIVALENT FRICTION COEFFICIENT AND MOVEMENT MODE
The relationship between the equivalent friction coefficient (H/L; see
Fig. 1) and the soil volume of landslides is summarized in Fig. 6, also
showing the three movement modes of landslide masses just described.
The equivalent friction coefficient is
in the range of μ = 0.1-0.9. As the soil
volume increases, the equivalent fric-
tion coefficient decreases, though this
tendency is not very conspicuous.
SOIL VOLUME AND MOVEMENT
MODE
The relationship between the soil
volume and travel distance (L2) of
landslides is summarized in Fig. 7,
including the three movement modes
described above. The well-known ten-
dency of travel distance of landslides
(in particular, that of debris slides) to
Fig. 4 - Distribution of
long-traveling
landslides
Fig. 5 - Classification of landslide
masses by movement mode
Fig. 6 - Relationship between equivalent friction coefficient and soil
volume of landslides
background image
N. USUkI & T. MIZUYAMA
790
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
and partly fluidized landslides, but on the scale of their
travel coefficients. The distribution of those landslides
and their travel coefficients are shown in Fig. 9.
become longer as the soil volume in-
creases is seen in Fig. 7. But it is also
apparent that the plots of fully fluid-
ized, partly fluidized, and hardly dis-
turbed landslides are widely distribut-
ed, implying that factors other than soil
volume should also be considered re-
garding the fluidity of landslide masses.
We examined slope gradient as a factor,
but found no clear relationship between
slope gradient and movement mode.
TRAVEL COEFFICIENT AND MO-
VEMENT MODE
The travel coefficients of our land-
slides are summarized in Fig. 8, along
with their classification in our threefold
scheme. The travel coefficient of fully
fluidized, partly fluidized, and hardly
disturbed cases is roughly Tr ≥ 0.5, 0.3 ≤
Tr < 0.9, and Tr ≤ 0.3, respectively. This
means that long-traveling landslides can
be classified to some extent based on the
travel coefficient and movement mode.
SOIL PROPERTIES OF LONG-
TRAVELING LANDSLIDES
DATA FOR STUDY
Although large amounts of data
are available on soils at the slip sur-
face, the amount of test data from land-
slide masses is very limited. However,
Niigata Prefecture, which is one of the
most landslide-prone prefectures in Ja-
pan, has accumulated a significant body
of soil test results from the 1478 major
landslides that occurred in the prefec-
ture during 1982-2004. We selected
those from 15 landslides whose travel
distance was measureable, whose soil
test results were available, and which
were not triggered by earthquakes or
artificial causes. Their travel distances
were 50-600 m and their travel coeffi-
cients were Tr = 0.2-6.0. Of those land-
slides, 13 had a travel coefficient of Tr ≥ 0.5, which
means they were fully fluidized. Therefore, we decided
to focus not on the separability between fully fluidized
Fig. 7 - Relationship between soil volume and travel distance of landslides
Fig. 8 - Relationship between travel distance and original horizontal dis-
tance of landslides
Fig. 9 - Histogram of travel coefficients
background image
SOIL PROPERTIES AND FLUIDITY OF LONG-TRAVELING LANDSLIDES
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
791
TRAVEL COEFFICIENT AND SOIL PROPERTIES
OF LANDSLIDE MASSES
i
wanaGa
(1983) summarized soil test data for
about 130 landslides in Niigata Prefecture in a data
base used as a standard source for the general physical
characteristics of landslide soils. To evaluate the soil
properties of landslides with large travel coefficients,
we obtained and analyzed data on the grain size and
liquid and plastic limits of the 15 Niigata landslides of
our study and compared these with the data on land-
slide masses from Iwanaga’s collection.
GRAIN SIZE DISTRIBUTION: We produced a ter-
nary grain size distribution figure from our landslide
data in accordance with the Japanese Unified Soil Clas-
sification System and compared it with the grain size
distribution of Iwanaga (Fig. 10). Iwanaga’s data are
mainly distributed in areas of the diagram representing
large proportions of clay and silt. In contrast,
landslides with a large travel coefficient are
distributed in the center area representing
more than 20% of gravel and sand, silt, and
clay. It is also apparent that as the percentage
of gravel and sand exceeds 20%, the travel
coefficient becomes larger, but as the gravel
and sand percentage exceeds 50%, the travel
coefficient becomes smaller again.
PLASTICITY: Using a similar soil clas-
sification system, we analyzed plasticity for
our landslide data and compared it with the
plasticity analysis of i
wanaGa
(1983) (Fig.
11). Line A is a boundary separating large
and small volumetric changes and perme-
abilities. Line B is a boundary separating
large and small compressibilities. Land-
slides with a large travel coefficient are dis-
tributed on or near Line A and Line B.
SOIL PROPERTIES OF LANDSLIDES
wITH A LARGE TRAVEL COEFFICIENT:
In Figure 10, the data of i
wanaGa
(1983)
are mostly for sediments containing a large
percentage of clay and silt, which are typi-
cal soils of landslide-prone areas. In our
data, landslides with a large travel coefficient
tend to contain a relatively large percentage
of gravel and sand. According to i
wanaGa
(1983), the Uonuma formation group in Nii-
gata Prefecture has the former type of grain
size distribution and tends to trigger collapse-
type landslides. The travel distance of collapse-type land-
slides is unknown, but their travel speed is probably faster
and the travel coefficient is probably larger than those of
ordinary landslides in view of their failure mode.
However, sediments from landslides with a large
travel coefficient, Tr ≥ 2.0, are distributed in the center
area that represents 20% or more of gravel and sand,
silt, and clay. From this, it may be said that the travel
coefficient becomes larger when the landslide mass
contains not only a large percentage of gravel and
sand but also a certain percentage of clay and silt.
MECHANICAL TESTS OF SOILS WITH
VARYING GRAIN SIZE DISTRIBUTIONS
TEST METHOD
Various factors affect the travel coefficient by control-
ling the movement mode of landslide masses. They
Fig. 10 - Data on a
triangle dia-
gram
Fig. 11 - Figure of plasticity
background image
N. USUkI & T. MIZUYAMA
792
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
RESULTS OF UNIAXIAL COMPRESSION TESTS:
The uniaxial compression test results show that the
strength of our experimental samples is in the order:
fine-grained soil 2 > finegrained soil 1 > intermedi-
ate soil > coarse-grained soil. Also, an intermediate
soil had the smallest modulus of deformation (E
50
),
a modulus indicating vulnerability to deformation. In
the test after 48- hour water immersion, all the speci-
mens tested (the fine-grained soil 2 specimen absorbed
so little water that we did not conduct this test on it)
lost their strength, particularly the specimens of coarse
and intermediate soil. These results suggest that inter-
mediate soil will lose its strength significantly during
the snowmelt season and after a long period of rain.
RESULTS OF TRIAXIAL COMPRESSION TESTS:
Figure 14 shows the results of triaxial compression
tests (Mohr’s circle). At compressions equivalent to
a depth of 5-10 m from the ground surface, interme-
diate soil had the smallest shear stress. But when the
equivalent depth was 15 m or more, fine-grained soil 2
had the smallest shear stress. From this, we infer that
landslide masses having a grain size distribution close
to intermediate soil are apt to fracture easily if they are
deformed under sliding loads. Fracture-prone landslide
masses, such as those composed of intermediate soil,
will be quickly fractured if a sliding movement occurs.
If a flow layer is formed on the slide plane, the frac-
tured mass will be incorporated into that flow layer.
include soil volume, slope gradient, water content, and
catchment area. The soil test results from Niigata Prefec-
ture suggest that soil properties, particularly mechanical
properties that vary with grain size distribution, have
an effect on landslide movements. To evaluate these ef-
fects of grain size distribution, we tested four types of
specimens with varied grain size distributions: physical
tests (grain size analysis, water content), consistency
tests (liquid limit, plastic limit, liquidity index), uniaxial
compression tests, and triaxial compression tests (un-
consolidated undrained conditions). Similar specimens
were also immersed in water for 48 hours to evaluate
any decrease in strength due to water immersion.
PREPARATION OF SPECIMENS
We performed grain size analysis using a sieve
based on the Japanese Industrial Standard (J1S-
A-1204) and the Japanese Geotechnical Society Stand-
ard (JGS0131), and produced four types of specimens:
coarse-grained soil, intermediate soil, fine-grained soil
1, and fine-grained soil 2. A core material used for dams
was used as the base material of the specimens. Undis-
turbed materials were used for the specimens of fine-
grained soil 1 and fine-grained soil 2. Coarse-grained
soil and intermediate soil for specimens were compact-
ed at 0.5 Ec, which is half of the standard energy, in
the compaction test to approximate natural conditions.
TEST RESULTS
The test results and consistency
characteristics of each specimen are
shown in Table 1.
GRAIN SIZE DISTRIBUTION AND
PLASTICITY:
Comparisons of the grain
size distributions and plasticity analysis
for our experimental samples with the
data of i
wanaGa
(1983) are shown in Figs.
12 and 13, respectively. In Fig. 12, the in-
termediate soil plots in the area of sedi-
ments from landslides having a large trav-
el coefficient, Tr ≥ 1.0. The fine-grained
soil 1 and fine-grained soil 2 specimens
plot in the area of ordinary landslides. The
coarsegrained soil specimen plots in the
area where landslide occurrence is unlike-
ly. But, in Fig. 13, all of these soils plot
roughly in the same area of the plasticity
diagram where ordinary landslides occur.
Tab. 1 - Results of physical tests and consistency characteristics
background image
SOIL PROPERTIES AND FLUIDITY OF LONG-TRAVELING LANDSLIDES
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
793
FLOW MODE OF LONG-TRA-
VELING LANDSLIDES
We examined the movement mode
of long-traveling landslides based on the
results of our soil mechanics tests. We
infer that longtraveling landslides occur
by the following mechanism if landslide
masses are fully water saturated.
1. A landslide starts to slide on the slide
plane.
2. Fracturing of the landslide mass be-
gins due to sliding loads and defor-
mation.
3. A flow layer forms below the sliding
landslide mass.
4. The fractured landslide mass is in-
corporated into the flow layer and the
layer continues to grow.
5. If sufficient water is supplied (inclu-
ding the water in the landslide mass),
the flow layer grows until it affects
the entire mass.
6. Due to the fracturing of the landslide
mass and its incorporation in the flow
layer, the landslide mass becomes
severely disturbed, losing its original
shape.
If a landslide mass is fully satu-
rated and there are no large topographic
changes (such as slope gradient), soil
properties close to those of an interme-
diate soil have an effect on fracturing of
the landslide mass (item 2 above), and
on development of a flow layer (item 4
above), and become one of the factors
that determine the travel coefficient.
Landslides with a large travel coef-
ficient do not retain their original shape;
their masses are severely disturbed and
enter a fully fluidized sliding mode. This is
probably because a landslide mass is frac-
tured as sliding advances and the fractured
mass is incorporated into the flow layer.
If grain size distribution of a land-
slide mass is close to that of an interme-
diate soil, the mass is prone to fracturing
and incorporation into the flow layer.
Landslides composed of intermediate soil
Fig. 12 - Data on a triangle diagram
Fig. 13 - Figure of plasticity
Fig. 14 - Results of the triaxial compression test
background image
N. USUkI & T. MIZUYAMA
794
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
and those with a large travel coefficient have low liquid
limits and low plasticity indexes and tend to be distrib-
uted around Line A of the plasticity analysis (see Fig.
11). If the water content is increased, these landslides
tend to liquefy and undergo a large volume change. Dur-
ing the snowmelt season or after long periods of rain, the
landslide mass will probably be saturated to a state close
to the liquid limit. Entering a liquid state means there is
a change in volume and the enlargement of voids. If a
sliding action is triggered in that state, the void ratio will
further increase due to the disturbance of the landslide
mass. Then, those voids will be filled with gravel and
sand in addition to liquefied clay and silt. The developed
flow layer will enter a mode like that of a debris flow,
producing a large travel coefficient.
On the other hand, if a landslide mass contains a
high percentage of clay and silt (such as fine-grained soil
1 and finegrained soil 2), it will not be easily fractured
even when sliding begins. Consequently, the landslide
mass will not be incorporated into the flow layer and the
layer will not grow. Because of its content of clay and
silt, the flow layer will show a behavior close to that of
a Bingham fluid, which is very viscous. If the flow layer
remains thin, or thinner than the flow layer of the plug
that has a smaller shear force than the yield strength,
the landslide mass will stop sliding and enter the partly
fluidized movement mode with little disturbance.
FUTURE PERSPECTIVES
There have been few studies on the movement
mode of longtraveling landslides, including their oc-
currence, flow, and deposition. Also, their soils have
not been extensively sampled. It would be helpful to
conduct field surveys, collect samples, perform soil me-
chanics tests, and accumulate data on this kind of sedi-
ment movement. It would also improve understanding
of the range of long-traveling landslides to investigate
the formation and development mechanism of a flow
layer that grows in a fractureprone intermediate soil.
REFERENCES
a
Ctual
s
tate
of
s
ediment
d
isasteRs
: 1983-1998 (1998) - SABO Technical Center (In Japanese)
H
su
, k.J. (1975) - Catastrophic debris streams generated by rockfall. Geol. Soc. Amer. Bull., 86: 129-140
H
su
, k.J. (1978) - Albert Heim: Observation on landslides and relevance to modern interpretation. In v
oiGHt
, B., ed., Rockslides
and Avalanches, Elsevier: 71-93
v
aRnes
d.J. (1987) - Slope movement types and processes. In: s
CHusteR
R. l. & k
Rizek
R. J. Ed., Landslides, analysis and
control. Transportation Research Board Sp. Rep., 176: Nat. Acad. oi Sciences: 11-33
H
unGR
o., e
vans
s.G., b
ovis
m., & H
utCHinson
J.n. (2001) - Review of the classification of landslides of the flow type.
Environmental and Engineering Geoscience, VII: 221-238.
J
aPanese
s
oCiety
of
s
oil
m
eCHaniCs
and
f
oundation
e
nGineeRinG
(edited) (1985) - Soil Engineering Library 27 - Prediction and
Countermeasures of Sediment-related Disasters: 27-64 (In Japanese)
l
andslide
C
ounteRmeasuRes
a
ssoCiation
(2003) - Round-table Talk on Landslides in Hokkaido (Part 2). Journal of Landslide
Engineering, 29(3) (87
th
issue): 34-47. (In Japanese)
m
izuyama
t., u
suki
n. & t
anaka
y. (2003) - Study of Long-traveling Landslides, Proc. of the 2003 JSECE Symposium: 184-
185 (In Japanese)
m
izuyama
t., u
suki
n. & t
anaka
y. (2004) - Study of Long-traveling Landslides (2) - Proc. of the 2004 JSECE Symposium:
238-239 (In Japanese)
n
iiGata
P
RefeCtuRe
(1989) - Guidebook on Geology in Niigata Prefecture: 92-93 (In Japanese)
o
yaGi
n., f
uJita
t., f
uRuya
t., H
ata
s., k
uRoda
k., n
akamuRa
s., n
akayama
y., s
Himizu
f., y
osHimatu
H., u
emuRa
t., H
iGaki
d. & y
aGi
H. (1991) - A draft presented at the Committee on Topographic and Geologic Terminology. Japan Landslide
Society (In Japanese)
s
assa
k. (2000) - Mechanism behind the Development of Fluidized Type Collapses, Proc. of Symposium at the Kansai Chapter, Japan
Landslide Society with the theme of “Occurrence, Movement, and Prediction of Fluidized Type Collapses”: 1-26 (In Japanese)
S
CHeideGGeR
a.e. (1973) - On the prediction of the reach and velocity of catastrophic landslides. Rock Mech., 5: 231-236.
s
oil
Q
uality
in
tHe
l
andslide
-
PRone
a
Reas
in
n
iiGata
P
RefeCtuRe
(1) (1983) - Journal of the Japan Landslide Society, 20(1):
28-36 (In Japanese)
s
oil
Q
uality
in
tHe
l
andslide
-
PRone
a
Reas
in
n
iiGata
P
RefeCtuRe
(2) (1983) - Journal of the Japan Landslide Society, 20(2):
5-14 (In Japanese)
background image
SOIL PROPERTIES AND FLUIDITY OF LONG-TRAVELING LANDSLIDES
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
795
s
tatistiCs
on
tHe
s
tate
on
l
andslides
(Part 2) (1975) - Reference No. 1121, Public Works Research Institute, Ministry of
Construction (In Japanese)
s
uRvey
R
ePoRt
on
l
andslide
-
Related
w
aRninG
and
e
vaCuation
(1989) - SABO Technical Center (In Japanese)
t
akei
a. (supervised) (1980) - Landslides, Collapses, and Debris Flows - Prediction and Countermeasures, Kajima Institute
Publishing Company: 192-226. (In Japanese)
u
suki
n., m
izuyama
t. & t
anaka
y. (2005) - Actual State of Long-traveling Landslides. Journal of the Japan Society of Erosion
Control Engineering (JSECE), 57(5): 47-52. (In Japanese)
Statistics