HKIE Transactions

ISSN: 1023-697X (Print) 2326-3733 (Online) Journal homepage: https://www.tandfonline.com/loi/thie20

Failure of Slopes and Soil Property

Characterisation

Kenji Ishihara JSGE JSCE ASCE JAEE

To cite this article: Kenji Ishihara JSGE JSCE ASCE JAEE (2002) Failure of Slopes and Soil
Property Characterisation, HKIE Transactions, 9:3, 8-19, DOI: 10.1080/1023697X.2002.10667886

To link to this article: https://doi.org/10.1080/1023697X.2002.10667886

Published online: 09 Apr 2013.

Submit your article to this journal

Article views: 60

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=thie20
 Failure of Slopes and Soil Property
Characterisation
KenjllSHIHARA JSGE JSCE ASCE JAEE
Department of Civil Engineering, Chua University

Some examples of landslides experienced in recent years are introduced with emphasis on evaluation of peak and residual
strengths of sails invalved in the slides. In one example, the peak strength and its interpretation are addressed in relatian to
triggering Of the large-scale landslide. In another example in terrestriai deposit flOW slides in gentle slapes are introduced
and interpreted from the concept of residual strength. Based on iaboratory tests on soils recovered from the sites Of these
landslides, characteristicfeatures Of material behaviour are highlighted putting emphasis on the conditions differentiating
between flow and non-flow type of deformation. For each of the cases stUdied, severai lessans learned are addressed,
indicating the importance of detailed investigation and insight into characteristics of soil behaviour relevant to the sites Of
landslides.
Keywords: Landslide, Stability Anaiysis, Loterai Displacement, Slape Failure, So/l Strength, Debris Flow, Underwater Slope
Instability

Introduction Types of Soil Failures

Deposits of sandy soils in slopes are always possessed of potentiality
to be put into conditions in Which movement is activated and continue
to progress largely in one direction once it is triggered by an external
agitation. This external force could be heavy rainfall, earthquakes,
accidental local slides, or artificial agitations such as pile driving or
vibration. No matter what the generic cause may be, once the flow-state
is stirred, massive soil tend to move downhill and bring about serious
consequence, which is more disastrous than that solely caused by the
triggering. This type of flow-failure could occur even over gentle slopes
where the stability in common sense is conceived not to be of serious
concern. Thus, the flow failure shOUld be envisioned as a phenomenon
which is differenl from that in which the movement is of limited amount.
The problem of slope stability may be thus classified into two types,
i.e., limited deformation type and unlimited deformation or flow type.
The difference is schematically illustrated in Fig 1. When the instability
is triggered by some external agency, then some amount of deformation
will take place anyway in the soil, accompanied by some cracks or
displacements in the range of 50 em to a few meters. However, the
displacement of such an order of magnitUde is not of much concern
because consequent damage would be minor and remedied with low
cost. However, if the flow-type displacement is triggered, consequence
would be trUly catastrophic as the moving soil mass rages over tile wide
area downhill.
It appears necessary, first of all, to recognize that there are generally
two stages in the progress of deformation at whIch failure is conceived
to occur in sandy soils under undrained conditions. The deformation
starts to develop SIgnIficantly at ttle stage when an external load induces
the shear stress WhIch is equal to or greater than the peak strength of
the soil. If the soil is in a state exhibiting dilative behaViour, the soil
tends to strain-harden and deformation will cease to develop further as
illustrated in the stress-path and stress strain diagrams of Fig 2. If the
soil of such characteristics exists in in-situ deposits, the deformation will
be of limited extent. Even when failure takes place, it will not progress
farther and consequent damage may not be disastrous. In the case of
soils being In a contractive or strain-softening state, failure will be equally
triggered when the externally induced stress becomes equal to the peak
strength. However, the deformation starts to develop extensively after
trIggering because of reduction In strength. If the 5011 of such characteristics
exists in field deposits, the deformation will be unlimited extent, leading
to complete slumping of fills or lateral spreading of the ground. This
type of deformation Is considered as the second-stage failure following
the triggering and called flow-type failure.
o o
.- c
"
,,--'"1:'-'-" ,

'"" """
',4,
,. ',
',
' _.L
L' .J.~

_. I ,
P~2' (0'1'0'))

Dens¢ sand
, ,
L1111lt~d dcfom1ation (Non-now!
q

_IN
,
"
,~

- ,
J .",.J ~ ~...•

,. fJ~--2'
, (.
al~"'3
• 1

Loose sand
, 1

Figure 2 - Dilative and Contractive Behaviour of Sandy Soils During

Figure 1 - Flow and Nonj/ow Deformations Of Sandy Soils in Slopes Undrained Shear

--.~----
- - - - - - - - _.. __. _ - - - - - - - - - - - - The Hong Kong Institution Of Engineers
 2550 m, involving about 35 million mJ of soils and rocks. The moving
slide mass shot the north-facing wall of the Denjo River canyon,

.
Stabilitl ana.Iysis
... No failure
disintegrated and burst into the air. Part of the shattered mass jumped
up through a height of about 70 m and sprayed over the ledge of the
for ailure terrace on the south. The main part of the mass was deflected toward
( External anllCV ) the southwest and plummeted down the canyon until it again Impinged
WISUS oc:ck $Irenath
~ on the east-facing waH of the terrace between the Denjo and Nigori RIvers.
Failure
Some fraction of the mass again surged up the west side of the valley

p()$\ -failure
...
instability llJ\alysis
( Gravily - driven )
{orce versus
residual strength L..I
I Row
Figure 3 - Classlficatian af Slape Stability Analysis
Analysis of slope stability is generally made by considering the peak
strength and therefore. it is the analysis to see if failure is Indeed triggered
or not. If a slope is identified to be composed of loose soils exhibiting
the contractive behaviour, analysis can also be made by envisioning
the state of soils deforming largely after a failure has been triggered
where tile mobilized strength is now smaller than the peak strength.
This type of analysis may be called post-failure instability analysis or
now analysis. A series of analysis envisioning these two stages of soil
deformation Is illustrated in a flow chart shown in Rg 3.
In the following. several examples of slope failures will be introduced
and reviewed from the context as above, and some points will be
highlighted which are important from perspectives of soil mechanics.
No flow
to a height of approximately 100 m and after spreading over the terrace,
it slid down into the neighboring valley of the N!gori River in the west.
It flowed further down the valley at probably a reduced speed and
converged at the confluence into the main debris flow, which had run
down along the zigzag path of the Denjo Valley. The main flow plummeted
further down southw~rds along the valley through a distance of about
9 km until it came to the point of another confluence where the Ohtaki
River streams toward the east. TIle debris flow partly deposited at this
location and blocked the stream drainage, thereby forming a dammed
Jake on the upstream side of the Ohtaki River. The debris flow traveled
eastward farther through a distance of 3 km, while depositing some portion,
until it came to a complete standstill at Korigase. The totai distance of
travel was 12 k.m and the debris flow claimed 9 casualties. Summarising
the accounts of eyewitnesses, the travel time for the debris flow is estimated
to have been on the order of 8 minutes, and therefore the average speed
was 80 km/h.
The schematic cross section along the entire leg of runout path of the
debris Is shown in Fig 8, where it may be seen that the course of the
debris trow Is divided into three parts; the upper reach where scouring
and grooving by the passing debris was predominant. the middle course

large-scale landslide In Naganoken-selbu

Earthquake of September 14. 1984, Japan

\~
General
A strong earthquake shook the central part of the Japan mainland at
8:48am on September 14, 1984 with its epicenter located at the southern
foot of Mount Ontake about 100km northeast of Nagoya, as shown In
Fig 4. The detailed map in the heavily affected area is shown in Fig 5.
By far the most spectacular in this earthquake was a cataclysmic debris
flow that originated on the south flank of the Mount Ontake and obliterated ,,,
the forest regIon of deeply incised valleys. The location of the debris ,
flow is shown in Fig 6, and its more detailed feature is demonstrated in o 5 10 km 1 "',,/
l J
Fig 7. The debris avalanche originated, as a huge landslide. from the
!
-'
south face of the mountain, between the altitudes of 1850 and

Figure 5 - Intensity Of Shaking Near the Epicenter (After Masaki et al.,

139'E 1985)
1
131E 1
136"E
31

37"N

KlyolOWl

'I
Figure 4 - Location of the Epicenter Of the Naganoken-seibu Earthquake
Of 1984 Figure 6 - Runout Route Of the Mt Ontake Debris Avalanche

TRANSACTIONS • Volume 9 Number 3

 neighbourhood were provided by Sakai (1985) as shown in Fig 10. The
cross sectlon in the longitudinal direction through A . A' is shown in
N
Fig 11. As can be seen in Figs 10 and 11, the headwall bluff about 130m
high is formed by steeply dipping fracture plane which cuts through several
layers of materials all from volcanic origin. Deposition of lava flow with
different compositions exists near the surface, underlaid by thin layeres
Denjo of scoria and pumice with yellowish colour. The pumice was overly
river consolidated, but because of its weakness in strength, it is deemed most
likely that the slide had been triggered by the failure of the pumice deposit
Underlying the pumice tephra there exists a breccia formation consisting
of cemented sand and gravel. The sliding surface downslope cuts through
the conglomerate and 'scoria deposit as indicated in Fig 11 Following
the earthquake, water was observed coming out of two springs on the
Nigori
rIver cliff of the Denjyo River in the most downslope part of the sliding area,
as indicated in Fig 10.

-+ Main river
~ Overflow
[] Deposition
Denuooted
§] area

~::::::I Hommocky
Qrea

Dammed
lake

Ohtaki Korigose
river

Figure 7 - Feature oj Flow oj the Debris Avalanche Figure 9 - Three Dimensional View oJ Mt Ontake Slide Zone

ConfluencE' Conflueoce End N

slide
( Grooving)
scraping
I ("'rHaL
depositIon
) I (~lt;on)
""9"''''

...... 1--4-AvatanChE',--I-·- - - Mud flow-----

.5~

lava

o L~1--2~~3-~4~...."S--e6~-7=---;6--e9:--,,"'O--f11,----,'''""'2
II
lava
D'istance (km) .•••
•••
,.
Scoria

Figure 8 - Profile of the Mt Onlake Debris Avalanche

•••
, ..
,,~ ~

Conglom-
erate
in which both scouring and deposition took place, and the gently sloping
part of the lower course where deposition of debris took place to a
!ill
PumicE'

maximum depth of 30 m in the riverbed. ~

Sond-
Landslide at Headwall SCarp
A large-scale landslide leading to the debris avalanche was triggered
.
;.-ave!

Spring

on the south flank of the Mount Ontake at a site about 2.5 km southeast
of the peak, as shown in Rg 5. The topmost part of the headwall scarp
was located at an altitude of 2550 m and the lowest elevation of the f".
;10- 'New ""'"
:Old scarp
o 300m
-J--J
slide scarp was at an altitude of 1850 m. The slide covered an area •
61

about 300 m wide and 1300 m long. A three-dimensIonal representation t :51.ide orea
of the slide area is given in Fig 9, where it is to be noticed that the
depth of sliding plane was almost 150 m at its deepest location. Figure 10 - Geological Regime In the Headwafl in Mt Ontake Landslide
The geological conditions over the basal failure surface and ;n its (Sakal, 7985)

m The Hong Kong Institution Of Engineers


mm
Lava
I2l
Pumic~
lID ~ to various degrees, pore water pressure measurement was not made
Conglom- Son(l-
trot.. gro~.. l
I:·:·J
-~
o
t
300m
'--J.._J
Figure 11 - Georogicaf Regime of Mt Ontake Slide Zone ;n a Longitudinal
cross Section after the Slide (After Sakai, 1985)
In-situ Soil Sampling
Undisturbed samples of soils were secured from the exposed failure surface
of the deposits which were considered to have been responsible for
triggering the slide. In the area of the headwall scarp at Mt Ontake, samples
were obtained from two places of pumice deposits with different altitudes.
The sampling site No 1 is considered to have had a depth of about
10 m from the original ground surface prior to the slide, whereas No 2
site had been at a depth of about 50 m from the original surface. Therefore,
the effective overburden pressures at these two sites are assumed to
have been approximately 200 kPa and 1000 kPa, respectively.
When performing sampling. columns of intact soil were trimmed and
tubes, ZS em in diameter and 14 em in length, were pushed gently into
the columns while expelling excess soil outside the tubes. The sample
was then sealed by paraffin to prevent moisture change. In the laboratory,
undisturbed specimens extruded from the sampling tubes were tested
both under static and dynamic loading conditions using a triaxial test
apparatus. Water content. void ratio, specific gravity and grain size
distribution curve were determined after completion of the tests for
representative specimens from each batch of samples. The results of several
physical tests are summarized in Table 1. The pumice specimens form
sites in the headwtll/ scarp at Mt Ontake are shown to have water content
as high as about 100% in their nearly saturated conditions, and the
material is identified to be low-plastic In its disturbed state. It Is surprising
to realize that the pumice with such a high porosity could have existed
in a deep deposit undergoing a high overburden pressure up about
1000 kPa. A typical gradation curve of the pumice is indicated in
Fig 12. Consolidated undrained triaxial tests (CU-test) were performed
.. :<00
m on undisturbed specimens to determine the static strength of pumices
under monotontc loading conditions. Because samples were saturated
during the undrained phase of load application, and the test results were
consistently interpreted on the basis of total stress concept.
For the specimens from No 2 site, in the Mt. Ontake headwall scarp, it
was considered desirable to employ a higher consolidation pressure in
__ the test, in (l(cordance with the high pressure which had eXisted before
the overburden was removed by the slide. However, because of the
limitation in the capacity of cell pressure application, the consolidation
pressure used was less than 500 kPa. ll1e results of the monotonic static
loading tests on samples from the No 2 site are shown in Fig 13 in terms
of stresses at failure plotted in the Mohr circles. The straight line enveloping
the Mohr circles is known to define a state of stress causing failure in
the soil sample. The failure envelopes thus determined from the static
tests are indicated by solid lines in Fig 13. The samples from No 2 site
were recovered about 8 months after the quake from intact deposits
exposed on the surface. However, because of a long-time exposure near
the surface, the samples appeared to have lost some of the water content
having been present at the time of the earthquake which had taken place
after a week of precipitation amounting to a total of about 160 mm. In
view of this, samples form No 2 site were saturated In the tests by
circulating de-(lired water for about 24 hours before the conduct of the
test. Samples were thus made to have a saturation ratio exceeding 90%
with a water content increased to about 110%. The results of tests shown
in fig 13 indicates that the angle of internal friction is taken as being
ep = 23° and the cohesion as (=1 SO kPa, on an average, although there
are some scatters in the test results.
Un<1'5hJrtxod ~pl" from !I'\fo
,"",a_It &corp 01 Ontokeo O"'lloncl'lfo
w.al~pd pun>ceo
101_11_122'" 0,_2·56
no12·S -15.1 kN/m'

Specllic Unil Water Liquid Plastlclly oL..L,,!ooo,--J.....,.;!;OO;-''-;600;;;!-..L.'600'" .Lu.L,::tooo;;;-LJ'!'200d~,,,,,,,:!c-.L'''!600

Site Gravlly Weight Content Limtl Index Co<>l;""'9 51."", ,CT (It·V,,",)

G, ro (%) W, Ip

Mt Ontake No 1 269 13.4 120 N, N, Figure 13 - Strength Charaderistics of Pumice from Mt Ontake Slide Area
headwaJi (NO 2sitel
scarp No 2 2.56 12.8-13.2 100-110 N, N,
Table 1 - Physical Properties Of the Pumices at 7Ivo Sites of Londsliding
Stability Analyses
Seismic stability analyses were conducted for the head scarp landsliding
1OC~~~~~~
~ Mt. Ontake
by using the pseUdo-static method Of slope stability analysis as follows,

-
<
-HeadwoU
- scarp
I
,•
0
(Pumice)
Fd ~
IW tan <I> + Col coso; I/(cos'o; (1 + tan 0; tan <l>/F,lI
(1 )
•
" so . I [wtano;+ a;, ,w]
--
•"
-•e" where W, I, a and denote, respectively, the total weight, the length and
~
the inclination of the sliding plane of each slice, as illustrated in Fig 14.
0 This formula is a modified version of the formula originally proposed
0001 0.01 ().1
Grain 5izf' (mm) by Janbu (1955). In this method, the effect of seismic loading is allowed
for by an equivalent statIc force determined as the produd of the maximum
Figure 12 - Typical Gradation Curve of Pumicefrom the Site Of Landsliding acceleration, arT\ilx' and the weight of the potential sliding mass.

------'--------------------m
TRANSACTIONS. Volume 9 Number 3
 The cross section shown tn Rg 11 was considered to represent the most
critical sliding surface on which the sliding was first triggered during
the earthquake (Sakai 1985). The analysis of slope stability was made,
therefore, for the profile modeled thereafter as shown in Ag 15. As
Indicated in Fig 15. the entire mass was divided into 16 slices for the
computationClJ scheme. The soil at the bottom of slices 1 to 11 was
allocated a strength parameter of ¢l = 500 because this part consists of
angular sand and gravel derived form volcanic Lava flow. The slices 12
to 14 were underlaid by the pumice which was identified to be a highly
porous soil directly associated with the catastrophic slide. In accordance
with the strength data shown in Fig 13, the angle of internal friction
was taken as (f) = 23° and the cohesion in dynamic loading. Co' was
assumed to be 300 kPa, a value twice as much as the static cohesion
indicated in Fig 13. The reason for thIs increased cohesion is based on
a number of dynamic loading tests that have ever been conducted on
cohesive solls (Ishihara, 1985). For the slices 15 to 16 the materIals consist
of fractured lava and conglomerate and, as such, are assumed to have
Figure 14 - Nototion of Slope Stobllity AnalysIs
an angle of internal friction of ell = 500. Unit weight of soil and rock in
each slice Is assumed equally to be 1, = 20 kN/m 1.
The factor of safety computed by Eq (1) is plotted versus the maximum
acceleration in Rg 16, where it can be seen that the fador of safety,
Fd , drops to unity when the acceleration becomes equal to about 280
gal. This value of equivalent acceleration may be considered reasonable
jUdging from the values estimated by several sources of information. It
may be mentioned that Initiation of the landslidlng can be Llscribed to
the presence of a thin layer of pumice deposit at great depths which
had probably been under near-saturated conditions due to the water
1 •• permeation from the heavy rainfall preceding the earthquake.

lessons Learned
One of the characteristic features of this slide was the fad that the slide
300m 0 took place along a deep-seated soil deposit which was located at a depth
I I l--l
of 100 m to 150 m. It is to be noted that the presence of a cohesive
soil at a considerable depth does not contribute much for maintaining
the stability of slopes. This tad can easily by understood if one considers
a simple case of a long slope in which the sliding surface is located at
Figure 15 - Praf/le of Sliding Usedfor Analysis (Mt Ontake Headwall) some depth, H, parallel to the slope surface. In static conditions, fador
of safety, F, can be obtained by

tan (f) c
F= +--,,-=~---::;:,--­ (2)
tan 0: 1, H cos ex sin 0.

Mt. Ontake slide where a is the inclination of the slope and Yt denotes the unit weight
2·0 Headwall scarp of the soil. The first term on the right-hand side indicates the contribution
."
Static factor from the frictional resistance and the second term the resistance due
"-,
of safety to cohesion component It Is to be noted that the resistance due to cohesion
•d
/ can be activated In a form of (/Yl H, and, therefore, if the depth of a
~

0
sliding surface, H, is large, the resistance by cohesion becomes negligibly
-• •
<T
.c
c small and the main resistance comes form friction. Therefore, when a
d
layer of soft 5011 exists at a great depth with an Inclination, ex, Which is
'·5
grater than the angle of internal friction of the soil, <1>, the factor of safety
'"c
c computed by Eq (2) could take a value less than unity for the case in
0
." which the value of el"{\ H becomes very small. In such a case, it is
--••
~

d
apparent that the stability of slopes could be seriously endangered. From
the considerations as above, one of the important lessons one could

-
0 1·0
learn from the tragedy of Mt Ontake landslide is the fad that wherever
there exists a deposit of clayey soils In slopes at a considerable depth,

-
L
0
the stability of soil and rock mass overlying the clayey soils must be
u
0
"-
considered meager. Therefore, It 15 cautioned that detection of any deposit
of cohesive soils by deep boring is badly in need whenever an important
construdion projed is under plan on mountain slopes.
Q.5l-.L-c':-:--'----,-'-'--,L----'-...,...~~-'--~-'-' Shown also in Ag 6 is the fador of safety in the static condition prior
o 100 200 300 400 500 600 to the earthquake which is obtained by using static cohesion.
Mox. acceleration Qmax (gal)
(=150 kPa, in Eq (1) with the condition amax = O. It may be seen that
the slope must have been in a stable condition prior to the earthquake
with a factor of safety of about 1.63, but lead to the catastrophic failure
Figure 76 - Factor Of Safety Versus Acceleration because of the additional dynamic force during the quake.

1Df-----------------------------------::Th~e:'-H::o-n-:-g=-K:::o-:-n=-g=-l:::n-::st:::itu=-t:::io=-n=-0:::1:::E::n::g:::in::ee=rs
Flow Slide in the Collapsible Loess Deposit in
Ex-soviet Tajik VOrzob
vaHey
General
In the suburb of Dushanbe, Tajikistan Republic of ex-USSR, an earthquake
-t
~
A90gal
Habobod S
of magnitude 5.5 took place on January 23, 1989. In this event, extensive l' L. .Dushanbe
liquefaction developed in the loess deposit of Aeolian origin In the gently ~ aSOgaL
c
sloping hilly terrain and led to a series of catastrophic landslides ~ ~gal
accompanied by a large-scale mudflow. In contrast to the hitherto known • .5
lOOgal
cases of liquefaction which have usually occurred in water-saturated sand
IUihok
deposits, the liquefaction in Tajik was unique and novelln that it occurred
unexpededly in a windlaid deposit of silt in a semi-arid region. The reasons
63()gal
for such liquefaction are thought to be the collapsible nature of highly
porous loessal silt Which had been wetted by irrigation water over the Cymbvlif A60gal
past years. The complete collapse of the loess structure due to the
o 10km
addItional action of the seismic shaking appears to have led to the LI~~-,--,-.JI
.c:.60gal
catastrophic landslide. In addition, the silt-sized soil constituting the loess
was of low plasticity and hence could easily slump and flow through a
distance as long as 2.0 km (Ishihara et al. 1990).
The location of the epicenter is shown in the map of Fig 1Z As shown
in Fig 18, the affected areas Is located on a relatively flat basin-like plain
whIch developed in front of the southern flank of the Pamir Mountain Figure 19 -Intensity Distribution Of the 1989 Jan 23 Earthquake
range. The epicentral area is located at Gissar near the junction of the
Kahirnighan River and the Halaka River as shown in Ag 19. landslide and Debris Flow
In the Village of Gissar, there were about 500 farmer's houses and bams
constructed of woods with adobe type walls. Even In the most severely
affected area, the degree of structural damage was moderate and limited
to small local area with Village houses and barns being partially destroyed.
The most striking feature of the damage was a series of landslides and
debris flow which took place over the gently sloping hilly terrain consisting
of windblown deposits 'loess'. There were four landslides In the affected
Chino
area of Gissar as indicated in plan view of Fig 20. One of the largest at
Savdi Okuli turned into a mudflow of vast scale and buried more than 100
Arabia
houses in 5 meters of mUd. An estimated 220 Villagers dies or were
Ri:a,dh
missing in the debris.
The largest and most cataclysmic phenomenon was the multiple slides
at Okuli which developed in a Slightly depressed area over hilly farmlands.
The plan view of the slide is shown in Rg 20. The genetic portion of
the slide extends westward over a length of about 1.5 km from the
escarpment at the east (see Fig 20). Taking into account the fact that
the slide encompasses an area about 850 m in frontage and 15 m In
depth on the average, the total volume of the soil mass is estimated to
be approxImately 20 million CUbic meters. At least two slides seem to
have been triggered independently from the hillsides on the north which
Figure 17 - Location of the Landslide in Tajik

• TQtl'lk~nt

•
-1-
_.¥l'OI....
.
• _,Rood
,~

1 : 11.oco
o \.Ok'"
L..I_~_-'I

Figure 18 - Location Of Dushanbe and Landslide Figure 20 - Plan View Of the Slide Area In Gissar

--~---=-~~---------------'m
TRANSACTIONS. Volume 9 Number 3
then merged into the main stream of the mudflow. There were many
traces of violent mud flow remaining on the exposed soils, indicating
evidence of liquefaction and consequent muddy flow of loess soil during
the earthquake (see Fig 21). The slides indicated by A and C in Fig 14
appear to have been initiated along the line of the water channel. It
appears likely that a small slide Initially Induced at the toe might have
retrogressed backwards over a distance of 15 km to the east and sideward
to the north as well as to the south. Over the rugged sliding area, the
original ground was broken into many blocks producing an extremely
irregular and hummocky surface. Amid broken blocks, there were spreads
of soil mud having presumably sprung out of the in-depth sliding zone,
suggesting that liquefaction of saturated loess soil occurred during the
earthquake. The sliding mass of the loess soil turned into a huge-scale
debris flow and traveled through a distance of about 2 km on nearly
flat surface of the ground, Near the distal tongue of the debris flow at
Okuli-poen, tens of farmer's houses and barns were buried in 2 to 5
meters of mud. A photograph in Fig 22 shows buried houses near the
end of he debris.
Several cone penetration tests were performed on the rugged surface,
at places 1 through 13 as indicated in Fig 20, by means of a portable
cone device. Typicai results of the cone tests are shown in Fig 23, where
it may be seen that in the sliding zone at depths of about 7 to 8 m,
qc-value was found to be about 1-5 kgf/cm 2 ,
Soli Condillons In the Affected Area
The area affected by the earthquake has a topography of gently sloping
hilly terrain, The wind laid material is silt and tan to light brown in colour.
The soil conditions in this area were investigated by the Geological Bureau
of Tajikistan Academy of Science in Dushanbe. The cross section along
the longitudinal direction of the slide (A-A' section in Fig 20) is presented
in Fig 24. It may be seen that the loess deposit covers the ground surface
to a depth of about 30 to 40 m, underlain by gravelly sand which varies
in thickness to a maximum of 200 m.
The results of investigations by the Geological Bureau of Tajikistan
Academy of Science show that the ground water table is located about
5 m from the ground surface, but the water content diminishes around
the depth of 20 m which is underiain by dry layers of the ioess. It is
also reported that, while the layer of loess about 3 m in thickness below
the ground water table has a water content of about 20%, the loess
layer below this level does possess a higher water content with a maximum
of 40%, On the other hand, the laboratory tests have shown that the
liquid limit and plastic limIt of the loess is around 30 and 20, respectively,
giving a value of about 10 for the plasticity index. Thus the natural water
content of the highly saturated loess is in excess of the liqUid limit. Fig
25 shows the grain size distribution curve of the loess, It may be seen
that the loess is comprised of 80% of silt and 15% of cray fraction, and
that there is no sand fraction.

Ok_Ill lotld4l1dt

1 Debril<.l.Okrnl _ _

r
800 m
,

"'"
Figure 21 - A View from the North over the Eastern Escarpment Of Okuli .-- ---
Slide
:..~ .-.-Jokm
"'"
m

..,
--~

-
.._...._, .. _- -- .. - ---- Figure 24 - Cross Section through A-A'

Clay Silt Sand

100 (

0 - 5i I t
-

Figure 22 - A Buried House in the Debris at Okuri - f-

GiSsor-,Tajik

o , , ,
~ , ~. ( g "::<tn"J
e'----'--._'"- P:~ , • 'I,• • (I;I/[m')
0.001 0.01 O. 1
4 8 10
" " " " "
,
,
, (,
N<>.O,'O J
,
, lb,
,"",II,ll,13
Particle size

, Okllti ili"'"
lGjik, USSR , O~ul; llld.
TaJ;k,Us.s.R.
Figure 25 - Gradation Curve Of the Loess 5011
•
, ,•
, .. -/'."'"'-'
_ ....t...
•,.,........ ....
~". "" ~
/'"-"--'
, Characteristics of loess Deposits
,
,• ~
It is well known that windblown deposits of soils characteristically possess
, ~~ vertical cleavage planes. It is also known (Clevenger, 1958; Dudley, 1970;
, Chen, 1987) that the windlaid loess is tan to light brown in color, light-
lO·
" weight and consists of matrix of cemented silt with a number of
interspersed macro and micro pores which are purported to be remains
of root holes. Generally, the loess deposit Is devoid of clear stratification
Figure 23 - Cone Penetration Test Results at Okuli Slide Area and is highly crumbly and brittle in a dry state, being easily crumbled

1IIf----- The Hong Kong Institution oj Engineers

and ground to powder by fingers. The loess is known to be composed
Wot.,. cOI1t.nt Pl!orrMClbi Ii ty
of sllt-sized particles of quartz and feldspars bonded together by a small
fraction of montmorillonite·type clay which is collapsible when wetted
• 0
101 ("10)
20 40 60
k
,n -.
(cm~d

10' ,W
~lth water. When maintained dry, it is reasonably strong and
w, w,
incompressible and the porous structure may persist even under a fairly · '"
I
~arge a.v.erburden pressure. However, once wetted, the loess may lose 5 ~
Its stabilIty. Because of Its highly porous nature, water can easily infiltrate !\J
into the pores, whereby breaking down the matrix structure. As a result
a significant amount of decrease in bulk volume and loss of shear strength
, ":",TV:/.:
0
~~~
solu ~
can take place upon water Inundation under sustained loads, leading rott>d
, zonr
to large settlements or failure of structures founded on such loess deposits
(Chen, 1987). This kind of characteristics is often referred to as hydraulic
•
•
S
"'r''/1
~;~: )};j

collapsibility.
.' 20 I
Another feature of engineering sIgnificance is the fact that disturbed loess ·, ..-. ..'..' . . - .. . . ..- . .
'.
• •
I
. ...' •.• •.' ' w, w,
Is of low-plasticity with its Atterberg limits plotting near the A-line in
the plasticlty chart. The plasticity index of the loess is generally around
·.' . . . ' .....•
• • •••••• •
2s
m

10. Thus, if collapse is triggered with a sufficient amount of water Invasion

in excess of the liqUid limit. the disturbed loess being devoid of cohesion
tends to easily slump and flow. It is known that any soil with a low
plasticity Index has a great potential to develop liquefaction and flow-
type failure (Ishihara and Kosekl, 1989).
With the above mentioned several characteristics of the loess taken
together, it Is expected with good reasons that the loess deposits in Gissar
area had been invaded with a large quantity of water prior to the
earthquake and narrowly on the verge of hydraulic collapse, and when
SUbjected to the seismic shock loading, it spontaneously developed
liquefaction and resulting mud flow.
Mechanism of Liquefaction and Debris Flow
The lands in the Gissar area had been left uncultivated throughout the
geological time. However, with the recent development of township and
increasing popUlation in the urban area of Dushanbe, the lands have
been cultivated and used as agricultural farmlands to produce wheat
and cotton. To mach the needs for water supply in the agricultural lands,
a network of water channels was constructed In the hilly area in Gissar.
Some sections of the water canal were lined with concrete but most
were unlined. The water for irrigation was pumped up to the hilltop,
and was distributed to each patch of farmlands from two large storage
tanks or. directly through a network of open canals. The water in the
canals had been leaking and penneating into the ground through the
vertical fissures In ~e loess deposit over the years and the large quantity
of water had been absorbed thereby filling up the pores in the loess.
As indicated schematically in Fig 26, the ground water table was found
to lie about 5 m from the ground surface but the water content of the
loess at a depth greater than about 20 m was found to be small. It was
also discovered that the water content in the layer between depths of
about 7 m and 17 m is greater than the liquId limit. Thus, the water
content might have probably been distributed throughout the depth of
the loess deposit as shown in Fig 26. At depths a few meters below the
ground water table, the pores of the loess were probably only partly
filled with water, but as the water pressure increased with increasing
depth, the pores probably became fully saturated, producing a state of
over-saturation- with a water content in excess of the liqUid limIt. The
exact depth of the vertical cleavage In the loess deposit is not known,
but the maximum depth at which vertical cracks can remain open may
be inferred by comparing the vertical overburden pressure with the uniaxial
compressional strength of the soil element, as illustrated in Fig 2Z Suppose
the vertical stress at depth Z given by y\ Z exceeds the uniaxial
compressional strength, Q." then the soil element at this depth will fail
whereby producing a large deformation in the lateral direction. If there
exists an open crack to this depth, It would be closed by this latE:ral
bulge. Therefore, the depth at which an open crack can remain may be
given by,
z~q,/Yt (3)
where "it is unit weight of the soil. The exact value of the unaxial
compressional strength of the loss at Gissar is not known, but in view
Figure 26 - Hydraulically Collapsible State Of Loessal Deposit
1) Condition for crack
., closing ..
y, Z >= qu
Z ~ qu /Yt.
Z
2) Uniaxial compressionat
strength of Loess, qu
qu. 20-40 l/m'
y,Z
y, = 1.5 t/m1
I
~
:.Z~15-25m
Figure 27 - Conditions Of (rack Development
of the cementation developed in the matrix structure, the strength may
be inferred roughly to be within the range of qu = 200-400 kPa. Then,
assuming the unit weight to be approximately 'Yt = 15 kN/m 2 • the depth
of crack penetration Is estimated to be about 15 to 25 m. ConsIdering
a dominant role played by the openness or closure of cracks, the gross
permeability of the loess deposit would probably have been distributed
through the depth as shown in Rg 26. 11 15 to be noted that, at the depth
between 15 and 20 m, the permeability coefficient decreases sharply,
prohibiting the water from migrating farther into the deeper portion of
the deposit. This would account for the fact that the measured water
content in the deeper deposit was small of the order of 10%. By retaining
a large quantity of water, the loess deposits between the depth of about
7 and 17 m appear to have been in a state of impending hydraulic collapse
even prior to the advent of the earthquake. In fact, it is reported that a
noticeable amount of ground settlements had be observed over the surface
of the farmland even prior to the advent of the earthquake. This fact
indicates that the hydraulic collapse had partially occurred in the loess
deposit in this area. However. since the topography in Gissar area is so
gentle, that the downslope component of the gravity force was not large
enough to cause land sliding and therefore the ground remained stable
as a whole before the occurrence of the earthquake. When the earthquake
occurred in this area, the saturated loess already in a precarious state
underwent shaking and total collapse of the loess structure was provoked,
leading to liquefaction of the silt-slzed material. Moreover, because of
the low-plasticity of the slit, the liquefied silt began to flow out even
on a nearly flat ground surface, carrying large masses of soil. In the
case of the slide in Okuli, the mudflow traveled through a distance of
2.0 km as indicated tn Fig 20. In the Okuli slide, the slide appears to
have started first in a small sedion adjacent to the toe, and a series of
sequential slides must have retrogressed from one section to another
until It reached the place of the final escarpment on the east. This type
of progressive failure Is likely to develop in soils which require only small

=T=-RA=-N~S~A~CT=/:-:O~N:-:S-'-~~0~/u-m-e-9-=-:-N~um-be-r:-:3------ -----------~---------------III
magnitude of driving force to cause failure. It is to be noted that no
landslide and mudflow in such a gentle slope would indeed take place
with such a large scale except by the shaking of an earthquake.

lessons learned -r- ··A

The wlndlald loess consisting maInly of silt·sized soli was likely to be K.P. ±()m
I H,W_l~::~
in a barely stable state on the verge of hydraulic collapse before the l.W~-:-

earthquake, and this led easily to a complete collapse in a form of

liquefaction upon being further subjected to seismic shaking. The complete
slumping and long-distance flowage of the mud on the nearly flat ground
was explained by the fad that the loess-forming slIt is of low plasticity
with a plasticity index of about 10. It is warned, therefore, that any silty
sand with a potential of hydraulic collapse is exposed to failure or flow,
Figure 28 - liIustration Of Key Points to be Defined in the Methad Of
if it is wetted or inundated with water and subjected to small agitation
Ground SurveyIng
such as earthquakes. The result of the post-stability analyses made for
four sites of landslides in Gissar indicated the mobilized undrained resIdual
strength of the silt to have been probably in the range between 2.0 and
15.0 kN/m 2 which apprOXimately coincides with similarly evaluated values
of residual strength in other case studies of flow failures.
..-
, t'
lateral Spreading of Waterfront Deposits In the 1995
Kobe Earthquake
General
The Hyogoken-Nambu earthquake, registering a magnitude of 7.2 (JMA),
occurred before down at 5:47am on January 17, 1995 and delivered a
very high level of devastating shock to densely popUlated area of Kobe-
Osaka corridor which forms a major Industrial heartland in the western
part of Japan main island. The total death toll was over 6,500 and more
than 300,000 people were left homeless. The very strong shaking
collapsed approXimately 150,000 houses and building. ignited fires,
wrecked elevated highways and railways, and destroyed ports and harbour Figure 29 - Location oj Cross Sedlons where Measurements were Mad
facilities. by the Method af Ground Survey
During the earthquake, liquefaction developed extensively over the areas
of reclaimed lands in the port and harbour district of Kobe city. In the
terrain behind the quaywalls, the liquefaction gave rise to an increase
•
in lateral force, and led to excessive movements of the caisson·type wall
as much as 3 to 4 m towards the sea.
The horizontal movement of the wall was accompanied by an equally
large amount of lateral spreading of ground soils which propagated
rearwards Inland. The ground rupture as it propagated backwards Is to
be considered as a phenomenon of engineering significance as it defines
the spatial extent of the ground distress to Which due consideration may
need to be given in relation to the design of foundations. In view of
this, attempts were made to investigate features of ground displacements •
and settlements caused by the lateral spreading of liquefied soil during
the earthquake in the wharf areas along the peripheries of Port Islands,
Rokko Island and Fukae Island in Kobe district.
• •
lateral Spreading
At a number of locations, measurements were made on the width of
open cracks and vertical offsets across the cracks on the ground surface Figure 30 - Detailed PrOfile oj the Ground Deformation at R·2 Section In
along an alignment in the direction perpendicular to the revetment line Rokko Island
(Ishihara et al. 1997). This method will be referred to as the ground
surveying. The scheme of survey on the ground surface is illustrated in constructed of graVity-type caisson walls with their inner compartments
Rg 28. The In-situ measurements started from a poInt C behind a quaywall filled with sand. A vast majority of the caisson boxes were placed on
until a datum point 0 is reached where there is no crack visible on the the stone mounds resting on the Masado beds, which were artificially
ground surface. By summing up the width of the crack openings placed after the soft clay In the original seabed had been removed. The
successively from the datum point towards the waterfront. the lateral material behind the caisson walls was also the Masado, the same soil
displacement at any point on the alignment was determined. It became as that used for the foundation bed. A typical result of displacement
thus possible to depld the distribution of the lateral displacement as a measurements is displayed in Rg 30 for the cross section R·2 in Rokko
function of distance inland from the waterfront where caisson-type Island where it can be seen that the caisson body did stoop, accompanied
quaywalls moved towards the sea. The locations where measurements by a lateral movement of 1.56 m and settlement of 0.96 m. The
were made in the area of Kobe are shown in Ag 29. They are all located dIsplacement Is shown to have extended inland as far backwards as
in the peripheries of man·made reclaimed island in which the soil consists 183.7 m. A soil profile near the point R·2 is shown in Rg 31. It may be
practically of the same materials, Ie, Masado derived from weathering seen that the man-made fills composed of Masado soil existed to a depth
of granite rocks. The revetments or quaywalls in tllis area were alt of about 20 m. liquefaction is considered to have occurred In this

11I1-----~--------------------------------7::::h;-e--:-:H;-o-ng-;K;-o-:n-:g-:ln-:s::t::itU-t~io::n:--:af;-E-;n=-g::in=-e=-e=rs
reclaimed fills. It is conceivable that there were dual factors which might
have acted to produce the lateral displacement of the caisson walls. These
are inertia force due to the intense shaking and softening of soils
surrounding the caisson caused by buildup of pore water pressure. It
may be roughly assumed that the effects of pore water pressure buildup
including liquefaction were, by and large, the same all over the waterfront
areas in Kobe port that were affected by the earthquake. However. the
effects of inertia force are deemed different depending upon the direction
to which the revetment line is oriented. As a matter of fad. the records
of accelerations secured during the earthquake showed motion
characteristics which are largely biased to the northwest·southeast direction.
Based on the records obtained nearby, the acceleration inferred to have
acted on the quaywalls oriented in two directions were established by
lnagaki et al. (1996) as indicated in Fig 32. It may be seen that the
horizontal acceleration against the north- or south-facing walls was of
order of 500 gal whereas an acceleration of the order of 250 .
300 gal is inferred to have acted against the east- or west-facing walls.
In view of this, all the data on the ground displacements measured by Figure 32 - Magnitude Of Acceleration Inferred to Have Acted on the
the method of the ground surveying at Port Island, Rokko Island and Quaywalf Facing Different Directions (lnagaki et 01., 1996)
Fukaehama were divided into two groups; one pertaining to the north-
or south-facing walls and the others to the east- or west-facing walls.
Fig 33 shows all the lateral displacements from the north- or south- 500
facing caisson walls plotted together, versus the distance from the line
of revetment. It is noteworthy that there is an acute drop in the North or south lacing
400 caisson waifs
displacement in the vicinity of the walls, which is indicative of pronounced
influence of inertia force in producing the large displacement near the
walls. Fig 34 shows all the data of the ground displacements obtained
at the sites where the quaywalls are oriented towards the east or west. f-2 - fl·z
P,8 ----. A·3·-··
As compared to similar data arrangements shown in Fig 33, the attenuation I"~ _._.. fl-. _._.-
,., - - - fl·~ _ . -
of lateral displacements is seen in Fig 34 being more gradual even in 1"·7 _ •• - fl'·! _.,_
Averag9 Cll0V9 -
the area in proximity to the walls. Taking the average cUlVes of attenuation
shown in Fig 33 and 34, the lateral displacement at any point normalized 100

to that at the waterfront is shown in Fig 35. It (an be seen in Fig 35

that the attenuation is more pronounced for the soil deposit retained o
_
'
- -
......... .. ..
~~~.~""~.~_:::=----~.~ ~ ...
by the north- or south-facing caisson walls as compared to the ground o SO 100 ISO 200
Distarl«i from the watertront (m)

5011 SPT ~"akle

typo Figure 33 - Lateral Displacements Versus the Distance from the
10 20 30
'" Waterfront (North- or South-facing Quaywaf/s)
T -2.4m

Reclaimed
deposll East or west ladng
~ca~'s,!:so~"~w~a'!l's,-_.
(Masado)

Rokko Island •
E 300
East $k.Idge
treatment 6119
J
"iii 200
f·t _
p_, ._
f-3 _ ••••

p.: _.-
p., _ ..-
P'·s-
1"-&-_ •••
p-·s _. __.
1\-1 _ . -
I\~ - ...

~ Average <;UfVll -

2S
100
B
o~~""~"c._.~"..~
SOy day

.. ·..
o 50 100 1W ~
Oistancelrom \he waterfront (m)

Figure 34 - Lateral Displacements Versus the Distance from the

Waterfront (East- or West-facing Quaywalls)

behind the walls facing the east or west direction. The reason for the
rapid attenuation in the former case is sImply the larger movements of
the quaywalls which were caused by a greater acceleration in the north-
south direction during the Kobe earthquake. As discussed above, the
ground movements in the backland retained by the east- or west-facing
walls appear to have been caused primarily by the liquefaction of the
Figure 31 - Soil Profile at a Site Near the Section R-2 and R~3 (Hanshin
soil deposit, and for this reason, it would be legitimate to choose the
Highway Authority, 1993)
-------,-------------------111
TRANSACTIONS. Volume 9 Number 3
 , ,
1.0
~
i
I
eUlor_
"i
E 0.5
I
!K!Ig _als
8
0"
~i NQ,ltI Qf toult!
~O
~
I
/ 2, pp. 'lfi7-190.
0,
0 so 100 ISO
Oislance lrom the _Iertron! 1m)
Figure 35 - Comparison 0/ Normalized Displacement Patterns in
Liquefied Deposits Retained by the Quaywafls Facing Different Directions
attenuation curve pertaining to the east- or west-facing walls for practical
purposes where estimate of lateral displacements Is needed considering
the quaywall movements less than 2.0 m with a ground shaking of the
order of 200-300 gal.
Lessons Learned
No matter whether it may be due to earthquakes or due to other causes,
if the quay wall in the harbour area is displaced laterally on the order
of 2.0 m, the possibility will be high that the lateral movement tends
to propagate inland and facilities installed on tile backland is exposed
to danger of damage. Thus, it would be necessary to pay due consideration
to Its damaging effects. It was shown in the in-situ survey that the lateral
movement could reach a place about 100m inland from the quaywalJ,
if the soil developed flow"type deformatIon with the quaywall displacement
of the order of 2.0 m.
Conclusive Remarks
Results of detailed studies were introduced for four cases of landslides.
It was shown that individual cases of landslides exhibited different
characteristic features in terms of soil behaviour particUlarly when the
soils are deformed largely.
In the first case in which the mountain slope was studied, the importance
was addressed for performing deep boring even ttuough stiff and hard
materIals with an aim to identify layeres of cohesive soils, if they happen
to exist. The existence of cohesion less soils at great depths was shown
to endanger the stability of slopes and once the failure is triggered, it
would bring about a catastrophic consequence because of a huge volume
of earth materials involved.
In tile second example described above, the collapsible soils were shown
to have a potential to become unstable if they are put in conditions
where water is Inundated. Failure of even gentte slopes could readily
take place, if there is a triggering agency such as seismic shaking. Once
it Is triggered, the slide tends to turn into a large-scale debris flow involving
a large volume of soils and brings about disastrous consequence.
In the last example quoted above, the importance was addressed for
considering the progression of soil movement backwards from the
wLlterfront which is caused by the lateral deformation of quaywalls. In
the backyard behind the quaywall, there may exist facilitles such as load-
unloading cranes, warehouses, storage tanks or chemical plants. These
industrial facilitates are exposed to danger of injuring if such ground
movement occurs.
Acknowledgements
ThIs text consists of an excerpt from several works and papers previously
pUblished. The author wishes to acknowledge much cooperative assistance
mr--------------------------------------T.:::h=-e=-H=-o::n::g=-::K:::o::ng:::-;/n::s:;tl:::·tu:::t:;:io::n:-o:::if-;E~n::g::in:::e::e=rs
offered by his colleagues. The author also wishes to express the mind
of his sincere thanks for the colleagues in Hong Kong as represented
by Professor C F lee for their kindness in Inviting and according me a
great honour to deliver this prestigious Peter lumb lecture.
References
1. Chen, z.v. (19871, 'Problems of regional salls', Proc. 8th Asian Regional
Conference on Soil Mechanics and Foundation Engineering, Kyoto, Japan, Vol.
2. Clevenger, W.A. (1958), 'Experiences wltn loess as foundatIon materiu]",
Transaction of ASCE, No. 2916, pp. 151·169.
3. Dudley, J.H. (1970), ~Review of collapsing soils", Journal of A5CE, SM3, pp.
200 925-947
4. IShihara, K., Yasuda, K. and Kato, M. (199n, "Characteristics of lateral Spreading
In Liquefied Deposits during the 1995 Hanshln-AwaJi Earthquake", Journal Of
Earthquake Engineering, Vol. 1, NO.1, pp. 23-55.
5. Ishihara, K. (1993), "Liquefaction and Flow Failure during Earthquakes,
Geotechnique, Vol. 23, pp. 315-415.
6. Ishihara, K., Okusa, S. Oyagl, N. and A. Ischuk (1990), 'liquefaction-Induced
FloW Slide In the Collapsible loess Deposit in Soviet TaJik', Soils and Foundations,
Vol. 30, NO.4, pp. 73'89.
7. IShihara, K, and Koseki, J. (1989), "CYCliC shear strength of fines·containing
sands-, Proc. Earthquake Geotechnical Engineering, 12th International Conference
on Soli Mechanics and Foundation Engineering. Rio de Janeiro. pp. 101-106.
8- Ishihara, K. (J986), 'Analysis of landslides during the 1984 Naganoken-seibu
Earthquake', Proc. of International Symposium on Engineerlng Geology Problems
in Seismic Areas, ttalian Group of IAEG, 8arl, Italy, vol. 2, pp. 175-215.
9. Ishihara, K. (1985), "Stability of Natural Deposits DurIng Earthquakes', Proe.
of 11 th International Conference on Soil Mechanics ancl Foundation Engineering.
san Franciscc, Vol. 2, pp. 321"376.
10. Inagakl, H., lai, S., Sugano, T., Yamazaki. H. and Inatomi, T. (1996), 'Perfonnance
of caisson type quay walls at Kobe Port', Soils and Foundations, Special Issue,
pp.119-136.
11. Janbu, N. (1955), "Application of Composite Slip surface for Stability Analysis',
Proc. European Conference on Stability of Earth Slopes, Stockholm, (31, PP
43-39,
12. Masaki, K., Tilniguchl, H., iida, K., Muramafsu, I, Taga, N. and Tagashl. Y. (1985),
'Intensity Distribution Based on Seismic Records of Ground' Motions, Damage
10 Houses and Ground, Overturns of Gravestones, and Questionnaire Survey',
Synthetic Investigation Of Earthquake and Damoge of the Nogonoken-seibu
Earthquake 0/1984, pp. 83-88. (in Japanese)
13. Sakai, J. (1985), "landslide Associated with the Westem Nagano Prefecture
Earthquake, 1984", SynthetiC Investigation Of Earthquake and Damage Of Ihe
Noganoken-seibu Earthquoke Of 1984, pp. 165-178. (in Japanese)
2nd Lumb Lecture - Introduction of the Speaker
Mrs lumb, Distinguished Guests, Ladies and Gentlemen,
First of all, I would like to express my heartfelt appreciation of the Lumb
lecture Committee under the chairmanship or Prof C F lee, for having
organised this distinguished event In memory of my most respected teacher
Professor Peter lumb.
Tonight it is a great pleasure to introduce our distinguished speaker of
the 2nd Lumb Lecture, Prof Kenji Ishihara.
Prof Ishihara is currenlly a Professor In the Civil Engineering Department
of Chua University in Tokyo and the Science University of Tokyo. He is
also the Professor Emeritus of the University of TOkyo. Prof Ishihara received
his education in Civil Engineering at the University of Tokyo, obtaining
his BS"degree in 1957, MS'degree in 1959, and PhD in 1963. During
1966 to 1967, he was a Visiting Research Associate at the University
of Illinois in Urbana USA. He has been affiliated with the University of
Tokyo since then, taking the post of professorship in geotechnical
engineering since 1977.
Prof Ishihara served for ISSMFE as secretary of Japanese National
Committee for the period of 7 years between 1970 and 1976 and as
Vice-President of Asian region of ISSMFE during the period of 1989-
1993. Between 1997-200'1 he was the President of the International
SOCiety.
Prof Isllihara's major research interest covers problems in 5011 dynamics
associated with earthquakes, such as liquefaction of sandy deposits, and
seismic stability of slopes and earth structures. He has written about 250
papers on the above subjects.
He has served on various occasions as consultant or adviser to UNESCO
projects in many countries including India, Iran, Yugoslavia, Mexico, Chile
and Ecuador. He is the author of a book 'Fundamentals of Soil Dynamic'
(1974) and textbook 'Soil Mechanics' (1988) both in Japanese, He recently
published in Oxford Press an English book entitled 'Soil Behaviour in
Earthquake Geotechnics',
Prof Ishihara has received many honours by being assigned on many
occasions to deliver lectures worldwide including the theme lecture in
11th ICSMFE in San Francisco and the 33rd Rankine Lecture of the British
Geotechnical Society in 1993. He acted as chairman of the Technical
Committee TC4 on Earthquake Geotechnical Engineering from 1985-1993.
He has also received honour by being awarded H B Seed Gold Medal
in 1998 from American Society of Civil Engineers. For his significant
contribution, title of Honorary Doctorate was conferred to him from
Technical University of Bucharest, Romania in 1995 and from Istanbul
Technical University, Turkey in 1999.
This evening, we are greatly honoured to have Prof Ishihara to present
the Second lumb Lecture on 'Failure of Slopes and Soil Property
Charaderlsation', In his lecture, some examples of landslides experienced
in recent years will be introduced with emphasis on evaluation of peak
and residual strengths of soils involved in landslides. This is very relevant
to HK slope problems. Without further ado, I would like to Invite you to
join me to welcome Prof Ishihara on stage to present his lecture. Thank
you.
Ir Raymond K S CHAN
2nd Lumb Lecture - Vote of Thanks
Mrs lumb, distinguished guests, ladies and gentlemen,
Prof Lumb was a fundamental thinker, he was concerned to get to the
root of an issue and that It should be properly understood. He was also
TRANSACTIONS' Volume 9 Number 3
a teacher and well respected by his students, many of whom <Ire in senior
positions in geotechnical engineering today.
Prof Ishihara is a very fitting speaker this evening. He too is a fundamental
thinker and his work is held in high esteem internationally.
The second Lumb lecture has concerned failures of slopes, and at what
a scale, 35 million cubic metres, three orders of magnitUde greater than
we have seen in Hong Kong, and a travel distance of 12 kms, about 10
times the travel distance of the Tsing Shan Flow slide.
We, in Hong Kong, spend a lot of time and effort in upgrading slopes
to prevent disasters. It is Important that unless we know how slopes can
fail we are missing the point.
Prof Ishihara has gIven us three cases histories and has highlighted 'lessons
learned'. They are there to stimulate our fundamental thinking. The
challenge is, when there are some 54,000 man made feature to inspect,
are we still thinking fundamentally?
Prof Ishihara, Peter lumb always gave us things to think about, you have
given us a very special lecture, we would all like to thank you.
Ir Dr John ENDICOTT
Kenji ISHIHARA
Prof Ken)1 Ishihara is currently a Professor in the
Department of Civll Engineering. Chua University, Tokyo
and the Science University of Tokyo. He Is also the
former President of the International Society for Soil
MechanIcs and Geotechnical Engineering (ISSMGE). His
major research Interest covers problems In soil dynamics
associated with earthquakes. He has been invited to
deliver many prestigious lectures worldwide, InCluding
the theme lecture at the 11 th ICSMFE in San Francisco
and the 33rd Rankine Lecture of the British Geotechnical
Society in 1993.