1097

EFFECTS OF EARTHQUAKE-INDUCED SETTLEMENT OF CLAY LAYER ON

THE GROUND SUBSIDENCE

Hiroshi MATSUDA1, Keiji SAKURADANI2 And Naoya EMOTO3

SUMMARY

At the Hyogo-ken Nanbu Earthquake on January 17, 1995, on the artificially reclaimed island, a
very large bulk of mud was ejected and the ground subsidence of 30-50cm or more was observed.
In this paper, to clarify the effect of earthquake-induced settlement of clay layer on the ground
subsidence, firstly the settlement-time relations for the separate layers in the ground which
measurement was started from one year before the earthquake are shown and the effects of the
earthquake on the settlement of clay layer are investigated. Secondly, the earthquake-induced
settlement of the alluvial clay layer in the Port Island was predicted. Thirdly, multi-directional
simple shear tests were carried out and the direction of cyclic shear on the settlement of clay layer
was observed. In conclusion, following results were obtained. 1) The earthquake-induced
settlement of alluvial clay layer of 16m thick reaches over 4cm. 2) When a clay is subjected to the
multi-directional shear strain, the larger the phase difference in the sinusoidal shear strains applied
from two directions, the smaller the effective stress and therefore the settlement in the
reconsolidation stage is affected by the phase difference.

INTRODUCTION

At the Hyogo-ken Nanbu Earthquake, since the large buildings constructed on the artificially reclaimed island
was supported by deep foundations, no serious damage to the buildings themselves appeared due to the ground
subsidence. However, the differential settlement between the buildings and the surrounding ground affected the
lifelines. Usually, it is considered that the earthquake-induced ground settlement is mainly caused by the
liquefaction of sandy soil layer [Matsuda, 1996, Lee, 1974, Pyke, 1975, Seed, 1972]. As for the settlement of
clay layer induced by the earthquake, it has been shown that the settlement reaches up to 5% in strain and even
for the overconsolidated clay the post-earthquake settlement is considerably large compared with the secondary
compression. [Ohara and Matsuda, 1988] The fact of the earthquake-induced settlement of clay layer has been
confirmed in the 1978 Miyagiken-oki Earthquake [Suzuki, 1984], in the 1957 Mexico Earthquake [Zeevaert,
1972], in the 1985 Mexico Earthquake [Jaime et al., 1987] and in the 1995 Hyogo-ken Nanbu earthquake
[Matsuda. 1997]. Especially, after the Hyogo-ken Nanbu Earthquake, the ground surface settled gradually with
time. This means that the settlement of clay layer increased by the earthquake, because the settlement induced by
the liquefaction of sandy soil takes place immediately after the earthquake. Therefore, the objective of this paper
is to clarify the effect of earthquake-induced settlement of clay layer on the ground subsidence in more detail.

1
Department of Civil Engineering, Yamaguchi University, Tokiwadai, Ube, Yamaguchi, Japan Email: hmatsuda@po.cc
2
Department of Civil Engineering, Yamaguchi University, Tokiwadai, Ube, Yamaguchi, Japan Email: yamaguchi-u.ac.jp
3
Maeda Co. Ltd., 4 Watanabe-Dori, Fukuoka, Japan Fax: +81-92-735-0064
 20
P1 1 2 1 2 6 8 6 10 1
0 3 4 5 7 9 Alluvial clay (Ma13)

10 -20

Elevation (O.P.m)
(m)
Pleistocene gravel
-40 layer (Dg1)

Elevation
-60 Upper Pleistocene
clay layer (Ma12)
-80
3 4 6
5 Dg2
P2 -100

7 Ma11
8 -120
0 500m
-140

-160 Ma10
0 2000m
Horizontal scale
Figure 1: Port Island. Horizontal scale

Figure 2: Soil profiles in Port Island.

POST-EARTHQUAKE SETTLEMENT-TIME RELATIONS FOR ALLUVIAL CLAY LAYER

Port Island is a reclaimed island, which is very near to the city center of Kobe, Japan. The soil profiles at the
different locations in Port Island are shown in Figure 2, in which the bore hole numbers are the same as those in
Figure 1. The depth of the original seabed at the location of Port Island was 10-13m. The alluvial clay layer of
10-20m thick (Ma13) lies beneath a reclaimed stratum and underlain by laminated gravel sand layers (Dg1)
which include thin silt and clay layers. The upper part of this laminated layer permits drainage from the alluvial
clay and also supports bearing piles. The upper Pleistocene clay layer of 20-30m thick (Ma12) lies beneath the
laminated layer. In Figure 2 soil profiles for Nos. 1, 10, 9, 6 and 8 show that toward the south, the surface of the
alluvial clay layer (Ma13) slopes and its thickness increases. Soil profiles Nos. 3, 4, 5, 6 and 7 show that the
thickness of the alluvial clay layer toward the east is almost the same, but there is a tendency that the thickness
of the upper Pleistocene clay layer (Ma12) increases. [Matsuda et al., 1996]

In Figure 3, the changes in the liquid limit wL, plastic limit wP, natural water content wn, void ratio e, unconfined
strength qu, consolidation yield stress Pc and compression index Cc are shown. The samples were obtained
around the center of the north part of the Port Island between 1975 and 1977 [Tanimoto, 1989], when the first
stage of the reclamation work had been finished. Now, the second stage of reclamation work is under
construction in the south part of Port Island.

w(%) (kN/m3) qu(kPa)

Void ratio ƒ tÁ Pc(kPa) Cc
0 20 40 60 80 1.2 1.4 1.6 1.8 2.0 15 16 17 18 5 10 15 100 200 0 1.0
-10

-15
Elevation (m)

Alluvial clay

-20
layer

-25

-30
wP wL
wn

Figure 3: Characteristics of alluvial clay on Port Island. [Tanimoto, 1989]

2 1097
 KH-1 Sub-hole(1) Sub-hole(2)

El
KP-15.8
-1.0
10.0 10

KP-32.6 Ma13 KP-33.3m

K-R, KH-1,
Subholes
5.0
-0.5 5

KP-60.2 Bottom of Ma9

(KH-1)
KP-82.9 Ma12 KP-87.10m 0.0 0

0.5 -5

Settlement(m)
Bottom of Ma13
Sub-hole(1)
1.0 Bottom of Ma12 -10

Sub-hole(2)
Settlement plates
1.5 Ground Surface
-15

KP-207.3 at+2.5m (K-R)

2.0 -20

Ma9 Earthquake
KP-235.2 17/1/1995
KP-247.70m 2.5 -25

600 800 1000 1200

Elapsed time (days)
Figure 4: Positions of
settlement plate. Figure 5: Settlement-time relations for the different depths in
Port Island.
At the point P2 in Figure 1, which is located in the second reclamation work site in Port Island, the settlements
were measured at four different depths as shown in Figure 4; Plate(1) was placed at the ground surface
(k.P.+2.5m), Plate(2) was placed at just under the alluvial clay layer Ma13 (K.P.-33.3m), Plate(3) was placed at
just under the upper Pleistocene clay layer Ma12 (K.P.-87.1m), Plate(4) was placed at just under the lower
Pleistocene clay layer Ma9 (K.P.-247.7m). The measurement of settlement was started on July 3rd 1992. In
Figure 5, the settlement-time relations for each settlement plate are shown. At the ground surface, the significant
settlement occurs by the earthquake. This is mainly due to the recompression of liquefied soil as mentioned
above. Figure 6 shows the settlement-time curves, in which the fixed line is the result by the Plate(1) at the
ground surface. It is seen that by the earthquake, the settlement of about 26 cm occurred. Curve I in Figure 6 is
the predicted result by considering that the earthquake didn’t occur and Curve II was obtained by moving Curve
I down to pass through the first data point obtained after the earthquake. When comparing the Curve II and the
observed curve, after the earthquake, the observed settlement is larger than the Curve II. This means that the
earthquake affects the settlement-time relations of clayey layer, because the settlement induced by the
liquefaction of sandy soil must occur in a very short period.

The earthquake-induced settlement of clay layer with different plasticity indices has been predicted based on the
simple shear test results [Matsuda and Nagira, 1998]. Figure 7 shows the relationships between the shear strain
amplitude and volumetric strain for the number of strain cycles of 200. The samples used in the test are Kaoline,
Yanai clay, Ariake clay, Onoda clay. The physical properties of these samples are shown in Table 1.

50
Curve II: Curve I was moved down to
pass through the first point measured
after the earthquake.
Settlement(cm)
(cm)

100
Curve I: Predicted by
Settlement

a hyperbolic curve.

150

Earthquake
17/1/1995 Ground Surface:
Observed curve.
200
600 800 1000 1200 1400
Elapsed time (days)

Figure 6: Settlement-time relations at the Port Island, Kobe. [Matsuda, 1997]

3 1097
 8
Table 1: Physical properties of clay Fitted line
Yanai clay
samples used in a series of
tests. 6
Onoda clay

Settlement in strain (%)

Yanai Ariake Onoda
Kaoline
clay clay clay

ƒÃv (%)
Ariake clay Ariake clay
Gs 2.718 2.661 2.629 2.628 4 Kaoline
Yanai clay
wL Kaoline
47.4 77.4 67.6 81.1 Onoda clay
(%)
n=200
wp 27.2 2 P.I.Pleistocene
31.0 28.9 17.0 P.I. Pleistocene
(%) layer (n=10)
layer(n=10)
Ip 16.4 48.4 50.6 52.9 P.I.Alluvial
P.I. Alluvial
layer (n=10)
layer(n=10)
0
0.01 0.1 1 10
ƒÁdyn (%)
Shear strain amplitude (%)

Figure 7: Volumetric shear strain induced by cyclic

shear strain.
Table 2: Settlement of alluvial clay layer in Port Island
induced by the 1995 Hyogo-ken Nanbu earthquake.

Component Thickness Settlement in Settlement

of (m) strain (%) (cm)
Earthquake
EW 0.21 1.67
NS 0.52 4.16

Since the details of the method to predict the earthquake-induced settlement have been shown [Matsuda et al.,
1998], for the alluvial clay layer in Port Island it is possible to calculate the settlement induced by the 1995
Hyogo-ken Nanbu Earthquake. In Table 2, the earthquake-induced settlement of alluvial clay layer for the EW
and the NS components of the earthquake motion are shown. The settlement of the alluvial clay layer is 4.2 cm
for the NS component and 1.7 cm for the EW component. Since the settlement of the ground surface which
occurred in a very short period was 26.8 cm as shown in Figure 1, the additional settlement of about 4 cm must
occur after the earthquake. The difference in the predicted settlement for the EW component and NS component
of the earthquake motion is very large. Therefore, the effects of direction of the earthquake motion on the post-
earthquake settlement are discussed in the following paragraph.

EFFECTS OF DIRECTION OF EARTHQUAKE MOTION ON THE SETTLEMENT OF CLAY

LAYER

The Hyogo-ken Nanbu earthquake of January 17, 1995 has been assigned a JMA magnitude of 7.2 by the Japan
Meteorological Agency. At the site designated as P1 in Figure 1, the seismographs were set up at four different
depths as shown in figure 8[GEORK,1995]. The seismoscope traces recorded at the depth of 16m are shown in
Figure 9. It is seen that the direction of the maximum acceleration is toward SE.

As mentioned above, the method to predict the post-earthquake settlement of clay layer has already been shown
[Matsuda and Ohara, 1990, Matsuda, 1997]. But the effect of direction of earthquake motion was not taken into
account in the prediction method. As for the earthquake-induced settlement of sandy layer, the effects of the
multi-directional loading have been investigated. [Pyke et al., 1975, Nagase and Ishihara, 1988] So, in this study
a multi-directional simple shear test apparatus as shown in Figure 10 was newly developed and a clay specimen
was subjected to the sinusoidal shear strain from two directions, which are perpendicularly intersecting at the
center of the specimen. In a series of tests, the shear strain amplitude and the phase angle of cyclic shear strain

4 1097
 0.0
Seismograph
Seismograph GWL GWL
No.4 G.L. No.4 G.L. Earth Fill
600
E(sand,
arth fill
gravel)

NS
10.0
Seismograph
Seismograph
(sand, gravel)
No.3: 16.0-16.8m
No.3 16.0-16.8m 500 PortPORT
Island
ISLANDGL-16m
GL-16m
20.0 Alluvial clay layer
Alluvial clay layer

Acceleration (Gal)
(Ma13)
(Ma13) 400

Seismograph
Seismograph No.2
30.0
Sandsand Pleistocene Layer
300
No.2: 32.0-32.8m
32.0-32.8m

40.0 Gravel
gravel Pleistocene
200
Sand sand layer
Depth (m)

50.0 100
Sandsand
60.0 0
Pleistocene
Pleistocene clay

70.0
clay layer
layer (Ma12) -100
(Ma12)
-200
80.0
Seismograph
Seismograph
Pleistocene layer
No.1 83.0-83.8m
No.1 83.0-83.8m Pleistocene
(gravelly sand) -300
90.0 layer
(gravelly sand) -400
100.0
-400 -300 -200 -100 0 100 200 300 400 500 600
Acceleration (Gal) EW

Figure 9: Seismoscope traces recorded during 1995

Figure 8: In situ condition where Hyogo-ken Nanbu Earthquake.
the seismographs were set
for each direction were changed and the effects of the phase difference of cyclic shear strain on the excess pore
water pressure and on the settlement were investigated.

Sample and Test Method

After the end of consolidation of a saturated specimen (60 mm in diameter, 200mm in height) by the
consolidation stress of σ vo =49 kPa, the cyclic shear strain which faces toward X-direction and Y-direction as
shown in Figure 11 was applied under the constant volume condition. The wave form of cyclic shear strain is
sinusoidal. In a series of tests, the phase difference θ in the cyclic shear strain of X-direction and that of Y-

Drainage Bellofram cylinder

e
Specimen Air pressure regulator
(75mm ,25mm h)
Ó Por pressure transducer
t
Fastener
V ti l d
Linear bearing
Membrane
Acrylic rings
Load cells
Y direction
LVDT
1518
X direction

Linear bearing
Specimen

Water tank
750

Figure 11: Direction of cyclic shear

strain.
815
Unit:mm

Figure 10: Multi-directional simple shear test

apparatus.

direction was changed as θ =0°, 20°, 45° and 90°. The sample used in this study was Kaoline as shown in Table

5 1097
 -1.0
X direction n=40 ¨Y direction n=60
ƒÁX=ƒÁY=0.78%

-0.8

ó v’ /ó vo
ă¢ƒÐv'/ƒÐvo'
-0.6

-0.4 X direction n=30 ¨Y direction n=70

ƒÁX=ƒÁY==0.72%

-0.2
ƒÐvo'=49kPa
Kaoline
0.0
0 20 40 60 80 100
Number of cycles n

Figure 12: Effect of direction of cyclic shear strain on the reduction in effective stress.

1. After the cyclic shear test, the specimen was reconsolidated under the vertical stress of σ vo =49 kPa. The
period of cyclic shear was fixed as 2.0s and the shear strain amplitude in the X-direction γ X and Y-direction γ Y
were changed in the range from 0.05 to 2.0%.

Decrease in Effective Stress Induced by Cyclic Shear

Before starting the series of tests, a special test was carried out, in which after the specimen was subjected to the
cyclic shear toward X-direction for 30 cycles, the additional cyclic shear toward Y-direction was applied for 70
cycles with the same shear strain amplitude as X-direction. During cyclic shear test, the volume of specimen was
kept constant and the change in the vertical stress ∆σ v was measured. In Figure 12, relationships between
∆σ v / σ vo and the number of cycles are shown. In the same figure, the results for the other number of cycles as
40+60 are also shown. When the direction of cyclic shear was changed, it is seen that the value of ∆σ v /σ vo
decreases. This means that the direction of cyclic shear have some effects on the excess pore water pressure
accumulated during the earthquake.

Typical records of shear strain-time relations for the multi-directional cyclic shear test for θ =90° are shown in
Figure 13. In Figure 14, traces of the cyclic shear strain for θ =0°, 20°, 45° and 90° at the 50th cycle are shown.
In the case of θ =0°, the trace forms the linear line and therefore this case corresponds to the simple shear test. In
the case of θ =90° the trace forms the circle line and this case corresponds to the gyratory shear test which was
called by Pyke et al., [1975].

In this study the shear strain amplitude and the phase difference during the cyclic shear test were changed as
mentioned above. The relations between ∆σ v / σ vo and the major shear strain amplitude γ M , for the number of
cycles n=50 are shown in Figure 15, where γ M is defined as the half of the major axis on the ellipse which is
traced by the shear strains in the X and Y directions. For example, when the cyclic shear strain amplitude toward
X-direction and that toward Y-direction are the same as γ X = γ Y , γ M is obtained as γ M = 2 ⋅ γ X = 2 ⋅ γ Y .

0.6
1.0 è=90
Shear strain in Y-direction (%)

X-direction Y-direction 0.4

è=20°
è=45°

0.5 0.2
Shear strain (%)

0.0
-0.2 è=0°

-0.4 n=50
-0.5
Kaoline
σv’=49kP
-0.6
Ð
ƒ vo'=49kPa Kaoline, K105 -0.6 -0.4 -0.2 0 0.2 0.4 0.6
-1.0
Shear strain in X-direction (%)
82 84 86 88 90
Time (s)

Figure 13: Typical records of shear strain-time relations Figure 14: Traces of the cyclic shear
in the multi-directional simple shear test. 6 strain for θ =0°, 20°, 45° and 1097
 -1.0

(%)
-0.8
ă¢ƒÐv'/ƒ¢ƒÐvo'
ƒÆ=20 ‹ ƒÆ=0 ‹ X or Y ƒÆ=90 ‹
X or Y direction ƒÆ=0 ‹

v
ó v’ /ó vo

X or Y

Volumetric strain å
-0.6 ƒÆ=45 ‹
ƒÆ=45 ‹ ƒÆ=0 ‹
ƒÆ=0 ‹ 4 ƒÆ=20 ‹

ƒÃv (%)
ƒÆ=20 ‹
-0.4 ƒÆ=90 ‹
ƒÆ=20 ‹ X or Y
ƒÆ=45 ‹ direction
ƒÆ=45 ‹ 2
-0.2
ƒÆ=90 ‹
Kaoline ƒÆ=90 ‹
n=100 Kaoline
0.0 ƒÐvo'=49kPa
0.01 0.1 1 10
0
ƒÁM(%)
Major shear strain amplitude ã M (%) 0.01 0.1 1 10
ƒÁdyn (%)
Major shear strain amplitude ã M

Figure 15: Relationships between shear strain Figure 16: Settlement induced by multi-directional
amplitude and the change in vertical cyclic shear strain.
effective stress.

Plots in Figure 15 are the observed results and the fixed lines are obtained by the following equation [Ohara and
Matsuda, 1988, Matsuda and Nagira, 1998].

−n
∆σ v' / σ vo = (1)
A ⋅γ M
m
+ n ⋅ γ M /( B + C ⋅ γ M )

where A, B, C and m are the experimental constants. In Figure 15, a close agreement between the observed plots
and the calculated lines was obtained. Although Equation (1) was derived for the simple shear test condition, it is
possible to apply to the multi-directional simple shear test results. Furthermore, it is seen that the larger the phase
difference, the larger the value ∆σ v /σ vo .

Settlement induced by the multi-directional cyclic simple shear

When the excess pore water pressure accumulated during the cyclic shear dissipates, the settlement occurs. In
Figure 16, relationships between the major shear strain amplitude γ M and the volumetric strain ε v for different
phases are shown. Data plots in the figure are observed results and curves were obtained by the following
equation.

Cdyn 1
εv = "og (2)
1 + eo 1 + ∆σ v' / σ vo

where eo is the initial void ratio and Cdyn is the dynamic compression index. In Figure 16, when the phase
difference increases, larger settlement occurs. For the post-earthquake settlement of sand layer, two opposite
results have been reported, the one is that the maximum shear strain induced during the irregular loading is
uniquely correlated with the volumetric strain during the reconsolidation [Nagase et al., 1988] and the other is
that the multi-directional shaking can apparently increase settlements to as much as three times the settlements
caused by shaking in one direction only [Pyke et al., 1975].

The results obtained in this study show that when a clay is subjected to the multi-directional shear strain, the
settlement in the reconsolidation stage is affected by the phase difference. Therefore the settlement of the alluvial
clay layer in Port Island as shown in Table 2 which was calculated for the one component of the earthquake
motion should be corrected. Then the larger result for the post-earthquake settlement would be obtained.

CONCLUSIONS

To clarify the effect of earthquake-induced settlement of clay layer on the ground subsidence, the settlement-
time relations for the separate layers in the ground were observed and by the multi-directional simple shear tests
the direction of cyclic shear on the settlement of clay was investigated. In conclusion, following results were
obtained.

7 1097
1) For the settlement of the alluvial clay layer in the Port Island induced by the Hyogo-ken Nanbu Earthquake,
the settlement over 4cm was predicted.
2) The effects of the direction of cyclic shear on the post-earthquake settlement are not negligible.
3) The larger the phase difference in the sinusoidal shear strains from two directions, the smaller the effective
stress.
4) When a clay is subjected to the multi-directional shear strain, the settlement in the reconsolidation stage is
also affected by the phase difference.

ACKNOWLEDGEMENT

The settlement-time relations at Port Island reported in this study were offered by the Kobe Municipal Office and
the accelerograms at the Port Island were supplied by the Committee of Earthquake Observation and Research in
the Kansai Area. The authors would like to express their gratitude for their help to this study.

REFERENCES

Jaime, A. P, Romo, M. P. and Jasso, M. R. (1987), ”Seismic Induced Settlement in a Building”, 8th Pan-
American Congress on Soil Mechanics and Foundation Engineering , pp 257-275.
Lee, K. L. and Albaisa, A. (1974), “Earthquake induced settlements in saturated sands”, Journal of the
Geotechnical Engineering Division, ASCE, 100, GT4, pp387-406.
Matsui, T. (1995), “The seabed ground, Kaiteijiban”, Japanese Geotechnical Society, Kansai Branch. (in
Japanese)
Matsuda, H. and O-hara, S. (1990), “Geotechnical Aspects of Earthquake-Induced Settlement of Clay Layer”,
Marine Geotechnology, 9, pp179-206.
Matsuda, H. and Shimizu, Y. (1996), “Seismic ground subsidence in near-shore reclaimed land”, The Kobe
Earthquake Geodynamical Aspects, Computational Mechanics Publications, Southampton, UK, pp73-98.
Matsuda, H. (1997), “Estimation of Post-Earthquake Settlement-Time Relations of Clay Layers”, Journal of
Geotechnical Engineering, 568/III, pp41-48. (in japanese)
Matsuda, H. and Nagira, H. (1998), “Estimation of the post-earthquake subsidence in clay layers”, 6th U.S.
National Conference on Earthquake Engineering, 134, pp1-11.
Matsuda, H. and Emoto, N. (1998), “Effect of direction of earthquake motion on settlement of clay layer”, The
10th Earthquake Engineering Symposium, Proceedings, 2, pp1427-1432. (in Japanese)
Nagase, H. & Ishihara, K. (1988), “ Liquefaction-induced compaction and settlement of sand during
earthquakes”, Journal of the Japanese Society of Soil Mechanics and Foundation Engineering, Soils and
Foundations, 28 1, pp65-76.
O-hara, S. and Matsuda, H. (1988), “Study on the Settlement of Saturated Clay Layer Induced by Cyclic Shear”,
Soils and Foundations 28, 3, pp103-113.
Pyke, R., Seed, H. B. and Chan, C. K. (1975) , “Settlement of sands under multidirectional shaking”, Journal of
the Geotechnical Engineering Division, ASCE, 101, GT4, pp379-398.
Seed, H. B. and Silver, M. L. (1972), “ Settlement of dry sands during earthquakes”, Journal of the Geotechnical
Engineering Division, ASCE, 98, SM4, pp381-397.
Suzuki, T. (1984), “Settlement of Saturated Clays under Dynamic Stress History”, Journal of the Japan Society
of Engineering Geology 25, 3, pp21-31. (in Japanese)
Tanimoto, K. and Tamura, K. (1989), “Settlement characteristics of coastal reclaimed land”, The construction
Engineering Research Institute Foundation, 31, pp229-243. (in Japanese)
The Committee of Earthquake Observation and Research in the Kansai Area (GEORKA), (1995), The
Distributed Seismic Records and Data. (in Japanese)
Zeevaert, L. (1972), Foundation Engineering for Difficult Subsoil Conditions, New York, Van Nostrand Reinhold
Company.

8 1097