Engineering Geology 266 (2020) 105461

Contents lists available at ScienceDirect

Engineering Geology
journal homepage: www.elsevier.com/locate/enggeo

Stability evaluation of slope subjected to seismic effect combined with T

consequent rainfall
⁎
You-Liang Chena, Geng-Yun Liua, , Ning Lia, Xi Dua, Su-Ran Wanga, Rafig Azzamb
a
Department of Civil Engineering, University of Shanghai for Science and Technology, 516 Jungong Rd, Shanghai 200093, PR China
b
Department of Engineering Geology and Hydrogeology, RWTH Aachen University, Lochnerstr.4-20 Haus A, D-52064 Aachen, Germany

A R T I C LE I N FO A B S T R A C T

Keywords: A large number of field investigations of earthquake damage have shown that sometimes slopes instability occurs
Slope stability as a result of a series of post-seismic ripple effects rather than of merely seismic load or ground motion, among
Seismic response which a combination of earthquakes and post-seismic rainfall is a typical contribution for slope instability. In this
Unsaturated infiltration paper, taking a slope in Southwest China as an example, failure mechanism and stability of slope under seismic
Plant reinforcement technology
load and heavy rainfall are analyzed with consideration of dynamic load and unsaturated infiltration using
GeoStudio software. Results indicate that, the acceleration of earthquake is sharply amplified at the top of the
slope, resulting in instantaneous tension stress, which causes tension cracks and therefore reduction in tensile
strength at the top of the slope. The existence of cracks contributes to the post-earthquake infiltration of rain-
water slows down the dissipation of excess pore water pressure. Slope instability is ultimately caused by per-
manent damage due to seismic load or ground motion and secondary damage on account for heavy rainfall.
Furthermore, the Plant Reinforcement Technology is put forward to reinforce the slope and satisfactory results
are obtained. Therefore, this research provides a theoretical basis to analyze and control slope stability in other
similar engineering geology problems.

1. Introduction rainfall have been reported in some cases. After the 1999 Taiwan
Earthquake, the area of landslides induced by heavy rains can reach
Earthquake triggered landslides have been studied in the seismic three times that of landslides directly induced by earthquakes (Lin
regions as a major type of geological disasters (Sepúlveda et al., 2005; et al., 2006). After the 2008 Wenchuan Earthquake, Beichuan area with
Chang et al., 2012; Jibson and Harp, 2016; Marc et al., 2017; Salinas- high seismic high intensity suffered the largest heavy rainfall with the
Jasso et al., 2019). However, a large number of field investigations of precipitation amount up to 250–350 mm, which led to further activity
earthquake damage show that slopes instability is a result of a series of of landslides and a large number of new landslides (Yin et al., 2009;
ripple effects caused by earthquakes rather than of merely seismic load Huang and Fan, 2013; Tang et al., 2009; Jun et al., 2018). Rainfall also
or ground motion (Lin et al., 2006; Chen and Hawkins, 2009; Liu et al., triggered many small-scale landslides from May to October after the
2013; Yano et al., 2019). Especially, seismic activity with consequent Kobe earthquake on January 17, 1995 in Japan (Kanaori and
rainfall pose a challenging threat to slope stability. Kawakami, 1996).
The crustal stress, stress field or failure at a fault that triggers Theoretical analysis (Bishop, 1955; Pantelidis and Griffiths, 2013a;
earthquakes will affect precipitation in the short term after an earth- Pantelidis and Griffiths, 2013b) have been performed to study the dy-
quake. A related effect of massive landslides after earthquakes will in- namic response and failure characteristics of slopes. Although the the-
crease a large number of dust and particles in the air, which are the best ories to assess slope stability is becoming increasingly mature, the
condensation nodules of water droplets; Furthermore, the seismic shock mechanism of slope instability under complex external forces remains
wave will continuously release energy into the air, causing air vibra- confusing. Shaking table model tests were conducted to investigate the
tions (shock waves) as well as a large number of condensation nodules failure modes as well as causes of landslides, and discuss the stage
and nodules over the earthquake area. Water vapor molecules collide failure characteristics of slopes under shaking table conditions (Hong
constantly, combine fully, which may eventually cause heavy rainfall et al., 2005; Lin and Wang, 2006; Hu and Chen, 2011). Aided with
(Huang and Fan, 2013). Slope instability due to earthquake-induced numerical simulation, a large number of practical slope instability cases

⁎
Corresponding author.
E-mail address: liugengyunsdau@sina.com (G.-Y. Liu).

https://doi.org/10.1016/j.enggeo.2019.105461
Received 1 April 2019; Received in revised form 15 November 2019; Accepted 18 December 2019
Available online 20 December 2019
0013-7952/ © 2019 Published by Elsevier B.V.
Y.-L. Chen, et al. Engineering Geology 266 (2020) 105461

have been studied (Iwata et al., 2013; Huang and Xiong, 2017) whereas Elevation / m Geological fault
there is usually no information about the actual pore pressure devel- 520
colluvium & talus accumulation
opment, especially the dynamic pore water pressure due to the earth- Severe weathering strata
quake, which largely limits the reliability of the calculation. 460
Mild weathering strata
Heavy rainfall is another main trigger of slope instability (Tang 1# (C#)
Bedrock
et al., 2009; Kimoto et al., 2013; Sasahara and Sakai, 2014). The in- 400
2#
filtration of pore water under rainfall elevates the free surface of the
3#
groundwater in the slope, weakens the rock and soil at the potential slip D# A#
340
surface, and thus reduces the stability of the slope, leading to the for- 4#
mation of the landslide. Actually, it is crucial to the slope stability that
the infiltration of rainfall in rock and soil mass of slope experienced a
280 5# (E#)

seepage unsaturated-to-saturated process, during which the physical B#

and mechanical properties of rock and soil in unsaturated areas change 220
continuously. 6#
In recent years, research on slope instability caused by rainfall have 160
been carried out (Okura et al., 2002; Moriwaki et al., 2004; Kimoto 0 75 150 225 300 375 450 525 565
Distance / m
et al., 2013). Typically, Fredlund et al. (1978) first took the influence of
unsaturated soil seepage into account in slope stability analysis. Alonso Fig. 1. Cross section of the slope used for modeling with location of control
et al. (1995) calculated and analyzed the stability of the slope with points.
consideration of the soil type, soil-water characteristic curve, rainfall
duration and infiltration intensity of the slope. Okura et al. (2002) and
rainfall in the area, and consequently the slope was saturated, which
Moriwaki et al. (2004) have successively carried out full-scale tests on
eventually caused instability and sliding. The landslide was formed in
mobile landslides and the results show that the increase of pore water
the early Quaternary collapsing or landslide deposits, mainly composed
pressure is a result from collapse of loose soil structure in the upper
of debris-rich clay that is loose and low permeable. The main scarp dips
slope while that is a result of a mix of soil compression and shearing in
to the southeast and develops crown cracks with length of tens to
the lower slope. Kimoto et al. (2013) concluded that rainfall infiltration
hundreds of meters in fault fracture zones. There are often several
can reduce matrix suction and increase permeability coefficient of un-
secondary scarps in the zones, forming complex rupture structures. The
saturated soil by numerical simulation.
geological profile of the slope and the arrangement of monitoring points
A large number of geology engineering practices show that the slope
are shown in Fig. 1.
instability is multifactorial among which seismic activity and heavy
rainfall are two major incentives, but their failure mechanisms are
completely different (Clague and Stead, 2012; Sorbino and Nicotera, 2.2. Hydrology
2013). Thus, the study on the mechanism of slope instability under
seismic load with subsequent rainfall infiltration and the corresponding Groundwater is Quaternary loose-rocks pore water, which occurs in
treatment or remediation measures urgently needs to be carried out sandstone and glutenite. It is seasonal temporary diving and mainly
from both theoretical and practical aspects. In this paper, taking a slope receives meteoric water infiltration and groundwater runoff recharge.
in Southwest China as an example, failure mechanism and stability of Precipitation variation along with seasons in this area is remarkable,
slope under continuous rainfall after earthquake are investigated with mostly concentrated from May to October. According to the data of the
consideration of seismic load and unsaturated infiltration using GeoS- meteorological stations in Sichuan Province from 2002 to 2016, as
tudio software. Further treatment measures are suggested as well. shown in Table 1, the annual average rainfall is 873.79 mm; the daily
maximum rainfall (24 h MX) reaches 420.5 mm; the maximum rainfall
2. Project overviews and calculating conditions is 105.4 mm in an hour; and the maximum continuous rainfall lasted
13 days.
2.1. Geological conditions
2.3. Calculating conditions
The slope, located in Sichuan Province, southwest China, has a
narrow and long strip-like topography that inclines from North to South The structure, lithology, and stress state of rock mass are the main
gradually. The terrain dips north with a slope of about 46–55°. Due to influence factors to the sliding mode of a slope. The failure mechanism
several earthquakes in the past, annual earthquake related effects like analysis for rock slope follows the principle of rock mass control.
landslides, slumps and barrier lakes can be observed. The underlying Combined with the actual working conditions of a slope in Southwest
bedrock is silty shale, limestone and dolomite. China, the instability mode of the slope is analyzed. And there emerges
The rock and soil mass of the slope is a layered structure, which an arc sliding surface with vertical cracks. It is of significance to study
mainly consists of the underlying bedrock (mostly silty shale, limestone and determine and confirm the potential sliding surface of the slope by
and dolomite), weathering zones and colluvium and talus accumulation research on the stability of mixed rock/soil slope based on limit equi-
(mostly clayey and sandy soil, with gravel and debris). According to the librium method. The automatic search of procedures in light of the
division of seismic zones in Sichuan Province, the slope lies in the slope sliding mode was adapted to determine the potential sliding
Anning River-Zemuhe seismic zone with estimated earthquake intensity surface.
of VIII and seismic acceleration (horizontal acceleration in this paper) For the slope in Southwest China, a model will be developed that
of 0.3 g. allows to investigate the main effects of the earthquake and consequent
The first sliding of the slope occurred in the late 1980s, about 78 km rainfall on the slope with consideration of dynamic load and un-
away from the epicenter (Magnitude 7. 2) and 54 km away from saturated infiltration using GeoStudio software Furthermore, the Plant
hanging or foot wall in the fault zone. After that, the slope slided again Reinforcement Technology, a biological remediation, is suggested and
under the influence of M8.0 Wenchuan earthquakes and aftershocks in how biological remediation would act in this case on the slope and
May 2008. The slope failed partially and ground cracks appeared on the improve strength are explored. Three types of working conditions were
slope surface. In January 2012, a further large-scale deformation oc- analyzed with consideration of seismic effect and consequent rainfall,
curred again. Several days before that event, there was continuous underground water, and slope reinforcement treatment using

2
Y.-L. Chen, et al. Engineering Geology 266 (2020) 105461

Table 1
Average rainfall and rainstorm statistics (Unit:mm).
Mouth 1 2 3 4 5 6 7 8 9 10 11 12

Average 34.3 36.9 126.7 284.7 361.6 406.7 384.7 262.7 144.7 54.7 33.5 32.4
24 h MX 52.3 71.2 111.4 145.4 154.3 358.9 403.0 422.7 341.9 323.2 157.6 62.6

1 1 1 2012 provided an opportunity to study the contribution of the earth-

quake to the occurrence of the subsequent, rainfall-induced landslides.
In this section, the variation law of the stress and displacement fields of
3 3 3 3
the slope under earthquake conditions was discussed using quake/w
3 3
s s s s module in GeoStudio. Furthermore, SEEP/W module based on satu-
rated-unsaturated seepage model is used to simulate slope stability
under the condition of post-earthquake rainfall through numerical
1 1 1
modeling. Subsequently, the failure mechanism is further explored.

3.1. Effect of additional horizontal dynamic stress.

Fig. 2. Stress state under natural state and seismic load: (a) Natural stress state;
(b-c) stress state under seismic action.
In this section, the representative elementary volume of the rock
slope is taken to discuss the stress state and failure law of the element
GeoStudio software. We will be creating the model to simulate the under cyclic load. The seismic load in slope is simplified as horizontal
working conditions in three stages as follows: (1) natural state plus load with the assumption that the initial stress of rock mass is caused by
earthquake conditions in the first stage; (2) post-earthquake rainfall gravity of rock masses and geological structure. Fig. 2 shows the stress
conditions based on the first stage of stress and displacement fields and state of rock and soil under natural state and seismic action. Eq. (1) is
underground water level in the second stage; (3) reinforcement condi- satisfied at the natural state:
tions after earthquake and rainfall in the third stage. Detailed analysis
σ1 = γh; σ3 = k 0 γh (1)
of the three conditions was carried out.
where σ1 is the vertical principal stress (major principal stress); σ3 is the
horizontal principal stress (the third principal stress); γ is the bulk
3. Stability evaluation of seismic effect combined with consequent
density; and k0 is the lateral pressure coefficient.
rainfall
When rock and soil mass are subjected to horizontal seismic load,
additional horizontal dynamic stress, divided into the same and oppo-
The damage of slope caused by earthquake mainly depends on the
site direction to horizontal principal stress, is acting compressive and
failure ground acceleration and the large fluctuation of pore water
dilative on the rock and soil mass due to cyclic character of the hor-
pressure along with earthquake. Ground acceleration (dynamic load)
izontal seismic load.
acts as additional load or force and increases driving forces, which
makes an increase in the sliding force along the sliding surface in po-
(1) When dynamic load and static horizontal stress act the opposite
tential sliding body. Besides, due to cyclic weakening, the potential
direction (have the opposite vector),
sliding surface enters the plastic state, which leads to the decline in
sliding resistance of slope body and the increase in the instability of σ1 = γh; σ3′ = σ3 − σs = k 0 γh − ρCs vs (2)
slope body. Traditional idea holds the view that shear failure caused by G
seismic load is a main contribution to slope failure under earthquake where: vs is shear wave velocity; Cs is the velocity of S wave, Cs =
ρ
with a similarity to that under static condition. However, based on the therein and G is the shear modulus; and σs is the additional horizontal
investigation of landslides caused by the Wenchuan earthquake, many dynamic stress.
landslides show that at the mountain ridge tensile fractures develop The horizontal seismic load causes a decrease in the horizontal
behind the ridge in the rear slope, and the slope below sheared off, principal stress in the rock and soil mass. At the time when
forming a unified high-speed sliding surface. (Yin et al., 2009; Jun k0γh > ρCsvs, that is σ3´ > 0, the decrease of the third principal stress
et al., 2018) Obviously, the traditional concept of slope seismic damage in the rock and soil mass causes shear failure. When third principal
only considers the shear failure without a consideration of the weak- stress turns negative, tensile failure occurs.
ening of slope due to the previous tensile failure. In fact, compared with
its compressive and shear strength, the tensile strength of rock and soil (2) When dynamic load and static horizontal stress act the same di-
is considerably lower. Under seismic load, the slope is more prone to rection (have the same vector),
tensile failure due to the cyclic loading. σ1 = γh; σ3′ = σ3 + σs = k 0 γh + ρCs vs (3)
After the earthquake, many cracks appeared on the slope surface,
and the heat inside the earth was released a lot. At the same time, some Additional horizontal dynamic stress is acting compressive and di-
water vapor underground entered the air along the crack, thus forming lative on the soil due to cyclic character of the horizontal seismic load.
a strong updraft on the ground together with dust and particles, so When σ3´ < σ1, the increase in third principal stress makes the soil
rainfall usually occurred after an earthquake. Research has indicated tend to stable. When σ3´ > σ1, the maximum principal stress and the
that the areas stricken by an earthquake would experience a prolonged minimum principal stress are reversed where the horizontal principal
influence (Huang and Li, 2014; Jun et al., 2018). Secondary hazards stress becomes the major principal stress. At this time, the increase of
triggered by the earthquake have greatly changed the land use and the major principal stress will lead to shear failure of rock and soil.
cover in the area, in particular, the serious natural vegetation loss and
degradation, which results in increased soil erosion and accumulated 3.2. Strength criterion and seepage equation
massive loose debris layer (Jun et al., 2018). Secondary disaster in-
duced by heavy rainfall after seismic activity increases the suscept- Fredlund et al. (1978) proposed the shear strength formula for un-
ibility to slope instability. The heavy rainfall event on January 17–22, saturated soils, which is expressed as follows:

3
Y.-L. Chen, et al. Engineering Geology 266 (2020) 105461

τf = c′ + (σ − ua) tan φ′ + (ua − u w ) tan φb (4) capacity of soil, so rainwater will all infiltrate. Therefore, the pressure
head at each point on the rainfall boundary should be satisfied: h ≤ 0.
where c' is cohesion; φ is the total normal stress on the failure surface;
ua and uw are net normal stress and matrix potential, respectively; φ' is (2) Runoff
the internal friction angle related to the net normal stress state variable
(σ − ua); and φb is the rate at which the shear strength increases with The rainfall boundary can be treated as the head boundary.
the matrix suction (ua − uw). Considering the inclination of the slope, surface water will flow fast.
A combination of stress state variables (σ − ua) and (ua − uw) with Therefore, the pressure head on the slope is 0, i.e., h = 0.
the pore gas pressure ua as a benchmark was adopted by Eq. (4), which At this time, the rainfall intensity is greater than the infiltration
is essentially an extension of the Mohr-Coulomb shear strength formula capacity of soil, and rainwater cannot infiltrate completely. Therefore,
for saturated soils. When the soil is saturated, uw is equal to the ua, so the discharge on the rainfall boundary should be satisfied:
the matrix suction (ua − uw) is equal to zero, achieving a transition
from unsaturated to saturated.
⎡k s k (h) ∂h + k s k (h) ⎤ n <p
Domenico and Schwartz (1997) provided a comprehensive theore- ⎢ ij r ∂x j
i3 r ⎥ i
tical review of groundwater flow through porous media. Saturated- ⎣ ⎦ Γ (8)
unsaturated porous seepage continuum equation is obtained on the According to the local weather station in Sichuan, the rainfall
assumption that fluid seepage in porous media follows the law of fluctuated little in five days period after the earthquake. 10e−6 m/s,
conservation of mass and Darcy's law (Lam et al., 1987; Gottardi and with a considerable similarity to the actual rainfall intensity, is chosen
Venutelli, 1993). as the precipitation boundary. The accumulated precipitation reached
∂ ⎛ ∂h ⎞ ∂ ⎛ ∂h ⎞ ∂h 86.4 mm in 24 h.
kx + kz = mwγw
∂x ⎝ ∂x ⎠ ∂z ⎝ ∂z ⎠ ∂t (5)

where x and z indicate the horizontal and vertical directions, respec- 3.3.3. Selection of calculation parameters
tively; h is the total hydraulic head, kx and kz are the permeability A strong motion record of the earthquake on May 23rd recorded by
coefficients function of unsaturated soil; γw is the unit weight of water; the seismic station was chosen to simulate the earthquake with 10 s and
mw is the coefficient of water volume change in respect of the change in a peak acceleration of 0.81 m/s2. The attenuation relations of ground
matric suction, which can be obtained from the slope of the soil-water motion were reported that the isoseismal lines of earthquakes above
characteristic curve; and t is the time. moderate intensity are mostly in the shape of ellipse in NWW direction.
The vectorial equation of soil unsteady seepage can be expressed as Thus, the constrained model of elliptical attenuation of earthquake
follows (John, 2008): intensity considering long and short axis was adopted in this work. On
the basis of the Seismic Hazard Analysis Package (ESE) which con-
[K]{H } + [M ]{H } = {Q} (6) siders the heterogeneity of seismic activity and is recommended by the
China Earthquake Administration(CEA), the time history curve of
where [K] is the unit eigenmatrix; {H} is the node head vector; [M] is
ground motion with the exceedance probability of 10% was set up.
the unit mass matrix and {Q} is the node flow vector.
Comprehensive analysis for the earthquake conditions was conducted
at each interval step (a time step of 0.02 s for modeling, totally 500
3.3. Boundary conditions and parameters
interval steps for a 10-s vibration). The revised simulated time series of
seismic ground motion with a seismic acceleration (horizontal accel-
3.3.1. Dynamic boundary conditions
eration in this paper) of 0.3 g is shown in Fig. 3
The dynamic boundary treatments included that: (1) the bottom
The calculation parameters were selected for stability evaluation in
boundary (i.e. the bottom of bedrock) was fixed horizontally and ver-
accordance with actual geological conditions as well as the modes of
tically; (2) in order to reduce the impact of seismic wave rebound in the
destruction in simulation areas. The engineering geological survey was
sliding body, the impedance boundary was set at the front and rear
carried out and the field direct shear test and field deformation test of a
boundary of the slope and displacement consolidation boundary verti-
rock mass were conducted. Then combined with relevant field reports,
cally and horizontally were set at lower part of the slope; (3) the top of
the material characterizations such as internal friction angle, cohesion
the slope was set as free boundary horizontally and vertically.
0.3
3.3.2. Seepage boundary conditions
Fixed boundary is often used in numerical simulation due to the
complexity of rainfall infiltration boundary. Actually, it is a constantly 0.2
changing boundary with rainfall duration, so runoff should be taken
Acceleration / m s-2

into account (Bandara et al., 2016; Cho, 2016). 0.1

(1) No runoff.
0.0
The rainfall infiltration boundary can be treated as the flux
boundary: -0.1

⎡k s k (h) ∂h + k s k (h) ⎤ n =p
⎢ ij r ∂x j
i3 r ⎥ i -0.2
⎣ ⎦ Γ (7)

where ks ij is the saturated permeability tensor; kr is the relative per-

-0.3
meability coefficient, which can be determined by the permeability 0 2 4 6 8 10
coefficient curve of unsaturated soil in unsaturated region; ni is the
normal direction vector of slope surface; and p is the rainfall intensity Seismic duration / s
on slope surface.
At this time, the rainfall intensity is less than the infiltration Fig. 3. Revised simulated time series of seismic ground motion.

4
Y.-L. Chen, et al. Engineering Geology 266 (2020) 105461

Table 2
Mechanical parameters for rock and soil mass structural surfaces.
Category of rock and soil Unit weight (KN/m3) Shear strength (C/(MPa)) Elastic modulus (MPa) Poisson's ratio (μ)

Before reinforcement After reinforcement Internal friction Angle φ/(°)

Colluvium & talus accumulation 18.9 20 43 31 767.28 0.27

Intense weathering 21.5 90 114 35 2496 0.3
Mild weathering 25.5 500 500 41 39,936 0.18
Geological fault 21 60 92 30 39,936 0.35
Bedrock 26.5 1200 1200 43 39,936 0.49

force, elastic modulus, and Poisson's ratio were determined syntheti- 1

cally. In this paper, Mohr-Coulomb elastic-plastic model is used and the 0.1
calculation parameters are shown in Table 2. 0.01
In addition, the soil-water characteristic curve (SWCC) representing
1E-3

Conductivity / m s-1
the relationship between water content and suction, and the hydraulic Severe weathering strata
conduction equation (HFC) representing the relationship between per- 1E-4
meability coefficient and suction are two important hydraulic char- 1E-5
acteristics in the analysis of unsaturated soil infiltration. Soil-water 1E-6 Colluvial deosit
(a)
characteristic curve according to VG model (Van Genuchten, 1980) is as 1E-7
follows: 1E-8
Fitting data
θs − θr 1E-9 Experimental data
⎧ θr + h<0
θ (h) = [1 + (α |h|)n]m 1E-10
⎨
⎩ θs h≥0 (9) 1E-11
0.01 0.1 1 10 100 1000
where θ is volumetric water content (%); θs is saturated water content Matric suction / kPa
(%); θr is residual water content (%); α and n are curve shape para-
meters, and the unit of α is m−1; m = 1–1/n. 0.6
The corresponding functional expression of the hydraulic conduc-
Volumetric water content / m3 m-3

tion equation is as follows (Mualem and Yechezkel, 1976; Ippisch et al., 0.5
2006)
n−1 n −m 2 0.4
⎧ k s {1 − (α |h|) [1 + (α |h|) ] } h<0
k (h) = [1 + (α |h|)n]m /2 Severe weathering strata
⎨ 0.3
⎩ ks h≥0 (10)
(b) Colluvial deosit
where ks is saturated permeability coefficient. 0.2
Based on the field data, the saturated permeability coefficients in
bedrock and mild weathering strata (including geological faults) are Fitting data
0.1
close to constant, 1.25e−6 and 1.0e−10 m/s, respectively. The saturated Experimental data
permeability coefficients in colluvium and talus accumulation and in-
tense weathering strata are 0.75 and 1.02 m/s, respectively whose soil- 0.0
water characteristic curve and permeability coefficient curve are esti- 0.01 0.1 1 10 100 1000
mated by Eqs. (9) and (10), as shown in Fig. 4. Matric suction / kPa
Fig. 4. Parameter curve for unsaturated permeability: (a) HFC; (b) SWCC.
3.4. Stability analysis of earthquake and consequent rainfall
action of earthquake, the seismic acceleration is sharply amplified at
Based on the Morgenstern-Price method, we calculated the slope
the top of the slope due to the topographic amplification effect, re-
stability for five selected areas, including underlying bedrock, weath-
sulting in an instantaneous tension load, which causes tension cracks
ering zones, colluvium and talus accumulation and so on by using
subsequently and weakening of the tensile strength at the top of the
GeoStudio software. In order to evaluate the slope stability on the two
slope. Fig. 5 (c) shows the distribution of the stress field at the end of
conditions, that is, natural state plus earthquake conditions in the first
the seismic loading. It can be seen that the maximum tensile stress near
stage; and post-earthquake rainfall conditions based on the first stage of
the geological fault at the top of the slope exceeds 0.5 MPa, causing a
stress and displacement fields and underground water level in the
large number of fine tensile cracks.
second stage, the cutting-in and cutting-out methods were adopted to
Fig. 6a and b show the pore water pressure distribution at different
examine the potential sliding surface position. Detailed analysis such as
stress states. The pore water pressure of slope increases continuously,
stress field and displacement field on the two conditions was carried
with both the increasing duration of seismic motion and the elevation of
out.
the observation point. In strata with poor permeability, such as bed-
rock, the excess pore-water pressure increases linearly. While in the
3.4.1. Analysis on stress field of slope under seismic activity poor weak strata with large permeability coefficient, which are usually
Comparing the initial effective stress distribution (Fig. 5a) with the located at the slope surface (monitoring point 1#, monitoring point 2#,
effective stress contours after Earthquake (Fig. 5b), the slope perfor- etc.), the excess pore water pressure increases linearly at first, and then
mance is reduced due to the lowered shear strength, especially in the tends to be stable gradually. It can be seen that the excess pore water
colluvium and talus accumulation at the foot of the slope with weak pressure around the geological faults at the foot and top of the slope is
rocks lithology. The effective stress decreases by > 50%. Under the

5
Y.-L. Chen, et al. Engineering Geology 266 (2020) 105461

450

400

350

Elevation / m
300 0s
(a) 4s
250 6s
10s
200

150
370 372 374 376 378 380 382 384
Pore-water pressure / kPa
35
M.P 1#
30 M.P 2#

Excess pore-water pressure / kPa

M.P 3#
25 M.P 4#
M.P 5#
20
M.P 6#
(b) 15

10

0
0 1 2 3 4 5 6 7 8 9 10
Seismic duration / s

Fig. 6. Pore-water pressure distribution: (a) Pore-water pressure curve at dif-

ferent elevations; (b) Excess pore-water pressure curve with seismic duration.

observed. Meanwhile, potential tension-shear slip surface gradually

come to form due to seismic load and the increment in the shear strain
of the slope increases (see Fig. 8b), which results in the decrease of X-
displacement of the slope with the increase of the slope height. The
maximum permanent displacement at the bottom of the slope, shows a
clear horizontal slip trend.
In the study of local seismic records, it is found that after the
Fig. 5. Stress distribution under seismic load: (a) Initial effective stress; (b) earthquake on February 3, 2012, several aftershocks occurred in the
Effective Stress after Earthquake; (c) Stress field after earthquake. next few hours, with magnitudes below 6. Although the peak accel-
eration of these aftershocks was small, some of them lasted for a long
relatively high where shear slip is more likely to occur. while, some even for several days. This paper interprets some after-
shock seismic records where the duration of seismic events varies from
10 to 30s, with an amplitude of 0.2 g. As shown A, B, C, D, and E in
3.4.2. Analysis on displacement field of slope under seismic activity Fig. 1, five monitoring points are selected, located at the foot of the
Fig. 7 shows the horizontal displacement field distribution of the slope, colluvium and talus accumulation, intense weathering layer and
slope in the first, third, fifth, eighth and tenth seconds (corresponding a, bedrock, respectively, to discuss the influence of the duration of
b, c, d, e, respectively) during the occurrence of seismic loading. During earthquake on the dynamic response of the slope.
the earthquake, the maximum horizontal displacement of the slope From the simulation results, the effect of duration of shaking on the
with seismic load exceeds 0.9 m, which is mainly located in the collu- peak acceleration response is not obvious, but it is of significance to the
vium and talus accumulation with poor lithology. It can also be seen displacement of the slope, as shown in Fig. 9. With the continuous in-
that in the first 2–3 s of the earthquake, the higher peak acceleration crease of the duration, the displacement of monitoring points increases
leads to the obvious displacement stratification phenomenon of bedrock significantly. When the duration increases to 20 s, the displacements at
with better lithology. This could also cause further amplification ef- A, C and D tend to converge, while the displacement at B colluvium and
fects. Then, with the decrease of peak acceleration, the amplitude of talus accumulation in the sliding body is still increasing. This indicates
seismic vibration decreases, and the displacement stratification phe- that with the continuous release of the energy during the earthquake,
nomenon gradually disappears (see Fig. 7e). the damage of soil in the slope is accumulating, but the slope with
As shown in Fig. 8, there is a significant vertical amplification effect better lithology, equipped with self-resistance to resist earthquake da-
and hysteresis effect that the closer the monitoring point is to the top of mage, is not susceptible to instability. It also shows that the slope
the slope, the greater the fluctuation range of the displacement values is

6
Y.-L. Chen, et al. Engineering Geology 266 (2020) 105461

Fig. 7. Development of displacement field over time under seismic load: (a) Displacement distribution after 1 s; (b) Displacement distribution after 3 s; (c)
Displacement distribution after 5 s; (d) Displacement distribution after 8 s; (e) Displacement distribution after 10s.

0.06 instability has a significant accumulation effect as a result of cyclic

M.P 2# weakening for rock and soil.
0.04 M.P 4#
M.P 6# 3.4.3. Stability analysis of slope under seismic activity
0.02 As parts of the rock and soil mass in the geotechnical structure
Shear strain

undergoes sudden strength damage, the stress will be readjusted and

0.00 distributed. The stress field and displacement field under stress redis-
(a) tribution after earthquake have been analyzed. Earthquake triggers
-0.02 tension area near the geological fault on the top of the slope and causes
a tensile stress field at the top of the slope, while shear strength re-
-0.04 duction occurs at the foot of the slope in colluvium and talus accu-
mulation. Furthermore, a certain permanent deformation of the slope
-0.06 occurs after the earthquake. Under seismic activity, the safety factor of
0 2 4 6 8 10 the slope has been changing always within the range of 1.185–1.43(see
Seismic duration / s Fig. 9). Fig. 10 also shows the safety factors of different sliding surfaces.
According to the results of numerical simulation, the safety factor of
0.3 most dangerous potential sliding surface after earthquake is 1.129. This
M.P 2# indicates that the slope only suffers from damage and no complete
0.2 M.P 4# failure or coherent slide develops, which agrees on the record for field
M.P 6#
X-relative displacement

observation.
0.1 Under seismic load, due to the effect of topographic amplification,
high displacement rates are usually located in the lower consolidated
0.0 debris and soils while the acceleration of earthquake is sharply ampli-
fied at the top of the slope. This results in instantaneous tensile stress,
(b)
-0.1 which causes tension cracks and decreases in tensile strength at the top
of the slope. The existence of cracks contributes to subsequent rain-
-0.2 water infiltration after earthquakes where the increase of pore water
pressure in slope leads to the reduction of effective stress in soil mass
-0.3 and the activation of potential sliding mass.

0 2 4 6 8 10 3.5. Stability analysis of slope on rainfall infiltration

Seismic duration / s
As shown in Fig. 11, owing to the difference between rainfall in-
Fig. 8. Deformation law for the slope under seismic load: (a) Shear strain curve;
tensity and permeability of rock and soil, a transient saturation zone is
(b) Relative X-displacement curve.
formed after a rainfall period, resulting in the rise of water level and the

7
Y.-L. Chen, et al. Engineering Geology 266 (2020) 105461

0.8 the colluvium and talus accumulation and intense weathering layers
#
(H = 425 m).Moreover, it can be seen that the matrix suction in the
0.7 M.P A unsaturated area of intense weathering layers begins to increase with
M.P B# the continuous rainfall, and the matrix suction tends to decrease at
0.6
M.P C#
X-Displacement / m

H = 415 m when the rainfall lasts to the fifth day.

0.5 M.P D# The process of generating and dissipating the excess pore water
M.P E# pressure adversely affects the slope. The excess pore water pressure
0.4 caused by the earthquake is the main factor causing the deformation
(a) and failure of the slope. The excess pore water pressure tends to dis-
0.3 sipate as time lapses when the rise of the groundwater level in the slope
during the rainfall will affect the dissipation and its execution. Here,
0.2
monitoring points A, D and E (see Fig. 1) are taken in the colluvium and
0.1 talus accumulation, intense weathering strata and bedrock, respec-
tively, and the effect of different rainfall duration on the variation of
0.0 excess pore water pressure is analyzed, as shown in Fig. 13. For the
10 15 20 25 30 35 40 colluvium and talus accumulation and intense weathering strata, due to
Seismic duration / s the high permeability, the excess pore water pressure dissipates rapidly.
However, the water level increases, the matrix suction of the soil de-
0.6 creases continuously with the rainfall infiltration, leading to an increase
M.P A#
in permeability. And rainfall infiltration makes a slower dissipation rate
0.5 M.P B# of excess pore water pressure, which has the greatest effect on the
0.4 M.P C# colluvium and talus accumulation on the slope surface. For the bedrock,
Y-Displacement / m

M.P D# due to the small permeability coefficient, the excess pore water pressure
0.3 dissipates so slowly, which rainfall infiltration has little effect on.
M.P E#
0.2 Fig. 14 shows the change of safety factor with rainfall. The safety
factor decreases with the continuous increase of rainfall, which may
(b) 0.1 lead to a failure. Among them, the critical slip surface will lose stability
on the fourth day of rainfall, and the range of landslide will expand
0.0 with the continuous rainfall.
-0.1 On the one hand, the infiltration of rainwater leads to an increase in
the water content in the unsaturated zone of the slope, which reduces
-0.2 the matrix suction, resulting in a decrease in the shear resistance of the
soil. Furthermore, it slows down the dissipation of excess pore water
8 12 16 20 24 28 32 36
pressure. On the other hand, the contact with the loose sliding soil
Seismic duration / s causes the soil to soften and the shear surface to penetrate, which
further accelerates the deformation and eventually destabilizes the
Fig. 9. Development of cumulative displacement over time under seismic
slope.
loading: (a) X-displacement curve; (b) Y-displacement curve.

4. Plant reinforcement technology and its application

increase of groundwater pressure, which negatively effects the slope
stability. Before the earthquake, the matrix suction (namely, negative
Plant Reinforcement Technology (PRT) is used to stabilize slope by
pore water pressure) increases linearly with slope height (see Fig. 12).
planting hierarchically fast-growing shrub branches or rhizomes in
With the infiltration of rainwater, the matrix suction of soil decreases,
slope to form reinforced soil. The mechanism of biotechnical slope
and the permeability coefficient increases, which accelerates the water
protection was studied from three aspect of mechanical function of
erosion into soil. On the first day of rainfall, the matrix suction in-
roots system, hydrologic effect of stem and leaf (includes dead branches
creases first, and then an inflection point appears near the junction of
and fallen leaves), and transpiration effect of plants. Furthermore, we

1.129
Safety Factor after earthquake Critical Slip Surface

1.40 Slip Surface 121

1.43
1.84 Slip Surface 96
1.86
1.13 Slip Surface 67
Slip Surface 41

1.45
1.40
Safety Factor

1.35
1.30
1.25
1.20
0 2 4 6 8 10
Seismic duration / s

Fig. 10. Development of FOS over time while earthquake.

8
Y.-L. Chen, et al. Engineering Geology 266 (2020) 105461

Geological Fault Initial water level

Severe Weathering Strata 1st Day
2nd Day
3rd Day
Colluvial Deosit & Talus Accumulation
4th Day
5th Day

Bedrock

Fig. 11. Development of slope water level during a 5-day rainfall period.

440 The root-soil composite medium system is similar to reinforced

concrete with relatively large tensile strength where tensile stress is
430 transferred to the surrounding soil through root-soil bond between
roots and soil. Wu et al. (1979) proposed a vertical root model for shear
420 strength of root-soil composite medium, assuming that the root surface
Elevation / m

is sufficiently frictional and restrained to prevent the root from being

pulled out. When shear stress is produced in soil, the dislocation dis-
410
initial state placement of the root causes the root to elongate and the tensile force Tr
1st Day to generate in the roots. The component of Tr along the tangent di-
400
2nd Day rection of the shear plane can directly resist the shear deformation
3rd Day while that along the normal direction can increase the normal stress on
390 the shear plane. Therefore, the increased shear strength of root-soil
5th Day
composites Δc is as follows:
380
-500 -400 -300 -200 -100 0 100 Δc = Tr (Ar / A)(sin θ + cos θ tan φ) (11)
Pore-water pressure / kPa where Tr is the average tensile force of the roots (N); Ar/A is the root
Fig. 12. Influences of rainfall on matrix suction. coverage rate on the shear plane; θ is the root inclination angle (°); and
φ is the internal friction angle (°). For the herb-root system and woody
plant root system, the value of Tr is different, which is discussed as
summarized the contribution of Plant Reinforcement Technology (PRT)
follow.
to slope stability in three ways: (1) anchorage of deep roots and re-
Firstly, for herb-root system, the roots of shallow rooted plants
inforcement of shallow roots; (2) rainfall interception and reduction in
twine in the surface of soil layer, forming a reinforced composite pro-
surface erosion; (3) change for the pore water pressure. (Gray and Sotir,
tection layer with a thickness of about 30 cm. Mechanical reinforce-
1996; Chen et al., 2007; Burylo et al., 2011). The cutting rhizome
ment effect of the root system is reflected in two aspects: (1) the in-
branches can play a reinforcing role immediately. Plant roots system
creased cohesion Δc of root-soil composites;(2) the wrapping effect of
can be divided into three types: herb-root system, horizontal and ver-
“string bag” on soil particles - the increase of cohesion prevents the
tical woody plant root system. The reinforcement mechanism of roots
shear failure of water flow and the formation of tension cracks in slope
system is different, including slope reinforcement of shallow roots and
surface. The wrapping effect of “string bag” weakens the dissociation of
slope anchoring of deep roots. Generally, slope reinforcement of
water flow on soil particles. The tension crack at the top of the slope is
shallow roots and hydrological effect of stem and leaf are for slope
formed by the joint action of the horizontal displacement from the top
erosion controlling; while slope anchoring of deep roots and tran-
of the slope to the free surface and the contraction of the surface soil
spiration effect of plants are for slope reinforcement. With the growth of
due to water loss. And the increased tensile strength Tr by the re-
the internal rhizome, the reinforcing role continues along with the slope
inforcement of plant roots is:
depth and the scope of reinforcing enlarges, which is especially suitable
m
for treating surface erosion of slope and shallow landslide. 1
Tr =
B
∑ Ti sin Θi
i=1 (12)
2
4.1. Mechanical function of root system where there are m roots in the soil with area B (cm ); T is the tensile
strength of single plant root; and Θi is the acute angle between the plant
The tensile strength of grass roots will restrict the deformation of root and the crack face (i = 1, 2, …, m).
soil, and the reinforcement effect is related to the density, strength and Besides, the strong vertical taproot of woody plants is deeply rooted
properties of grass roots. The root-soil interaction can be regarded as a into the soil layer. Vertical deep roots passing through the weak layer or
pattern of fiber-reinforced soil without strength and size differences sliding surface strengthen the by means of anchoring and shaft friction,
among roots being considered. The effect of fiber reinforcement on the which acts a bit like the bolt or anti-slide pile. The root system is
strength of composite media can be illustrated by the equivalent con- simplified as a full-length adhesive bolt with the taproot as the axial
fining pressure theory and apparent cohesion (Burylo et al., 2011). and the lateral roots as the branch to calculate the anchorage force of
Root fiber improves the shear strength of soil mainly by trans- vertical taproot, which focus on segments of roots with a diameter >
forming the shear stress in soil into the tensile stress of roots through 1 mm at the depth arranging from h1 to h2 under the slope surface. The
the friction of the root-soil interface (Waldron, 1977; Shewbridge and maximum anchorage force provided by roots in the depth range of h1-h2
Sitar, 1990), as shown in Fig. 15. can be deduced as follow:

9
Y.-L. Chen, et al. Engineering Geology 266 (2020) 105461

8
1st Day
3rd Day 12 1st Day

Excess pore-water pressure / kPa

Excess pore-water pressure / kPa

5th Day 3rd Day
6 10 5th Day

8
(a) 4
6
(b)
4
2
2

0 0
0 2 4 6 8 10 12 14
2 4 6 8 10 12
Time /Day
Time / Day
30
1st Day
Excess Pore-water Pressure / KPa

3rd Day
25 5th Day

20
(c)

15

10
2 4 6 8 10 12
Time /Day

Fig. 13. Effect of rainfall on dissipation of excess pore water pressure in different strata: (a) for the colluvium and talus accumulation; (b) for the intense weathering
strata; (c) for the bedrock.

Critical Slip Surface Deformed Roots

2.0
Slip Surface 121 Undeformed Root
1.8 Slip Surface 96
Slip Surface 67
1.6 Slip Surface 41
Factor of safety

1.4
Tr Shear Zone
1.2

1.0

0.8

0.6 Fig. 15. Mechanical effect of root anchorage.

0 1 2 3 4 5
Rainfall duration / Day 4.2. Transpiration effect of plants

Fig. 14. Effect of rainfall duration on the factor of safety.

Woody plants with strong taproot, deeply embedded in the soil, can
absorb the pore water in deep of the slope. Plants exert tension of
h2
Tr = 2πμ ∫h 1
P (z ) Q (z ) z dz
(13)
1–2 MPa on pore water through photosynthesis and transpiration be-
fore wilting (Fredlund and Rahardjo, 1993). Plant roots continuously
absorb water from the soil, making water in the soil and pore water
where μ is the static friction coefficient between roots and soil. P (z) and
pressure decrease. In-situ monitoring shows pore water pressure in soil
Q (z) are determined by field measurement and then fitting data. More
is generally negative within the effected area of woody plants with
detailedly, the vertical roots are divided into n segments horizontally
strong taproot. The decrease of pore water pressure enhances the shear
within the extension range of roots. The number of roots Ni and the
strength of soil, which is calculated quantitatively by the expression of
average radius Ri of roots are obtained for any segment [i, i + 1]
unsaturated shear strength with consideration of root reinforcement as
(1 ≤ i ≤ n-1) by statistical measurement. And then, P (z) and Q (z) are
follows:
obtained by fitting Ri and Ni, respectively.
τf = c′ + Δc + (σ − ua) tan φ′ + (ua − u w ) tan φb (14)

10
Y.-L. Chen, et al. Engineering Geology 266 (2020) 105461

Water level after roots-reinforce 5. Conclusions

Water level without reinforce
In this paper, taking a slope in Southwest China as an example,
Initial water level slope stability and failure mechanism under continuous rainfall after an
earthquake are analyzed with consideration of seismic load, un-
saturated infiltration and improved plant reinforcement technology
1.6 using GeoStudio software. Summarizing the results from the different
1.5 investigations, the following conclusions are drawn:
1.4
1.3 1.25 1. Under seismic load, high displacement rates are usually located in
1.2
the lower consolidated debris and soils. Topographic amplification
1.20
causes tension stress at the peak, which triggers tension cracks and
1.1
Dynamic safety factor 1.15 decreases in tensile strength at the top of the slope.
1.10 2. The existence of cracks contributes to post-earthquake rainwater
Safety factor in rainfall
1.05 infiltration which slows down the dissipation of excess pore water
pressure. Slope instability is ultimately caused by soil weakening
0.0 2.5 5.0 7.5 1
10.0 2 3 4 5
due to seismic load and reduction in dissipation rate of pore water
Time / s Time / Day
pressure.
Fig. 16. With subsequent earthquake triggered rainfall for a plant reinforced 3. Plant Reinforcement Technology can significantly improve the sta-
slope. bility of the slope and is proved effective by selectively planting in
weak or damaged areas for dealing with the damage and instability
of slopes subjected to seismic activity and subsequent heavy rainfall.
4.3. Hydrologic effect of stem and leaf

Data availability
The erosion effect of water flow on the slope surface is weakened, or
the slope surface is protected from erosion and water-proof erosion
Some or all data, models, or code generated or used during the study
through weakening the splash erosion of raindrops and inhibiting the
are available from the corresponding author by request.
erosion of runoff on the slope surface. Runoff is reduced by 68% and
Data in this paper are presented in the form of Figs. 3–15 and
98% while the sediment produced by erosion is reduced by 95% and
Fig. 18, which are available from the corresponding author by request if
98%, respectively for slopes with vegetation cover of 60% and 100%
the source data needed.
with rainfall intensity of 0.81 mm/min and duration of 30 min (Wang
et al., 2004; Li et al., 2004). In the model, the hydrologic effect of stem
Declaration of Competing Interest
and leaf was quantitatively described as the decrease of runoff on the
slope surface.
The authors declare that they have no known competing financial
In the simulation, the surface layer of colluvium and talus accu-
interests or personal relationships that could have appeared to influ-
mulation and part of intense weathering layer are used as vegetation
ence the work reported in this paper.
growth layer. And deep-rooted shrub Quercus humilis is planted near
the geological faults in the stress-stretching area at the top of the slope
Acknowledgements
and in the colluvium and talus accumulation. Herbal shallow-rooted
herb vetiver was used on the downslope surface. The parameters of the
This work was supported by the National Natural Science
simulation are selected according to the Morris and Stormont (1997),
Foundation of China (Grant No. 51608323) and Natural Science
Zornberg et al. (2003) and Burylo et al. (2011). The root diameter is set
Foundation of Shanghai (Grant No. 16ZR1423300). The authors would
0.19–1.48 mm, and the density is 0.5 kg/m3. LAI = 1 is selected for
like to thank Dr. Rafig Azzam, Dr. Ning Li and Dr. Suran Wang for
shrubs and root depth is 0.8 m; LAI = 2 is selected for herbs, and root
helpful discussions, research team under Dr. Youlian Chen and Ning Li
depth is 0.2 m.
for financing this study.

References
4.4. Analysis of strengthening effect of plant reinforcement

Alonso, E., Gens, A., Lloret, A., Delahaye, C., 1995. Effect of rain infiltration on the sta-
Fig. 16 shows the variation of safety factor of the slope and the bility of slopes. In: Proceedings of the First International Conference on Unsaturated
water level change caused by post-earthquake rainfall through plant soils/unsat’95/paris/france/6–8 september 1995. 1.
reinforcement treatment. It can be seen that the safety factor of the Bandara, S., Ferrari, A., Laloui, L., 2016. Modelling landslides in unsaturated slopes
subjected to rainfall infiltration using material point method. Int. J. Numer. Anal.
landslide increases due to the increase of cohesion of the slope soil Methods Geomech. 40 (9), 1358–1380. https://doi.org/10.1002/nag.2499.
layer, which is not only reflected in the rainfall condition, but also in Bishop, A.W., 1955. The use of the slip circle in the stability analysis of Earth slopes.
the seismic condition where the stability of the slope increases sig- Geotechnique 5 (1), 7–17. https://doi.org/10.1680/geot.1955.5.1.7.
Burylo, M., Hudek, C., Rey, F., 2011. Soil reinforcement by the roots of six dominant
nificantly. On the one hand, the root system directly enhances the
species on eroded mountainous marly slopes (southern alps, France). Catena 84
stability of the slope due to the improved cohesion, and greatly restricts (1–2), 0–78. https://doi.org/10.1016/j.catena.2010.09.007.
the formation and development of the crack at the top of the slope as a Chang, K.T., Lin, M.L., Dong, J.J., Chien, C.H., 2012. The hungtsaiping landslides: from
ancient to recent. Landslides 9 (2), 205–214. https://doi.org/10.1007/s10346-011-
result of enhanced tensile strength of roots system; On the other hand,
0293-5.
the root system reduces the rainfall infiltration in the weak stratum of Chen, H., Hawkins, A.B., 2009. Relationship between earthquake disturbance, tropical
the slope, and strengthens the surface runoff, causing significant re- rainstorms and debris movement: an overview from Taiwan. Bull. Eng. Geol. Environ.
duction of pore pressure values, and acceleration of the excess pore- 68 (2), 161–186. https://doi.org/10.1007/s10064-009-0209-y.
Chen, J.B., Kong, L.W., Zhao, Y.L., Lu, H.B., 2007. On influence factors of suction and
water pressure dissipation, which substantially improves the stability of deformation of unsaturated soil under evaporation and transpiration effect. Rock Soil
the slope. Thus, it comes to a conclusion that the plant reinforcement Mech. 28 (9), 1767–1772 (in Chinese).
technology is effective in dealing with slope stability under seismic load Cho, S.E., 2016. Stability analysis of unsaturated soil slopes considering water-air flow
caused by rainfall infiltration. Eng. Geol. 211, 184–197. https://doi.org/10.1016/j.
due to earthquakes and rainfall. enggeo.2016.07.008.

11
Y.-L. Chen, et al. Engineering Geology 266 (2020) 105461

Clague, J.J., Stead, D. (Eds.), 2012. Landslides: Types, Mechanisms and Modeling. induced landsliding based on seismological parameters. Nat. Hazards Earth Syst. Sci.
Cambridge University Press. 17 (7), 1159–1175. https://doi.org/10.5194/nhess-17-1159-2017.
Domenico, P., Schwartz, F.W., 1997. Physical and chemical hydrogeology. Phys. Chem. Moriwaki, H., Inokuchi, T., Hattanji, T., Sassa, K., Ochiai, H., Wang, G., 2004. Failure
Hydrogeol. 28 (6). https://doi.org/10.2134/jeq1999.00472425002800060047x. processes in a full-scale landslide experiment using a rainfall simulator. Landslides 1
Fredlund, D.G., Rahardjo, H., 1993. Soil Mechanics for Unsaturated Soils. Wiley, New (4), 277–288. https://doi.org/10.1007/s10346-004-0034-0.
York. https://doi.org/10.1002/9780470172759. Morris, C.E., Stormont, J.C., 1997. Capillary barriers and subtitled covers: estimating
Fredlund, D.G., Morgenstern, N.R., Widger, R.A., 1978. The shear strength of unsaturated equivalency. J. Environ. Eng. 123 (1), 3–10. https://doi.org/10.1061/(asce)0733-
soils. Can. Geotech. J. 15 (3), 313–321. https://doi.org/10.1139/t78-029. 9372(1997)123:1(3).
Gottardi, G., Venutelli, M., 1993. Richards: computer program for the numerical simu- Mualem, Yechezkel, 1976. A new model for predicting the hydraulic conductivity of
lation of one-dimensional infiltration into unsaturated soil. Comput. Geosci. 19 (9), unsaturated porous media. Water Resour. Res. 12 (3), 513–522. https://doi.org/10.
1239–1266. https://doi.org/10.1016/0098-3004(93)90028-4. 1029/wr012i003p00513.
Gray, D.H., Sotir, R.B., 1996. Biotechnical and Soil Bioengineering Slope Stabilization: A Okura, Y., Kitahara, H., Ochiai, H., Sammori, T., Kawanami, A., 2002. Landslide fluidi-
Practical Guide for Erosion Control. John Wiley & Sons. zation process by flume experiments. Eng. Geol. 66 (1), 65–78. https://doi.org/10.
Hong, Y.S., Chen, R.H., Wu, C.S., Chen, J.R., 2005. Shaking table tests and stability 1016/S0013-7952(02)00032-7.
analysis of steep nailed slopes. Can. Geotech. J. 42 (5), 1264–1279. https://doi.org/ Pantelidis, L., Griffiths, D.V., 2013a. Stability of earth slopes. Part I: two dimensional
10.1139/T05-055. analysis in closed-form. Int. J. Numer. Anal. Methods Geomech. 37 (13), 1969–1986.
Hu, X., Chen, Z.L., 2011. Dynamic response of horizontal bedded rocky slopes under https://doi.org/10.1002/nag.2118.
earthquakes and influence of ground motion parameters. Appl. Mech. Mater. 90–93, Pantelidis, L., Griffiths, D.V., 2013b. Stability of earth slopes. Part II: three dimensional
570–573. https://doi.org/10.4028/www.scientific.net/AMM.90-93.570. analysis in closed-form. Int. J. Numer. Anal. Methods Geomech. 37 (13), 1987–2004.
Huang, R., Fan, X., 2013. The landslide story. Nat. Geosci. 6 (5), 325–326. https://doi.org/10.1002/nag.2116.
Huang, R., Li, W., 2014. Post-earthquake landsliding and long-term impacts in the Salinas-Jasso, Jorge A., Ramos-Zuñiga, Luis G., Montalvo-Arrieta, Juan C., 2019. Regional
Wenchuan earthquake area, China. Eng. Geol. 182, 111–120. https://doi.org/10. landslide hazard assessment from seismically induced displacements in Monterrey
1016/j.enggeo.2014.07.008. Metropolitan area, Northeastern Mexico. Bull. Eng. Geol. Environ. 78 (2),
Huang, Y., Xiong, M., 2017. Dynamic reliability analysis of slopes based on the prob- 1127–1141. https://doi.org/10.1007/s10064-017-1087-3.
ability density evolution method. Soil Dyn. Earthq. Eng. 94, 1–6. https://doi.org/10. Sasahara, K., Sakai, N., 2014. Development of shear deformation due to the increase of
1016/j.soildyn.2016.11.011. pore pressure in a sandy model slope during rainfall. Eng. Geol. 170 (4), 43–51.
Ippisch, O., Vogel, H.J., Bastian, P., 2006. Validity limits for the van Genuchten-Mualem https://doi.org/10.1016/j.enggeo.2013.12.005.
model and implications for parameter estimation and numerical simulation. Adv. Sepúlveda, Sergio A., Murphy, W., Jibson, R.W., Petley, D.N., 2005. Seismically induced
Water Resour. 29 (12), 1780–1789. https://doi.org/10.1016/j. advwatres.2005.12. rock slope failures resulting from topographic amplification of strong ground mo-
011. tions: The case of Pacoima Canyon, California. Eng. Geol. 80 (3-4), 336–348. https://
Iwata, N., Yoshinaka, R., Sasaki, T., 2013. Applicability of the seismic response analysis doi.org/10.1016/j.enggeo.2005.07.004.
using multiple yield model for discontinuous rock. Int. J. Rock Mech. Min. Sci. 60 (8), Shewbridge, S.E., Sitar, N., 1990. Deformation-based model for reinforced sand. J.
196–207. https://doi.org/10.1016/j.ijrmms.2012.12.039. Geotech. Eng. 116 (7), 1153–1170. https://doi.org/10.1061/(asce) 0733-
Jibson, R.W., Harp, E.L., 2016. Ground Motions at the Outermost Limits of Seismically 9410(1990)116:7(1153).
Triggered Landslides. Bull. Seismol. Soc. Am. 106 (2), 708–719. https://doi.org/10. Sorbino, G., Nicotera, M.V., 2013. Unsaturated soil mechanics in rainfall-induced flow
1785/0120150141. landslides. Eng. Geol. 165, 105–132. https://doi.org/10.1016/j.enggeo. 2012.10.
John, K., 2008. Seep Modeling With SEEP/W2007 Version[R]. Geoslope International 008.
Ltd., Calgary, Canada, pp. 266–268. Tang, C., Zhu, J., Li, W.L., Liang, J.T., 2009. Rainfall-triggered debris flows following the
Jun, X., Ming, W., Kai, L., Coulthard, T.J., 2018. Modeling sediment movement and wenchuan earthquake. Bull. Eng. Geol. Environ. 68 (2), 187–194.
channel response to rainfall variability after a major earthquake. Geomorphology Van Genuchten, M.T., 1980. A closed-form equation for predicting the hydraulic con-
320, 18–32. https://doi.org/10.1016/j.geomorph.2018.07.022. ductivity of unsaturated soils 1. Soil Sci. Soc. Am. J. 44 (5), 892–898. https://doi.org/
Kanaori, Y., Kawakami, S.I., 1996. The 1995 7.2 magnitude Kobe earthquake and the 10.2136/sssaj1980.03615995004400050002x.
Arima-Takatsuki tectonic line: implications of the seismic risk for Central Japan. Eng. Waldron, L.J., 1977. The shear resistance of root-permeated homogeneous and stratified
Geol. 43 (2–3), 61–82. https://doi.org/10.1016/S0165-1250(97)80006-5. soil. Soil Sci. Soc. Am. J. 41 (5), 843–849. https://doi.org/10.2136/sssaj1977.
Kimoto, S., Oka, F., García, E., 2013. Numerical simulation of the rainfall infiltration on 03615995004100050005x.
unsaturated soil slope considering a seepage flow. Geotech. Eng. 44 (3), 1–13. Wang, Rui-gang, Yan, Shu-wang, Wei-dong, Deng, 2004. Analysis of seepage stability of
Lam, L., Fredlund, D.G., Barbour, S.L., 1987. Transient seepage model for saturated-un- high-filled embankment slope due to rainfall infiltration [J]. China J. Highway
saturated soil systems: a geotechnical engineering approach. Can. Geotech. J. 24 (4), Transp. 17 (4), 25–30.
565–580. https://doi.org/10.1139/t87-071. Wu, T.H., McKinnell III, W.P., Swanston, D.N., 1979. Strength of tree roots and landslides
Li, Zhu-long, Ding, Xiao-jun, Shu-zeng Shu-zeng, Zhao, 2004. Environment protection in on Prince of Wales Island, Alaska. Can. Geotech. J. 16 (1), 19–33.
application of asphalt mixture [J]. J. Traffic Transport. Eng. 4 (4), 1–4. Yano, A., Shinohara, Y., Tsunetaka, H., Mizuno, H., Kubota, T., 2019. Distribution of
Lin, M.L., Wang, K.L., 2006. Seismic slope behavior in a large-scale shaking table model landslides caused by heavy rainfall events and an earthquake in northern Aso
test. Eng. Geol. 86 (2–3), 118–133. https://doi.org/10.1016/j.enggeo. 2006.02.011. Volcano, Japan from 1955 to 2016. Geomorphology 327, 533–541. https://doi.org/
Lin, C.W., Liu, S.H., Lee, S.Y., Liu, C.C., 2006. Impacts of the Chi-Chi earthquake on 10.1016/j.geomorph.2018.11.024.
subsequent rainfall-induced landslides in Central Taiwan. Eng. Geol. 86 (2–3), Yin, Y., Wang, F., Sun, P., 2009. Landslide hazards triggered by the 2008 wenchuan
87–101. https://doi.org/10.1016/j.enggeo.2006.02.010. earthquake, sichuan, china. Landslides 6 (2), 139–152. https://doi.org/10.1007/
Liu, S.H., Lin, C.W., Tseng, C.M., 2013. A statistical model for the impact of the 1999 Chi- s10346-009-0148-5.
Chi earthquake on the subsequent rainfall-induced landslides. Eng. Geol. 156, 11–19. Zornberg, J.G., Lafountain, L., Caldwell, J.A., 2003. Analysis and design of evapo-
https://doi.org/10.1016/j.enggeo.2013.01.005. transpirative cover for hazardous waste landfill. J. Geotech. Geoenviron. 129 (5),
Marc, O., Meunier, P., Hovius, N., 2017. Prediction of the area affected by earthquake- 427–438. https://doi.org/10.1061/(ASCE)1090-0241(2003)129:6(427).

12