Professional Documents
Culture Documents
To cite this article: Adarsh S. Chatra , G. R. Dodagoudar & V. B. Maji (2017): Numerical modelling
of rainfall effects on the stability of soil slopes, International Journal of Geotechnical Engineering,
DOI: 10.1080/19386362.2017.1359912
than the high permeability soil in terms of stability. Furthermore, Capillary forces above the groundwater table saturate the soil to
the rate of decrease in factor of safety (FOS) values for advanced short distance and above the capillary fringe unsaturated zone
patterns in rainfall is found to be more than the other rainfall pat- exists. The unsaturated zone exists till soil experiences satura-
terns. However, their study was restricted to pore-water pressure tion due to rainfall or flooding. The mathematical statement that
and FOS variation with time. Ali et al. (2014) investigated the describes the flow of water in a variably saturated soil is obtained
effects of different boundary conditions (slope angle, drainage by combining the statement of moisture conservation and the
of layer) on the failure time and failure depth subjected to vari- simplified momentum equation which is given by Darcy’s law
ous rainfall intensities. They concluded that the shallow failure (Chow, Maidment, and Mays 1988). The equation governing the
was observed for more permeable drained boundary, whereas one-dimensional flow of water in unsaturated soils is attributed
the deeper failure was observed for less permeable drainage to Richards (1931) and is given as
boundary. Xue et al. (2016) conducted coupled and uncoupled
𝜕𝜃 𝜕 𝜕𝜃
( )
analyses on the effect of long duration rainstorms on red-clay = D(𝜃) + k(𝜃) (1)
slopes in China. This study has indicated that the settlement is 𝜕t 𝜕z 𝜕z
observed at the slope surface for short duration rainstorm and where k(θ) = permeability of soil, θ = volumetric water content,
for longer duration rainstorm, non-uniform swelling of the slope z = elevation head, t = time, D(θ) = k(∂ω/∂θ) = soil water diffusiv-
is observed. No significant variation in pore-water pressure and ity and ω = suction head. Two unknown parameters in Equation
FOS values are detected for both the coupled and uncoupled (1) which need to be determined are permeability and diffusiv-
analyses. ity. These parameters are functions of volumetric water content
The study presented in this paper focuses on correlation which depend on the matric suction. The Richard’s equation is
Downloaded by [University of Florida] at 07:48 04 August 2017
between the stability of soil slopes and rainfall intensity and dura- highly non-linear and the soil parameters are dependent on the
tion. Uniform rainfall patterns are used in the numerical analysis flow variables which makes it difficult to obtain the analytical
of soil slopes with different durations and soil conditions. The solution. The matric suction is defined as the difference between
effects of hydraulic characteristics of soil, rainfall intensity and the pore air and pore-water pressures. The relationship between
duration on the stability of soil slopes of different densities viz. the water content in a soil and the pore-water pressure can be
loose, medium and dense are studied in detail. The main rainfall expressed in the form of volumetric moisture content vs. matric
events are applied to the slopes right after the application of the suction, known as the soil–water characteristic curve (SWCC).
antecedent rainfall. The antecedent rainfall is applied to simulate A widely used representation of the hydraulic characteristics of
the initial conditions prevalent just before the occurrence of the the unsaturated soils is the set of closed-form equations formu-
major rainfall events. This will be the case in most of the resid- lated by van Genuchten (1980), which are based on the capillary
ual soil slopes during rainfall in tropical and sub-tropical areas. model of Mualem (1976). The SWCC, the diffusivity and the
The variations in pore-water pressure in slopes under different permeability functions are expressed as
rainfalls are computed using finite difference analysis (FDA) of ( )
transient water flow through unsaturated soils. These pore-wa- 𝜃 − 𝜃r
ter pressures are then used along with the strength reduction
Θ = ( ) (2)
𝜃s − 𝜃 r
technique to calculate the FOS using FDA. All the finite dif-
ference computations have been performed using FLAC. The (1 − m) ks 1∕ − 1∕
( )
hydro-mechanically coupled transient analysis is required to be soils are obtained from the transient flow analysis. Then the FOS
formulated in a consistent manner. The two-dimensional finite calculations are performed utilising the SSR technique by pro-
different analysis for the transient water flow through unsatu- gressively reducing the shear strength of the material (cohesion
rated–saturated soils has been adopted earlier in the field of and friction) to bring the slope to a state of limiting equilibrium.
hydrology. In this study, the rainfall-induced instability analy- The soil slope is modelled using Mohr-Coulomb failure criterion.
sis of slopes is performed using a finite difference code, FLAC The SSR technique requires many trial runs with a trial factor of
(FLAC Manual 2011). The governing differential equations for safety Ftrial with the c′ and φ′ adjusted as given by
subsurface flow are expressed as
� 1 �
ctrial =
Ftrial
c (10)
𝜕qis
[ ] [ ]
𝜃 𝜕Ps 𝜕𝜃s 𝜕𝜀
n s + = − + 𝜃s (5)
Ks 𝜕t 𝜕t 𝜕xi 𝜕t [ ]
1
𝜙�trial = tan−1 tan 𝜙� (11)
[ ] [ g
] Ftrial
𝜃g 𝜕Pg 𝜕𝜃g 𝜕qi 𝜕𝜀
n + = − + 𝜃g (6) The Ftrial value at which slope fails can be found using bracketing
Kg 𝜕t 𝜕t 𝜕xi 𝜕t
and bisection. Initially, upper and lower brackets are established.
where qis and qig = fluid flow of water and air, respectively, Ks The simulation for first Ftrial converges is considered as lower
and Kg = bulk modulus of water and air, respectively, θg = fluid bracket and non-convergence of Ftrial is taken as upper bracket.
saturation of air, Ps and Pg = pore-water and pore-air pressures, Further, a point midway between the convergence and non-con-
Downloaded by [University of Florida] at 07:48 04 August 2017
respectively, n = porosity of soil and ε = volumetric strain. vergence are tested. If the simulation converges, the new value
The slope stability computations are carried out in the pres- is updated to the lower bracket and if the simulation does not
ent study using shear strength reduction (SSR) technique in converges; the upper bracket is replaced with the new value. The
FLAC. The shear strength contribution coming from the neg- process is repeated until the difference between the upper and
ative pore-water pressure in unsaturated soil is considered in lower brackets is less than a particular value.
the stability analysis. For this purpose, Bishop’s effective stress
formula is used (Bishop 1959):
( ) ( ) Stability analysis using FLAC
𝜎 � = 𝜎 − P g + 𝜒 Pg − P s (7) Slope geometry and boundary conditions
where σ′ = effective normal stress, σ = total normal stress and In this study, the slope geometry, mechanical and hydraulic
χ = parameter which is function of soil type and degree of sat- properties that are given in Cai and Ugai (2004) are considered.
uration and it’s values varies between zero and unity. The shear Figure 1 shows the homogeneous slope of silty sand having spe-
strength of soil is expressed as cific slope geometry and boundary conditions. The slope angle
of 33.7°, slope height of 10 m and initial height of groundwater
𝜏max = c � + 𝜎 � tan 𝜙� (8) table from the bottom of the slope is taken as 4 m in the analysis.
where τmax = shear strength of unsaturated soil, c′ = effective The boundary conditions utilised for the seepage analysis are also
cohesion of soil, ϕ′ = angle of internal friction of soil shown in Figure 1. The three main rainfall events with intensities
Substituting Equation (7) in Equation (8) we get the following (i.e. boundary fluxes), I = 0.1 ks, 1 ks and 10 ks are applied to the
equation: surface of slope right after the application of a particular ante-
cedent rainfall during the stability analysis of rain-induced slope
instability. Along the sides of the slope below the water table, a
( ) ( )
𝜏 max = c � + 𝜎 − Pg tan 𝜙� + 𝜒 Pg − Ps tan 𝜙� (9)
boundary condition equal to total head is applied.
Traditionally, slope stability analysis is analysed using limit equi-
librium approach. Over the years, finite element/finite differ- Initial conditions
ence method with shear strength reduction (SSR) technique has
been applied to the practical slope stability analysis (Zienkiewicz, Firstly, an antecedent total rainfall of 2000 mm prior to the
Humpheson, and Lewis 1975; Ugai 1989; Dawson, Roth, and main rainfall event is applied to the surface of slope for one
Drescher 1999; Griffiths and Lane 1999). Studies have shown that year in order to simulate the antecedent groundwater condi-
the SSR technique is a reliable and robust approach to assess the tions in the slope section. The pore-water pressure, degree of
safety factor of the slope and locating the corresponding critical saturation, shear strain, displacement, etc. are evaluated at the
slip surface. One of the main advantages of the SSR technique end of the main rainfall events. The plots corresponding to the
is that the failure surface is found automatically through the above parameters are drawn against the depth from the slope
zones within the material where shear stresses overcome the surface at the selected locations in the top and bottom portions
shear strength of the material. The SSR technique does not have of the slope sections, Section x–x and y–y (Figure 1). Initially,
the concept of slices, hence there is no need for the assumption the pore-water pressure distribution near the soil surface, after
about inter slice side forces. the application of antecedent rainfall prior to the main rainfall
The analysis of the rain-induced slope instability is carried event, is different for different soils depending on the denseness
out using FLAC in plain strain mode. The pore-water pressures of the soil. Bowles (1997) suggested the equivalent density and
needed in the calculation of shear strength of the unsaturated friction angle for different relative densities for the granular soils
4 A. S. CHATRA ET AL.
having very loose to very dense state. The density values adopted Table 1. Soil properties of the silty sand.
in the present study are taken from Bowles (1997). Lee and Singh Soil ks (m/s) γd (kN/m3) n (−) φ′ (°) ψ (°)
(1971) proposed an approximate relation between the relative Loose 1 × 10−6 15 0.4339 30 0
Downloaded by [University of Florida] at 07:48 04 August 2017
Soil properties
The three different soil types namely loose, medium and dense
are considered to the study the response of silty sand slope sub-
jected to three different main rainfall events. The properties of
all these soil types are given in Table 1, in which ks = saturated
permeability, γd = dry unit weight, n = porosity, φ′ = friction angle
and ψ = dilation angle.
The shear strength and soil hydraulic properties as given in
(Cai and Ugai 2004) are kept constant for all the cases to ensure
that the changes in stability of slopes are only attributed to the
changes in denseness of the soil. Table 2 gives the engineering
and hydraulic properties of the soil slope. Figure 2 shows the soil
water characteristic curve of soil.
Results and discussion Figure 2. The soil water characteristic curve of soil (Cai and Ugai 2004).
Figure 3. (a) Variation of pore pressure at crest for Section x–x with I = 0.1 ks, (b) Variation of pore pressure at Section y–y, 2 m from toe with I = 0.1 ks, (c) Variation of pore
pressure at crest for Section x–x with I = 1 ks, (d) Variation of pore pressure at Section y–y, 2 m from toe with I = 1 ks, (e) Variation of pore pressure at crest for Section x–x
with I = 10 ks, (f) Variation of pore pressure at Section y–y, 2 m from toe with I = 10 ks.
x–x and y–y are provided. The FOS values with varying rainfall pore-water pressures with rainfall duration (i.e. time) for a rain-
duration are obtained and provided in the form of plots. The plots fall intensity, I = 0.1 ks. For the slope section with loose soil con-
depicting the critical slip surfaces are also given along with the dition [see Figure 3(a)], at the crest of the slope (Section x–x),
shear strain increment. the pore pressure increases gradually from −5.6 kPa at 0 h rainfall
The following abbreviations are used in the figures: L = loose to −0.90 kPa at 84 h rainfall. For the slope section with medium
soil, M = medium soil, D = dense soil, pp = pore-water pres- soil condition, the variation in pore pressure is not observed
sure, FOS = factor of safety, Fi/Fs = initial FOS/FOS during at a depth of 2 m from the crest of the slope. For the slope sec-
analysis, sat = degree of saturation, ssi = shear strain increment, tion with dense soil conditions, no change in pore pressure is
Xdisp = Horizontal displacement, Ydisp = Vertical displacement. observed all along the depth of the slope at Section x–x. The
These abbreviations are used in the figures: for example, L_0h_ pore pressure remains constant for dense soil throughout the
pp_0.1ks_xx can be taken as – Loose soil with rainfall intensity duration of the rainfall (84 h). This is mainly due to the lower
of 0.1 ks depicting pore-water pressure at Section x–x of Figure permeability and porosity thereby restricting the infiltration of
3. Similarly all other abbreviations are used accordingly and are water in the slope.
self-explanatory. Figure 3(b) shows the variation of pore pressure for the case
Suradi and Fourie (2014) conducted the parametric studies of loose, medium and dense soil conditions at Section y–y, 2 m
on the Jabiru slope, Australia, considering uniform rainfall inten- from the toe with I = 0.1 ks. The variations in pore pressures at
sities. It is noted that the upper bound (I = 4 ks) to lower bound the surface of the slope near to toe for the loose, medium and
(I = 0.25 ks) intensities are sensitive to the rate of decrease in dense soil conditions are −3, −4 and −4.4 kPa, respectively, at
FOS towards minimum FOS. It is concluded that the rainfall time t = 0 h. At time t = 84 h, the pore pressure values for the
intensities higher than the upper bound do not increase the rain- loose, medium and dense soil conditions are 0, 0.82 and −3.7 kPa,
water infiltration, whereas the rainfall intensities lower than the respectively.
lower bound cannot saturate the surface soils. The three rainfall Figure 3(c)–(f) depict the variations of pore pressure with
intensities considered in this study are representative of the lower depth for I = 1 ks and I = 10 ks, respectively, for all the three
and upper bound intensities occurring in the present study area. soil conditions. The difference in magnitude of pore pressure
variations for I = 1 and 10 ks is negligible. The pore pressure
variation at the crest of the slope and at a depth of 4 m from
Effect of rainfall on pore-water pressure generation
the crest (Section x–x) for the loose soil condition having I = 1
An antecedent total rainfall of 2000 mm (I = 6.43 × 10−8 m/s) ks increases gradually from −5.6 to −6 kPa at t = 0 h to 0.3 and
is applied to the slope section up to one-year period in order 21 kPa at t = 17 h, respectively. The gain in pore pressure at the
to simulate the antecedent groundwater conditions prior to the slope surface and at a depth of 4 m from the crest of the slope
main rainfall event. Figure 3(a) and (b) show the evolution of (Section x–x) are 100 and 128%, respectively, and is responsible
6 A. S. CHATRA ET AL.
for the reduction of the FOS of slope to 0.92. It can be stated that shown in Figure 4(a)–(d). The rate of decrease in FOS values for
a rain storm of intensity, I = 1 ks occurring for 17 h of duration the loose and medium soils is rapid till 48 and 12 h, respectively,
is responsible for inducing instability in the slope with loose soil for the rainfall intensity of 0.1 ks [see Figure 4(a)]. For the case
conditions. This particular rainfall event can be considered as a of dense soil slope, the FOS values remain unchanged over the
triggering event for the slope under consideration. entire duration of the rainfall and for the medium soil slope; the
The pore pressures at the surface of the slope near to toe at FOS values remain constant after 12 h of rainfall.
time t = 0 h are −3.5, −3.9 and −4.4 kPa, respectively, for the slope Figure 4(b) and (c) show the variations of FOS values with
sections with loose, medium and dense soil conditions. After the time for the intensity of rainfall equal to 1 and 10 ks, respectively.
application of rainfall intensity of 1 ks, at time t = 17 h, the pore For both the intensities, the magnitude and trends of FOS val-
pressures are observed to be 0.47, 0.46 and −3.6 kPa, respec- ues with time is similar. It can be observed that for I = 1 ks and
tively, for the slope sections with loose, medium and dense soil I = 10 ks, the variations of pore pressures [see Figure 3(b) and
conditions. It is seen from Figure 3(c) and (d) that for the slope (c)] and FOS values are almost same. For further analysis, the
section with loose soil condition, there is 113.5% increase in the rainfall intensities, I = 0.1 ks and I = 1 ks are only considered for
pore pressure and for the medium soil the increase is 111.6%. the loose and medium soil slope sections. For the slope section
The increase in pore pressures at Sections x–x and y–y further with loose soil condition, a FOS of 1.7 was obtained at the end
increases the wetting front depth with time thereby bringing the of antecedent rainfall duration. For the same slope with rainfall
slope to the point of instability. In the case of dense soil slope, intensity, I = 1 ks, at the end of 17 h of rainfall, the FOS value
no major differences in the pore pressure values are observed. reduced to 0.92 thereby inducing the instability in the slope. This
This is attributed to the lower permeability and porosity of the situation can be considered as a rain-induced slope failure for
Downloaded by [University of Florida] at 07:48 04 August 2017
Figure 4. (a) Variation of factor of safety with time for I = 0.1 ks, (b) Variation of factor of safety with time for I = 1 ks, (c) Variation of factor of safety with time for I = 10 ks,
(d) Normalised factor of safety with time for I = 0.1 and 1 ks.
INTERNATIONAL JOURNAL OF GEOTECHNICAL ENGINEERING 7
three different rainfall intensities. Figure 4(d) shows the nor- applied rainfall intensity of 0.1 ks (i.e. 1 × 10−8 m/s) is smaller
malised safety factors for the three slope sections with three than the antecedent rainfall intensity of 6.43 × 10−8 m/s which
different rainfall intensities. For the intensity of rainfall equal results in the decrease of the degree of saturation.
to 0.1 times, the saturated permeability of the loose, medium A variation of degree of saturation with rainfall duration is
and dense soil slopes, the reduction in FOS values is 15, 10 and shown in Figure 5(c) and (d) for the rainfall intensity equal to
0.7%, respectively. saturated permeability of the soil. The degree of saturation along
The decrease in FOS values for the slopes with loose, medium Section x–x reduces from 15% at the surface to 7% at a depth of
and dense soil conditions for I = 1 and 10 ks is 45, 17 and 0.7%, 1 m for the loose soil at t = 0 h. There after the degree of satura-
respectively, from their corresponding initial values. It is seen tion remains constant along the depth. At t = 17 h of rainfall, the
from the results that a negligible decrease in the FOS value is soil along the depth is completely saturated. At rainfall duration
noted for the case of dense soil slope. The normalised FOS of of 17 h, the increase of pore-water pressure caused the slope to
0.55 for the loose soil slope comes under the failed state as this fail by fully developing the critical slip surface. For I = 1 ks, the
value corresponds to a FOS of less than 1. degree of saturation for the medium soil condition is 98% at
t = 0 h up to a depth of 1 m from the crest of the slope. For the
remaining depth 1 to 4 m, the degree of saturation decreases
Effect of rainfall on degree of saturation
from 98 to 10%. At t = 17 h, the rain water infiltrated all along
Figure 5(a) and (b) show the variations of degree of saturation the depth of the slope thereby fully saturating the slope. Even
for loose and medium soil slopes with time for I = 0.1 ks. The after the saturation, the medium soil slope is stable due to lesser
degrees of saturation near the crest and at Section y–y, 2 m from increment in pore-water pressure [see Figure 4(c)].
Downloaded by [University of Florida] at 07:48 04 August 2017
the toe of the slope for the loose soil condition at time t = 0 h are
15.7 and 14.7%, respectively, and after 84 h of rainfall, the final
Strain response of slope during rainfall
degrees of saturation observed at the crest and toe of the slope
are 81 and 82.4%, respectively. For the medium soil condition, The rainfall infiltration which leads to reduction in matric suc-
a decrease in degree of saturation is observed for I = 0.1 ks at tion also changes the shear strength and unit weight of the soil.
t = 0, 84 h up to the depth of 1 m from the crest of the slope. The The plots of shear strain increment with time for loose soil slope
Figure 5. (a) Variation of degree of saturation at crest for Section x–x with I = 0.1 ks, (b) Variation of degree of saturation at Section y–y, 2 m from toe with I = 0.1 ks, (c)
Variation of degree of saturation at crest for Section x–x with I = 1 ks, (d) Variation of degree of saturation at Section y–y, 2 m from toe with I = 1 ks.
8 A. S. CHATRA ET AL.
Downloaded by [University of Florida] at 07:48 04 August 2017
Figure 6. (a) Variation of shear strain increment at crest for Section x–x with I = 0.1 ks, (b) Variation of shear strain increment at Section y–y, 2 m from toe with I = 0.1 ks,
(c) Variation of shear strain increment at crest for Section x–x with I = 1 ks, (d) Variation of shear strain increment at Section y–y, 2 m from toe with I = 1 ks, (e) Shear strain
increment after 12 h of rainfall for loose soil slope, I = 1 ks, (f) Shear strain increment after 17 h of rainfall for loose soil slope, I = 1 ks.
Figure 7. (a) Variation of horizontal displacement at crest for Section x–x with I = 0.1 ks, (b) Variation of horizontal displacement at Section y–y, 2 m from toe with I = 0.1 ks,
(c) Variation of horizontal displacement at crest for Section x–x with I = 1 ks, (d) Variation of horizontal displacement at Section y–y, 2 m from toe with I = 1 ks.
INTERNATIONAL JOURNAL OF GEOTECHNICAL ENGINEERING 9
are shown in Figure 6(a)–(d) for I = 0.1 and 1 ks. Figure 6(c) (i.e. slope with I = 0.1 ks is more compared to the medium soil slope.
I = 1 ks) depicts the variation of the shear strain increment along For the case of loose soil slope, the magnitudes of horizontal
Section x–x at t = 0 and 17 h which allow the identification of displacements are much more for I = 1 ks as compared to 0.1 ks.
the most active shear zones for loose soil condition. The subse- The maximum values of horizontal displacements are observed
quent increase in rainfall duration helps in the development of for loose soil slopes all along the depth section x–x.
progressive failure mechanism [see Figure 6(e) and (f)]. For the The larger horizontal displacements of the loose soil slopes
case of I = 0.1 ks, no major changes in shear strain increment is are responsible for the actual mass movement involved in the
observed for both the loose and medium soil slopes rain-induced landslides. The actual landslide mass to be involved
Figure 6(e) depicts the plot of shear strain increment after in the mass movement can be evaluated after the location of
12 h of rainfall for the loose soil slope with I = 1 ks. It can be the critical slip surface along with the magnitudes of horizontal
seen from the figure that the slope is still stable. With increas- displacements and nodal velocities.
ing rainfall duration, distinct shear bands are formed leading
to failure along a critical slip surface. Large irreversible shear Vertical deformation response of slope during rainfall
deformations have occurred after 17 h of rainfall at the upper
portion of the slope. The most probable slip surface originated The variations of vertical displacements at Sections x–x and y–y,
from the toe of the slope and moved towards the top surface as 2 m from the toe of the slope for the loose and medium soil
shown in Figure 6(f). slopes are depicted in Figure 8(a)–(d) for I = 0.1 and 1 ks. Figure
8(a) and (b) show the distribution of vertical displacements at
Sections x–x and y–y for I = 0.1 ks for the loose and medium soil
Downloaded by [University of Florida] at 07:48 04 August 2017
Figure 8. (a) Variation of vertical displacement at crest for Section x–x with I = 0.1 ks, (b) Variation of vertical displacement at Section y–y, 2 m from toe with I = 0.1 ks, (c)
Variation of vertical displacement at crest for Section x–x with I = 1 ks, (d) Variation of vertical displacement at Section y–y, 2 m from toe with I = 1 ks.
10 A. S. CHATRA ET AL.
magnitude of vertical displacements which are usually evaluated and vertical displacement are established using the non-linear
by the coupled hydro-mechanical analysis. regression. Equation (14) provides the relationship between vs
and Ydisp for the loose soil condition with I = 1 ks for the rainfall
duration of t = 0 to 17 h. Equation (15) gives the relationship
Relationship between seepage velocity and vertical
between vs and Ydisp for the loose soil condition with I = 0.1 ks
displacement
for the rainfall duration of t = 0 to 17 h. Similarly, Equation (16)
In FLAC, the fluid flow equations are expressed in terms of pres- gives the relationship between vs and Ydisp for the loose soil con-
sure in lieu of head. Therefore, the Darcy’s law can be expressed dition with I = 0.1 ks for the rainfall duration of t = 17 to 84 h,
in terms of pressure gradient as which is a rational fit, correlates better with the FLAC results.
𝜕p The correlations are given as below:
vs = 𝜅 (13)
𝜕y 2
vs = 430.0881 − 1040.17 Ydisp + 770.3924 Ydisp
where vs = seepage velocity, κ = mobility coefficient and 𝜕p∕𝜕y 3 4 (14)
− 159.77 Ydisp + 7.69238 Ydisp
= rate of change of pore pressure with respect to depth in a slope.
The objective is to develop a relationship between the seepage
velocity and vertical displacement (Ydisp) evaluated considering 2
vs = 7.38252 + 67.7801 Ydisp − 1174.12 Ydisp
(15)
different rainfall intensities (i.e. I = 0.1 ks, and I = 1 ks) and dura- 3 4
+ 6496.222 Ydisp − 10435 Ydisp
tions at a depth of 4 m from the crest for the loose and medium
soil slope sections.
Downloaded by [University of Florida] at 07:48 04 August 2017
( )
The first step is to develop the expression for the pore pres- 1
sure as a function of depth for different rainfall intensities and
vs = 2 (16)
0.10992 + 0.14684 Ydisp − 0.4282 Ydisp
durations. The relationship is of fourth-order polynomial which
fits better with the FLAC results. Further, the rate of change of where Ydisp is expressed in mm units and vs is in mm/h.
pore pressure at a depth of 4 m is calculated for different rainfall Equations (17) and (18) provide the relationship between vs
intensities and durations. Using Equation (13), seepage veloc- and Ydisp for I = 1 and 0.1 ks, respectively, for the medium soil
ity is evaluated for the loose and medium soil conditions for conditions at a depth of 4 m from the crest. The correlations
I = 0.1 and 1 ks. The relationships between the seepage velocity are:
Figure 9. (a) Geometry and finite difference mesh of instrumented site (Ng et al. 2003), (b) Comparison of measured and simulated values of pore pressure distribution at
a depth of 0.6 m from berm (R2), (c) Comparison of measured and simulated values of pore pressure distribution at a depth of 1.2 m from berm (R2).
INTERNATIONAL JOURNAL OF GEOTECHNICAL ENGINEERING 11
Table 3. Engineering and hydraulic properties of the soil (Ng et al. 2003). the present study have greater promise in the practical stability
Soil properties Measured values analysis of the rain-induced earth slope instability. Actual rainfall
Mechanical parameters data can be used in the simulation of pore pressure generation
Unit weight (kN/m3), γd 15.2 thereby giving the information about the stability of the slope on
Elastic modulus (kPa), E 10000 real-time basis and the same can be utilised in the early warning
Poisson’s ratio, υ 0.33
Effective cohesion(kPa), c′ 16.7 system. This information can be used to advantage in the land-
Effective friction angle, degrees 28.7 slide hazard assessment. Based on the results obtained for the
Hydraulic parameters chosen slope section, soil properties and rainfall intensities, the
Saturated permeability (m/s), ks 3.6 × 10−7
Scaling suction (kPa), P0 5 following conclusions are drawn:
Shape parameter, m 0.1
• During rainfall infiltration, the pore pressures increased
due to the dissipation of matric suction and increase in the
2 groundwater table corresponding to the antecedent condi-
vs = 2.07098 − 3.21279 Ydisp + 44.15854 Ydisp tions of the slope. For the slope with loose soil condition,
(17)
3
− 206.368 Ydisp 4
− 293.7865 Ydisp the increase in pore-water pressure is more as compared
to the slopes with medium and dense soil conditions for
all the rainfall intensities. The rate of change in pore-water
pressure is faster for the slope with loose soil condition
( )
1
vs = 2 (18) followed by the slopes with medium and dense soil condi-
Downloaded by [University of Florida] at 07:48 04 August 2017
Disclosure statement
Limitation No potential conflict of interest was reported by the authors.
The present study has considered uniform rainfall events and
specific soil types for the slope and accordingly the numerical ORCID
simulations have been conducted using the FLAC. However, it is
possible to include different rainfall patterns with non-uniform Adarsh S. Chatra http://orcid.org/0000-0002-3595-888X
rainfall intensities in the numerical simulation of the rain-in-
duced slope instability. References
Ali, A., J. Huang, A. V. Lyamin, S. W. Sloan, and M. J. Cassidy. 2014.
List of symbols “Boundary Effects Rain-Induced Landslide.” Computers and Geotechnics
61: 341–354.
c′ Cohesion of soil Bishop, A. W. 1959. “The Principle of Effective Stress.” Teknisk Ukeblad 106
Downloaded by [University of Florida] at 07:48 04 August 2017
(39): 859–863.
c′trial Trial cohesion of soil Bowles, J. E. 1997. Foundation Analysis and Design. International ed.
D(θ) Soil water diffusivity Singapore: McGraw-Hill.
Dr Relative density Cai, F., and K. Ugai. 2004. “Numerical Analysis of Rainfall Effects on Slope
Stability.” International Journal of Geomechanics 4 (2): 69–78.
E Young’s modulus Chen, L., and M. H. Young. 2006. “Green-Ampt Infiltration Model for
Ftrial Trial factor of safety Sloping Surfaces.” Water Resources Research 42: 1–9.
FDA Finite difference analysis Chow, V. T., D. R. Maidment, and L. W. Mays. 1988. Applied Hydrology.
New York: McGraw-Hill.
FOS Factor of safety
Dawson, E. M., W. H. Roth, and A. Drescher. 1999. “Slope Stability Analysis
k(θ) Permeability of soil by Strength Reduction.” Geotechnique 49 (6): 835–840.
ks Saturated permeability (hydraulic conductivity) Dodagoudar, G. R., and G. Venkatachalam. 2001. “Rain-Induced Failures
kr Relative permeability in Residual Soil Slopes.” Journal of Nepal Geological Society 22: 49–54.
FLAC Manual. 2011. Fluid-Mechanical Interaction. Minneapolis: Itasca
κ Mobility coefficient Consulting Group
Ks Bulk modulus of water Ganapathy, G. P., K. Mahendran, and S. K. Sekar. 2010. “Need and Urgency
Kg Bulk modulus of air of Landslide Risk Planning for Nilgiris District, Tamil Nadu State,
India.” International Journal of Geomatics and Geosciences 1 (1): 29–40.
m van Genuchten parameter Gasmo, J. M., H. Rahardjo, and E. C. Leong. 2000. “Infiltration Effects on
n Porosity of soil Stability of a Residual Soil Slope.” Computers and Geotechnics 26: 145–
P0 Scaling suction 165.
van Genuchten, M Th. 1980. “A Closed-Form Equation Predicting the
Pg Pore-air pressure
Hydraulic Conductivity of Unsaturated Soils.” Soil Science Society of
Ps Pore-water pressure America 44: 892–898.
qis Fluid flow of water Griffiths, D. V., and P. A. Lane. 1999. “Slope Stability Analysis by Finite
qig Fluid flow of air Elements.” Geotechnique 49 (3): 387–403.
Green, W. H., and C. A. Ampt. 1911. “Studies on Soil Physics: Flow of Air
SSR Shear strength reduction and Water through Soils.” Journal of Agricultural Science 4: 1–24.
Rc Relative compaction Hendron, A. J., and F. D. Patton. 1987. “The Vaiont Slide – A Geotechnical
t Rainfall duration, time Analysis Based on New Geologic Observations of the Failure Surface.”
Engineering Geology 24: 475–491.
vs Seepage velocity Iverson, R. M. 2000. “Landslide Triggering by Rainfall Infiltration.” Water
Xdisp Horizontal displacement Resources Research 36 (7): 1897–1910.
Ydisp Vertical displacement Jaiswal, P., C. J. van Westen, and V. Jetten. 2011. “Quantitative Assessment
of Landslide Hazard along Transportation Lines Using Historical
z Elevation head
Records.” Landslides 8: 279–291.
θ Volumetric water content Lee, K. L., and A. Singh. 1971. “Relative Density and Relative Compaction
ω Suction head of Soils.” Journal of the Soil Mechanics and Foundations Division 97 (7):
Θ Relative degree of saturation 1049–1052.
Lumb, P. 1962. “Effect of Rain Storms on Slope Stability.” Proceedings of the
θs Saturated volumetric water content Symposium on Hong Kong Soils, Hong Kong, 73–87.
θr Residual volumetric water content Lumb, P. 1975. “Slope Failures in Hong Kong.” Quarterly Journal of
θg Fluid saturation of air Engineering Geology 8: 31–65.
Mualem, Y. 1976. “A New Model for Predicting the Hydraulic Conductivity
ε Volumetric strain of Unsaturated Porous Media.” Water Resources Research 12: 513–522.
τmax Shear strength of unsaturated soil Muntohar, A. M., and H. J. Liao. 2010. “Rainfall Infiltration: Infinite Slope
φ′ Friction angle of soil Model for Landslides Triggering by Rainstorm.” Natural Hazards 54:
967–984.
ϕ′trial Trial friction angle of soil
INTERNATIONAL JOURNAL OF GEOTECHNICAL ENGINEERING 13
Ng, C. W. W., and Q. Shi. 1998. “A Numerical Investigation of the Stability Richards, L. A. 1931. “Capillary Conductions of Liquids through Porous
of Unsaturated Soil Slopes Subjected to Transient Seepage.” Computers Mediums.” Journal of Applied Physics 1 (5): 318–333.
and Geotechnics 22 (1): 1–28. Suradi, M., and A. Fourie. 2014. “The Effect of Rainfall Patterns on the
Ng, C. W. W., L. T. Zhan, C. G. Bao, D. G. Fredlund, and B. W. Gong. 2003. Mechanism of Shallow Slope Failure.” Aceh International Journal of
“Performance of an Unsaturated Expansive Soil Slope Subjected to Science and Technology 3 (1): 1–18.
Artificial Rainfall Infiltration.” Geotechnique 53 (2): 143–157. Tsaparas, I., H. Rahardjo, D. G. Toll, and E. Leong. 2003. “Infiltration
Philip, J. R. 1991. “Hillslope Infiltration: Planar Slopes.” Water Resources Characteristics of Two Instrumented Residual Soil Slopes.” Canadian
Research 27 (1): 109–117. Geotechnical Journal 40: 1012–1032.
Pradel, D., and G. Raad. 1993. “Effect of Permeability on Surficial Stability Ugai, K. 1989. “A Method of Calculation of Global Safety Factor of Slopes
of Homogeneous Slopes.” Journal of Geotechnical Engineering 109 (1): by Elasto-Plastic FEM.” Soils and Foundations 29 (2): 190–195.
62–70. Xue, K., B. Ajmera, B. Tiwari, and Y. Hu. 2016. “Effect of Long Duration
Rahardjo, H., E. C. Leong, and R. B. Rezaur. 2008. “Effect of Antecedent Rainstorm on Stability of Red-Clay Slopes.” Geoenvironmental Disasters
Rainfall on Pore-Water Pressure Distribution Characteristics in 3 (12): 1–13.
Residual Soil Slopes under Tropical Rainfall.” Hydrological Processes 22: Zienkiewicz, O. C., C. Humpheson, and R. W. Lewis. 1975. “Associated
506–523. and Non-Associated Visco-Plasticity and Plasticity in Soil Mechanics.”
Rahimi, A., H. Rahardjo, and E. Leong. 2011. “Effect of Antecedent Rainfall Geotechnique 25 (4): 671–689.
Patterns on Rainfall-Induced Slope Failure.” Journal of Geotechnical and
Geoenvirontmental Engineering 137 (5): 483–491.
Downloaded by [University of Florida] at 07:48 04 August 2017