Application of Capillary Barrier Systems for Slope Stabilization Under Extreme Rainfall: A Case Study of National Highway 10, India

Abstract

Global warming has led to an increase in extreme rainfall events, which often result in landslides, posing significant threats to infrastructure and human life. This study evaluated the effectiveness of the Capillary Barrier System (CBS) in enhancing slope stability along a vulnerable section of India’s National Highway 10 (NH10) during maximum daily rainfall. The GEOtop model was employed to conduct water balance simulations and obtain the pore–water pressure (PWP), which was then used to calculate the Factor of Safety (FoS). Results showed that CBS effectively delayed the rise in PWP, leading to lower peak values and smaller areas of very high and high risk levels. Spatial distribution mapping further confirmed that CBS minimized very high risk zones. At three historical landslide points, CBS slopes generally maintained FoS values above 1, demonstrating enhanced stability and improved resilience to extreme rainfall. These findings highlight the potential of CBS as a viable strategy for slope reinforcement in regions susceptible to heavy rainfall.

Full text

Citation: Cheng, Y.; Li, Y. Application
of Capillary Barrier Systems for Slope
Stabilization Under Extreme Rainfall:
A Case Study of National Highway 10,
India. Infrastructures 2024,9, 201.
https://doi.org/10.3390/
infrastructures9110201
Academic Editor: Francesca Dezi
Received: 24 September 2024
Revised: 30 October 2024
Accepted: 8 November 2024
Published: 10 November 2024
Copyright: © 2024 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Article
Application of Capillary Barrier Systems for Slope Stabilization
Under Extreme Rainfall: A Case Study of National Highway
10, India
Yusen Cheng 1and Yangyang Li 1,2,*
1Suzhou Industrial Park Monash Research Institute of Science and Technology, Monash University,
Suzhou 215000, China; e0792203@u.nus.edu
2Department of Civil Engineering, Monash University, 23 College Walk, Clayton,
Melbourne, VIC 3800, Australia
*Correspondence: yangyang.li@monash.edu
Abstract: Global warming has led to an increase in extreme rainfall events, which often result in
landslides, posing significant threats to infrastructure and human life. This study evaluated the
effectiveness of the Capillary Barrier System (CBS) in enhancing slope stability along a vulnerable
section of India’s National Highway 10 (NH10) during maximum daily rainfall. The GEOtop model
was employed to conduct water balance simulations and obtain the pore–water pressure (PWP),
which was then used to calculate the Factor of Safety (FoS). Results showed that CBS effectively
delayed the rise in PWP, leading to lower peak values and smaller areas of very high and high risk
levels. Spatial distribution mapping further confirmed that CBS minimized very high risk zones. At
three historical landslide points, CBS slopes generally maintained FoS values above 1, demonstrating
enhanced stability and improved resilience to extreme rainfall. These findings highlight the potential
of CBS as a viable strategy for slope reinforcement in regions susceptible to heavy rainfall.
Keywords: Capillary Barrier System; extreme rainfall; slope stability; slope stabilization; GEOtop;
physical model
1. Introduction
In recent times, global warming has significantly altered weather patterns, leading
to an increase in both the frequency and intensity of extreme precipitation events world-
wide
[1,2]
. Such extreme rainfall events have been identified as a significant trigger for slope
failures and landslides in many countries, including Australia [
3
], China [
4
,
5
], India [
6
,
7
],
Italy [
8
,
9
], and Singapore [
10
,
11
], especially in mountainous and hilly regions. When rain-
fall infiltrates the soil, it decreases the matric suction, which refers to the difference between
pore–air pressure and pore–water pressure (PWP) in unsaturated soils and contributes
to the soil’s shear strength. The decrease in matric suction leads to a reduction in shear
strength and an increase in PWP, making the slope more susceptible to failure. The impact
of such events is expected to become more severe due to ongoing climate change [
12
]. Con-
sequently, there is an increasing need for effective slope stabilization measures, particularly
in areas of critical importance, such as around highways and residential zones, to prevent
catastrophic failures and protect human life and infrastructure.
Traditional slope stabilization techniques, such as retaining walls, soil nailing, and
drainage systems, have been employed to reinforce slopes. While effective in some sce-
narios, these methods can be expensive, environmentally disruptive, and may not always
provide satisfactory protection under extreme rainfall conditions [
13
]. The Capillary Barrier
System (CBS) is a geotechnical engineering technique designed to reduce water infiltration
and enhance slope stability [
14
,
15
]. It consists of a fine
−
grained soil layer placed over a
coarse
−
grained soil layer. Under unsaturated conditions, the coarse
−
grained layer exhibits
Infrastructures 2024,9, 201. https://doi.org/10.3390/infrastructures9110201 https://www.mdpi.com/journal/infrastructures

Infrastructures 2024,9, 201 2 of 20
lower permeability than the fine
−
grained layer, creating a barrier effect that restricts water
flow only in fine
−
grained soil. During rainfall, the infiltrated water can be retained within
the fine
−
grained soil or, depending on the slope of the terrain, flow laterally toward the
toe of the slope [
16
]. The accumulated water can be removed through processes such as
evaporation, transpiration, and lateral drainage [
17
]. This mechanism prevents excessive
rainwater from infiltrating into the residual soil of the slope, thereby enhancing slope
stability and reducing the risk of landslides. However, once the fine
−
grained soil becomes
saturated, the capillary barrier effect disappears, leading to a breakthrough where water
infiltrates the coarse−grained and residual soil layers [15].
Many studies have demonstrated the effectiveness of CBS in reducing water infiltration
during rainfall events, thereby contributing to slope stabilization [
18
–
20
]. Scarfone et al. [
21
]
modeled the behavior of a 6
−
meter
−
high slope, demonstrating that CBS effectively main-
tains higher soil suction and lower saturation levels, thus preventing rainfall
−
induced
slope instability. Gao et al. [
22
] employed the Green
−
Ampt model and the Janbu method
to analyze the infiltration process and stability of inclined capillary barrier covers under
rainfall conditions, validating via a case study using an experimental slope model measur-
ing 270 cm in length and 60 cm in width. In addition, some researchers have compared
CBS with other slope stabilization methods to further demonstrate its effectiveness. For
instance, Rahardjo et al. [23] conducted a comparative study between CBS and vegetative
slopes, showing that CBS maintained lower PWP within the slope during both low
−
and
high
−
intensity rainfall events, thereby maintaining higher soil suction and stability than
vegetated slopes. Similarly, Li et al. [
24
] compared CBS with vegetative covers using Vetiver
grass, indicating that the Factor of Safety (FoS) for CBS slopes was significantly higher than
that of slopes reinforced with vegetation.
These studies have provided valuable insights into the mechanisms and effectiveness
of CBS, but they predominantly focus on small
−
scale applications at specific locations, as
well as laboratory experiments and numerical modeling based on idealized slope models.
However, research on CBS’s application at a regional scale remains relatively scarce, despite
its importance for ensuring the stability and safety of large infrastructure projects, such
as roadside slopes along critical highways. Few studies have evaluated the impact of CBS
on slope stability across larger and more complex terrains. Furthermore, there is a limited
focus on the spatial characteristics of slope stability variations with CBS implementation.
Therefore, this study aims to evaluate CBS’s performance in a regional context under
extreme rainfall. By utilizing physical modeling and simulating maximum daily rainfall
events, this research analyzes the variations in PWP and FoS, as well as further investigating
the spatial distribution of slope stability with and without CBS application along a critical
section of the selected highway. The outcomes are expected to provide valuable insights
into the broader application of CBS for regional slope stabilization and landslide risk
mitigation.
2. Study Area
India’s National Highway 10 (NH10) is a crucial transportation route, extending
approximately 174 km and connecting cities such as Siliguri in West Bengal and Gangtok in
Sikkim. NH10 plays a significant role in the regional economy and transportation network.
This study focuses on a specific section of NH10, extending from the Kalimpong Dansong
Forest in the north to the Sitong Forest in the south, with a distance of around 29.5 km
alongside the Tista River. A 3 km buffer zone around this segment has been established
as the study area (Figure 1). This 3 km buffer zone was determined based on a global
rainfall
−
induced landslide catalog [
25
,
26
]. This selected buffer can effectively include
most of the documented landslide points along the NH 10 segment (Figure 1), allowing for
the slope stability assessment by considering the highway’s surrounding slopes that have
historically been susceptible to landslides.

Infrastructures 2024,9, 201 3 of 20
Infrastructures 2024, 9, x FOR PEER REVIEW 3 of 23
The study area mainly encompasses a complex mountainous region. The elevation in
this area is highly varied, ranging from around 100 m to 1410 m above sea level, increasing
from south to north. The geomorphology is characterized by rugged mountains, steep
scarp faces, and river valleys that deepen into steep gorges towards the trunk rivers [27].
The geological feature of the study area is defined by the Daling Group, including the
Reyang and Gorubathan Formations, which are mainly composed of metamorphic rocks
like phyllites, quarites, and schists [28]. According to the Köppen climate classification
[29], the study area falls under the subtropical highland climate (Cwb). The annual pre-
cipitation ranges from 2000 to 4000 mm, mostly during the monsoon season between June
and September [30]. The intense and prolonged rainfall during the monsoon periods fre-
quently triggers landslides, highlighting the necessity for effective slope stabilization
measures to maintain the stability and reliability of this vital transportation corridor.
Figure 1. Overview of the study area. (a) The full extent of India’s NH10. (b) Study section of NH10
with a 3 km buffer zone.
3. Methods
3.1. Rainfall Paerns and Topographical Data
Extreme rainfall refers to rainfall events that exceed the typical levels for a particular
region over a specific period. Many studies have used daily precipitation that exceeds a
certain statistical threshold to represent extreme daily rainfall [31,32]. Due to the chal-
lenges of precisely determining this threshold, this study used the maximum daily rainfall
of 304 mm/day, equivalent to 12.67 mm/hour, recorded in 1968, as a representative meas-
ure to simulate extreme rainfall events [33]. According to the India Climate & Energy
Figure 1. Overview of the study area. (a) The full extent of India’s NH10. (b) Study section of NH10
with a 3 km buffer zone.
The study area mainly encompasses a complex mountainous region. The elevation in
this area is highly varied, ranging from around 100 m to 1410 m above sea level, increasing
from south to north. The geomorphology is characterized by rugged mountains, steep scarp
faces, and river valleys that deepen into steep gorges towards the trunk rivers [
27
]. The
geological feature of the study area is defined by the Daling Group, including the Reyang
and Gorubathan Formations, which are mainly composed of metamorphic rocks like
phyllites, quartzites, and schists [
28
]. According to the Köppen climate classification [
29
],
the study area falls under the subtropical highland climate (Cwb). The annual precipitation
ranges from 2000 to 4000 mm, mostly during the monsoon season between June and
September [
30
]. The intense and prolonged rainfall during the monsoon periods frequently
triggers landslides, highlighting the necessity for effective slope stabilization measures to
maintain the stability and reliability of this vital transportation corridor.
3. Methods
3.1. Rainfall Patterns and Topographical Data
Extreme rainfall refers to rainfall events that exceed the typical levels for a particular
region over a specific period. Many studies have used daily precipitation that exceeds a
certain statistical threshold to represent extreme daily rainfall [
31
,
32
]. Due to the challenges
of precisely determining this threshold, this study used the maximum daily rainfall of 304
mm/day, equivalent to 12.67 mm/h, recorded in 1968, as a representative measure to simu-
late extreme rainfall events [
33
]. According to the India Climate & Energy Dashboard [
34
],

Infrastructures 2024,9, 201 4 of 20
the Darjeeling region, where the study area is located, typically exhibits a shallow water
level between 2 and 5 m below ground level (mbgl) during both the pre
−
monsoon and
post
−
monsoon seasons. In this study, we selected the median value of 3.5 m and assumed a
spatially homogeneous groundwater table to represent the initial conditions. Topographical
data for the study area were obtained from the ASTER Global Digital Elevation Model
(GDEM) V003 dataset [
35
]. A Digital Elevation Model (DEM) map with a resolution of 30
m was created within the study area, along with corresponding slope and aspect layers
(Figure 2).
Infrastructures 2024, 9, x FOR PEER REVIEW 4 of 23
Dashboard [34], the Darjeeling region, where the study area is located, typically exhibits
a shallow water level between 2 and 5 m below ground level (mbgl) during both the
pre−monsoon and post−monsoon seasons. In this study, we selected the median value of
3.5 m and assumed a spatially homogeneous groundwater table to represent the initial
conditions. Topographical data for the study area were obtained from the ASTER Global
Digital Elevation Model (GDEM) V003 dataset [35]. A Digital Elevation Model (DEM) map
with a resolution of 30 m was created within the study area, along with corresponding
slope and aspect layers (Figure 2).
Figure 2. Maps used in GEOtop modeling: (a) DEM; (b) slope angle; (c) aspect.
3.2. Soil Properties and CBS Seings
The soil properties for this study were collected through a comprehensive literature
review, focusing on previous research conducted within or near the study area [36,37].
The soil water characteristic curve (SWCC) used in the referenced paper is characterized
by the Gardner equation [38] (Formula (1)). However, since the GEOtop model used in
this study for water balance simulations employs the Van Genuchten equation [39] (For-
mula (2)), a curve−fiing procedure was performed to convert the soil properties from the
Gardner format to the Van Genuchten format (Figure 3). This conversion was conducted
using a non−linear least square fiing method via the curve_fit function from the SciPy
library in Python. The procedure optimized the parameters αVG and nVG by minimizing
the difference between the VWC values calculated by the Gardner and Van Genuchten
equations. The data used for the fiing corresponded to matric suction values ranging
from 10−2 kPa to 103 kPa. While there may be differences between the two equations, the
close alignment of the two curves ensures the overall behavior of soil properties is well
represented. The fiing performance was evaluated using R2 and RMSE, with values of
0.9949 and 0.0099, respectively, indicating a strong fit between the curves. The residual
soil properties, assumed to be homogeneous, used in this study are presented in Table 1.
𝜃=𝜃+󰇛𝜃−𝜃
󰇜e (1)
where 𝜃 represents the volumetric water content (VWC) under a specific soil suction 𝜓,
𝜃 is the residual volumetric water content, 𝜃 is the saturated volumetric water content,
and 𝛼 is a shape parameter.
𝜃=𝜃+𝜃−𝜃

󰇛1+󰇛𝛼𝜓󰇜󰇜 (2)
where 𝛼 is related to the inverse of the air entry suction, 𝑛 is related to the pore size
distribution of the roil, and 𝑚 =1−1/𝑛.
The shear strength of unsaturated soil can be calculated by Formula (3) [40].
Figure 2. Maps used in GEOtop modeling: (a) DEM; (b) slope angle; (c) aspect.
3.2. Soil Properties and CBS Settings
The soil properties for this study were collected through a comprehensive literature
review, focusing on previous research conducted within or near the study area [
36
,
37
].
The soil water characteristic curve (SWCC) used in the referenced paper is characterized
by the Gardner equation [
38
] (Formula (1)). However, since the GEOtop model used
in this study for water balance simulations employs the Van Genuchten equation [
39
]
(F
ormula (2)
), a curve
−
fitting procedure was performed to convert the soil properties from
the Gardner format to the Van Genuchten format (Figure 3). This conversion was conducted
using a non
−
linear least square fitting method via the curve_fit function from the SciPy
library in Python. The procedure optimized the parameters
α
VG and nVG by minimizing
the difference between the VWC values calculated by the Gardner and Van Genuchten
equations. The data used for the fitting corresponded to matric suction values ranging
from 10
−2
kPa to 10
3
kPa. While there may be differences between the two equations, the
close alignment of the two curves ensures the overall behavior of soil properties is well
represented. The fitting performance was evaluated using R
2
and RMSE, with values of
0.9949 and 0.0099, respectively, indicating a strong fit between the curves. The residual soil
properties, assumed to be homogeneous, used in this study are presented in Table 1.
θ=θr+(θs−θr)eαGψ(1)
where
θ
represents the volumetric water content (VWC) under a specific soil suction
ψ
,
θr
is the residual volumetric water content,
θs
is the saturated volumetric water content, and
αGis a shape parameter.
θ=θr+θs−θr
1+(αVG ψ)nV G mVG (2)
where
αVG
is related to the inverse of the air entry suction,
nVG
is related to the pore size
distribution of the roil, and mVG =1−1/nVG .

Infrastructures 2024,9, 201 5 of 20
Infrastructures 2024, 9, x FOR PEER REVIEW 5 of 23
τ=𝑐󰆒+󰇛σ−𝑢󰇜tan ϕ󰆒+󰇛𝑢−𝑢󰇜tan ϕ (3)
where τ is the shear strength, 𝑐󰆒 is the effective cohesion, ϕ󰆒 is the effective friction an-
gle, ϕ is the angle of shearing resistance with respect to matric suction, 󰇛σ−𝑢󰇜 repre-
sents the net normal stress, and 󰇛𝑢−𝑢󰇜 represents the matric suction.
In addition, the selection of CBS parameters is critical in the numerical analysis to
simulate the reinforced slope and assess its effectiveness. In this study, a comprehensive
literature review was conducted, and CBS parameters were selected based on the research
by Rahardjo et al. [16]. The selected CBS consisted of 60 cm thick soil, with 40 cm of
fine−grained soil and 20 cm of coarse−grained soil. According to unsaturated soil mechan-
ics principles, this CBS configuration can theoretically create an effective capillary barrier
effect due to the difference between the hydraulic properties of fine− and coarse−grained
soils. The fine−grained layer possesses higher unsaturated permeability than the
coarse−grained layer within the relevant matric suction range. This difference in permea-
bility facilitates lateral rainwater flow within the fine−grained layer at the top, limiting
water infiltration into the underlying coarse−grained layer and residual soil, thereby en-
hancing slope stability. This CBS design aligned well with the study’s aim to mitigate wa-
ter infiltration during rainfall, which is critical for maintaining slope stability along the
vulnerable section of NH10. Notably, the SWCC parameters of CBS in the reference study
were described using the Fredlund and Xing equation (Formula (4)) [41].
𝜃=𝐶󰇛𝜓󰇜⋅𝜃
ln 𝑒+󰇡𝜓
𝑎󰇢



(4)
where 𝐶󰇛𝜓󰇜 is a correction factor typically set to 1 [42] and 𝑎, 𝑛, and 𝑚 are fiing
parameters related to the air−entry value of soil.
To ensure consistency with the GEOtop model, a curve−fiing procedure similar to
that for residual soil was also applied to convert these parameters to the Van Genuchten
format (Figure 4). For fine−grained soil, the R
2
was 0.9994 and the RMSE was 0.0045. For
coarse−grained soil, the R
2
was 0.9951 and the RMSE was 0.0085. The CBS properties used
in this study are also summarized in Table 1.
Figure 3. Curve fiing for residual soil: conversion from the Gardner equation to the Van Genuchten
equation.
Figure 3. Curve fitting for residual soil: conversion from the Gardner equation to the Van
Genuchten equation.
Table 1. Summary of residual soil and CBS (coarse−grained soil and fine−grained soil) properties.
Parameters Residual Soil Fine−Grained Soil Coarse−Grained Soil
Saturated hydraulic
conductivity (Ks)1.32 ×10−5m/s 1.2 ×10−6m/s 4.0 ×10−3m/s
Saturated water content (θs) 0.3962 0.387 0.437
Residual water content (θr) 0.0779 0.0179 0.0106
αG0.271 kPa−1− −
aFX −10 kPa 0.2 kPa
nFX −5 6
mFX −1.2 1.2
αVG 0.8899 kPa−18.78 ×10−2kPa−14.5042 kPa−1
nVG 1.9761 3.6714 3.8868
mVG 0.4940 0.7276 0.7427
Effective cohesion (c′) 0.65 kPa 0 0
Unit weight (γ)17.24 kN/m319.0 kN/m320.0 kN/m3
Saturated unit weight (γsat)20 kN/m325.393 kN/m327.762 kN/m3
Effective friction angle (φ′) 30◦34◦35◦
The shear strength of unsaturated soil can be calculated by Formula (3) [40].
τ=c′+(σ−ua)tan ϕ′+(ua−uw)tan ϕb(3)
where
τ
is the shear strength,
c′
is the effective cohesion,
ϕ′
is the effective friction angle,
ϕb
is the angle of shearing resistance with respect to matric suction,
(σ−ua)
represents the
net normal stress, and (ua−uw)represents the matric suction.
In addition, the selection of CBS parameters is critical in the numerical analysis to
simulate the reinforced slope and assess its effectiveness. In this study, a comprehen-
sive literature review was conducted, and CBS parameters were selected based on the
research by Rahardjo et al. [
16
]. The selected CBS consisted of 60 cm thick soil, with 4
0 c
m
of fine
−
grained soil and 20 cm of coarse
−
grained soil. According to unsaturated soil
mechanics principles, this CBS configuration can theoretically create an effective capil-
lary barrier effect due to the difference between the hydraulic properties of fine
−
and
coarse
−
grained soils. The fine
−
grained layer possesses higher unsaturated permeability
than the coarse
−
grained layer within the relevant matric suction range. This difference in
permeability facilitates lateral rainwater flow within the fine
−
grained layer at the top, lim-
iting water infiltration into the underlying coarse
−
grained layer and residual soil, thereby

Infrastructures 2024,9, 201 6 of 20
enhancing slope stability. This CBS design aligned well with the study’s aim to mitigate
water infiltration during rainfall, which is critical for maintaining slope stability along the
vulnerable section of NH10. Notably, the SWCC parameters of CBS in the reference study
were described using the Fredlund and Xing equation (Formula (4)) [41].
θ=C(ψ)·θs
nlnhe+ψ
aFX nFX iomFX (4)
where
C(ψ)
is a correction factor typically set to 1 [
42
] and
aFX
,
nFX
, and
mFX
are fitting
parameters related to the air−entry value of soil.
To ensure consistency with the GEOtop model, a curve
−
fitting procedure similar to
that for residual soil was also applied to convert these parameters to the Van Genuchten
format (Figure 4). For fine
−
grained soil, the R
2
was 0.9994 and the RMSE was 0.0045. For
coarse
−
grained soil, the R
2
was 0.9951 and the RMSE was 0.0085. The CBS properties used
in this study are also summarized in Table 1.
Infrastructures 2024, 9, x FOR PEER REVIEW 6 of 23
Figure 4. Curve fiing for CBS: conversion from the Fredlund and Xing equation to the Van Genuch-
ten equation. (a). Fine−grained soil. (b). Coarse−grained soil.
Table 1. Summary of residual soil and CBS (coarse−grained soil and fine−grained soil) properties.
Parameters Residual Soil
Fine−Grained
Soil
Coarse−Grained
Soil
Saturated hydraulic conductivity
(Ks) 1.32 × 10−5 m/s 1.2 × 10−6 m/s 4.0 × 10−3 m/s
Saturated water content (θs) 0.3962 0.387 0.437
Residual water content (θr) 0.0779 0.0179 0.0106
αG 0.271 kPa−1 − −
aFX − 10 kPa 0.2 kPa
nFX − 5 6
mFX − 1.2 1.2
αVG 0.8899 kPa−1 8.78×10−2 kPa−1 4.5042 kPa−1
nVG 1.9761 3.6714 3.8868
mVG 0.4940 0.7276 0.7427
Effective cohesion (c′) 0.65 kPa 0 0
Unit weight (γ) 17.24 kN/m3 19.0 kN/m3 20.0 kN/m3
Saturated unit weight (γsat) 20 kN/m3 25.393 kN/m3 27.762 kN/m3
Figure 4. Curve fitting for CBS: conversion from the Fredlund and Xing equation to the Van Genuchten
equation. (a). Fine−grained soil. (b). Coarse−grained soil.
3.3. GEOtop Model
GEOtop is a physically based distributed hydrological model used for water bal-
ance simulation that incorporates topographical, meteorological, and soil data [
43
]. It
employs a three
−
dimensional (3D) grid system to represent the spatial variability within
the study area, enabling the simulation of water movement through multiple heterogeneous
soil layers.

Infrastructures 2024,9, 201 7 of 20
The core of GEOtop’s water balance simulation is based on the 3D Richards’ equation
(Formula (5)) [44], which models unsaturated water flow in the soil.
∂θ(x,t)
∂t=∇·(K∇(z+ψ)) + S(x,t)(5)
where xrepresents the position, trepresents the time,
θ
is the volumetric water content, K
is the hydraulic conductivity, zis the gravitational head, ψis the soil water pressure head,
∇·(K∇(z+ψ))
represents the flux divergence of water per unit volume, and Srepresents
the water exchanges between atmosphere and soil, including evaporation and transpiration
processes.
In this study, the GEOtop model was configured with eight soil layers. A field survey
conducted by Gupta and Chattaraj [
27
] in the Kalimpong Dansong Forest, situated in
the southern part of the study area along NH10, identified very hard rock at a depth of
approximately 5 to 8 m below the surface. Based on this observation, layer 8 (located at a
depth of 5 m and below) was set as an impermeable rock layer to represent a boundary that
prevents further vertical water movement (Figure 5). This configuration also simplified the
simulations by assuming that deeper layers do not significantly contribute to water flow,
thereby allowing for a focus on surface and subsurface water movement, where shallow
landslides typically occur. The CBS was used to replace the top three layers (layer 1 to
layer 3) of the original slope. The simulation covered a 48 h time series with a 1 h time
step. It started with a 24 h maximum daily rainfall event with an intensity of 12.67 mm/h,
followed by another 24 h dry period to observe water drainage and recovery processes
within the soil layers. Using the modified CBS parameters, the water balance simulation
was repeated under the same rainfall patterns. Pore–water pressure maps were generated
on an hourly basis under two scenarios (i.e., with and without CBS) at the center of each
soil layer throughout the simulation (Figure 6). Details regarding the configuration of soil
and precipitation parameters can be found in Appendix A.
Infrastructures 2024, 9, x FOR PEER REVIEW 8 of 23
Figure 5. Schematic diagrams of the original slope and CBS slope. The red dots indicate the center
of each layer where the PWP values are calculated. (a). Original slope with eight residual soil layers.
(b). CBS slope with fine− and coarse−grained soil layers in the top three layers.
Figure 6. The workflow of the GEOtop model.
3.4. Factor of Safety (FoS) Calculation
To quantitatively assess the slope stability, the Factor of Safety (FoS) was calculated.
Given that extreme rainfall events may lead to ponding at the slope surface, the FoS cal-
culation distinguished between ponding and non−ponding conditions. The ponding was
determined by comparing the PWP to the calculated soil depth; if the PWP exceeds the
soil layer’s thickness, the condition was identified as ponding; otherwise, it was identified
as the non−ponding condition.
Figure 5. Schematic diagrams of the original slope and CBS slope. The red dots indicate the center of
each layer where the PWP values are calculated. (a). Original slope with eight residual soil layers.
(b). CBS slope with fine−and coarse−grained soil layers in the top three layers.

Infrastructures 2024,9, 201 8 of 20
Infrastructures 2024, 9, x FOR PEER REVIEW 8 of 23
Figure 5. Schematic diagrams of the original slope and CBS slope. The red dots indicate the center
of each layer where the PWP values are calculated. (a). Original slope with eight residual soil layers.
(b). CBS slope with fine− and coarse−grained soil layers in the top three layers.
Figure 6. The workflow of the GEOtop model.
3.4. Factor of Safety (FoS) Calculation
To quantitatively assess the slope stability, the Factor of Safety (FoS) was calculated.
Given that extreme rainfall events may lead to ponding at the slope surface, the FoS cal-
culation distinguished between ponding and non−ponding conditions. The ponding was
determined by comparing the PWP to the calculated soil depth; if the PWP exceeds the
soil layer’s thickness, the condition was identified as ponding; otherwise, it was identified
as the non−ponding condition.
Figure 6. The workflow of the GEOtop model.
3.4. Factor of Safety (FoS) Calculation
To quantitatively assess the slope stability, the Factor of Safety (FoS) was calculated.
Given that extreme rainfall events may lead to ponding at the slope surface, the FoS
calculation distinguished between ponding and non
−
ponding conditions. The ponding
was determined by comparing the PWP to the calculated soil depth; if the PWP exceeds the
soil layer’s thickness, the condition was identified as ponding; otherwise, it was identified
as the non−ponding condition.
For non−ponding conditions, the FoS was calculated using Formula (6) [45,46].
FoS =tan ϕ′
tan δ+c′−ψ(Z,t)γwtan ϕ′
γZsin δcos δ(6)
where
ϕ′
is the effective friction angle,
c′
is the effective cohesion,
ψ(Z,t)
is the PWP at
depth
Z
and time
t
,
γ
and
γw
are the unit weights of soil and water, respectively, and
δ
is
the slope angle.
For ponding conditions, Formula (7) [47] was used.
FoS =c′+Z(γsat −γw)cos2βtan ϕ′
γsatZsin βcos β(7)
where γsat is the saturated unit weight of soil.
4. Results
4.1. Simulation Results of Pore–Water Pressure (PWP)
Over the elapsed time, differences in PWP between the original slope and the CBS
slope were observed. The depths of 0.8 m, 1.25 m, and 1.75 m represented the central points
of soil layers 4 to 6 (Figure 5), as well as the depths in which the GEOtop model outputs
simulated PWP results. Layers 1 to 3 corresponded to the CBS design, while layers 7 and 8
were below the groundwater table and assumed to be saturated. Therefore, the variations
in PWP were analyzed specifically at the depths of 0.8 m, 1.25 m, and 1.75 m (Figure 7) to
capture the impact of CBS.
As rainfall began, the PWP in the original slope rapidly increased and reached close to
0 kPa at depths of 0.8 m and 1.25 m, indicating the approach of the saturated condition.
At 1.75 m, while the increase was also obvious, it did not reach saturation within the
simulation period due to the gradual infiltration of rainfall through the upper layers before
impacting the deeper soil. In contrast, the CBS slope showed a delayed response across
all depths, with a few hours’ lags before the PWP rises, especially noticeable in the deeper

Infrastructures 2024,9, 201 9 of 20
layers, including 1.25 m and 1.75 m. This difference indicated the CBS’s effectiveness in
impeding water infiltration and mitigating the increasing rate of PWP.
Infrastructures 2024, 9, x FOR PEER REVIEW 10 of 23
Figure 7. Variations in the average PWP for the original and CBS slopes at various depths.
4.2. Simulation Results of Slope Stability
In this study, the slope stability was categorized into four distinct categories based
on the FoS values (Table 2), following the guidelines provided by Rahardjo et al. [48]. In
engineering practice, a FoS of 1.5 or higher is generally required to ensure safety and is
considered to represent a low−risk condition. Conversely, a FoS value smaller than 1
means that the driving force exceeds the resisting force, indicating instability and a very
high risk. In addition, the values between 1 and 1.5 typically define high risk and moderate
risk levels, allowing for a more detailed assessment of slope stability. These risk categories
are applicable to a wide range of slopes, especially in regions where rainfall significantly
threatens slope stability like the study area. Figure 8 illustrates the percentage of pixels in
each category varies with time and depth for both the original slope and CBS slope (A
more detailed table can be found in Appendix B).
Table 2. Slope stability categories based on FoS values.
Slope Stability Categories FoS Range
Very high risk FoS ≤ 1
High risk 1 < FOS ≤ 1.25
Moderate risk 1.25 < FOS ≤ 1.5
Low risk FOS > 1.5
In the very high risk category (Figure 8a), the percentage of pixels in the original
slope increased rapidly, particularly at depths of 0.8 m and 1.25 m. At the 0.8 m depth, the
original slope showed a steep rise at around the 18 h mark after the rainfall started, peak-
ing at approximately 24 h with a percentage over 12%. This rapid increase indicated the
quick response of shallow soil layers to rainfall infiltration, leading to evident instability
conditions. At the 1.25 m depth, however, the rise in very high risk pixels occurred later,
mainly after the maximum daily rainfall event had ended (around the 24 h mark). This
delayed increase, which eventually reached nearly 14%, was driven by the continued
Figure 7. Variations in the average PWP for the original and CBS slopes at various depths.
After the end of the rainfall, the PWP at the depths of 1.25 m and 1.75 m continued
to increase. This increase can be attributed to the continued water infiltration from the
upper layers, as the extreme rainfall event typically has a prolonged effect on the deeper
soil layers. In the CBS slope, a similar upward trend was found in these two layers, and a
noticeable phenomenon was that the post
−
rainfall increase in PWP was more significant
than that of the original slope at 1.25 m. This was due to the delaying effect of CBS on water
infiltration, making the water reach this layer mainly after the end of rainfall. However,
although the increase was more apparent after the end of rainfall, the overall PWP increase
in the CBS slope throughout the simulation period remained smaller than that of the
original slope across all depths. In addition, for the shallower layer (0.8 m), the PWP in
both types of slopes showed a decreased tendency during the post
−
rainfall period due to
the continuous water infiltration, which aided in gradually returning the shallow soil to
the unsaturated condition.
4.2. Simulation Results of Slope Stability
In this study, the slope stability was categorized into four distinct categories based
on the FoS values (Table 2), following the guidelines provided by Rahardjo et al. [
48
]. In
engineering practice, a FoS of 1.5 or higher is generally required to ensure safety and
is considered to represent a low
−
risk condition. Conversely, a FoS value smaller than
1 mea
ns that the driving force exceeds the resisting force, indicating instability and a very
high risk. In addition, the values between 1 and 1.5 typically define high risk and moderate
risk levels, allowing for a more detailed assessment of slope stability. These risk categories
are applicable to a wide range of slopes, especially in regions where rainfall significantly
threatens slope stability like the study area. Figure 8illustrates the percentage of pixels
in each category varies with time and depth for both the original slope and CBS slope (A
more detailed table can be found in Appendix B).

Infrastructures 2024,9, 201 10 of 20
Table 2. Slope stability categories based on FoS values.
Slope Stability Categories FoS Range
Very high risk FoS ≤1
High risk 1 < FOS ≤1.25
Moderate risk 1.25 < FOS ≤1.5
Low risk FOS > 1.5
Infrastructures 2024, 9, x FOR PEER REVIEW 11 of 23
water infiltration from the upper layers, demonstrating the prolonged effects of rainfall
on the deeper soil layers even after the rainfall had stopped. For the deeper 1.75 m layer,
an increase in very high risk pixels did not become apparent until 12 h after the end of the
rainfall (around the 36 h mark), indicating a more delayed response. By contrast, the CBS
slope showed a noticeably later increase and a lower maximum percentage of very high
risk pixels across all depths. Especially at the 1.25 m and 1.75 m depths, the CBS slope did
not exhibit a significant rise.
For the high−risk category, the trend in the original slope was similar to the very high
risk category in terms of growth paern. The CBS also demonstrated its effectiveness in
delaying the increase in high−risk pixels and maintaining lower peak values, but the max-
imum percentage of pixels in this category was higher across all depths compared to the
very high risk category (Figure 8b). Additionally, the increase in the high−risk category
occurred earlier and reached its maximum much faster. After reaching the peak, a down-
ward trend was observed; this was also found in the moderate−risk category (Figure 8c).
This decline was characterized by a sharp decline followed by a more gradual decrease.
The sharp decline could be aributed to specific areas that have transited into higher risk
categories as the rainfall continues, driven by increased PWP. After the end of rainfall, the
variations in risk became more complex. For the shallower soil, especially at 0.8 m depth,
the drainage process and continued water infiltration to deeper layers led to a reduction
in its PWP, thereby enhancing slope stability in the upper layers. This can be confirmed
by Figure 8d, where an increase in the proportion of low−risk pixels was observed about
3 h after the end of rainfall. In contrast, for deeper soils such as the 1.75 m depth, the
continued infiltration after the end of rains can still lead to increased risk. However, this
increase was less pronounced in the CBS slope than in the original slope, indicating that
the CBS effectively slowed the risk increase. In addition, the CBS slope retained a larger
proportion of low−risk areas, particularly for the 1.25 m and 1.75 m depths, showing al-
most no sharp reduction in low−risk areas during the rainfall event.
Figure 8. Variations in the percentage of pixels for each risk category between the original slope and
CBS slope at various depths. (a). Very high risk. (b). High risk. (c). Moderate risk. (d). Low risk.
In the very high risk category (Figure 8a), the percentage of pixels in the original
slope increased rapidly, particularly at depths of 0.8 m and 1.25 m. At the 0.8 m depth,
the original slope showed a steep rise at around the 18 h mark after the rainfall started,
peaking at approximately 24 h with a percentage over 12%. This rapid increase indicated
the quick response of shallow soil layers to rainfall infiltration, leading to evident instability
conditions. At the 1.25 m depth, however, the rise in very high risk pixels occurred later,
mainly after the maximum daily rainfall event had ended (around the 24 h mark). This
delayed increase, which eventually reached nearly 14%, was driven by the continued water
infiltration from the upper layers, demonstrating the prolonged effects of rainfall on the
deeper soil layers even after the rainfall had stopped. For the deeper 1.75 m layer, an
increase in very high risk pixels did not become apparent until 12 h after the end of the
rainfall (around the 36 h mark), indicating a more delayed response. By contrast, the CBS
slope showed a noticeably later increase and a lower maximum percentage of very high
risk pixels across all depths. Especially at the 1.25 m and 1.75 m depths, the CBS slope did
not exhibit a significant rise.
For the high
−
risk category, the trend in the original slope was similar to the very
high risk category in terms of growth pattern. The CBS also demonstrated its effectiveness
in delaying the increase in high
−
risk pixels and maintaining lower peak values, but the

Infrastructures 2024,9, 201 11 of 20
maximum percentage of pixels in this category was higher across all depths compared
to the very high risk category (Figure 8b). Additionally, the increase in the high
−
risk
category occurred earlier and reached its maximum much faster. After reaching the peak,
a downward trend was observed; this was also found in the moderate
−
risk category
(Figure 8c). This decline was characterized by a sharp decline followed by a more gradual
decrease. The sharp decline could be attributed to specific areas that have transited into
higher risk categories as the rainfall continues, driven by increased PWP. After the end
of rainfall, the variations in risk became more complex. For the shallower soil, especially
at 0.8 m depth, the drainage process and continued water infiltration to deeper layers
led to a reduction in its PWP, thereby enhancing slope stability in the upper layers. This
can be confirmed by Figure 8d, where an increase in the proportion of low
−
risk pixels
was observed about 3 h after the end of rainfall. In contrast, for deeper soils such as the
1.75 m depth, the continued infiltration after the end of rains can still lead to increased
risk. However, this increase was less pronounced in the CBS slope than in the original
slope, indicating that the CBS effectively slowed the risk increase. In addition, the CBS
slope retained a larger proportion of low
−
risk areas, particularly for the 1.25 m and 1.75 m
depths, showing almost no sharp reduction in low−risk areas during the rainfall event.
4.3. Spatial Risk Distribution
To analyze the spatial distribution of each slope stability category, this study mapped
the risk distribution at a depth of 0.8 m for the original slope and CBS slope at four different
rainfall stages (Figure 9). At 12 h of rainfall, almost the entire study area, whether for
the original or the CBS slope, was classified as low risk. However, at 24 h of rainfall, the
differences between the two types of slopes became evident. Due to the CBS’s ability to
delay the increase in PWP and reduce water infiltration, only a minimal portion of the CBS
slope exhibited very high risk. In contrast, the original slope had a widespread distribution
of very high risk areas throughout the study region, indicating a transition to unstable
conditions under extreme rainfall. As the rainfall and infiltration continued, the CBS slope
showed an increase in very high risk areas, gradually expanding in spatial distribution.
Despite this growth, the extent of very high risk areas in the CBS slope remained smaller
compared to the original slope. The spatial pattern of these high
−
risk areas became similar
between the two types of slopes, but the CBS’s mitigating effect was evident in the reducing
regions affected by the rainfall. The CBS slope allowed areas that would otherwise be
classified as very high risk and high risk under extreme rainfall conditions to remain in
the moderate
−
risk or even low
−
risk categories. After 24 h of rainfall end, the spatial
distribution of risk showed little visual difference from the pattern observed 12 h earlier.
This character aligns with the results in Figure 8, where reductions in high risk and very
high risk categories, as well as increases in low
−
risk areas, were highly gradual during
this period, suggesting that the recovery of the slope stability after extreme rainfall was a
slow process.
To further validate the effectiveness of the CBS, three historical landslide points were
selected in the northern, central, and southern parts of the study area (Figure 10 and Table 3)
based on the global rainfall
−
induced landslide catalog [
25
,
26
] for detailed analysis. For
each point, a 100 m radius buffer zone was established, and the variations in the mean FoS
within these buffers were calculated (Figure 11). The three landslide points demonstrated
similar patterns in the FoS variations across different depths and rainfall timesteps.
While the risk distribution at 1.25 m showed a similar trend to that of 0.8 m, the
distribution at 1.75 m exhibited a more significant delayed risk increase, with more apparent
differences between the CBS slope and the original slope emerging at 12 to 24 h after the
end of rainfall. The detailed risk distribution maps for the 1.25 m and 1.75 m depths can be
found in Appendix C.

Infrastructures 2024,9, 201 12 of 20
Infrastructures 2024, 9, x FOR PEER REVIEW 13 of 23
Figure 9. Risk distribution at a depth of 0.8 m for original slope and CBS slope at different rainfall stages.
Figure 9. Risk distribution at a depth of 0.8 m for original slope and CBS slope at different
rainfall stages.
5. Discussion
5.1. Effectiveness of CBS in Enhancing Slope Stability
The results of this study, including statistical analysis and spatial risk mapping, have
demonstrated the effectiveness of the CBS in reducing the extent and severity of slope
instability risk, thereby enhancing slope stability under extreme rainfall. It significantly
reduced the proportion of areas categorized as very high and high
−
risk, with the CBS
slope showing lower peak proportions and a more gradual increasing rate. In contrast, the
CBS slope maintained a higher proportion of the low
−
risk area after extreme rainfall across
all depths, indicating a more stable slope condition throughout the event. These findings
align with the theoretical basis of CBS, which relies on the differing hydraulic properties of
fine
−
and coarse
−
grained soils to create a barrier effect. By impeding water infiltration,
CBS can mitigate rapid increases in PWP, as demonstrated in Section 4.1, thus maintaining
higher FoS values and reducing areas categorized as high−risk or very high risk.
Table 3. Information of the three selected landslide points.
Landslide Points Longitude Latitude
Minimum FoS Value for Each Depth
0.8 m 1.25 m 1.75 m
OS CBS OS CBS OS CBS
North point 88.4308 27.0713 0.86 1.00 0.88 1.23 0.87 0.87
Central point 88.4301 27.0257 0.78 0.97 0.92 1.23 0.78 0.78
South point 88.4481 26.9722 0.83 0.93 0.82 1.16 0.90 0.91
Notes: OS indicates the original slope. CBS indicates the CBS slope.

Infrastructures 2024,9, 201 13 of 20
Infrastructures 2024, 9, x FOR PEER REVIEW 14 of 23
5. Discussion
5.1. Effectiveness of CBS in Enhancing Slope Stability
The results of this study, including statistical analysis and spatial risk mapping, have
demonstrated the effectiveness of the CBS in reducing the extent and severity of slope
instability risk, thereby enhancing slope stability under extreme rainfall. It significantly
reduced the proportion of areas categorized as very high and high−risk, with the CBS
slope showing lower peak proportions and a more gradual increasing rate. In contrast,
the CBS slope maintained a higher proportion of the low−risk area after extreme rainfall
across all depths, indicating a more stable slope condition throughout the event. These
findings align with the theoretical basis of CBS, which relies on the differing hydraulic
properties of fine− and coarse−grained soils to create a barrier effect. By impeding water
infiltration, CBS can mitigate rapid increases in PWP, as demonstrated in Section 4.1, thus
maintaining higher FoS values and reducing areas categorized as high−risk or very high
risk.
To further validate the effectiveness of the CBS, three historical landslide points were
selected in the northern, central, and southern parts of the study area (Figure 10 and Table
3) based on the global rainfall−induced landslide catalog [25,26] for detailed analysis. For
each point, a 100 m radius buffer zone was established, and the variations in the mean FoS
within these buffers were calculated (Figure 11). The three landslide points demonstrated
similar paerns in the FoS variations across different depths and rainfall timesteps.
Figure 10. The location of each landslide point.
In general, the CBS slope demonstrated an obvious stabilization effect for all three
selected historical landslide points, as indicated by consistently higher minimum FoS
Figure 10. The location of each landslide point.
Infrastructures 2024, 9, x FOR PEER REVIEW 15 of 23
values compared to the original slope across all depths (Table 3). Specifically, at the depths
of 0.8 m and 1.25 m, the original slopes of all three landslide points exhibited a notable
decrease in the FoS during the rainfall event, with values dropping below 1, indicating a
very high risk. In contrast, the CBS−reinforced slopes demonstrated significantly im-
proved slope stability. At these depths, the FoS for the CBS slopes mostly remained above
1, illustrating that the CBS effectively prevented the slopes from reaching critical risk cat-
egories under extreme rainfall conditions. Furthermore, the implementation of the CBS
led to smaller variation ranges of FoS values, meaning that the difference between the
maximum and the minimum of FoS values was smaller than that of the original slopes.
This reduced variability indicated a more stable slope condition and a more controlled
and resilient response to extreme rainfall events. At the 1.75 m depth, however, both the
original and CBS slopes across the three landslide points showed no significant differences
in FoS.
These findings further confirmed the CBS’s effectiveness in enhancing slope stability
under extreme rainfall conditions. The improvement in FoS values across different depths
demonstrated its ability to mitigate landslide risks, particularly at sites that are susceptible
to failure.
Table 3. Information of the three selected landslide points.
Landslide Points Longitude Latitude
Minimum FoS Value for Each Depth
0.8 m 1.25 m 1.75 m
OS CBS OS CBS OS CBS
North point 88.4308 27.0713 0.86 1.00 0.88 1.23 0.87 0.87
Central point 88.4301 27.0257 0.78 0.97 0.92 1.23 0.78 0.78
South point 88.4481 26.9722 0.83 0.93 0.82 1.16 0.90 0.91
Notes: OS indicates the original slope. CBS indicates the CBS slope.
Figure 11. Variations in the FoS for each landslide point. (a). North point. (b). Central point.
(c). South point.

Infrastructures 2024,9, 201 14 of 20
In general, the CBS slope demonstrated an obvious stabilization effect for all three
selected historical landslide points, as indicated by consistently higher minimum FoS values
compared to the original slope across all depths (Table 3). Specifically, at the depths of
0.8 m
and 1.25 m, the original slopes of all three landslide points exhibited a notable decrease
in the FoS during the rainfall event, with values dropping below 1, indicating a very high
risk. In contrast, the CBS
−
reinforced slopes demonstrated significantly improved slope
stability. At these depths, the FoS for the CBS slopes mostly remained above 1, illustrating
that the CBS effectively prevented the slopes from reaching critical risk categories under
extreme rainfall conditions. Furthermore, the implementation of the CBS led to smaller
variation ranges of FoS values, meaning that the difference between the maximum and
the minimum of FoS values was smaller than that of the original slopes. This reduced
variability indicated a more stable slope condition and a more controlled and resilient
response to extreme rainfall events. At the 1.75 m depth, however, both the original and
CBS slopes across the three landslide points showed no significant differences in FoS.
These findings further confirmed the CBS’s effectiveness in enhancing slope stability
under extreme rainfall conditions. The improvement in FoS values across different depths
demonstrated its ability to mitigate landslide risks, particularly at sites that are susceptible
to failure.
5.2. Recommendations for Future Work
Despite the overall findings demonstrating that the selected CBS design significantly
enhanced slope stability, the analysis revealed certain aspects that require further inves-
tigation. First, at depths of 1.25 m and 1.75 m, a slightly higher initial percentage of very
high risk pixels in the CBS slope compared to the original slope at this depth was observed
in each risk category (Figure 8). This unusual phenomenon might be attributed to the
CBS properties rather than the negligible differences in matric suction. Specifically, the
CBS material had a higher unit weight compared to the residual soil, contributing to a
lower initial FoS value as the FoS is influenced by the product of unit weight and depth
(Formulas (6) and (7)). The deeper layers, as well as the higher unit weight of the CBS
material compared to the residual soil, may initially exhibit lower FoS values, even before
significant infiltration occurs. This reflects the CBS’s material properties and influences
on risk distribution regarding soil depth, highlighting the importance of appropriately
selecting CBS material configurations to avoid introducing new, human−induced risks.
In addition, for the three historical landslide points, although CBS showed obvious
stabilization effect, the FoS at these landslide points still did not exceed 1.5 (the threshold
for low risk) even after CBS implementation. This outcome may also be attributed to the
CBS design, where the selected parameters and layering settings may not be optimal.
However, this study did not include a comparative study of different CBS designs. The
primary aim was to assess the effectiveness and general impact of CBS on slope stability
rather than to optimize CBS designs, which would involve more extensive simulations
and require a broader range of CBS parameters to comprehensively evaluate different
designs. Future research can expand on these findings to explore more tailored solutions
within this area. Researchers have shown that CBS performance can vary significantly
depending on the soil properties used in CBS design. For instance, Qian et al. [
49
] evaluated
different material combinations and concluded that the hydraulic conductivity and grain
size distribution of the soil layers play a significant role in determining the CBS’s ability
to prevent water infiltration. Vishnu and Bharat [
50
] developed a numerical model to
assess the impact of soil hydraulic characteristics on CBS performance, emphasizing the
importance of selecting optimal material combinations to maximize the CBS’s effective-
ness. Therefore, selecting optimal CBS material and adjusting layers’ combinations could
maximize the effectiveness of CBS under different conditions. Future work could focus on
optimizing CBS material properties and configuration to suit specific study areas, thereby
further enhancing slope stability and mitigating any potential initial risk caused by higher
unit weight or other CBS materials’ properties.

Infrastructures 2024,9, 201 15 of 20
Moreover, employing CBS on a regional scale also needs to consider the economic and
environmental issues. For example, Rahardjo et al. explored the use of recycled crushed
concrete in CBS and found that this material could effectively reduce rainwater infiltration,
providing an environmentally sustainable solution [
51
]. Future research could also explore
the application of sustainable materials to make CBS eco−friendly and cost−effective.
6. Conclusions
This study focused on a section of India’s NH10 highway to evaluate the effectiveness
of the CBS in enhancing slope stability on a regional scale. By employing the GEOtop model
for water balance simulation, PWP was obtained and utilized for FoS calculation. Then, the
statistical analysis was conducted, and the spatial distribution of slope stability categories
was mapped to assess the stabilization effect of the CBS. Three historical landslide points
in the northern, central, and southern parts of the study area were selected for further case
analysis of the FoS variations at different depths.
Based on the results, several key findings were revealed. The CBS slope effectively
modulated PWP variations during and after extreme rainfall events by slowing the water
infiltration process, leading to a gentler increase in PWP and lower peak values, especially
in the layers of 1.25 m and 1.75 m, where the CBS prevented rapid soil saturation and
mitigated the prolonged effects of rainfall. Additionally, the CBS significantly reduced
the area classified as very high risk and high
−
risk regarding the slope stability category,
retaining a greater proportion of the area in the low
−
risk category. Spatial distribution
mapping further supported this finding, as after 24 h of rainfall, the original slope showed
widespread very high risk zones, while the CBS slope exhibited only minimal areas in this
category, facilitating the transition of more regions to moderate
−
and low
−
risk categories.
Furthermore, the analysis of the FoS variations at the three selected historical landslide
points highlighted that the original slope all experiences a drop in FoS values below 1,
indicating a high likelihood of slope failure. In contrast, the CBS slope maintained FoS
values generally above 1, as well as a smaller range of FoS variation, demonstrating
enhanced stability and a more resilient response to rainfall. These findings emphasized
the CBS’s capacity to enhance slope stability and mitigate landslide risks under extreme
rainfall conditions along the NH10.
As this study did not include a comparative analysis of different CBS designs or
alternative stabilization methods, future research could explore the performance of varying
CBS settings, including different soil parameters and layering designs, to assess their effec-
tiveness under diverse conditions. This would enable a more comprehensive assessment
of the practical feasibility of various CBS designs, identifying the most cost
−
effective and
applicable solutions for large and complex regions.
Author Contributions: Conceptualization, Y.C. and Y.L.; methodology, Y.C. and Y.L.; software, Y.C.
and Y.L.; validation, Y.C. and Y.L.; formal analysis, Y.C.; investigation, Y.C. and Y.L.; resources, Y.C.
and Y.L.; data curation, Y.C.; writing—original draft preparation, Y.C.; writing—review and editing,
Y.C. and Y.L.; visualization, Y.C.; supervision, Y.L.; project administration, Y.L. All authors have read
and agreed to the published version of the manuscript.
Funding: This research received no external funding.
Data Availability Statement: The original contributions presented in the study are included in the
article; further inquiries can be directed to the corresponding author.
Acknowledgments: The authors would like to thank to Monash University, Australia, for the
invaluable support provided throughout this research. The author would also like to acknowledge
Saranya Rangarajan from the School of Civil and Environmental Engineering, Nanyang Technological
University, Singapore, for her assistance with the software used in this research.
Conflicts of Interest: The authors declare no conflicts of interest.

Infrastructures 2024,9, 201 16 of 20
Appendix A
Table A1. Configuration of soil parameters in the SoilParFile, a file providing the soil parameters,
used for the GEOtop simulation. The parameters are described in Sections 3.1 and 3.2. Layers 1
to 3 were set as residual soil properties and CBS properties, respectively, to obtain two groups of
simulation outcomes for comparison.
Dz
(Layer
Thickness,
in mm)
Kh
(Lateral
Hydraulic
Conductivity,
in mm/s)
Kv
(Normal
Hydraulic
Conductivity,
in mm/s)
res
(Residual Water
Content)
sat
(Saturated Water
Content)
a
(Alpha
Parameter of
Van Genuchten)
n
(N parameter of
Van Genuchten)
v
(M parameter of
Van Genuchten)
200 1.2 ×10−31.2 ×10−30.017894 0.387 0.000878 3.671354 0.727621
200 1.2 ×10−31.2 ×10−30.017894 0.387 0.000878 3.671354 0.727621
200 4 4 0.010575 0.437 0.045042 3.886801 0.742719
400 1.32 ×10−21.32 ×10−20.0779 0.3962 0.008899 1.976090 0.493950
500 1.32 ×10−21.32 ×10−20.0779 0.3962 0.008899 1.976090 0.493950
500 1.32 ×10−21.32 ×10−20.0779 0.3962 0.008899 1.976090 0.493950
3000 1.32 ×10−21.32 ×10−20.0779 0.3962 0.008899 1.976090 0.493950
5000 1.32 ×10−51.32 ×10−50.0779 0.3962 0.008899 1.976090 0.493950
Table A2. Configuration of meteorological parameters in the MeteoFile, a file providing the meteo
forcing data, used for the GEOtop simulation. These data mainly involve hourly precipitation,
described in Section 3.1. The “Date” field can be set arbitrarily but must represent a continuous 48 h
time series.
Date Precipitation
(mm/h) Date Precipitation
(mm/h) Date Precipitation
(mm/h) Date Precipitation
(mm/h)
16/10/2013
07:00 12.67 16/10/2013
19:00 12.67 17/10/2013
07:00 017/10/2013
19:00 0
16/10/2013
08:00 12.67 16/10/2013
20:00 12.67 17/10/2013
08:00 017/10/2013
20:00 0
16/10/2013
09:00 12.67 16/10/2013
21:00 12.67 17/10/2013
09:00 017/10/2013
21:00 0
16/10/2013
10:00 12.67 16/10/2013
22:00 12.67 17/10/2013
10:00 017/10/2013
22:00 0
16/10/2013
11:00 12.67 16/10/2013
23:00 12.67 17/10/2013
11:00 017/10/2013
23:00 0
16/10/2013
12:00 12.67 17/10/2013
00:00 12.67 17/10/2013
12:00 018/10/2013
00:00 0
16/10/2013
12:00 12.67 17/10/2013
01:00 12.67 17/10/2013
12:00 018/10/2013
01:00 0
16/10/2013
14:00 12.67 17/10/2013
02:00 12.67 17/10/2013
14:00 018/10/2013
02:00 0
16/10/2013
15:00 12.67 17/10/2013
03:00 12.67 17/10/2013
15:00 018/10/2013
03:00 0
16/10/2013
16:00 12.67 17/10/2013
04:00 12.67 17/10/2013
16:00 018/10/2013
04:00 0
16/10/2013
17:00 12.67 17/10/2013
05:00 12.67 17/10/2013
17:00 018/10/2013
05:00 0
16/10/2013
18:00 12.67 17/10/2013
06:00 12.67 17/10/2013
18:00 018/10/2013
06:00 0

Infrastructures 2024,9, 201 17 of 20
Appendix B
Table A3. Comparison of FoS categories between the original slope and CBS slope at various depths
and rainfall stages.
Rainfall
Stages
Slope Stability
Categories
0.8 m 1.25 m 1.75 m
Original Slope CBS Slope Original Slope CBS Slope Original Slope CBS Slope
Natural state
Very high risk 0.00% 0.00% 0.00% 0.00% 0.53% 0.67%
High risk 0.00% 0.00% 0.10% 0.20% 4.71% 5.34%
Moderate risk 0.00% 0.00% 1.46% 2.09% 11.12% 11.82%
Low risk 100.00% 100.00% 98.44% 97.71% 83.64% 82.17%
6 h of rainfall
Very high risk 0.00% 0.00% 0.00% 0.00% 1.67% 1.91%
High risk 0.00% 0.00% 0.10% 0.20% 7.36% 8.00%
Moderate risk 0.00% 0.00% 1.47% 2.10% 13.25% 13.77%
Low risk 100.00% 100.00% 98.43% 97.70% 77.72% 76.32%
12 h of rainfall
Very high risk 0.00% 0.00% 0.00% 0.01% 2.05% 2.31%
High risk 0.00% 0.00% 0.11% 0.21% 7.88% 8.46%
Moderate risk 0.00% 0.00% 1.50% 2.14% 13.51% 14.04%
Low risk 100.00% 100.00% 98.39% 97.65% 76.56% 75.19%
18 h of rainfall
Very high risk 0.51% 0.00% 0.00% 0.01% 2.35% 2.58%
High risk 10.59% 0.14% 0.12% 0.23% 8.18% 8.78%
Moderate risk 22.35% 6.04% 1.60% 2.22% 13.64% 14.16%
Low risk 66.55% 93.82% 98.28% 97.54% 75.83% 74.47%
24 h of rainfall
Very high risk 10.11% 0.79% 0.08% 0.02% 2.56% 2.80%
High risk 21.04% 15.00% 5.62% 0.28% 8.42% 8.99%
Moderate risk 18.50% 20.49% 18.66% 2.52% 13.79% 14.28%
Low risk 50.35% 63.71% 75.63% 97.18% 75.23% 73.93%
6 h after
rainfall ends
Very high risk 12.00% 4.66% 5.56% 0.07% 2.76% 3.00%
High risk 19.80% 18.13% 22.54% 0.48% 9.28% 9.18%
Moderate risk 17.71% 19.54% 20.84% 5.23% 16.39% 14.35%
Low risk 50.49% 57.67% 51.06% 94.22% 71.56% 73.47%
12 h after
rainfall ends
Very high risk 11.25% 4.67% 10.59% 0.14% 3.21% 3.19%
High risk 19.30% 17.73% 22.41% 1.30% 13.26% 9.34%
Moderate risk 17.69% 19.53% 18.60% 12.01% 20.53% 14.42%
Low risk 51.77% 58.07% 48.40% 86.54% 62.99% 73.05%
18 h after
rainfall ends
Very high risk 10.64% 4.41% 12.77% 0.21% 4.50% 3.34%
High risk 18.96% 17.26% 21.38% 2.94% 17.63% 9.47%
Moderate risk 17.70% 19.48% 18.11% 16.28% 21.63% 14.51%
Low risk 52.69% 58.85% 47.75% 80.58% 56.24% 72.68%
24 h after
rainfall ends
Very high risk 10.16% 4.18% 13.66% 0.29% 6.22% 3.48%
High risk 18.70% 16.88% 20.91% 4.65% 20.79% 9.58%
Moderate risk 17.73% 19.41% 17.87% 18.60% 21.14% 14.58%
Low risk 53.42% 59.54% 47.56% 76.46% 51.85% 72.37%

Infrastructures 2024,9, 201 18 of 20
Appendix C
Infrastructures 2024, 9, x FOR PEER REVIEW 20 of 23
Appendix C
Figure A1. Risk distribution at a depth of 1.25 m for the original slope and CBS slope at different rainfall stages.
Figure A1. Risk distribution at a depth of 1.25 m for the original slope and CBS slope at different
rainfall stages.
Infrastructures 2024, 9, x FOR PEER REVIEW 21 of 23
Figure A2. Risk distribution at a depth of 1.75 m for the original slope and CBS slope at different rainfall stages.
Figure A2. Risk distribution at a depth of 1.75 m for the original slope and CBS slope at different
rainfall stages.

Infrastructures 2024,9, 201 19 of 20
References
1.
Fischer, E.M.; Knutti, R. Observed Heavy Precipitation Increase Confirms Theory and Early Models. Nat. Clim. Change 2016,6,
986–991. [CrossRef]
2.
Myhre, G.; Alterskjær, K.; Stjern, C.W.; Hodnebrog, Ø.; Marelle, L.; Samset, B.H.; Sillmann, J.; Schaller, N.; Fischer, E.; Schulz,
M.; et al. Frequency of Extreme Precipitation Increases Extensively with Event Rareness under Global Warming. Sci. Rep. 2019,
9, 16063. [CrossRef] [PubMed]
3.
Suradi, M.; Fourie, A.B.; Saynor, M.J. An Experimental and Numerical Study of a Landslide Triggered by an Extreme Rainfall
Event in Northern Australia. Landslides 2016,13, 1125–1138. [CrossRef]
4.
Wen, Y.; Gao, P.; Mu, X.; Li, M.; Su, Y.; Wang, H. Experimental Study on Landslides in Terraced Fields in the Chinese Loessial
Region under Extreme Rainfall. Water 2021,13, 270. [CrossRef]
5.
Zhou, C.; Huang, W.; Ai, D.; Xu, H.; Yuan, J.; Kou, L.; Luo, X. Catastrophic Landslide Triggered by Extreme Rainfall in Chongqing,
China: July 13, 2020, Niuerwan Landslide. Landslides 2022,19, 2397–2407. [CrossRef]
6.
Martha, T.R.; Roy, P.; Govindharaj, K.B.; Kumar, K.V.; Diwakar, P.G.; Dadhwal, V.K. Landslides Triggered by the June 2013
Extreme Rainfall Event in Parts of Uttarakhand State, India. Landslides 2015,12, 135–146. [CrossRef]
7.
Das, T.; Rao, V.D.; Choudhury, D. Numerical Investigation of the Stability of Landslide
−
Affected Slopes in Kerala, India, under
Extreme Rainfall Event. Nat. Hazards 2022,114, 751–785. [CrossRef]
8.
Roccati, A.; Faccini, F.; Luino, F.; Ciampalini, A.; Turconi, L. Heavy Rainfall Triggering Shallow Landslides: A Susceptibility
Assessment by a GIS−Approach in a Ligurian Apennine Catchment (Italy). Water 2019,11, 605. [CrossRef]
9.
Lazzari, M.; Piccarreta, M. Landslide Disasters Triggered by Extreme Rainfall Events: The Case of Montescaglioso (Basilicata,
Southern Italy). Geosciences 2018,8, 377. [CrossRef]
10.
Rahardjo, H.; Nistor, M.-M.; Gofar, N.; Satyanaga, A.; Xiaosheng, Q.; Chui Yee, S.I. Spatial Distribution, Variation and Trend of
Five−Day Antecedent Rainfall in Singapore. Georisk Assess. Manag. Risk Eng. Syst. Geohazards 2020,14, 177–191. [CrossRef]
11.
Rahardjo, H.; Zhai, Q.; Satyanaga, A.; Li, Y.; Rangarajan, S.; Rahimi, A. Slope Susceptibility Map for Preventive Measures against
Rainfall−Induced Slope Failure. Urban Lifeline 2023,1, 5. [CrossRef]
12. Gariano, S.L.; Guzzetti, F. Landslides in a Changing Climate. Earth−Sci. Rev. 2016,162, 227–252. [CrossRef]
13.
Jiang, W.; Tan, Y. Heavy Rainfall
−
Related Excavation Failures in China during 1994 to 2018: An Overview. Eng. Fail. Anal. 2021,
129, 105695. [CrossRef]
14. Ross, B. The Diversion Capacity of Capillary Barriers. Water Resour. Res. 1990,26, 2625–2629. [CrossRef]
15. Stormont, J.C. The Effectiveness of Two Capillary Barriers on a 10% Slope. Geotech. Geol. Eng. 1996,14, 243–267. [CrossRef]
16.
Rahardjo, H.; Kim, Y.; Satyanaga, A. Role of Unsaturated Soil Mechanics in Geotechnical Engineering. Int. J. Geo
−
Eng. 2019,10, 8.
[CrossRef]
17.
Satyanaga, A.; Rahardjo, H.; Hua, C.J. Numerical Simulation of Capillary Barrier System under Rainfall Infiltration in Singapore.
Numer. Simul. Capill. Barrier Syst. Rainfall Infiltration Singap. 2019,5, 43–54. [CrossRef]
18.
Tami, D.; Rahardjo, H.; Leong, E.-C.; Fredlund, D. A Physical Model for Sloping Capillary Barriers. Geotech. Test. J. 2004,27,
173–183. [CrossRef]
19.
Rahardjo, H.; Santoso, V.A.; Leong, E.C.; Ng, Y.S.; Hua, C.J. Performance of an Instrumented Slope Covered by a Capillary Barrier
System. J. Geotech. Geoenvironmental Eng. 2012,138, 481–490. [CrossRef]
20.
Scarfone, R.; Wheeler, S.J.; Smith, C.C. Numerical Modelling of the Application of Capillary Barrier Systems for Prevention of
Rainfall−Induced Slope Instability. Acta Geotech. 2023,18, 355–378. [CrossRef]
21.
Scarfone, R.; Wheeler, S.J.; Smith, C.C. Numerical Study of the Application of Capillary Barrier Systems for Prevention of
Rainfall
−
Induced Slope Instabilities. In Proceedings of the 4th European Conference on Unsaturated Soils (E
−
UNSAT 2020),
Lisboa, Portugal, 24–26 June 2020; Volume 195.
22.
Gao, C.; Zhu, Y.; Zhang, Y. Stability Analysis of the Inclined Capillary Barrier Covers under Rainfall Condition. Buildings 2022,
12, 1218. [CrossRef]
23.
Rahardjo, H.; Krisnanto, S.; Leong, E.C. Effectiveness of Capillary Barrier and Vegetative Slope Covers in Maintaining Soil Suction.
Soils Rocks 2016,39, 51–69. [CrossRef]
24.
Li, Y.; Satyanaga, A.; Rahardjo, H. Characteristics of Unsaturated Soil Slope Covered with Capillary Barrier System and
Deep−Rooted Grass under Different Rainfall Patterns. Int. Soil Water Conserv. Res. 2021,9, 405–418. [CrossRef]
25.
Kirschbaum, D.B.; Adler, R.; Hong, Y.; Hill, S.; Lerner
−
Lam, A. A Global Landslide Catalog for Hazard Applications: Method,
Results, and Limitations. Nat. Hazards 2010,52, 561–575. [CrossRef]
26.
Kirschbaum, D.; Stanley, T.; Zhou, Y. Spatial and Temporal Analysis of a Global Landslide Catalog. Geomorphology 2015,249, 4–15.
[CrossRef]
27.
Gupta, A.; Chattaraj, R. The Challenges Faced in Maintaining the Sevoke
−
Rangpo Section of NH
−
10. In Proceedings of the
228th Mid−Term Council Meeting of IRC, Siliguri, West Bengal, India, 14−15 June 2024; pp. 91–109.
28.
Dey, S.; Das, S. Landslide Susceptibility Mapping Through Hyperparameter Optimized Bagging and Boosting Ensembles: Case
Study of NH
−
10, West Bengal, India. In Geo
−
Environmental Hazards Using AI
−
enabled Geospatial Techniques and Earth Observation
Systems; Choudhury, T., Koley, B., Nath, A., Um, J.
−
S., Patidar, A.K., Eds.; Advances in Geographic Information Science; Springer
Nature Switzerland: Cham, Switzerland, 2024; pp. 123–140, ISBN 978-3-031-53762-2.

Infrastructures 2024,9, 201 20 of 20
29.
Arnfield, A.J. Koppen Climate Classification. Available online: https://www.britannica.com/science/Koppen%E2%88%92
climate%E2%88%92classification (accessed on 16 September 2024).
30.
Department of Industry, Commerce & Enterprises Government of West Bengal Climate of West Bengal. Available online:
https://wbindustries.gov.in/Climate.html (accessed on 16 September 2024).
31.
Shaffie, S.; Mozaffari, G.; Khosravi, Y. Determination of Extreme Precipitation Threshold and Analysis of Its Effective Patterns
(Case Study: West of Iran). Nat. Hazards 2019,99, 857–878. [CrossRef]
32.
Chen, Y.; Chen, X.; Ren, G. Variation of Extreme Precipitation over Large River Basins in China. Adv. Clim. Change Res. 2011,2,
108–114. [CrossRef]
33.
Das, S.; Pandit, K.; Sarkar, S.; Kanungo, D.P. Stability and Hazard Assessment of the Progressive Zero Landslide in the Kalimpong
Region of Darjeeling Himalaya, India. Geotech. Geol. Eng. 2024,42, 1693–1709. [CrossRef]
34.
State
−
Wise Ground Water Levels in India. Available online: https://iced.niti.gov.in/climate%E2%88%92and%E2%88%92
environment/water/ground%E2%88%92water%E2%88%92levels (accessed on 7 September 2024).
35.
NASA/METI/AIST/Japan Spacesystems and U.S./Japan ASTER Science Team ASTER Global Digital Elevation Model V003
2019. Available online: https://lpdaac.usgs.gov/products/astgtmv003/ (accessed on 16 September 2024).
36.
Chawla, A.; Sarkar, K.; Abhishek, R.; Chawla, S.; Pasupuleti, S.; Mishra, S. A Geotechnical Approach to Compare Different Slope
Stabilization Techniques for Failed Slope in the Darjeeling Hills, India. Environ. Earth Sci. 2023,82, 376. [CrossRef]
37.
Dikshit, A.; Satyam, N.; Pradhan, B. Estimation of Rainfall
−
Induced Landslides Using the TRIGRS Model. Earth Syst. Environ.
2019,3, 575–584. [CrossRef]
38.
Gardner, W.R. Some steady
−
state solutions of the unsaturated moisture flow equation with application to evaporation from a
water table. Soil Sci. 1958,85, 228–232. [CrossRef]
39.
van Genuchten, M.T. A Closed
−
Form Equation for Predicting the Hydraulic Conductivity of Unsaturated Soils. Soil Sci. Soc. Am.
J. 1980,44, 892–898. [CrossRef]
40.
Fredlund, D.G.; Morgenstern, N.R.; Widger, R.A. The Shear Strength of Unsaturated Soils. Can. Geotech. J. 1978,15, 313–321.
[CrossRef]
41. Fredlund, D.G.; Xing, A. Equations for the Soil−Water Characteristic Curve. Can. Geotech. J. 1994,31, 521–532. [CrossRef]
42.
Leong, E.C.; Rahardjo, H. Permeability Functions for Unsaturated Soils. J. Geotech. Geoenvironmental Eng. 1997,123, 1118–1126.
[CrossRef]
43.
Rigon, R.; Bertoldi, G.; Over, T.M. GEOtop: A Distributed Hydrological Model with Coupled Water and Energy Budgets.
J. Hydrometeorol. 2006,7, 371–388. [CrossRef]
44. Richards, L.A. Capillary Conduction of Liquids Through Porous Mediums. Physics 1931,1, 318–333. [CrossRef]
45. Taylor, D.W. Fundamentals of Soil Mechanics. Soil Sci. 1948,66, 161. [CrossRef]
46. Iverson, R.M. Landslide Triggering by Rain Infiltration. Water Resour. Res. 2000,36, 1897–1910. [CrossRef]
47. Graham, J. Methods of Stability Analysis. In Slope Instability; Wiley: New York, NY, USA, 1984; pp. 171–215.
48.
Rahardjo, H.; Li, Y.; Satyanaga, A. The Importance of Unsaturated Soil Properties in the Development of Slope Susceptibility Map
for Old Alluvium in Singapore. E3S Web Conf. 2023,382, 06009. [CrossRef]
49.
Qian, T.; Huo, L.; Zhao, D. Laboratory Investigation into Factors Affecting Performance of Capillary Barrier System in Unsaturated
Soil. Water. Air. Soil Pollut. 2010,206, 295–306. [CrossRef]
50.
Vishnu, G.; Bharat, T.V. Influence of Soil Hydraulic Characteristics in Modelling Capillary Barrier Systems. In Proceedings of the
Geo
−
Congress 2022, Charlotte, NC, USA, 20–23 March 2022; American Society of Civil Engineers: Charlotte, NC, USA, 2022;
pp. 265–273.
51.
Rahardjo, H.; Santoso, V.A.; Leong, E.C.; Ng, Y.S.; Tam, C.P.H.; Satyanaga, A. Use of Recycled Crushed Concrete and Secudrain in
Capillary Barriers for Slope Stabilization. Can. Geotech. J. 2013,50, 662–673. [CrossRef]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.