Abstract

This study presents assessments of the spatial variability of a soil landslide in the Three Gorges Reservoir Area in China. The results challenge the idea that variability of permeability within a landslide is secondary. Consequently, it cannot be assessed through single representative values. It was found that its variability spans two orders of magnitude, a consequence of it being influenced by both morphology and regolith properties. Identification of zones that display comparable behavior can be done through assessment of remote-sensing images collected by drones, as growth patterns of vegetation correspond well with zones that displayed similar permeability coefficients within the landslide body. Furthermore, extensive variability of permeability was caused by a diverse range of seepage mechanisms, namely pipe network flow, interface network seepage and pore water seepage. Thus, the hydraulic behavior of large landslide bodies is remarkably complex.

INTRODUCTION

It is a common practice to assess the hydraulic behavior of slopes and waste bodies considering characteristic permeability coefficient values for the whole formation. Usually, they are established after performing a single field test, or through inversion of seepage profiles considering overall stability numerical simulations (Yang et al. 2016).

However, recent studies indicate that the spatial variability of the permeability coefficient within a slope affects its stability (Chang & Cheng 2014; Cho 2014; Dou et al. 2015; Chen et al. 2017). Consequently, there is widespread interest in understanding it. First, Xu & Wang (2010) considered multivariable analyses of numerical models aimed at reproducing observed meso-seepage fields. They were further advanced by Chen et al. (2016), who proposed low-scale estimates of permeability based on observed properties of rock–soil mixtures on site. Likewise, assessment of time histories of water head data at several locations has been helpful for assessing the spatial variability of permeability (Alzraiee et al. 2017). Furthermore, direct field assessments of the variability of permeability have also been done, particularly, double-ring water injection tests (Chen et al. 2012; Dong et al. 2017). Those results show that permeability showcases outliers beyond three to four times the average value within a single slope.

The aim of this study is to present a methodology to understand the hydraulic behavior of large soil landslides. Its scope is to highlight the variability of the permeability coefficient by proposing a way to identifying regions where it shows comparable values. The methodology firstly uses remote sensing images of vegetation to characterize sections that showcase similar hydraulic properties. Then, their surface permeability coefficient is grossly assessed through field double-ring permeability tests. Afterward, extensive field inspections allow for first assessments of diverse flow processes within the landslide, making use of water table monitoring and scanning electron microscope images of samples. Therefore it allows for first assessments of the hydraulic behavior of landslide bodies.

The Three Gorges Reservoir Area in China is particularly well suited for assessing the proposed methodology. Landslides are extensive on its borders (Yin et al. 2016; Huang et al. 2017) while hydraulic flow boundary conditions are well constrained due to the management of the reservoir levels. This allows for control of externalities that can compromise results. Among all unstable bodies on the riverbank, the Shiliushubao landslide is of interest, due to its large volume (more than 10 million m3), complex hydraulic characteristics and material composition involving landslide bodies.

The results of this study show that the permeability coefficient within large soil landslides presents large variations, making assessments considering single representative values highly dubious. This complexity leads to the interaction of diverse flow process at both micro- and meso- scales, leading to varied hydraulic behavior.

STUDY AREA

The Three Gorges Project is the largest hydro-junction project in the world. More than 4,200 landslides have occurred within the area surrounding the Three Gorges Reservoir in central China as a consequence of its complex geological environment and heavy rainfall (Yin et al. 2010; Jiao et al. 2014; Sun et al. 2017; Zhao et al. 2017). Particularly, its riverbanks have been significantly affected by the operational oscillation of the water level of the reservoir, leading to seepage on steep border slopes which has caused slope failures (Luo et al. 2009; Yin et al. 2016).

The Shiliushubao landslide is located on the north bank of the Yangtze River, 1.5 km downstream of Badong County Town, Hubei Province. The landslide significantly pre-dates (more than a century before) the reservoir, and is mostly made of soil and regolith. The trailing edge elevation of the Shiliushubao landslide is 350 m, while the shear outlet elevation of the front edge is 60 m. Its total body length reaches 550 m, while its transverse width spans between 350 and 470 m. Thus, it encompasses an area of almost 0.25 km2, while its average thickness is almost 50 m. Consequently, its volume has been estimated at 1,180 × 104 m3. Its average slope angle is 26° with a strike of N 188°E. The landslide body is comprised of a mixture of rock and soils originated in the Quaternary, as a result of weathering and disaggregation of red mudstone, siltstone, sandstone and limestone-green marl from the Triassic Badong Formation (Luo et al. 2010; Li et al. 2018). An overview of the landslide is presented in Figure 1.

Figure 1

Localization and morphology of the Shiliushubao landslide and layout of monitoring points and field tests.

Figure 1

Localization and morphology of the Shiliushubao landslide and layout of monitoring points and field tests.

METHODS

Use of remote sensing for qualitative assessment of the permeability coefficient

The large surface of the Shiliushubao landslide makes single representative assessments of the permeability coefficient unreliable. In this study, remote-sensing images were used to identify locations that can be considered to have comparable permeability coefficients. Fortunately, the study area was subjected to a reforestation initiative which led to the cultivation of fruit trees all over the landslide surface. As similar species were deployed, the difference in their growth patterns became a successful proxy of permeability conditions. Particularly, copious growth was observed in places where the permeability is large; and in contrast, locations where the permeability was low showcased stunted plant growth, as access to nutrients was curtailed, as shown in Figure 2. This led to the definition of three distinct zones with expected different permeability coefficients. Thus, 12 double-ring tests were performed, two per each identified block (Shi et al. 2016).

Figure 2

Regional division of surface permeability of the Shiliushubao landslide: (a) growth of fruit trees on block b1; (b) growth of fruit trees on block b2; (c) growth of fruit trees on block b3.

Figure 2

Regional division of surface permeability of the Shiliushubao landslide: (a) growth of fruit trees on block b1; (b) growth of fruit trees on block b2; (c) growth of fruit trees on block b3.

Figure 3

Characteristic photos of surficial material in the Shiliushubao landslide: (a) and (b) typical photos of surface matter on block b2; (c) typical photos of surface matter on block b1; (d) typical photos of surface matter on block b3; (e) typical photos of surface matter on the first-grade slope; (f) typical photos of surface matter on the second-grade slope.

Figure 3

Characteristic photos of surficial material in the Shiliushubao landslide: (a) and (b) typical photos of surface matter on block b2; (c) typical photos of surface matter on block b1; (d) typical photos of surface matter on block b3; (e) typical photos of surface matter on the first-grade slope; (f) typical photos of surface matter on the second-grade slope.

Assessment of permeability within the landslide body

Detailed inquiry of the permeability coefficient within a large landslide body is challenging. The most reliable way to do it is through extensive water injection tests, which provide meaningful point estimates of this property. Therefore, sampling at several locations is required to obtain useful results. However, this requires complex drilling and deployment of support infrastructure (tanks, pumps, pipes) that make this approach unpractical. This study is based on a limited number of borehole water injections and pump cross-section tests (eight at four locations h1, h2, h4 and h5 in Figure 1) supplemented with passive monitoring of hydrological conditions. Consequently, firstly the front permeability of the Shiliushubao landslide has been studied through borehole water injections and borehole water pumping tests (h1, h2) in the course of drilling following local standards (China Geological Survey 2012). This allows for the computation of four hydraulic gradients among the four monitoring holes and the reservoir water level monitoring point considering times where precipitation was scant and oscillations on the reservoir level were minimal.

The permeability coefficient can be estimated according to flow continuity, as the inflow and outflow of the landslide can be assumed to be equal under stated conditions. As flow cross-sections can be taken as roughly uniform, groundwater velocity V in the landslide can be approximately considered constant, leading to the following relationship: k2-1i2-1 = k4-2i4-2 = k5-4i5-4 (k represents the permeability coefficient, i represents the hydraulic gradient).

Study on seepage characteristics

Before proceeding with the field seepage assessment, electrical resistivity tomography (ERT) (Perrone et al. 2014) tests and drilling data were employed to characterize the landslide body. The ERT was conducted along two perpendicular lines spanning 300 and 280 m, involving 60 and 56 measuring and receiving points, with a spacing of 5 m between each electrode (Figure 1), while the reservoir was close to full capacity. Resistivity profiles were found through inversion of data collected in the field. This long exploration line passes through the borehole section with the aim of finding correlations between characteristics of in situ geomaterials and electric resistivity measurements.

Afterward, a large-scale seepage test was performed at the steepest descent of the landslide into the reservoir. A prismatic pit 1.5 × 1 × 0.5 m deep was excavated 0.3 m from the edge of the ridge. Then water was constantly fed in to ensure a constant water table of at least 0.1 m. within it. Finally, samples along diverse locations in the landslide were collected and subjected to electron microscopy (SEM).

RESULTS

Characteristics of the surface permeability of the landslide

Analysis of images from unmanned aerial vehicles (UAVs)

Assessment of overall geomorphological characteristics of the landslide led to the definition of four topological areas, denominated as the water level fluctuation zone, the first-grade slope, the platform, and the second-grade slope. Furthermore, a more detailed subdivision of the platform was made considering the behavior of introduced fruit-tree specimens. In region b2 in Figure 1, plants display stunted growth with almost no foliage, while they were healthy elsewhere; posterior electrical resistivity measurements indicated a high water content in the locations of b2. Thus, the poor growth of the fruit trees can be explained by low permeability at these locations, which, weakens root activity and eventually leads to hypoxic respiration. Assessment of the behavior of vegetation within a landslide body could therefore be useful for investigating, in a coarse manner, water permeability across it.

Analysis of field investigation

Large-scale assessment of UAV images allowed for the definition of closer inspections at each one of the zones previously identified. Surface details of each region are shown in Figure 3. The following was observed for each identified region.

Platform-b2: Poor water conductivity was verified in the field. In particular, a large 4 × 2 m trench full with 1 m of water was found in it. Also, there were 14 small pits surrounding the trench, showcasing several depths of stagnant water. Moreover, these pits were left after the plants within them had died. Therefore, permeability below a certain depth is clearly very poor, making strata almost impervious to water. The local lithology in this area is primarily comprised of brick-red mudstone with a small amount of strongly weathered purple-red argillaceous siltstone. It can be easily fragmented into a powder by finger kneading. Furthermore, its particles are on average less than 2 cm.

Platform-b1: The surface lithology consists of gray-white marl and gray-yellow sandstone, which contains a small amount of mildly weathered siltstone and mudstone. They are made by fragments not easily crushable by hand, larger than 2 cm.

Platform-b3: The surface lithology consists of moderately weathered debris from purple-red argillaceous siltstone, which contains a small amount of brick-red mudstone. These particles can be fragmented via finger kneading. Furthermore, gravel particles are larger, with a diameter of more than 2 cm.

Slope: The surface lithology of the first-grade is similar to what is observed in platform b3. The lithology of the surface debris of the second-grade slope is more complex, and several lithologies are almost exposed. Clearly, change in the slope angle led to more segregation of conforming materials, and consequently, an improvement of permeability.

It is possible to establish a link between surface lithology and the permeability coefficient within a landslide body. Clearly, as mudstone is particularly susceptible to weathering and fragmentation, its disaggregation leads to the formation of fine particles that eventually reduces permeability within healthy blocks. In contrast, marl showcases a significantly higher abrasion and compression resistance, thus its weathering produces larger gravel particles that allow for a higher permeability coefficient. The strength of the siltstone ranges between these two materials, and consequently its weathering will lead to regolith that will display a permeability coefficient that is expected to range between what is observed for values associated with homogeneous marl and mudstone regolith.

In-site double-ring permeability tests

In-site double-ring tests confirm qualitative assessments based on UAV images and field inspections. The variability of observed permeability is summarized in Table 1. However, the variability in results is unexpected. There is a divergence of more than 70 times among sampled values, the most critical characteristic being the surface morphology. Permeability is remarkably high in the first-grade slope and the water level fluctuation zone of the landslide that is periodically submerged by the reservoir lake, while diminishing noticeably on the second-grade slope and the platform between them. Even within the relatively geomorphologically uniform platform, low values can be only 1/11 of the largest. Thus, considering large-scale uniform permeability regions for hydrogeological and stability assessments is not well founded.

Deep permeability characteristics of the landslide

Permeability assessment through borehole tests

Besides assessing permeability properties on the surface, deep assessments were effected in boreholes 1 and 2 at depths of 145 and 160 m. Results are presented in Table 2. The most significant trend observed is the difference in permeability in borehole 2. A depth gap of just 15 m leads to a difference of more than one order of magnitude at this particular location, while at borehole 1 more moderate variations are observed.

Table 1

Permeability coefficient of surficial layers of the Shiliushubao landslide

NumberLocationPermeability coefficient (m/d)Average permeability coefficient (m/d)
Water level fluctuation zone 5.70 5.52 
5.34 
First-grade slope 9.0 9.172 
9.344 
Platform-b1 1.33 1.411 
1.492 
Platform-b2 0.115 0.12 
0.125 
Platform-b3 0.531 0.52 
0.509 
Second-grade slope 0.225 0.233 
0.241 
NumberLocationPermeability coefficient (m/d)Average permeability coefficient (m/d)
Water level fluctuation zone 5.70 5.52 
5.34 
First-grade slope 9.0 9.172 
9.344 
Platform-b1 1.33 1.411 
1.492 
Platform-b2 0.115 0.12 
0.125 
Platform-b3 0.531 0.52 
0.509 
Second-grade slope 0.225 0.233 
0.241 
Table 2

Permeability coefficient at depth of front edge of the Shiliushubao landslide

NumberLocationPermeability coefficient (m/d)Remarks
h1 – 145 m 9.028 Water injection test 
h1 – 160 m 7.539 
3.458 Pumping test 
h2 – 145 m 8.639 Water injection test 
h2 – 160 m 0.501 
NumberLocationPermeability coefficient (m/d)Remarks
h1 – 145 m 9.028 Water injection test 
h1 – 160 m 7.539 
3.458 Pumping test 
h2 – 145 m 8.639 Water injection test 
h2 – 160 m 0.501 

It is clear that permeability presents significant variations along all spatial directions. However, complexities related to performing tests at depth make its assessment along the vertical direction particularly challenging. Measurements presented in this study are sparse, yet highlight clearly the complexity of the issue at hand.

The behavior of the groundwater table in boreholes

Time-histories of groundwater level were collected at several locations within the landslide body (Figure 4). The groundwater table at well h1 follows closely (without lag) changes in the reservoir level, while groundwater levels in borehole h2 show some slight lag before July 16 and then match reservoir levels afterward. The groundwater level at borehole h4 only resembles trends in reservoir levels slightly. Finally, records at borehole h5 denote that groundwater levels are insensitive to variations in reservoir level, but rather are highly conditioned by rainfall.

Figure 4

Variation of groundwater level and reservoir water level of the Shiliushubao landslide.

Figure 4

Variation of groundwater level and reservoir water level of the Shiliushubao landslide.

Figure 5

Behavior of the water table when the level of the reservoir (a) descends and (b) ascends for several time-lapses.

Figure 5

Behavior of the water table when the level of the reservoir (a) descends and (b) ascends for several time-lapses.

Thus, variations of the water level reach 30, 28, 9 and 14 m for boreholes h1, h2, h4, and h5 respectively. It is clear that water level changes become milder as the distance to the reservoir edge increases. Likewise, it must be stressed that the water level at the back scarp of the landslide is high. This is due to sub-surficial water flow from the mountain behind it. Thus, most of the landslide body, and particularly, its potential shear band, can be expected to be steadily saturated.

Observed hydraulic gradients showcase extensive variability, as shown in Figure 5. Between boreholes h1 and h2, the gradient is 0.079 while between h4 and h2 it is 0.288; while between boreholes h4 and h5, it is 0.128, almost half the previous value. According to the relationship between hydraulic gradient ratio and permeability coefficient, the deep permeability of the middle and rear parts of this landslide body can be obtained.

Based on all collected results in this study it is possible to outline trends of the permeability coefficient along the body of the Shiliushubao landslide (Figure 6).

Figure 6

Spatial distribution of permeability coefficient within the Shiliushubao landslide.

Figure 6

Spatial distribution of permeability coefficient within the Shiliushubao landslide.

Seepage characteristics of the landslide body

The internal structure of the landslide body

Electrical resistivity measurements indicate a heterogeneous structure without clearly identifiable layers (Figure 7). Comparison with samples taken from boreholes is reliable, allowing for geo-materials classification based on resistivity measurements. Areas where it is less than 50 Ω·m are conformed by a mix of residual soil and regolith of mudstone and siltstone, while values larger than 200 Ω·m indicate the presence of marl–soil matrices. However, a direct correlation with the permeability coefficient is poor. Although resistivity is largest in the vicinity of borehole h4, permeability is the lowest around it. Likewise, the largest observed permeability is noticed around boreholes h1 and h2, but resistivity there is among the lowest observed within the landslide body. Therefore, electrical resistivity is helpful for identifying geomaterials but does not provide representative results for directly assessing permeability.

Figure 7

Comparison of electrical measurement results and borehole cores of the Shiliushubao landslide: (a) longitudinal electrical measurement results of the Shiliushubao landslide; (b) transverse electrical measurement results of the Shiliushubao landslide.

Figure 7

Comparison of electrical measurement results and borehole cores of the Shiliushubao landslide: (a) longitudinal electrical measurement results of the Shiliushubao landslide; (b) transverse electrical measurement results of the Shiliushubao landslide.

Seepage system of the landslide body

Seepage within the landslide is driven by both large- and micro-scale water flow. This is illustrated in Figure 8, where several stages of the large-scale seepage tests are presented. Firstly, water leaks out from natural drains within the slope, just 2 minutes after the start of the test. After about 25 minutes of adding water, the section begins to discharge water from a soil–rock interface (Figure 8(b)). Finally, water flow becomes steady across the whole surface up to full saturation (Figure 8(c)).

Figure 8

The large-scale seepage section in the site: (a) pipe network seepage; (b) interface network seepage; (c) pore seepage, namely full section seepage after saturation; (d) the results of SEM of the earth–rock interface (gap display, magnified 50 times).

Figure 8

The large-scale seepage section in the site: (a) pipe network seepage; (b) interface network seepage; (c) pore seepage, namely full section seepage after saturation; (d) the results of SEM of the earth–rock interface (gap display, magnified 50 times).

Large-scale flow is governed by the connection of large openings between uneven soil and rock particles, as large hard gravel prevents a tightly packed arrangement of the matrix, thus allowing for openings. This is mostly observed in soil–rock masses where marl regolith is predominant. This explains the occurrence of springs during the rainy season in the first-grade slope.

Mesoscale flow is due to the small-scale roughness of the contact between soil and gravel particles. Figure 8(d) displays a scanning electron microscope (SEM) image magnified 50 times of a sample taken from the landslide. It is clear that there are micro-gaps between gravel and soil particles that allow for water percolation. As cementation processes advance, these gaps become clogged and the flow is constrained to pores within the aggregates, thus diminishing overall permeability. Therefore, this seepage mechanism is dominant in mixtures of soil with regolith from siltstone and mudstone.

Micro-scale flow is a common form of pore seepage in soil. It is notable that the relatively high permeability of the front of the Shiliushubao landslide is dominated by pipe network seepage, while the middle and later part is dominated by interface network seepage combined with pore seepage.

DISCUSSION

This study shows that assessment of vegetation performance based on images collected by drones is a viable tool for large-scale characterization of the permeability coefficient within a large landslide. Furthermore, better results can be obtained through the use of data in the infrared range, as it will be more correlated with the performance of diverse vegetal species in a more detailed manner (Xie et al. 2008). Therefore, further studies should explore the use of multi-spectral images for characterization of the hydraulic behavior of large geological formations.

Furthermore, it is clear that the permeability coefficient is not conditioned only by lithology. The historical geological processes leading to combination and fragmentation of rock foundation material are relevant for the hydraulic behavior of regolith. Thus properties at both micro- and meso-scale play a significant role, and they can be captured sparsely with single geophysical tests. This explains why ERT measurements correlated poorly with observed permeability coefficients in the field. Likewise, the variability of the permeability coefficient is extensive, considering both location within the landslide and material properties. In particular, places close to the surface in contact with the reservoir display average permeability values that are six times more than what is observed in the back scarp. Also, regolith from marl showcases permeability coefficients more than 11 times what is observed for regolith from mudstones. Consequently, our results indicate that the stochastic nature of permeability should not be neglected in assessing the hydraulic behavior of large landslides, making probabilistic approaches a pressing need (Zhang 2002).

CONCLUSIONS

This study introduces a framework to assess the variation of permeability within landslide bodies. It integrates remote sensing, field inspections, electrical resistivity measurements and results of permeability and seepage tests. It is observed that the permeability of a landslide body is conditioned by the interaction of its morphology and its surface lithology. It was found that the Shiliushibao landslide showcases variations of permeability of almost two orders of magnitude. Seepage behavior also presents extensive variability. On the surface in contact with the reservoir, water table changes are synchronous with variations in its level, However, in locations slightly further away, hysteresis increases sizably. Eventually, water levels within the landslide body become independent of oscillations in the reservoir, being rather conditioned by rainfall.

Seepage in the rock–soil mixture within the landslide body is the result of three different mechanisms: pipe network seepage, interface network seepage, and pore seepage. The first one is predominant in the contact interface between the landslide and the reservoir, while the second and third become more predominant at locations increasingly away from it. This explains observed trends in seepage hysteresis and variations on the coefficient of permeability.

Although electrical resistivity measurements were helpful in characterizing different geological materials within the landslide, their usefulness for direct assessment of permeability was poor. This is a consequence of the role of the effects of the spatial distribution of geomaterials and their geological history. Both condition seepage mechanisms that cannot be inferred directly from the results of these geophysical tests.

The issue of how change in reservoir level affects the permeability of landslide bodies remains open. Assessments made within this study do not make it possible to discerned if observed permeability pre-dates the reservoir, or is a consequence of it. This is a relevant issue for assessing the stability of reservoir banks, and thus is worthy of further study.

ACKNOWLEDGEMENTS

This work was financially supported by the National Natural Science Foundation of China (Grant Nos. 41977255, 41521002), the Sichuan Science and Technology Program (Grant No. 2019YJ0403), and the Key Scientific Project of TGR (Grant No. 000121 2015C C60 005). The authors wish to thank the anonymous referees for their helpful suggestions and constructive comments, which have contributed greatly to improving the quality of the manuscript.

REFERENCES

REFERENCES
Chang
K. T.
Cheng
M. C.
2014
Estimation of the shear strength of gravel deposits based on field investigated geological factors
.
Eng. Geol.
171
,
70
80
.
DOI: 10.1016/j.enggeo.2013.12.014
.
Chen
Z. H.
Chen
S. J.
Chen
J.
Sheng
Q.
Min
H.
Hu
W.
2012
In-situ double-ring infiltration test of soil–rock mixture
.
J. Yangtze River Sci. Res. Inst.
29
(
4
),
52
56
.
DOI: 10.3969/j.issn.1001-5485.2012.04.012 (in Chinese)
.
Chen
Z. C.
Luo
X.
Liu
K.
Ye
L. Z.
Zuo
C. Q.
2016
Seepage characteristics and mitigation measures of a gravel soil landslide
.
Rock Soil Mech.
37
(
3
),
813
818 + 849
.
DOI: 10.16285/j.rsm.2016.03.025 (in Chinese)
.
Chen
Z. H.
Huang
J. H.
Qin
W. T.
Yang
C. L.
2017
Effects of the spatial variability of saturated permeability on the slope stability
.
J. Chongqing Univ
.
40
(
3
),
59
69
.
DOI: 10.11835/j.issn.1000-582X.2017.03.007 (in Chinese).
China Geological Survey
2012
Handbook of Hydrogeology
, 2nd edn.
Geological Press
,
Beijing, China
.
Dong
H.
Huang
R. Q.
Luo
X.
Luo
Z. D.
Jiang
X. Z.
2017
Spatial distribution and variability of infiltration characteristics for shallow slope of gravel soil
.
Chin. J. Geotech. Eng.
39
(
8
),
1501
1509
.
DOI: 10.11779/CJGE201708018 (in Chinese)
.
Dou
H. Q.
Han
T. C.
Gong
X. N.
Qiu
Z. Y.
Li
Z. N.
2015
Effects of the spatial variability of permeability on rainfall-induced landslides
.
Eng. Geol.
192
,
92
100
.
DOI: 10.1016/j.enggeo.2015.03.014
.
Jiao
Y. Y.
Song
L.
Tang
H. M.
Li
Y. A.
2014
Material weakening of slip zone soils induced by water level fluctuation in the ancient landslides of Three Gorges Reservoir
.
Adv. Mater. Sci. Eng.
2014
,
202340
.
DOI: 10.1155/2014/202340
.
Li
H. J.
Xu
Q.
He
Y. S.
Deng
J. H.
2018
Prediction of landslide displacement with an ensemble-based extreme learning machine and copula models
.
Landslides
15
(
10
),
2047
2059
.
DOI: 10.1007/s10346-018-1020-2
.
Luo
X. Q.
Wang
F. W.
Zhang
Z. H.
Che
A. L.
2009
Establishing a monitoring network for an impoundment-induced landslide in Three Gorges Reservoir Area, China
.
Landslides
6
,
27
37
.
DOI: 10.1007/s10346-008-0140-5
.
Luo
X. Q.
Sun
H.
Tham
L. G.
Junaideen
S. M.
2010
Landslide model test system and its application on the study of Shiliushubao landslide in Three Gorges Reservoir Area
.
Soils Found.
50
(
2
),
309
317
.
DOI: 10.3208/sandf.50.309
.
Perrone
A.
Lapenna
V.
Piscitelli
S.
2014
Electrical resistivity tomography technique for landslide investigation: a review
.
Earth-Sci. Rev.
135
,
65
82
.
DOI: 10.1016/j.earscirev.2014.04.002
.
Shi
D. M.
Wang
W. L.
Jiang
G. Y.
Peng
X. D.
Yu
Y. L.
Li
Y. X.
Ding
W. B.
2016
Effects of disturbed landforms on the soil water retention function during urbanization process in the Three Gorges Reservoir Region, China
.
Catena
144
,
84
93
.
DOI: 10.1016/j.catena.2016.04.010
.
Sun
G. H.
Yang
Y. T.
Cheng
S. G.
Zheng
H.
2017
Phreatic line calculation and stability analysis of slopes under the combined effect of reservoir water level fluctuations and rainfall
.
Can. Geotech. J.
54
(
5
),
631
645
.
DOI: 10.1139/cgj-2016-0315
.
Xie
Y.
Sha
Z.
Yu
M.
2008
Remote sensing imagery in vegetation mapping: a review
.
J. Plant Ecology
1
(
1
),
9
23
.
DOI:10.1093/jpe/rtm005
.
Xu
W. J.
Wang
Y. G.
2010
Meso-structural permeability of S-RM based on numerical tests
.
Chin. J. Geotech. Eng.
32
(
4
),
542
550
(in Chinese)
.
Yang
H.
Tang
M. G.
Cai
G. J.
He
J. B.
Cai
X. Y.
2016
Research on deformation mechanism of Erdaohe landslide in Three Gorges Reservoir Area
.
Sci. Tech. Eng.
16
(
8
),
66
71
(in Chinese)
.
Yin
Y. P.
Wang
H. D.
Gao
Y. L.
Li
X. C.
2010
Real-time monitoring and early warning of landslides at relocated Wushan Town, the Three Gorges Reservoir, China
.
Landslides
7
(
3
),
339
349
.
DOI: 10.1007/s10346-010-0220-1
.
Yin
Y. P.
Huang
B. L.
Wang
W. P.
Wei
Y. J.
Ma
X. H.
Ma
F.
Zhao
C. J.
2016
Reservoir-induced landslides and risk control in Three Gorges Project on Yangtze River, China
.
J. Rock Mech. Geotech. Eng.
8
(
5
),
577
595
.
DOI: 10.1016/j.jrmge.2016.08.001
.
Zhang
D. X.
2002
Stochastic Methods for Flow in Porous Media: Coping with Uncertainties
.
Academic Press
,
San Diego, CA, USA
.
Zhao
N. H.
Hu
B.
Yi
Q. L.
Yao
W. M.
Ma
C.
2017
The coupling effect of rainfall and reservoir water level decline on the Baijiabao landslide in the Three Gorges Reservoir Area, China
.
Geofluids
2017
,
3724867
.
DOI: 10.1155/2017/3724867
.