Landslide Disaster Mitigation in Three Gorges Reservoir, China

Geo-hazard study results for the Three Gorges reservoir of Yangtze River are presented at three geographical scales: first, at the largest scale, is the geo-hazard investigation and evaluation of the whole reservoir area, which involves 19 counties (54462 km²); secondly, at a mid-level scale, is the initiation mechanism of complex slopes (landslides) in the Three Gorges section of the Yangtze River (about 4000 km²); lastly, at the small scale, research on the fan-shaped slope of Badong county new town was studied to ascertain its geologic characteristics, landslide initiation factors, and geo-environmental quality.

Through working on the classification of the reservoir bank slopes and their deformation mechanism, failure conditions, countermeasure on various types of slope failures, stability analysis and assessment of the bank slopes and large landslides, and monitoring and forecasting, this chapter summarizes the bank slope stability evaluation for the purpose of Three Gorges reservoir dam construction. The exploration and research methods for the evaluation of bank slope stability before the dam construction can be found in this chapter.

The Three Gorges reservoir area is characterized by widely distributed strata of the Badong formation, in which large-size landslides and deep-reaching loosely consolidated formations are likely to occur. Therefore, it is significant to reveal the mechanism and patterns of the large-size landslides that occur in the Badong formation for a better understanding of the development of the nature of deformation and process of formation of the deep-trending loose-slope stratum in the Wushan, Fengjie, and Badong renewal city zones. In this chapter, the geological characteristics of the rock mass of the Badong formation are analyzed on the basis of a systematic explanation of the rock mass strata, space variation of the lithological combinations as well as the space variation of stratum thickness, and structural deformation of the Badong formation. To demonstrate the basic law of long-term deformation of the Badong formation slopes and the patterns of later stage reformation and landslide evolution, the authors use the Huangtupo landslide as an example, as it is a typical failure in the Badong formation.

The locality features of high steep slope and dangerous rockmass in the section of a 63 km-long valley in the Great Three Gorges area are described by authors on the basis of the site investigation and geological mapping result of many field surveys. There are 12 high, steep dangerous rockmasses in the area which are described, as is the monitoring situation. The high and steep slope in the valley section is controlled by stratum lithology and locality feature, and hazard of dangerous monomer rock is mainly subject to the lithology and fracture combination, along with human activity.

When the impoundment of the Three Gorges reservoir reaches a water level of 175 m, the reservoir banks along the main river course and its tributaries will have a total length of 5,311 km. Among them, the rock banks are 213 km, and the banks of unconsolidated soil and rock mixture are 249 km. Bank collapse is one of the main engineering geological problems in the Three Gorges reservoir area. In this chapter, using an extensive amount of data on bank collapse obtained through geological investigation, field measurement of the bank collapse parameters, and simulation analysis, we systematically explored the bank collapse types, parameters, and bank collapse prediction methodology. In the Three Gorges reservoir area, the typical bank collapses are wash and abrasion, toe-erosion collapse, rock break-off and slides, and landslide. The evolutionary processes, the distribution features, and the corresponding geological conditions of these bank collapse types are analyzed as well. Eigenvalue angles of the rock and soil masses, including underwater deposit eigenvalue angle θ, wash–abrasion eigenvalue angle α, and steady above-water eigenvalue angle β, are fully discussed in the chapter. Based on the exploration of the correlation between bank collapse prediction parameter and the composition of the rock and soil mass and the correlation between mechanical features and the hydrodynamic conditions in the reservoir, an innovative prediction method that is adaptive to mountain–river reservoirs like Three Gorges reservoir area, Bank-Slope Structure Prediction Method (BSSPM), is proposed. The analogical study of three water storage reservoirs for many years, a comparison of the predicted results with that of the actual happenings, shows that BSSPM has a good adaptation for bank collapse prediction of mountain–river reservoirs.

The research area in this chapter is the Three Gorges area from Yunyang to Wushan county, which is about 100 km long and 40 km wide and covers some 4,000 km² in total. Forty percent of landslides in the Three Gorges area happened in this region, and 205 landslides have been recorded with relatively complete data.

On the basis of the analysis of historic landslide hazard impacts to local society and economy, disaster land status quo, and features in the Three Gorges area several technological methods, structural measures, and social and ecological mitigation measures have been implemented for the Xintan landslide disaster. The Xintan landslide is a typical landslide in the area and is used as an example for the general treatment methods of landslide reclamation and engineering in the Three Gorges Reservoir area. These methodologies can be applied in land treatment projects which are about to be put into place in other areas on a large scale. These methods have great potential to significantly ease the tension between people and land issues, will help to promote the successful outcome of the construction of the Three Gorges Project, and will solve the migration and settlement problems of the displaced people who have become migrants, as a result of being relocated from their lands. These methods can solve the sustainable development problems of the social economy in the Three Gorges Reservoir areas.

The first impoundment of the Three Gorges Dam reservoir in China started from a water-surface elevation of 95 m on June 1, 2003 and reached 135 m on June 15, 2003. Shortly after the water level reached 135 m, many slopes began to deform and some landslides occurred. The Qianjiangping landslide is the largest one; it occurred on the early morning of July 14, 2003, and caused great loss of lives and property. Field investigation revealed that although failure occurred after the reservoir reached 135 m, the stability of the slope was already reduced by pre-existing, sheared bedding planes. To study the mechanism of the rapid motion of this reactivated landslide, two soil samples were taken from a yellow clay layer and a black silt layer in the sliding zone and a series of ring shear tests were conducted on the samples. One series of ring shear tests simulates the creep deformation behavior, while the other series simulates different shear rates. Conclusions drawn from analysis of the ring shear tests indicate that the mechanism of the rapid motion of the reactivated landslide was caused by the rate effect of the black silt layer during the motion phase after the creep failure. The yellow clay layer did not play any important role in the rapid motion in the 2003 event.

The Qianjiangping landslide occurred following the first impoundment of the Three Gorges reservoir and a prolonged rainfall. To evaluate the quantitative effects of the reservoir impoundment and rainfall on the landslide, a sensitive analysis was performed with particular consideration to the landslide’s unique characteristics and the seven parameters susceptible to the interactions between the landslide’s materials and water from the two sources. These parameters include unit weight of the landslide’s material, groundwater level above the slip zone, shear strength components of the upper and lower sections of the landslide’s slip zone, and uplift pressure of the reservoir water. Results of this study show that the factor of safety (FS) of the landslide was the most sensitive to the cohesion of shear strength of the slip zone’s lower section, and the least sensitive to groundwater level above the slip zone. It is found that the effect of shear strength reduction of the slip zone’s lower section was the most crucial in the landslide’s initiation. The landslide resulted from the combined influences of the reservoir impoundment and rainfall. The role of the reservoir impoundment in the landslide was dominant with a contribution percentage about 61.3%, whereas rainfall may have been the factor triggering the landslide.

Water is the most sensitive factor affecting the mechanical characteristics of rock and soil mass. The increase in the water ratio of unsaturated soil mass reduces both instantaneous strength and long-term strength of soils and changes the soil’s creep property. In view of the close correlation of rock and soil mass with water action and effects over time, this chapter presents the concept of “unsaturated soil creep” based on the theories of unsaturated soil mechanics and creep soil mechanics. An unsaturated soil creep triaxial apparatus has been researched and developed based on the basic principles used in the application of unsaturated soil triaxial apparatus and triaxial creep apparatus. The law of the impact of matrix suction on the creep characteristics of the soil mass in the Qianjiangping landslide in the Three Gorges area has been applied, and an unsaturated soil creep model that can reflect the impact of matrix suction has been established through tests using the apparatus. The research result has facilitated a new approach for analyzing the long-term strength and deformation of unsaturated soils and predicting the long-term stability of the landslide under the action of reservoir water fluctuation or rainfall.

The Shuping landslide was reactivated by the initial impoundment of the Three Gorges Dam Reservoir, China, in June 2003. For the purposes of landslide disaster mitigation in the reservoir area and identification of landslide movement and deformation caused by reservoir level changes, a monitoring system mainly consisting of drum-style extensometers was installed in the eastern part of the Shuping landslide. Systematical monitoring was started in August 2004 with installation of 13 extensometers above the waterline after the initial impoundment. In August 2006, 11 more drum-style extensometers were installed above the high waterline (175 m), and 5 flexible extensometers were installed along a longitudinal section in the elongation to low waterline. In August 2007, three more drum-style extensometers were installed to make the whole monitoring line connected. In this chapter, the monitoring results from August 2004 to May 2008 are presented, and the deforming of the Shuping landslide caused by both reservoir level changes and rainfall is examined.

Many gentle dip landslides have taken place in Wanzhou, located in the Three Gorges Reservoir area. In order to study the mechanism of the gentle dip landslides, the authors selected the Anlesi landslide as a typical gentle dip landslide to study in detail. Field investigations show that the slip zones of the Anlesi landslide formed from a white mudstone in Jurassic red strata by compressive stress. The X-ray diffraction and infrared ray analysis reveal that the main mineral components of the slip zone are composed of montmorillonite, illite, feldspar, and quartz. A set of tests were conducted on the slip zone specimens to obtain the physicomechanical characteristics. Test results show that the slip zone soils are silty clay, of medium swelling potential, as shear strength becomes very low once the slip zone attracts water to saturation.

The main factors contributing to the gentle dip landslide mechanism are incompetent beds, recent tectonic activities, and intensive rainfall. Several stable, continuous, and thick incompetent beds exist in Jurassic red strata in Wanzhou. The integrity of incompetent beds was compromised under tectonic stress. The recent tectonic activities caused shear failure along the incompetent beds and joints in the sandstone. With the effect of intensive rainfall, water permeates to the incompetent beds along tectonic fissures, resulting in swelling of the soil material and high hydrostatic pressure in fissures of the strata. Therefore, the slopes are prone to slide along the incompetent beds.

The Bailuxi River is located in the mountainside of Dabashan Mountain and to the west side of the watershed of the Daninghe River and the Duhe River. It is at the southeast edge of an area of heavy rainfall in the centre of the Qinba mountainous region. For reasons of its climatic environment and the existence of a sufficient supply of sediments and large reserves of such in the river reaches, the human engineering and economic activity have become more and more intensive, resulting in an increase in hazardous environmental and geological conditions for mud-rock flows to occur in the Bailuxi River. Since the mud-rock flows there constitutes a direct threat to the inhabitants’ life and property safety in Bailu Town and its bank downriver of the Daninghe River, attention and concern for the hazardous condition are necessary. Through investigation, it has been found that the mud-rock flow channel of the Bailuxi River consists of the main mud-rock flow channel and its branches, the Yangjiawan branch of the mud-rock flow channel being the largest, the most dangerous, and potentially the most damaging one. The branch channel is a frequent mud-rock flow channel at a small scale, but during certain conditions, a large mud-rock flow at a considerably large scale could possibly occur. The primary channel flow of a large-sized scale is of low frequency and it is only triggered by large-sized mud-rock flows occurring in the branch channels. The deposit in the area of Bailu is the accumulation of the mud-rock flows that have erupted previously.

The Majiagou landslide is located on the left side of the Zhaxi-he River, a tributary of the Yangtze River. There is a large old landslide there originally, on which three secondary landslides were formed. The Majiagou landslide is one of the secondary ones. After the water level of the Three Gorges Reservoir rose from 95 to 135 m in June 2003, the Majiagou landslide occurred including a 20 m- long, 3–5 cm wide, (locally 10 cm-wide) fissure at its back that occurred in a time-frame of 3 months. The implication is that the reservoir impounding reactivated the landslide. The slide mass of the Majiagou landslide is composed of a thin layer of silty clay on the top with low permeability and a thick layer of angular pebbles as the main portion beneath the silt clay which has high permeability. The rate of water level fluctuation is between 0.6 and 4.0 m/d with regard to the altitude between 145 and 175 m in the reservoir during its normal operational state. Under these conditions, the FEM method was applied to simulate the groundwater changes in the slide mass, coinciding with the reservoir water level fluctuation. The results suggest that the groundwater level almost changes with the reservoir water level simultaneously within the slide mass, which means that the groundwater gradient is very gentle. Therefore, the effect of the reservoir water level fluctuation on the landslide stability is mainly by the action of buoyant force. Without considering dam failure, the seepage force is so little as to be negligible. The stability of the landslide in cases of different water levels is then calculated using the Morgenstern-Price method, and the results show that the factors of safety decrease with the water level rise. As the water level increases to 165 m, the factors of safety are at the minimum value, increasing with the water level rise. The minimum value reflects that the landslide is in a critical state, so stabilizing design was applied using stabilizing piles as well as a water drainage system. The project was completed in early 2006, and the water level has risen to 156 m in October 2006. Monitoring data illustrated that there is no further deformation of the landslide and the stabilizing piles, so the stabilizing work is effective.

The Huangtupo slope is one of the most noted large-scale slopes with geologic hazards that cause problems related to the residential safety of immigrants in the reservoir area of the Three Gorges Project, Yangtze River, China. The gravitational process, blended with tectonic deformation, large-scale covering of loose debris, and long-term surficial mass movement, complicates the Huangtupo slope and gives rise to a lack of consensus on the slope nature and stability. Characterization of the structural geometry of the slope deformation and reconstruction of its development history are believed to be pivotal in understanding what has happened and what will happen to the slope. Based on a thorough field investigation combined with an electrical resistivity survey, a three-stage model involving mass rock creep–primary landsliding–partial reactivation is proposed. The first stage follows the incision of the Yangtze River along the axes of the Guandukou syncline, when Huangtupo experienced long-term gravitational deformation referred to as ‘mass rock creep’. Mass rock creep in the Huangtupo slope can be classified as two basic processes: toppling and deep-seated creep. Toppling mainly occurs on the exterior part of the slope and is characterized by inclination, sliding, and segmentation of brittle deformation of cleavage rocks. Deep-seated creep occurs predominantly in the interior part of the slope and brittle–ductile flow folding accompanying low-angle shearing is its representative deformation. A primary, large-scale landslide occurred in the second stage as a subsequence of previous mass rock creep. The landslide was ca. 4×107 m³ in volume with elevations of 640 m a.s.l. at the head and of 80 m a.s.l. at the foot. Significant evidence for the landslide is the variation of attitudes of structural foliation and lineation among outcrops in the slope. The last stage began subsequent to the original sliding; surficial and partial reactivation on the Huangtupo landslide plays a leading role in this stage. Two sliding events in this area in 1995 were ascribed to the partial reactivation mainly due to rainfall, water level fluctuation of the Yangtze River, as well as human activity. It was suggested that analogous failures in 1995 would continue on the Huangtupo landslide and become even more frequent under the combined effect of human activity and reservoir filling.

Hydrogeological conditions of bank slopes in the Three Gorges Reservoir will change greatly due to the fluctuation of water level in the reservoir, which will affect the stability of the bank slopes. The probable failure mode and mechanism of the Xietan landslide is presented using a physical model in order to test varying conditions of water level fluctuation. The result will give insight as to the failure modes and mechanisms of landslides that have engineering geological conditions similar to those of the Xietan landslide.

The Xintan landslide occurred on June 12, 1985 on the left bank of the Yangtze River, 27 km upstream from the Three-Gorge Dam. It has a large volume of about 20 million m³ and totally destroyed the Xintan Town. Detailed geological investigation has been conducted since 1968, and a precise monitoring system was established in 1977, so that the early warning was issued before the 1985 event. In this chapter, the geological condition, monitoring data of the landslide deformation used for early warning are presented, with an analysis of the mechanism of the large-scale landslide.

The Xintan landslide is a large-scale landslide that occurred in Xiling Gorge, the last gorge upstream of the Three Gorges in the Yangtze River. It occurred on May 12, 1985. Through intensive monitoring and field investigation, the occurrence of this landslide was predicted successfully, and human loss was avoided in Xintan Town, which is located at the toe part of this landslide. In this chapter, the deformation property is described, and the monitoring data are presented to show the prediction process.

The Xintan landslide, a fast-moving landslide that occurred on June 12, 1985, is located on the north bank of the Yangtze River at Xintan Town, Zigui County, Hubei Province, China. Investigations showed that this landslide travelled at a speed of about 20 m/s and induced a huge water wave with wave run-up of 54 m on the opposite shore. Back-analysis of water waves generated by the Xintan landslide is used to determine the friction angle of the sliding zone in a state of movement. Newton’s second law and the basic principles of kinematics are used to obtain landslide velocity by the method of back-analysis, and the volume conversion law and the viscous force formula in flow fields are applied to calculate the initial height of the water wave. In terms of continuity equation, movement equation of transient flow and water head loss theory in open channels in hydrodynamics, the propagation process of the landslide surge is divided into two stages: the sharp-decay stage and the slow-decay stage. It is assumed that the sharp-decay stage is a kind of exponential decay and the slow-decay stage is a kind of water head loss with propagation distance in an open channel. Wave run-up on the opposite shore is calculated with the consideration of slope angle and run-up azimuth. It is concluded that an appropriate value of the reduction factor of friction angle of the sliding zone is between 0.8 and 0.9, and the results have significance for the selection of the friction angle of the sliding zone in a state of movement and calculation of landslide velocity.

Reinforcement measures are inevitably taken for most landslides in the reservoir area of Three Gorges, China. The reinforcement designs are often based on numerical analysis and land-use planners’ experience, which will make them conservative in most cases. Obtaining a global optimum design which can guarantee both the landslide stability and lower costs of remedial works is a significant problem. This chapter presents a new method for the optimization of reinforcement design for landslides that is an integration of evolutionary artificial neural network algorithm, numerical analysis and GIS. Artificial neural network is used to build the nonlinear relationship between the parameters of reinforcement measures, factor of safety (FOS), and engineering cost; and network structure is optimized by genetic algorithm. Numerical analysis is used to create training and learning examples, and GIS technique plays the role of result visualization and data provision for numerical analysis. During the optimization procedure, the engineering cost is taken as a fitness function of genetic algorithm and FOS is regarded as a constraint condition. Accordingly, the optimized parameters of reinforcement measures, such as pile space, length, and the geometry and size of slide-prevention piles, will be obtained. Based on these optimized parameters and the intelligent prediction model, the reinforcement design will directly lead to more economical engineering costs. Finally, the reinforcement design will be visualized in a three-dimensional strata model of a landslide, which can contribute to land-use planners’ synthetic decision-making. In addition, some integrating geological sections and local strata model can be generated in the GIS platform which will be used for numerical analysis for landslide stability. The proposed method is applied to the Muzishu landslide in the reservoir area of Three Gorges, China. The results provide a satisfactory optimum design which makes it possible to significantly reduce engineering costs.

Fractal phenomena are a naturally occurring characteristic of landslides. In the process of landsliding, a series of traces of the rupture surface on the ground surface extends progressively, as various types of slope failure occur. After this enlarging of the surface of rupture is completed, mass movements increase the volume of material because the displaced material dilates. After landslide occurrence, geometrical features of the rupture surface appearing on the ground surface are one way of providing information to evaluate slope instability. Correspondingly, fractal dimensions of the landslide boundary trace can reach a maximum in view of its geometry, which indicates that the landslide is in a critically stable state. In the case of changes of external factors, slide masses are likely to reactivate. This chapter addresses fractal dimensions of the trace of a landslide outline and its application in evaluating slope instability. In the vicinity of Badong County Town on the south shore of the Three Gorges Reservoir, fractal dimensions of 11 landslides with respect to boundary trace were calculated using a box-counting method. Six landslides were chosen as a case study to distinguish the relationship between fractal dimensions of boundary trace and slope instability. Compared with previous work in calculating the factor of safety, a landslide has the potential to reactivate if the value of fractal dimensions is between 1.4 and 1.5. A landslide is stable overall but minor slope failures are likely to occur in small areas, if the value of fractal dimensions is 1.1–1.3. A value of fractal dimensions of 1.0 indicates that the landslide is in a stable state. In summary, landslides with larger fractal dimensions in view of the boundary trace were evaluated as not being in a stable state.

This chapter evaluates the stability of the Huanglashi landslide in the Three Gorges of the Yangtze River. By means of the uncertainty evaluation model for the Huanglashi landslide, this chapter calculates and evaluates the stability of the landslide at present and after water storage, and then analyzes the stability after the completion of the Three Gorges project. The stability analysis can be used for predicting the stability tendency and the prevention of future disasters.

Remote sensing data provide not only information about landslide characteristics, such as landslide locations, scarp shapes, but also the related environmental factors, such as lithology, tectonic structure and land cover, which have significance for landslide occurrence. This chapter aims to use remotely sensed data to extract lithology information, and to relate this to landslide incidence. A variety of digital image techniques were applied to remote sensing data, and several image lithostratigraphy units were delineated by the identification elements such as weathering manifestations, drainage patterns, weathering, erosion characteristics, and surface morphology. Landslide location was identified by image interpretation and from field survey data. Other landslide-related factors, such as altitude and slope, were derived from digital elevation models. A method of generalized likelihood ratio was then utilized to analyze the relationships between landslide occurrence and lithology factors. Based on these data, causal factors, contributing to landslide occurrence, have been combined into a binary logistic regression model, and landslide probabilities are then calculated cell by cell. The results state that the rate of regression model classification has successfully increased from 71.5 to 92.6% of the overall slopes after “lithostratigraphical unit” variable was added. The study shows that remote sensing data can provide a quantitative causative factor source for landslide hazard assessment.

Monitoring and early warning is an important measure for geological disaster mitigation, the goal being to avoid or reduce major casualties and huge property losses. The information obtained for landslide deformation process is valuable for analysis and for the understanding of deformation law and landslide process. Real-time monitoring can achieve automatic continuous monitoring data collection, and this timely information is critical for analysis and for decision-making to avoid lengthy response times which can delay emergency measures, causing serious consequences. A more rapid response capability has been regarded as important by more and more people and has become the inevitable trend for monitoring of geological disasters. The research results are systematically summed up by the Research and Demonstration for Technology and Method on Geologic Hazard Monitoring and early warning investigations of the Ministry of Land and Resources P.R.C. The goals are to demonstrate the components and benefits of real-time monitoring, including implementing the methods of single-chain construction of the real-time system, the subsequent monitoring network, data collection system, data transmission system, and the data processing and information dissemination system. The goal is also to introduce and demonstrate the construction and implementation of the system and to show measures to deal with failure. This work will have implications for the installation of future real-time monitoring systems.

With the development and expansion of landslide research, quantitatively conducting typical landslide hazard assessment is demonstrated to be an effective method in assessing landslide hazard. This study sets up an index system for landslide hazard assessment by means of intensity factors and probability factors of landslide occurrence. Based on the information entropy method, subjective weight was calculated for 21 typical landslides in Wushan County, located in the Three Gorges Reservoir area, and a qualitative typical landslide hazard degree assessment model was established. This model was then used in evaluating the hazard degree of the landslides in Wushan County, and the results were contrasted and confirmed through field investigation. The accuracy rate of the investigation is as high as 90.4%. As a result, this model demonstrates a useful application of a new methodology and provides a basis for future warning and hazard evaluations of a typical landslide.

In a case study of the Baota landslide in the impoundment area of the Three Gorges project, the conceptual model of groundwater systems in this large-scale landslide was established by integrating the data from hydraulics, physical–chemical characteristics, environmental isotopes, and temperatures of the groundwater in the landslide. The conceptual model of groundwater systems includes the recharge, seepage, discharge condition, and trend of groundwater which is in turn controlled by the environmental factors (e.g., rainfall, reservoir impoundment, and artificial surface water recharge). This is one of the keys to predict the slope stability and to apply countermeasures to mitigate landslides.

After the Three Gorges reservoir impoundment water level changed, the groundwater dynamic stress field and the rock mass stress field changed considerably. As the water level changes, wave erosion as well as rain storm, causes bank collapse to often occur along the reservoir area. This chapter aims to find the mechanism causing bank collapse and apply this type of information using models which can then describe these mechanisms for later research. In order to find out the relationships between bank collapse and bank width, slope angle, property of bank soil, and intensity of rainfall, some modeling tests were carried out and the results showed that the width of the bank is the main factor affecting bank collapse. Another conclusion is that the process of bank collapse can be characterized as time-dependent. It is proposed that the process be divided into four characteristic phases: surface erosion, shallow erosion, deep erosion, and termination of slope movement. Deep erosion is the key phase of the four, for which the velocity and scale of bank collapses can be calculated.