Abstract
Landslide risk assessment is considered an important objective in minimizing damage caused by this extreme event. The biggest challenge of assessing landslide risk is predicting the likelihood of a catastrophic event. To evaluate this possibility, studies have focused on building integrated maps of assessment, using predictive models such as testings, scale models, numerical models, and actual field observations. This article summarizes the research achievements and applications achieved in the past ten years of the Institute of Transport Science and Technology in the landslide risk assessment in Vietnam.
You have full access to this open access chapter, Download chapter PDF
Similar content being viewed by others
Keywords
- Landslide
- Risk assessment
- Potential
- Susceptibility
- Reactive
- Testing
- Ring shear apparatus
- Monitoring
- Early warning
1 Introduction
Vietnam is located on the Indochinese peninsula in Southeast Asia, with a population of over 90 million people. The average annual rainfall from 1400 to 2400 mm receipt in most places of the country, with some places reaching over 4000 mm Vietnam regularly suffer from 7 to 10 tropical storms accompanied by floods and landslides. Over the past two decades, natural disasters in Vietnam have caused more than 13,000 casualties and over US$6.4 billion in property damage, of which landslides are considered the most catastrophic natural disasters.
Vietnamese Government Strategy for landslide reduction to mountainous areas and central highlands is “proactively prevent natural disasters”, for which landslide risk evaluation is one of the key solutions. The Institute of Transport Science and Transportation (ITST) under the MOT was established in 1956 as an institution for science and technology research and application for the transport sector. For more than 35 years ITST has been involved in the landslide field and proposed several solutions for scientific research and new technology application in disaster prevention, response, and mitigation to specific strategy mentioned above. Since the ODA Project named “Development of Landslide Risk Assessment Technology along Transport Arteries in Vietnam” began in November 2011, the development of new landslide risk assessment technology and its application to forecasting, monitoring, and disaster preparedness of landslides in Vietnam has been raised to a new level and contributed to geo-disaster reduction along main transport arteries and on residential areas. This paper is a summary of the achievements in Landslide Risk Assessment in the tropical zone of Vietnam that contributes to the mitigation of natural disaster vulnerability over the last 10 years.
2 Concept of Landslide Risk Assessment
The term “risk” combines the concepts of threat to life, difficulty and danger of evacuating people and property in the event of a disaster, potential structural damage and house value housing, social disruption, crop loss, and destruction of public facilities.
In the framework of this paper, the concept of risk is determined from the point of view of the International Center for Geohazards (ICG) of the Norwegian Geotechnical Institute (Uzielli 2009) as follows: R = HxVxE (1) Where: R (Risk) is the possibility of loss due to the accident occurring. H (Hazard) is the likelihood of a catastrophic event occurring. The value of H will be determined between 0 and 1, where H = 0, corresponds to no events occurring, and H = 1 corresponds to certain events. H has no units. V (Vulnerability): When an accident occurs, injuries may be caused to people, their production, and daily life. V is understood as people’s ability and material, social, economic, environmental, cultural, institutional, and political assets to be damaged due to disasters. V has no units and ranges from 0 to 1, where V = 0 corresponds to no injury and V = 1 corresponds to complete injury. E (Value of Vulnerable Elements): Vulnerable elements include people, assets (houses, traffic structures, vehicles, crops, livestock…), livelihood activities, and the environment. E is defined in monetary units (for economic damage assessments) or person units (for human damage assessments).
Along with the concept, risk assessment methods are also increasingly diverse. These methods can be grouped in two directions of assessment: direct—qualitative and indirect—quantitative (through a set of indicators), specifically: sociological investigation method, integrated cartographic method, etc. combination and index method. Each assessment method has its advantages and disadvantages.
For the field of landslides, due to the diverse nature of scale, type, and large distribution space, when considering the concept of risk, most scientists focus on assessing the probability of occurrence of hazard event (H) on an area scale and evaluate the risk for specific landslides.
Assessing the overall landslide risk for an area involves answering the questions of what “types of landslides occur”, “where they did occur”, “where is landslide susceptibility area”, and “how likely they are to reactivate”. The best tool to assess landslide risk for an area is the development of integrated landslide maps related such as landslide distribution maps, landslide classification maps, landslide susceptibility assessment maps, or reactive assessments of landslides that have occurred. Based on integrated maps, the planning, land use, response, and damage reduction of the phenomenon are determined. Due to a large number of occurrences, the assessment of the probability of occurrence of each landslide in an area is usually not covered. However, risk assessment for the specific landslide case is mentioned for landslides with special significance such as large scale, important located places. Models for prediction at different levels such as laboratory tests, small-scale simulation models, numerical models, and monitoring and evaluation of characteristic parameters on Real models are used to study and evaluate the shifting mechanism through the expression of related parameters. Based on result predictions from these models, the mitigation measures are determined.
3 Achievements in Landslide Risk Assessment
3.1 Landslide Identification and Integrated Landslide Mapping
3.1.1 Landslide Identification and Landslide Topographic Area Mapping
As mentioned above, the first input when assessing landslide risk is to recognize the type and location in which they are distributed. When a landslide occurs, it will change the topography of the area and leave easily recognizable and distinguishable signs on the land surface from the surrounding environment. Most of these signs are related to changes landform of the topography of the area. To identify the phenomenon of landslides, it is necessary to have a view from above and the surrounding area of the overall slope failure to identify the features and interpret the changing process of these features over time. In the recognition and interpretation of aeronautical images, a three-dimensional view of the deformation of a slope can be established through stereoscopic observations of pairs of aerial images for aeronautical interpretation. Through this stereoscopic image, an overall picture of the land can be observed; Landform classification, land cover classification, and vegetation classification can be determined; The area of land deformed by the landslide can be cleared (Miyagi 2013).
Based on knowledge of geomorphology, characteristics of vegetation cover, drainage conditions, signs of surface displacement in the form of landslides such as scarps; irregular or hummocky topography below scarps, at the body; bare linear tracks oriented downslope; fresh rock exposure; fresh rock accumulation at the slope base; disordered vegetation and disarranged drainage… are realized.
These signs are analyzed and compared with an identification catalog of characteristics of landslide types such as Rotational slide, Translational slide, Compound slide, Debris slide, Debris Flow, complex slide, etc., from which the type, features, and evolution of the past displacement process can be determined, interpreted and recognized. However, the aerial photograph is a similar image. The interpretation of topographic information from vegetation images was used as a method of information collection. Depending on the image quality and scale, some positions or features within the slider block are not clearly defined. In such cases, it is necessary to conduct an actual field survey. The information on landslide features is recorded as a landslide identification information layer. The landslide distribution topographic map is created by the combination between the landslide identification information layer and the topographic information layer.
The method of building topography maps by aerial photo interpretation in Vietnam has been applied since early. However, the landslide inventory maps have only been carried out and published after 2000. A typical example of Landslide identification and mapping is a landslide inventory map along HCM road (Central region of Vietnam), which has been carried out and published in 2016.
Figure 1 is the landslide distribution map of the Kham Duc area, in Viet Nam—one of the 6 first sheets of landslide inventory maps. Based on the analysis of more than 100 black and white aerial photographs on a scale of 1/33,000, ninety-six landslides were recognized, with 5 types of landslides including (1) Rotational sliding; (2) (Translational sliding; (3) Complex; (4) Debris sliding; (5) Debris flow. Of the five types, the first three are identified by topographic and topographic form characteristics; the other two types are identified by the topographical features of the sliding body.
Currently, remote sensing image technology is being increasingly developed and provides better quality land surface images for analysis, but landslide identification based on interpretation of images taken from aircraft is still considered the basis for landslide identification, which has been carried out in Vietnam.
3.1.2 Application of Landslide Identification Using Unmanned Aerial Vehicles (UAV) in National Road Survey of Vietnam
Remote sensing imaging technology using unmanned aerial vehicles (UAVs) for landslide terrain recognition has become common in the field of landslide disaster prevention in many countries, but its application for survey and design of prevention and mitigation works in Vietnam has not been popularized. Research and application of remote sensing imaging technology using unmanned aerial vehicles (UAV) in landslide surveys of roads in Vietnam is a requirement set by the Ministry of Transport (MOT) to ITST the research team.
The experiment was conducted at the road connecting Lao Cai-Sapa, National Highway 4D, Highway 7-Nghe An Province, and Ha Long City—Quang Ninh Province using a UAV Phantom IV. The images were taken from the UAV with a larger than 70% overlap area. Classifying of photos, and handling blurry and noisy images were carried out with specialized software. The overall image of the target area was built by combining coordinates of UAV photos in the same GIS geographic coordinate system and correcting through the coordinates of the ground control point (GCP). The orthogonal image model and Digital surface model (DSM) were built from cloud point technology and automatically recognize similar points. Photograph of the area and contour maps, which were constructed from DSM-assisted landslide identification. Figure 2 depicts the development of a landslide topographic map of the section KM 119-Km121, National Highway 4D, Vietnam.
Through the study and evaluation of the type of UAV, the method of photography and interpretation analysis, the advantages and disadvantages of the method compared with conventional survey and images taken by UAV equipment, developing the landslide occurred map was considered feasible.
The advantages of this method are low cost, short implementation time, and high access capacity to hazardous locations where landslides are occurring. For the preliminary landslide survey, the 1:1000 scale landslide topographic map is suitable. The detail of the GSD ground image is proportional to the flight height, (the lower the flight height, the higher the detail image we can get). To build the landslide topographic map with a scale of 1/500 to 1/1000, UAVs fly heights within 50 to 150 m are suitable. To identify the microfeature of the sliding block such as cracks, scrap, etc., a GSD of at cm/pixel is required. However, the biggest drawback of this method is the uneven vegetation cover on the soil surface. In such areas, the Digital Elevation Model (DEM) will be built from the DSM after removing the vegetation cover elevation or comparing it with the available topographic elevation documents.
Although there are still limitations, this topographic survey method still has undeniable outstanding advantages of creating a panoramic view of the landslide object and short implementation time. This method is also safe for operators, and especially suitable for emergency response when a new landslide occurs.
3.1.3 Landslide Mapping through Alos World 3D Mapping (AW3D) and Google Earth
Landslide identification through stereo pair aerial photograph interpretation is considered a primitive technology for the construction of landslide distribution maps. However, in Vietnam, this source of air photo data is limited in quantity, image scale as well as shooting time. With the rapid development of remote sensing imaging technology, taking pictures of the entire surface of the earth has become common and available. Digital images with three-dimensional coordinate information of the land surface in various scale formats through services such as NASA’s SRTM can be used as an alternative to stereo pair aerial photographs by traditional technology in building stereoscopic images.
The landslide identification map of 70 km along National Highway 7—Nghe An Area—Vietnam was conducted to set up a test based on the above idea. Experimentation with 3 representative locations was conducted from AW3D data with 10 m contour generation. The comparison of the landslide interpretation results of the sliding block such as landslide boundary, main scarp, and landslide body with the actual ones of the landslide through the field survey showed similarities.
The AW3D data with a 5 m grid size, has almost the same level of detail as the 1/2000 scale aerial photograph used. Topographic maps with contours generated from the above DSM data combined with visualizations from Google Earth are used for analysis and identification of landslides along the route. Different viewing angles to Google Earth and Landform from topographic maps with contour lines from 5 to 10 m have helped identify small to medium-scale landslides with a width of 50–200 m (Fig. 3).
Landslide mapping through Digital 3D World map (AW3D) and Google Earth offers many advantages over traditional aerial analysis, but it still has its weaknesses due to the use of DSM for analysis. In some areas, “fake” main scrap may be identified where there is a sudden change in vegetation cover with high height.
3.1.4 Assessing Landslide Reactivation Potential and Risk Evaluation Mapping for Occurring Landslides
The next issue with assessing landslide risk for an area after the landslides are identified is to compare their susceptibility to reactivation. The ability to reactivate a landslide depends on 3-factor groups including Trigger factors (rain phenomena, earthquakes), Characteristics of the sliding mass (type of motion, and geometrical and geomorphological features), and Internal factors (geology, hydrology). The trigger factor group is a passive factor and the sensitivity to the landslide reactivation is proportional to the magnitude and frequency of the phenomenon that their potential did not study in this research. Excluding the trigger factors, when assessing the risk and building the risk assessment map for humid tropical areas like Vietnam, 2 group factors landslide characteristics and internal factors have been considered.
The groups of feature elements of the landslide are considered from the perspective of geomorphology and divided into sub-groups according to the point of view of Miyagi et al. 2004 including (1)The micro landforms of the soil body as an aspect of the characteristics of movement; (2) The boundary of major landslide components as an aspect of the time process and (3)The landslide topography and the adjoining environment as an index of geomorphic setting. In each sub-group, the related geomorphological member factors that affect the reactive ability of the sliding block were mentioned and classified by a score (Fig. 4).
As for the internal factors, the geological factors mentioned include geologic age, bedrock lithology and structure, surficial geology, and level of weathering. Each geological factor is classified into different features or levels that describe the susceptibility to landslide reactivation and are characterized by a certain number of scores (Fig. 5).
Based on the discussion, the assessment of each factor affecting the reactivity is divided into different levels which are specified by a certain weight score value using the AHP model. The susceptibility of different landslide sites to reactivation depends on the total AHP score, in which geomorphology will account for 44.44% and geology will account for 55.56%.
Figure 6 depicts a landslide risk map in the area of Thanh My—one of 6 landslide risk maps, which were developed based on the identification map and risk assessment method which is mentioned above. For these maps, the landslide reactivation potential is divided into 4 levels AA: very high, A high, B medium, and C low. This landslide risk assessment map is one of the first and most important ones in Vietnam for evaluation of the landslide reactive susceptibility to occurring landslides.
3.1.5 Landslide Potential Assessment and Landslide Susceptibility Mapping
For Landslide Risk Evaluation on natural slopes where landslides have not occurred, an assessment of the sensitivity to vulnerability of the slope by sliding was carried out.
The basic principle for evaluation is based on “past and present are the keys to the future” (Vaners 1984; Carrara et al. 1991). Under this hypothesis, future landslides are likely to occur under the same geological, geomorphological, and hydrogeological conditions as having occurred with previous landslides. The conditions that led to the landslide can be used to identify possible landslides in similar areas in the future.
On that basis, surveys and statistics on the occurrence of landslide phenomenon were conducted for the study area. The Landslide distribution map was developed for the target area. The main causative factors related to the occurrence of sliding block displacements such as slope angle (geomorphology), type of rock, fault density (geology), distance to the road, land used (human being), and precipitation (climate) are selected for assessment. The affection of causative factors to each occurred landslide was discussed and recorded.
Figure 7 is a score comparison matrix of causative factors and the classes within each factor, as required by the AHP method.
The causative factor maps were developed on the same scale from available data, in which each one was classified for assessment. The number and density of occurred landslides for each class are estimated by superimposing the landslide distribution map with the causative factors map in the GIS system. The sensitivity to a landslide of each class zone of causative factor maps is calculated and evaluated through the value of the number and density of occurred landslide. The weight value point was given to each causative factors map depending on their sensitiveness and their classes then an analytical hierarchical process (AHP) is used to combine these maps for landslide susceptibility mapping.
The landslide distribution map and Landslide susceptibility map of the corridor area, along HCM road from Quang Tri to Kontum in the center of Viet Nam are created from overlap landslide causative factors maps with calculated Eigenvectors from the AHP model.
According to the Landslide susceptibility map presented in Fig. 8, susceptibility to land sliding was divided into 4 classes very high, high, moderate, and low. The calculation showed that 82.66% of over 600 occurred landslides in the distribution map fall in the high and very high susceptible zone.
For each study area, the accuracy of the Landslide susceptibility map depends on the data and basic parameters selected to calculate and assess the risk of landslide hazards presented in the form of maps in the GIS system. The selection of input map data to calculate landslide hazard risk depends on three main factors, namely “relevance”, “availability”, and “scale” of the map.” Relevance” shows the close relationship of the main factors related to landslides in the study area. “Availability” refers to the available data and the data likely to be obtained in the study. And “scale” of the input map refers to the map scale of the factors affecting the landslide process that will be used to calculate the risk of landslide hazard for the study area. The accuracy of the map is characterized by the number of landslides occurring over a highly susceptible zone.
3.1.6 Digital Database Structure on Topography, Geology, Hydrology, and Landslide Identification for Landslide Mitigation in Survey and Design Work
Based on researching and applying digital databases for the management of landslides and disaster prevention in the world, an orientation to create and apply a database for storing, managing, and proposing landslide mitigation solutions is created. The built-in landslide data layer is considered a separate layer in the structure of many different data layers in the same geographic coordinate system GIS, which is known as Big Data. The research objective is to build a database structure on landslides including the information on topography, geology, and hydrology relating to landslide identification for landslide response and mitigation for traffic in the mountain zone. The data structure consists of two parts including Assessing the level of risk and Countermeasure selection.
3.1.7 Assess the Level of Risk
Besides spatial and temporal parameters as well as time and type of movement, relevant factors are divided into groups such as the Topographic group (length, height, natural slope gradient angle on the upper slope, distance from the foot of the slope to the road, and a number of steps on the slope), Geology group (slope materials, degree of weathering, the depth and strike of the material layer, cracks; and Hydrogeological and hydrogeological groups (daily rainfall intensity and signs of groundwater). The survey sheet for landslide collecting data is presented in Fig. 9. Each factor is again assigned a certain weight depending on its importance to the degree of susceptibility to displacement.
The weighted model (AHP) is then used for the classification and evaluation of landslides inbuilt stores, in which The total score (maximum 100) of the risk factors for each landslide site is divided into 5 levels Very High Risk, High Risk, Medium Risk, Low Risk, Low Risk, Very Low Risk with scores as 100–80, 80–60, 60–40, 40–20 and 20–1, respectively. (Thanh et al. 2020).
Three special factors of sliding blocks mentioned as Direction of discontinuity, Wethearing, and Groundwater are used to evaluate the level hazard of sliding blocks. Depending on the presence of these factors during the survey, the hazard level was divided into five levels, with the level increasing gradually. Table 1 presents the principles of assessing the hazard level of Landslides.
3.1.8 Countermeasure Selection
Based on assessing the level of risk and level of Hazard, the response and mitigation solutions are considered and selected. A detailed survey of the selected locations is conducted for further evaluation and countermeasures selection. Three groups of countermeasures include the impact mitigation Control works; Prevention works and monitoring and early warning work are given and recommended for applicability according to characteristics and recommended levels of danger. Depending on the actual situation of each landslide, suitable solutions will be selected.
The successful application of research results for building a database for National Highway NH4D, NH 6, and NH 7 in Vietnam is the first step for applying Big data in managing landslides along the roads.
3.2 Landslide Simulation and Testing
3.2.1 Development of Untrained Dynamic-Loading Ring Shear Apparatus
The dynamic-loading ring shear apparatus is Prof. Sassa’s Invention in 1992(DPRI-3). ICL 2 is the sixth generation device that was developed to donate it to Vietnam together with the technological transfer.
Like other shearing apparatus, ICL 2 was designed for the Physical simulation of landslides. It can reproduce stress on the potential sliding surface within a slope during rainfall, earthquakes, and both. It can also measure the shear displacement and pore pressure, as well as the mobilized shear resistance during shearing (Figs. 10 and 11).
The ICL-2 has six main components, including (1) Computer control system, (2) Main control unit for control and monitoring, (3) Instrument box, (4) Power supply box, (5) Pore-water pressure control unit with backpressure unit, and (6) de-aired water and vacuum system. The overview of ICL-2 is shown in Fig. 10.
The important improvement of ICL 2 compared to previous models is the sample box with 100 mm inner diameter, 142 mm outside diameter, and the largest sample height of 52 mm to create a shearing area of 79.79 cm2 and Max. shear speed (cm/s) of 50 cm/s and the Max. Pore water pressure and the Max. normal stress up to 3000 Kpa. ICL2 is capable of simulating undrained shear state, pore-water pressure control state, and monitoring earthquake state.
After ICL-2 donated version has been completed, it has been tested by Vietnamese long-term and short-term trainees for around two years. All trouble sources were solved, and all trainees had the confidence to use ICL-2. Manuals via video and written materials were developed for use by future trainees. The video manual made by ITST is included in Landslide Dynamics: ISDR-ICL Landslide Interactive Teaching Tools.
3.2.2 Mastering the Use of the ICL2 and Applying it to Study Rapid Landslides in Vietnam
The ICL2 ring shear apparatus is single-piece production equipment that is relatively complex and highly costly. To efficiently operate the device and apply it to study rapid landslides in Vietnam, understanding its structure, comparing its performance with other devices, conducting practical tests, and building operating instructions were required by the MOT.
After 2 years of experience on the ICL2 device, long-term and short-term Vietnamese trainees made some comparative comments on the improvement of ICL compared to previous devices as following (Fig. 12).
The new high-pressure undrained ring shear apparatus (ICL-2) is used to simulate the undrained shear behavior of soil under all kinds of loads such as minimum normal stress: 300 kPa (corresponding to a depth of 15 m) to a maximum of 3000 kPa (corresponding to a sliding surface depth of 100–200 m), at the maximum shear speed of 50 cm/s. It can create a large shear displacement (such as 5–20 m shear displacement) necessary to measure the shear strength during the motion of landslides. It is quite different from the direct shear test and the triaxial test with shear displacement being less than a few centimeters to 10 centimeters.
Another advantage is ICL 2’s ability to create pore pressure and monitor up to the Undrained test. High-speed landslides are the result of high-pore water pressure mobilized during motion. To measure pore water pressure during movement, a prerequisite is the ability to hold the pore pressure generated in the shear box/cell. This feature cannot apply to Direct shear, Triaxial tester, and conventional ring shear tester.
The new high-pressure undrained ring shear apparatus (ICL-2) with more advanced features than the old ones. The normal stress loading system is the biggest improvement of the ICL-2 device compared to the DPRI devices. The normal stress loaded from a long frame above the shear box in DPRI devices is replaced by pulling the center axis of the shear box on the ICL-2. The deformations of the preload during stress changes are very small on ICl-2, giving better resistance to drainage. The next difference in the ICL and DPRI series of machines is the waterproof rubber ring. The rubber ring of all DPRI devices is firmly attached to the cutter box. The thickness of the adhesive layer must be uniform and the height of the rubber ring top faces of the inner and outer rings should be equal to maintain the undrained condition. They must have an extremely smooth surface, which requires high technology to be manufactured. In the ICl series, the rubber ring is compressed by a Teflon ring and held in place by a metal ring with screws attached. It is therefore not necessary to be as smooth as the rubber rings used in the DPRI series of machines.
The difference between the ICL-1 and ICL-2 devices is the normal stress and shear speed. The maximum shear speed of the ICL-1 device is 5.4 cm/s and the maximum shear speed of the ICL-2 device is 50 cm/s. The maximum normal stress and resistance to drainage of ICl-1 are 1 MPa while that of ICL-2 is 3 MPa.
Currently, for other common tests, it is not possible to determine the residual shear resistance value of the soil. So when it is only necessary to test the soil in a non-destructive state, other common tests such as Direct shear, Triaxial tester as well as Conventional ring shear tester can be used. After failure, it is necessary to determine the parameters of the soil using the ring shear test.
ICL2 ring shear apparatus plus LS-RAPID simulation software are effective tools to study the mechanism of movement and assess the risk of large-scale landslides.
3.2.3 Elucidation of the Initiation Mechanism of the Landslide
ICL2 was applied by Lam Huu QUANG and other Vietnamese engineers to test the samples taken from drilling in the Hai van Station and succeeded to test those samples. This is the world’s first successful application of the undrained dynamic-loading ring-shear apparatus to test samples taken from the potential sliding surface found from the drilled cores in the precursor stage of Landslides.
This research applied the undrained dynamic loading ring shear apparatus ICL-2 to drill-core samples from the precursor landslide. Samples for ring shear tests were taken from sandy soil layers of cores found at depths of ~21, ~31, and ~ 50 m, which were considered as possible landslide slip surfaces.
The inclinometer monitoring detected slight movement at about 50 m depth, where the boundary between highly weathered granitic rock and weathered granitic rock was identified. The test was carried out for this in 3 models: Drained Speed Control Test, Undrained Monotonic Stress Control Test, and Pore Water Pressure Control Test (Figs. 13, 14 and 15).
3.2.4 Development of Hazard Assessment Technology for the Precursor Stage of Landslides (Using ICL Result and LS Rapid)
The dynamics of post-failure motion of the Haivan landslides and the development of hazard assessment technology for the precursor stage of landslides have been carried out to predict its influence on the national railway which locates on the toe of the landslide body. After considering the level of risk, the slip surface at the 50-m-depth was selected as the target (Fig. 16).
LS-Rapid is an integrated Landslide Simulation Model, which can simulate the initiation and motion of landslides triggered by earthquakes and rains. Calculations are made using the steady-state shear resistance measured by the undrained ring shear apparatus.
According to Fig. 15, a vertical column is considered within a landslide mass. The model calculated the discharge of X and Y direction (M, N) and the height (h) of soil mass by assuming that the balance of all forces acting on this column (Self-weight (W), Seismic forces, Lateral pressure, Shear resistance including the effect of water pressure) will accelerate the soil mass (m) by acceleration (a) on the horizontal plane(1) and the discharge flowing into the column is the same with the change of the height of soil (2).
The landslide dynamic parameters obtained from the ring shear test of the 50-m-deep sample were used in an integrated numerical simulation model LS-RAPID. The simulation result (Fig. 17) gave the critical pore-pressure ratio for landslide occurrence, the landslide’s likely maximum speed, total volume, and depth of landslide debris that could cover the railway. The displacement activation parameters of the sliding block and the predictive model results contribute to the consideration of establishing a railway safety system at Hai Van through monitoring and early warning.
3.3 Monitoring Sensitive Parameters on Landslide Sites and Developing an Early Warning System
3.3.1 Investigation of Active Landslide for Selection of Pilot Monitoring Site
To evaluate risk and develop early warning systems based on landslide monitoring, the selection of a suitable site is very important for study and monitoring equipment setting. With criteria required for study such as large scale, deep seat, and active landslide, the Haivan landslide had been considered the most suitable site (Fig. 18).
Hai Van is a particularly important area on the North-South railway line, located in the central region. Hai Van railway station is located at a low position on the mountainside of Bach Ma mountain range overlooking the sea. The site has a particularly complex geological structure with a strongly weathered granite origin. Every year, this location receives a very large amount of rain (over 3000 mm/year). Slope damage has occurred along the railway line through this area, causing significant damage to Vietnam Railways Corporation. The area is still a potentially high-risk area for landslides.
In terms of geomorphology, the ridge area behind Hai Van Railway Station has the topography of a landslide that happened in the past. Over time, with the weathering and erosion, the growth of plants in the area has erased part of the landslide identity. Aerial photo interpretation and topographic maps of the area indicate that the area has a topographical shape of a deep-seat landslide with a wide range (1.5 km width, 1.6 km height). The top of the area has a horseshoe shape with a large main sliding scarp which was divided into 2 sub-scarps.
The actual site survey results lead to the conclusion that this area has a large sliding block—a product of strongly weathered granite that has moved from the original position of the slope, leaving large and small scarps on the upper slope, and creating a landslide body at the bottom of the slope. The later internal magnetic analysis also confirmed the above statement (Fig. 19).
The landslide body was divided into secondary sliding blocks with widths from 50 m to 100 m. Figure 18 (left) shows a Landslide identification map in the Hai Van area. The Secondary sliding blocks B1, B2, B3, and B4 are located on the body of a deep seat landslide, in which reactive signals such as small deformation cracks were found along sections A-A of the B3 block. The national railway runs through the body deep seat landslide and sliding block body. The surface displacement broke the structure of the protective stone roof protection structure and pushed the railway out of its horizontal position by 20 m in 2007.
Drilling and sampling of the intact coaxial cavity were carried out for the geological survey. At the center of the landslide body, three boreholes were made with a depth of 30, 60, and 80 m. The results of stratigraphic distribution indicate that the main geological products are granite and weathered granite with different degrees of weathering. From a depth of 0 to 51 m, the geology is a gravel-sand mixture caused by strongly weathered granite and unfinished granite cores. Notably, from 0 to 12 m, the degree of weathering is very intense, to which some weathered part easily breaks by pushing it with one’s finger.
From 51 to 54 m depth, a layer of weathered fine Granite with white granite composed of feldspar, quartz, muscovite and trace biotite solid with crack was found. From 54 to 80 m, the geology of the area is granite bedrock with the distribution of leucocratic migmatite, granite, gneiss, and melanocratic migmatite rocks respective to the increasing depth.
The experience of landslide identification, the result of site surveying, drilling, and sampling at Hai Van identified the overview of large-scale landslides with a potential depth of slip surface of 51 m and secondary sliding blocks with a potential depth of slip surface of 10–12 m. For risk evaluation, the evidence of the slip surface needs to be made clear before testing its samples for the Elucidation of the initiation mechanism and the dynamics of post-failure motion of the targeted landslides. Hai Van landslide is considered a standard example for surveying to evaluate landslides, especially large-scale landslides in Vietnam.
3.3.2 SP Drilling Method, Slide Surface Assessment, Measurement of the Groundwater Table
The purpose of the geological survey at Hai Van is not only to collect samples for testing but also to install monitoring devices in the borehole to assess the state of the sliding block, accurately detect the slip surface, and monitor groundwater level.
The Hai Van Landslide area’s geological features include deeply weathering rocks, and expectations were that it would be difficult to do drilling work and retrieve high-quality core samples there. The SP method was adopted as it made it possible to do drilling work with the drilling equipment and personnel in Vietnam. The difference from conventional drilling methods is that this drilling method uses sleeve-incorporating double core barrel and diamond bits, which make it possible to extract the core from loose layers as gravel and sand mixed with gravel. At the same time, this technology of drilling was attempted to transfer to Vietnam.
Geological drilling for the sliding surface survey at Hai Van was carried out at the center of the landslide body with a depth of 30, 60, and 80 m. Lithologic description, Drilling samples, and position of drilling holes on the profile section of the Haivan landslide are presented in Fig. 20. The objective of drilling for the sample is to take uniform drilling cores for detecting the depth of slip surface, aquifer, fractured and weathered geological layers as well as bedrock through observation and evaluation of stratigraphic distribution layer from the cores as well as to mechanically test parameters of soil and rock such as grain size, water content, volumetric weight, limit Atterberg, shear strength. Survey drilling results indicate 2 suspected zones at 10–12 m depth, where its uniaxial compressive strength and level weathering change, and at 50–51 m depth, where a weathering boundary between crushed sand and gravel and hard granite bedrock appeared. These zones are believed to have a high potential for locating slips surface of the Haivan landslide.
For determining the slip surface, relying on the results of the geological drilling survey is inadequate. The inclinometer device installed in the 80 m depth borehole to evaluate the horizontal deformation is an effective method to confirm the displacement position of the sliding block in depth.
The inclinometer horizontal displacement measurement results for Hai Van Landslide as an example are presented in Fig. 20, which showed 2 zones in depth (10–12) and (50–51 m) corresponding to the level conversion of weathering from strong to weak weathering and from weaky weathering and hard granite bedrock, respectively. The horizontal displacement results confirm that the slip surface prediction based on the geological core drilling results is reasonable (Fig. 21).
For groundwater level monitoring, the Groundwater pressure gauges were installed in 30 m and 60 m depth boreholes. Monitoring Position of groundwater on the profile section of Haivan landslide and Initial Monitoring data of rainfall, the groundwater table is presented in Fig. 22. Initial Monitoring data showed the sensitive relationship between groundwater table and rainfall, which is considered a potential cause to trigger the landslide movement.
3.3.3 Development of the Integrated Automatic Monitoring System for Rainfall-Groundwater-Slope Movement
For the study of the actual performance of the sliding block, the monitoring system is set up to monitor the movement of the sliding block through representative indicators such as precipitation, surface displacement, displacement within the sliding block, and groundwater level to verify in-room studies and early warning establishment.
In the Hai Van area, an integrated, continuous monitoring system, time synchronization, and automatic data transmission have been established. The location of each device is designed depending on the characteristics of each. Device. The number of devices is shown in Table 2 below. The layout diagram of monitoring equipment of the Long span Extensometer of the monitoring equipment is shown in Fig. 23. The entire system of equipment is powered by a solar battery system. Field monitoring data can be transmitted directly and in real time to the monitoring center in Hanoi.
The results of integrated monitoring between accumulated rainfall, groundwater height, and actual surface displacement are established in Fig. 24. The signs of the increase in displacement are consistent with the increase in groundwater level which was predicted.
3.3.4 Principle for Landslide Early Warning System Establishment
Landslide early warning is generally understood as the recognition of the beginning of sliding block movement through the values of the predicted motion trigger parameters. To implement early warning effectively, it is necessary to have a reliable landslide forecasting model and the monitoring and measurement of trigger parameters accurately. Depending on the landslide characteristics and forecasting model, the monitoring indicators will be selected and divided into different warning levels such as safe, dangerous, and especially dangerous.
With the advancement of technology, monitoring and measuring with accuracy using modern equipment for relevant parameters such as surface displacement, groundwater level, deformation, and precipitation can be carried out in an integrated and real-time manner. The arrangement of the Hai Van monitoring system is a vivid and complete example of this monitoring system.
The results of topographical, geological and, groundwater surveys at Hai Van have shown that the Hai Van Landslide is a large sliding block with a sliding surface at depth (50–51 m) and its landslide body was divided into 4 secondary blocks with shallow slide surface as (10–12 m). The results of the deformation measurement in the borehole also show signs of displacement at both 51 m and 10 m depths. Therefore, to establish an early warning system, the displacement possibility of both sliding blocks was considered.
To determine the sensitivity parameters to the displacement of large-scale, deep seat landslide, the soil sample of the drilled core at 51 m depth was tested, using ring shear ICL2, and its results were used in the integrated numerical simulation model LS-RAPID.
Results of susceptibility assessment of the precursor stage of a threatening landslide at Haivan Railway Station, Vietnam have shown that the pore-water pressure corresponding to the height of groundwater in the borehole can be used as an important monitoring parameter for early warning (Quang et al. 2018). The value of the height of the groundwater level rising 14 m to the borehole mouth in Fig. 25 is the key value that activates the sliding block to start moving.
Landslide experiments on flume slope with artificial rains and in the lab were conducted to study the displacement mechanism of the shallow sliding block at Hai van, and to predict the sensitivity parameters for the displacement of the sliding block to establish early warning systems.
Soil samples were taken from the body of the large slip block at the site for simulation experiments with different rainfall levels. During the experiment, the relationship between the displacement of the slope surface and the change in groundwater level of the soil layer was continuously monitored and recorded.
Before the slope is demolished, the displacement velocity of the soil mass increases very quickly. The image of the Landslide experiment on natural slopes with artificial rains from the beginning stage and ending stage of the moving process is presented in Fig. 26. The sliding block gradually changes state from a steady state, with small cracks appearing, to cracks opening, to rapid movement down the slope, and at the end of the movement. The measured displacement velocity is proportional to the precipitation and the slope displacement. Figure 27 presents a result of displacement, accumulated precipitation, and the reverse velocity value over time. For this test, the inverse of displacement velocity by time was created, and accumulated rainfall at the inverse velocity of Zero was used as a sensitive parameter for displacement prediction.
Due to the characteristics of Vietnam’s North-South railway going through the affected area of large and secondary landslides, the movement of one or both landslides at the same time will hurt the railway. Under the same triggering conditions due to heavy rain, the hurt in the borehole and accumulated rainfall are used as key parameters for early warning.
3.4 Other Achievements
3.4.1 ITST-ICL Guideline for Landslide Risk Assessement
ITST-ICL guideline for landslide risk assessment is a set of standardized documents intended to guide the survey, assessment, and prediction of landslide risk.
This set of guidelines is the result of the Technical Cooperation Project named “Development of Landslides Risk Assessment Technology along Transport Arteries in Vietnam”, which was carried out by cooperation of ICL and ITST.
The Guideline is in both English and Vietnamese, and was published in 3 volumes: Volume 1-Site survey and landslide risk assessment mapping; Volume 2-Testings, simulations, and software for landslide risk assessment and Volume 3-Monitoring and Experiment landslides by flume combined with artificial rain. The content of the integrated set of guidelines for landslide risk assessment is divided into the following 6 sections with 33 guidelines (GLs), including: (1) Site Mapping and Forecasting, (2) Material Testing, (3) Monitoring, (4) Model testing and (5) Software application. Each guideline is coded as GL xx—2016, compiled in the format of Vietnamese national standards with content covering the name of the guide, abstract, the scope of application, terms, definitions, the main body of the guideline, conclusions, and references. The number and name of the guidelines have presented the Table 3 and Fig. 28.
The development of the landslide integrated guideline is not only useful for research and implementation of the project but also serves as an important reference for students and researchers in landslide risk assessment as well as consultants in assessing the risks of landslides and proposing mitigation solutions. The next step of the project dissemination strategy is to upgrade these guidelines to the basic ITST standard and then to become the Vietnam National Standard (TCVN) to strengthen the development of applied technology nationwide for the prevention and mitigation of landslides in Vietnam.
3.4.2 Developing High-Quality Human Resources for Landslide Research through International Cooperation Projects SATREPS Project
The project “Develop Landslide Risk Assessment Technology along Transport Arteries in Vietnam” is the SATREPS project, formed in 2008, within the framework of the program “Science and technology policy” established by the Ministry of Foreign Affairs through the Japan International Cooperation Agency (JICA) and the Ministry of Education, Culture, Sports Science and Technology (MEXT) through the Japan Science and Technology Agency (JST). The project implementation period is 5 years from 2011 and ends at the end of 2016. The overall objective of the project is to socialize the development of landslide risk assessment technology and develop an early warning system that contributes to the safety of the traffic system in particular and the safety and development of the mountainous community in Vietnam. The project is implemented by the Institute of Transport Science and Technology (ITST) and the International Landslide Society (ICL).
The basic research and training achievements of the project include:
-
1.
For Human Resource Development for Landslide Research: After finishing the project, 3 PhDs and 5 Masters were trained in Japan and 05 other research students are continuing to study next year.
-
2.
To Landslide risk mapping: Six landslide inventory and risk evaluation maps along Ho Chi Minh road with a length of 60 km, scale (1: 12,000), and Hai Van area map had been established using aerial photo interpretation and site investigation. An evaluation of the sensitivity map for 150 km along the Ho Chi Minh Trail was developed using the AHP method. (Luong et al. 2016; Tien et al. 2016). Aerial photo analysis, combined with pictures taken by UAV to identify landslide spots along Highway 7, (25 kilometers) from Muong Xen to Tam Quang was chosen as the subjects for the setup identifier map (Dung et al. 2016). Recognition conditions of pre-soil through analysis model digital surface (DSM) of the extent of forest cover and others have been developed.
-
3.
To In-room experiment and landslide simulation: to develop investigation and simulation technology of landslide phenomenon, able to apply for slip surface deeper 100 m. (Landslides Magazine—Vol.11, No. 5 in 2014.) The developed ring shear apparatus (ICL II) was revised in 2014–2015 based on the experiences of testing by Vietnamese short-term trainees as well as long- term trainees. (Quang et al. 2018; Tien et al. 2017). Shifting mechanisms at Hai Van Samples were taken from the ground and the drilled cores at various depths in the Hai van Landslide were tested using ICL-2 and computer simulation was conducted based on the measured parameters by Vietnamese researchers. Adding the function of simulating tsunamis generated by earthquakes is one of the targets of JST research.
-
4.
Setting up an early warning and monitoring system at Hai Van station: Selection Two Landslides and Three boreholes were drilled and sampled in their original form; An integration monitoring system including rain gauge, extensometer, inclinometer, total stations, and GNSS are developed here. An automatic data transmission system from Haivan to office projects in ITST and a display system have been set up. This monitoring system can allow monitoring on-site and from real-time remote expressions of sliding blocks.
Landslide chutes, recording systems, and pore water pressure sensors are provided and adapted for ITST experiments. Results of this contribute to the landslide early warning principle.
3.4.3 SAKURA Science Program
The program is implemented by Tohoku Gakuin University and the Institute of Transport Science and Technology under the sponsorship of the Japan Science and Technology Agency (JST). The duration of the program is from 2017 to 2018. The goal of the program is to strengthen the scientific research capacity of young engineers under 40 years old to participate in learning and exchanging Japanese experience and technology on disaster prevention.
Outcomes from the program:
The Sakura Science Program provided young ITST engineers and researchers with initial knowledge and understanding of landslides through attendance at an international symposium, lectures on landslides, and special visits to the typical large-scale landslides in Japan. At the above landslide locations, modern solutions, and new types of works to reduce the impact of landslides on traffic, infrastructure, and people were introduced by Japanese experts. Young researchers from ITST have investigated and collected landslide data by UAV combined with ground control point measurement by the GNSS system. 15 young Vietnamese engineers and researchers under the age of 40 participated in the program in Japan for 10 days in 2017 and 2018.
3.4.4 GRASS ROOT Project
Grass Root Project Named “Capacity Building of Local Community for Slope Disaster Risk Reduction (Slope DRR). It was implemented under the cooperation of Kurihara City, Advantechnology Co., Ltd., Lao Cai Province under the support of JICA (Miyagi et al. 2020). The project time is 2019 to 2023. The Institute of Transport science and technology participated in the project as an expert and coordinating agency. Project Purpose is to help residents’ disaster prevention organization can carry out disaster prevention and evacuation plan in the pilot commune of Lao Cai province under cooperation with the administration,
The project’s output includes:
Basic understanding of the community territory: shares disaster reduction power strengthening-related information, and understands the disaster risk on site.
Support the resident activity: The education program on grasping disaster risk and evacuation drill is carried out by the collaboration between Japan, Lao Cai DARD, and the local community.
Development of activities in the local community: Management and improvement of the Slope DRR such as maps, manuals, and evacuation drills will be carried out by the leadership of the Lao Cai side.
Cooperation and deployment:Japan, the Lao Cai DARD, and relevant ministries and agencies are provided with the knowledge of this project as an example.
The basic information for disaster risk reduction is established through (1) the map of landslide terrain distribution in a large area, the scale of 1/25000), the landslide risk reduction baseline map slopes for pilot communes and the vicinity of 1/5000, and (2) interactive DRR maps based on UAV images and user manuals were developed with the main contribution of ITST researchers. The above project is a perfect combination of researchers, managers, and local communities for landslide reduction.
4 Conclusions
Landslide risk assessment is one of the fundamental issues in the Vietnamese government’s “proactively prevent natural disasters” strategy for natural disaster prevention, response and mitigation solutions for mountainous areas and central highlands. Over the past 10 years, with the efforts of Vietnamese scientists and the support of international cooperation projects and research programs, the Institute of Transport Science and Technology has made important strides in assessing landslide risk.
The overall landslide risk assessment for the area can be done through established and modern landslide identification as well as the development of integrated maps for the classification, distribution, reactivation, and landslide potential assessment. Through the principle of assessment, an application that builds a database on landslides along the roads has been carried out, contributing actively to the management of the road system in mountainous areas.
For landslide risk assessment with specific landslides, especially for deep, large-scale landslides, site topographical and geological surveys define landslide geometrical features. The determination of the slip surface and landslide state can be done by SP drilling and installation borehole inclinometer. Testing samples taken at slip surface by ICL2 ring shear apparatus and application of LS Rapid simulation can assist in the understanding of the initiation mechanism of the Landslide and the dynamics of post-failure motion of the targeted landslides for the prediction. Early warning capabilities can also be realized through synchronous and real-time monitoring systems of the affected parameters using accurate monitoring equipment.
To facilitate the replication and application of research results, an integrated set of guidelines for landslide risk assessment has been developed in Vietnamese. This guideline is gradually being upgraded to a basic standard and national standard.
The achievements for landslide risk assessment are not only useful for natural disaster prevention, response, and mitigation in Vietnam but can also be applied to humid tropical areas with similar conditions.
References
Abe S, Tien DV, Ha DN, Hoshide T, Nishitani T, Miyagi T (2018) Topography and landslides in weathered granitic rock areas – Hai Van landslide in Central Vietnam. Landslides 15:1675–1689
Carrara A, Cardinali M, Detti R, Guzzetti F, Pasqui V, Reichenbach P (1991). GIS techniques and statistical models in evaluating landslide hazard. In Earth surface processes and landforms, 5, 16 (pp. 427–445). New York: John Wiley & Sons
Do NH, Asano S, Ochiai H, Goto S, Huynh TB, Nguyen KT (2021) Determining the position of the sliding surface of rainfall-induced landslides at Hai Van station, Viet Nam during rainy season in 2016. In: 60th Domestic Conference on Japan Landslide Society, pp 170–171
Dung ND, Miyagi T, Luong LH, Hamasaki E, Hayashi K, Tien DV, Daimaru H, Abe S (2016) Trial of landslide topography mapping using ALOS W3D data – case study along the National Road no. 7 in Central Vietnam –. Trans, Japanese Geomorph. Union 33-1:127–140
Luong LH, Miyagi T, Tien PV (2016) Mapping of large scale landslide topographic area by aerial photograph interpretation and possibilities for application to risk assessment for ho chi Minh route, Viet Nam - trans, Japanese Geomorph. Union 33-1:105–126
Luong LH, Miyagi T, Tien PV, Loi DH, Hamasaki E, Abe S (2017) Landslide evaluation in center provinces of Vietnam – WLF 4
Miyagi T (2013) TXT-tool 1.081–2.1 landslide mapping through the interpretation of aerial photographs
Miyagi T, Prasada GB, Tanavud C, Potichan A, Hamasaki E (2004) Landslide risk evaluation and mapping—manual of landslide topography and risk management. Report of the National Research Institute for Earth Science and Disaster Prevention No. 66, pp 75–137
Miyagi T, Thanh NK, Tien DV, Luong LH, Viet QV (2020) Slope disaster risk reduction map as a communication tool for community-based DRR in Japan & Vietnam
Quang LH, Loi DH, Sassa K, Takara K, Ochiai H, Dang K, Abe S, Asano S, Ha DN (2018) Susceptibility assessment of the precursor stage of a landslide threatening Haivan Railway Station, Vietnam. Landslides 15:309–325
Sassa K, He B (2012) TXT-tool 3.081-1.1 landslide initiation mechanism. In: ICL, landslide teaching tools. ICL, Tokyo, pp 205–214
Thanh NK, Miyagi T, Shinobu S, Tien DV, Luong LH, Ha DN (2020) Developing Recognition and Simple Mapping by UAV/SfM for Local Resident in Mountainous Area in Vietnam—A Case Study in Po Xi Ngai Community, Laocai Province. Understanding and Reducing Landslide Disaster Risk pp 103–109
Tien DV, Miyagi T, Abe S, Hamasaki E, Yoshimatsu H (2016) Landslide susceptibility mapping along the ho chi Minh route in Central Vietnam: AHP approach applied to a humid tropical area-. Trans, Japanese Geomorph. Union 33-1:79–104
Tien DV, Khang NX, Sassa K, Miyagi T, Ochiai H, Vinh HD, Quang LH, Dang K, Asano S (2017) Results of a technical cooperation project to develop landslide risk assessment technology along transport arteries in Vietnam (IPL-175). Workshop on World Landslide Forum - Advancing Culture of Living with Landslides, pp 411–417
Uzielli M (2009) Quantitative Estimation of Vulnerability to Landslides: the VIS framework, International Centre for Geohazards (ICG), Norwegian Geotechnique Institute
Vaners D (1984) Landslide hazard zonation: a review of principles and practice. UNESCO, Paris
Acknowledgments
This report is a reconstruction, addition, and correction of research results of The project “Develop Landslide Risk Assessment Technology along Transport Arteries in Vietnam”, Which had been implemented from 2011 to 2016 and other post-project ones.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
Copyright information
© 2023 The Author(s)
About this chapter
Cite this chapter
Van Tien, D. et al. (2023). Landslide Risk Assessment in the Tropical Zone of Vietnam as a Contribution to the Mitigation of Natural Disaster Vulnerability. In: Alcántara-Ayala, I., et al. Progress in Landslide Research and Technology, Volume 2 Issue 1, 2023. Progress in Landslide Research and Technology. Springer, Cham. https://doi.org/10.1007/978-3-031-39012-8_13
Download citation
DOI: https://doi.org/10.1007/978-3-031-39012-8_13
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-39011-1
Online ISBN: 978-3-031-39012-8
eBook Packages: Earth and Environmental ScienceEarth and Environmental Science (R0)