Mechanisms of rainfall-induced landslides and interception dynamic response: A Case Study of the Ni changgou Landslide in Shimian, China

doi:10.21203/rs.3.rs-3132547/v1

Download PDF

Article

Mechanisms of rainfall-induced landslides and interception dynamic response: A Case Study of the Ni changgou Landslide in Shimian, China

https://doi.org/10.21203/rs.3.rs-3132547/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 18 Jan, 2024

Read the published version in Scientific Reports →

You are reading this latest preprint version

Geological hazards, especially landslides and mudslides, are frequent in Caoke County, Sichuan Province, China. In September 2022, a large-scale mudslide occurred in Ni changgou with continuous rainfall, and it caused great damage to the surrounding roads and downstream village buildings. The mechanical parameters of the soil were obtained through a basic investigation of the landslide characteristics of Nichanggou. Upon that, the finite element-discrete element method was used to reconstruct the three-dimensional numerical model of the landslide on the right bank of Nichanggou, and the initiation mechanism of rainfall on landslide and the formation of debris flow impact dam process were simulated. Furthermore, the pore pressure, stability coefficient as well as displacement of the landslide body were analyzed, it turned out that under long-term continuous rainfall conditions, the slope safety factor decreases with the increase of rainfall time by analyzing the pore pressure, stability coefficient and displacement of the landslide body, which exist under different conditions of heavy and extreme rainstorms. Then the laws affecting landslide initiation and damage characteristics could be found out. The duration of landslide movement is about 200 s, the maximum average velocity of the landslide reaches 4.85 m/s, and the average movement distance is close to 500 m. These measured data matches well with that simulated landslide mound morphology.In addition, as the downstream debris flow washout is part of the initiating material source, the topography and debris flow are simplified accordingly, the whole process of impact, climbing height and back silting of the debris flow impact blocking structure is reproduced. The coupled method shows the whole process of debris flow from formation, to blocking, The results show desired effect on the mudflow prevention and management.

Physical sciences/Engineering

Physical sciences/Engineering/Civil engineering

Rainfall

Landslide initiation mechanism

Interception dynamic response

Debris flow

Numerical simulation

Landslides are usually caused by earthquakes or heavy rainfall and are often accompanied by a large number of casualties and extensive damage[1,2,3]. Due to the huge scale, unpredictability, short duration and destructiveness of landslide hazards, research on the initiation mechanism and sliding characteristics of landslides should be given great attention in this regard[4]. Considering the intensity of rainfall, the watershed area of the pile, the depth of the pile, and the number of hydraulic conductivities,Wu-Young [5] et al. analyzed the destabilization of post-earthquake loose piles and its initiation. Haruka Tsunetaka [6] et al. concluded that continuous duration of the rainfall threshold and the rainfall pattern were more likely to trigger the sliding of landslides, and the entrainment of the piled sediments facilitated the development and propagation of the subsequent landslides surge and propagation. Under the effect of continuous strong rainfall, Huilin Bai [7] et al. argued that the pore water content of the loose landslide body should increase rapidly in a short period of time, and the pile would present a flowing state with much higher permeability, at which time the loose landslide body would turn easy to become a landslide under the action of gravitational potential energy of the gully. Taeun Kang [8] et al. believed the water generated by rainfall infiltrated into the stratum, leading to an increase in subsurface runoff. As a result, the concentration of the sediment-water mixture and the viscosity of the water flow decreased, which accelerated the generation of landslides plus thus the occurrence of slides. Li Jing [9] revealed the damage evolution law of the soft and weak soil interlayer under the influence of rainfall. which could be concluded as follows: rainfall showed a significant degradation damage effect on the rheological properties of the weak interlayer, and the shear strain of the weak interlayer increased continuously with the increase of the number of rainfall, the degree of rheological rate fluctuation also expanded continuously, and the rheological time-dependent characteristics were obviously enhanced. Finally, the rainfall would bring about a damage circulation system inside the weak sandwich, which accelerated the damage deterioration of the weak sandwich and leaded to sliding.

Landslide initiation can form mudflow by sliding along the channel. in this sense, the establishment of barrage dams, drainage channels and other control works are positively useful for the control of debris flow[10]. At present, lots of research works have been conducted in this regard, either at home or abroad. Among which, Dong Yaogang et al. [11] used an indoor flume test to simulate the process of mudflow on three different types of check dams in the indoor flume tests, of which, and the vibration acceleration of mudflow was significantly reduced after the addition of check dams, and the addition of drainage culverts to check dams could significantly reduce the destructive effect of mudflow. Based on hydrodynamic theory, a parametric correction to the calculation equation of existing empirical debris flow impact force peak was made by Li Bailong [12], and several comprehensive suggestions for designing the dam were also put forward considering the effect of debris flow on the dam body, rushing height, and climbing height. Additionally, in order to study the flow pattern, retention, and regulation characteristics of debris flow through check dam, Zhou Wenbing et al. [13] implemented several physical model tests, from which they found the opening size and debris flow accumulation density counted as the main factors affecting the degree of blockage of the dam, and the reduction rate of peak debris flow was related to the degree of blockage of the dam, and the barrier dam could adaptively regulate the debris flow under different conditions. More importantly, Dong Yuan et al. [14] studied the dynamic processes of debris flow accumulation density, channel gradient, dam body and debris flow velocity through experiments, and the results showed that the dam could effectively trap and regulate debris flow, which is believed to provide technical reference for further practical engineering.

In this contribution, a continuous-discontinuous coupling method is used to analyze the changes of slope water pressure and displacement of the slope under different rainfall conditions for numerical simulation, which reveals the mechanism of landslide sliding, and the state of landslide sliding is simulated by converting the seepage force to discrete element nodes, and finally, the impact process of debris flow and dam is also simulated, revealing the whole process of landslide sliding under the influence of rainfall until impacting with the dam.

2.1Overview of the landslide Area

Nichanggou landslide, a first-class tributary of the left bank of Tian Wan river, is located in group 2 of He Ping village, Cao Ke township, Shimian county (Fig. 1a), which is a first-class tributary of the left bank of Tian Wan river, about 5km from Cao Ke township, Asbestos county in the north, about 59km from Asbestos Shimian county. The geographic coordinates of the landslide center are given as: 102°5′44.79″E, 29°23′33.99″N. There is a village road passing in the middle reaches of Nichanggou, and the traffic conditions are general.

The overall slope direction is 82°, the upper part of which is a steep slope with a slope of about 50°, the middle part is a landslide pile area with a topographic slope of about 45°, and the lower part is a landslide area with a steep topographic slope of 60–70° (Fig. 1b). In the middle of the slope, the thickness of the pile is 5-10m, and the composition of the pile is powdery clay with crushed stones. as shown in Fig. 2, This shows that the Ni changgou landslide with an elevation between 0 m to 850 m.

2.2Geomorphology

The stratigraphic lithology of the study area range is shown in Fig. 3, which is widely distributed with metamorphic rocks (Tzg) black cloud schist of the Upper Triassic Zagashan Formation. The lithology is grayish-white, made up of massive marble and occasionally interspersed with quartzite and quartz schist, while piling up a large amount of loose material, and if heavy rains occur frequently, landslides will appear along the channel to form mudslides.

2.3Landslide characteristic analysis

This landslide is not a newly-formed area, as a small-scale landslide has occurred before the rainfall, and heavy rain triggered a large-scale avalanche slide in the upper part of the landslide. It is the loose soil layer and barite on the upper part of the slope make up the cause of the landslide, along the contact surface between the surface layer of strong weathering zone and the intact bedrock landslide, the shear outlet is located in the middle of the slope. The landslide body is 1060m long, 180m wide on average, 1-8m thick, with a volume of 576,000 m³. According to the deformation characteristics of the landslide, it is divided into three sub-areas: slip source area, accumulation area and collapse area (Fig. 4). Figure 5 shows the photo of the siltyclay in the landslide site.

As a result of heavy rainfall, the landslide slid along the trench, and under the influence of heavy rainfall and surface water flow, the landslide flowed into the downstream trench (Fig. 6a.), which also had a small amount of accumulation (Fig. 6b.) and eventually caused serious damage to the surrounding farmland, vegetation, houses and infrastructure (Fig. 6c).

2.4 Rainfall Characteristics

The climate of the study area is mountainous with subtropical monsoon climate as its basal zone, where is alpine boreal climate in the plateau above 3500m above sea level, and subtropical monsoon climate below 3500m. The average annual precipitation for several years is 1200.9mm (displayed in Fig. 7), but it is unevenly distributed in space and time as the precipitation is mainly concentrated in May to September, accounting for 86.4% of the annual precipitation; meanwhile, there is more rainfall in the mountains than in the river valleys, and it mostly occurs as heavy rain or showers. Multi-year average rainfall statistics for the study area is shown in Table 1.

Table 1

Multi-year average rainfall statistics for the study area.
Month	Rainfall (mm)
January	1.3
February	5.0
March	14.2
April	51.3
May	85.4
June	227.0
July	277.6
August	281.5
September	205.6
October	38.7
November	12.3
December	1.0
All year	1200.9

3.1. Study on the mechanism of landslide initiation by rainfall

Rainfall infiltration is a continuous dynamic process, which performs a significant impact on the landslide initiation, while the intensity of rainfall affects the speed of landslide initiation. Rainfall triggered landslide initiation can be divided into two stages. In the first stage, the infiltration intensity of the soil is higher than the rainfall intensity, thus the rainfall will not form seepage, it belongs to the completely unpressurized state. In the second stage, the infiltration intensity of the soil is less than the rainfall intensity, the soil absorption rate is higher, rainwater will produce runoff on the surface, then the pressurized state occurs. The greater the rainfall intensity, the smaller the rainfall infiltration rate, the longer the duration of rainfall affecting induced landslides, and the greater the water content leading to soil destabilization.

Generally, the discrete element method is widely used in landslide initiation related research works, compared with traditional mechanics, discrete elements have the advantage of simulating large deformation, adopting discrete elements, considering the effect of rainfall, modeling with equivalent depth of infiltration [15], and applying flow-solid coupling [16]. In the process of discrete element simulation of landslides, the calculation of stress resulted from the excessive amounts of elements often causes to large deviations. Therefore, the method of continuous-discontinuous coupling calculation [17] can be used to simulate the landslide initiation, of which the soil is treated as a continuous medium while the damaged soil is regarded as discrete elements. The infiltration of rainwater is quite difficult for the discrete element simulation, however, it is much easier in the continuous medium simulation. If the two are combined, it can simulate both the landslide initiation caused by rainfall and the damage effect of continuous medium under rainfall in an effective way. The results of this study are believed to provide a valuable reference for the application of continuous-discontinuous numerical simulation in geotechnical engineering.

3.2. Rainfall model

In this contribution, we propose a continuous-discontinuous coupling model with the relative processes of different rainfall amounts, of which the saturated and unsaturated pairs of infiltration flows are considered in order to continue the simulation of the landslide initiation process under the influence of rainfall after the completion of rainfall infiltration, and to save the results of rainfall calculations into another storage node that opened within the cell.

The rainfall loads are transformed into equivalent forces applied to the continuous model and the discontinuous model[18]. The simulation of the model is divided into rainfall process and landslide initiation process. Since rainfall is a long period process, deformation is usually not considered in the analysis [19], the theory of saturated-unsaturated seepage is employed instead to record the pore pressure field and safety factor. The landslide initiation process allows the model to deform freely, during this process, the areas where damage may occur get discretized, it gives possibility to perform a continuous-discontinuous coupling computational analysis for rainfall landslides.

The rainfall process can be simulated according to saturated-unsaturated theory, of which the unsaturated permeability coefficient is generally used to reflect the change process with soil-water characteristic curve. According to the relationship between soil water content and permeability coefficient, the Van Genuchten (VG) model [20] between them can be represented as follows:

$${\theta }={\theta }_{r}+\frac{{\theta }_{s}-{\theta }_{r}}{[1+{\frac{p}{\text{a}})}^{{n}^{{\prime }}}{]}^{{m}^{{\prime }}}}$$

Where θ denotes the volumetric water content of the soil; ${\theta }_{r}$ refers to residual volumetric water content of the soil; ${\theta }_{s}$ is the saturated volumetric water content of the soil; $p$ represents the pore water pressure (Pa) of the soil; $\text{a}{,m}^{{\prime }},{n}^{{\prime }}$are the fitting parameters, and generally, $\text{a}$ takes the value of 100, ${m}^{{\prime }}$ takes the value of 1, and ${n}^{{\prime }}$takes the value of 2.

Considering the relationship between volumetric water content and saturation in the soil, ${\theta }=\text{n}\text{s}$, where $\text{n}$ is the porosity and $\text{s}$ refers to the saturation. The relationship between negative pore water pressure and saturation can be formulated by bringing in the above equation.

$$\text{s}={s}_{r}+\frac{1-{s}_{r}}{[1+{\frac{p}{\text{a}})}^{{n}^{{\prime }}}{]}^{{m}^{{\prime }}}}$$

Where $\text{s}$ means saturation; ${s}_{r}$ denotes residual saturation, and the rest of parameters mean the same as above. The relationship between pore water pressure permeability coefficient ${k}_{w}$ and soil saturation permeability coefficient ${k}_{s}$ an be specified as follows:

$${k}_{w}=\frac{{k}_{s}}{1+\text{a}[{\frac{({u}_{a}-{u}_{w}}{{p}_{w}\text{g}}\left)\right]}^{n}}$$

Where ${u}_{a}$ denotes the intra-pore gas pressure, ${u}_{w}$ denotes the pore water pressure, ${u}_{a}-{u}_{w}$ negative pore water pressure absolute value, ${p}_{w}$ denotes the fluid density, g denotes the acceleration of gravity, a, n is the fitting parameter, for general soil body a take the value of 0.1, n take the value of 2.

From the above formula, it can be seen that as the absolute value of pore pressure gradually decreases, that is, the water content gradually increases, the permeability coefficient gradually increases, thus it is considered that the permeability coefficient reaches its maximum value when the soil is saturated and remains unchanged.

3.3. Continuous-Discontinuous Equivalence Method

An equivalent simplification is needed with the saturated-unsaturated infiltration calculations, it aims to apply the rainfall effects to the particles of discontinuous mechanics. First of all, the increase of the gravitational acceleration in the soil, that results from rainfall infiltration, must be considered. The soil gravity ρ affected by rainfall can be calculated through the following equation.

$${{\rho }={{\rho }}_{d}+{\theta }·{\rho }}_{w}$$

The equation ${{\rho }}_{d}$ represents the dry weight of soil; ${\theta }$ refers to the volumetric water content; ${{\rho }}_{w}$ denotes the weight of water. In saturated soil, the direction of seepage force keeps consistent with the direction of seepage [15] as the presence of head pressure difference can produce seepage force on the soil, while the effect of seepage force in unsaturated soils is neglected. The equivalent infiltration force ${\text{F}}_{f}$ applied in the saturated region is:

$${\text{F}}_{f}=J·V={{\gamma }}_{w}·i·V$$

The variable $J$ is the infiltration force; $V$ denotes the volume of the unit grid or particle cell; ${{\gamma }}_{w}$ represents the heaviness of water (N/m³ ), $i$ refers to the hydraulic gradient, and its magnitude value is the ratio of the difference between the water potential at two points in the soil and its infiltration distance. The magnitude of the buoyancy force on the particle under the action of saturation can be calculated by the following equation:

$${\text{F}}_{f}=pgV$$

In the saturated state of the soil, there is a certain degree of discounting in its physical and mechanical parameters due to the softening effect. In the discontinuous model, the effects of rainfall loading on the increase in soil weight, seepage and buoyancy forces, and parameter discounting are considered, and the equations are (4)-(6). In the continuous model, since it is not the main area where debris flow initiation occurs, only the effects of soil weight increase and parameter discounting are considered to ensure the reasonableness of the calculation efficiency.

3.4. SPH-FEM theoretical model (blocking dynamic response model)

Landslides are initiated along gullies to form mudflows, which are usually composed of large amounts of sediment and rock and are excited to born under extreme climatic conditions such as heavy rainfall in valleys. Studies on modern debris flow have evolved from traditional single-discipline and single-method studies to an interdisciplinary combination and multi-method research. Several representative studies are specified as follows. A new coupled SPH-DEM method by Wang Zhichao [21] was proposed to model the coupled motion problem of fluids and solid particles, which considered the capillary and viscous interactions between the fluid and particles. Similarly, Xu Wenjie [22] used SPH-DEM coupling to simulate the 1963 Vajont reservoir landslide surge in Italy, and it turned out that the final accumulation pattern of the calculated landslide is more consistent with the actual accumulation topography. Also, a SPH-DEM coupling method by lin xiaohui et al [23] was employed to simulate the occurrence and propagation of discrete soil landslide surge waves, with which the landslide scale and surge velocity correlation curve were developed. The coupling methods consist of smooth particle hydrodynamic method (SPH), finite element method (FEM), of which the SPH represents the fluid, while the FEM unit refers to the topography and structure. For finite element and discrete element, the SPH-FEM coupling model is a better choice to simulate the contact process between debris flow and rigid structural units.

In this contribution, an experience based SPH-FEM model is used to simulate the stopping response of the debris flow to the dam. The fluid in the debris flow is calculated using SPH, and the debris flow trench as well as barrier structure are carried out using FEM, upon which a complex dynamic interaction model of particle-fluid-structure and a simplified trench dam model can be established to illustrate the process of debris flow impact with the dam.

3.5. Model Establishment

The calculation model is based on the Nichanggou landslide model, and the slope calculation model is divided by hexahedral mesh, with 9009 nodes and 43888 cells in total(Fig. 8a). In order to analyze the infiltration damage in the upper, middle, and lower parts of the landslide, 9 points are now implanted and the location of the monitoring points (Fig. 8b). Nichanggou landslide with the micro-mechanical parameters shown in Table 2.

Table 2

Landslide-related parameters.
Parameters	Landslide	Bedrock
Modulus of elasticity(Mpa)	3.0	3.8
Poisson's ratio	0.3	0.35
Density(kg/m3)	2105	2650
Cohesion(kpa)	10.3	10.5
Friction angle(°)	26.5	28
Penetration rate	1.81×10^-8	1.0×10^-8

Landslides can easily cause mudflows along trenches, and the impact of mudflows on control engineering structures is indeed an important consideration. At present, the impact effect of mudflow on the dam body belongs to a complex process due to the complex topography and large grid calculation. In the most research literature, it is simplified by selecting the SPH-FEM method and using rigid body cells to facilitate the simulation on the bottom of the trench as well as the rock bodies on both sides(Fig. 9). Considering the mudflow material source of this area is mainly rainfall, the mudflow fluid is simulated by SPH particles, with a total of 9219 particles. The gravitational acceleration used in the calculation is 9.8 m/s². For the debris flow channel, the simulation is performed with rigid shell material, which can significantly reduce the time of simulation, and the shell cells are used this turn, with a total of 5000 shell cells. The relevant parameters are shown in Table 2.

3.6. Calculated Solutions

To induce landslide initiation mechanism with rainfall, this contribution simulates rainfall landslides by establishing continuous-discontinuous coupling. After the calculation of initial seepage, the calculation results are used in the coupling model, and a changing process of seepage field is derived from different rainfall amounts, and continuous-discontinuous coupling landslide analysis under rainfall conditions is carried out on the targeted landslide. The initial pore pressure field is set at the boundary of both sides of the model(Fig. 10), and then the saturation-disaturation calculation is used to stabilize the generation. The pore pressure and safety coefficient changes are determined by the coefficient parameters that given by different rainfall amounts. Rainfall is classified according to the meteorological classification standard in China: 24h rainfall < 10mm is considered as light rain; 24h rainfall between 10 ~ 24.9mm is considered as medium rain; 24h rainfall between 25 ~ 49.9mm is considered as heavy rain; 24h rainfall between 50 ~ 99.9mm is considered as heavy rain; 24h rainfall between 100 ~ 249.9mm is considered as heavy rain; 24h rainfall ≥ 250mm is defined as very heavy rainfall, considering the different conditions of rainfall, the selection is heavy rainfall and very heavy rainfall is shown as Table 3.

Table 3

Simulation of parameters related to heavy rainfall.
Rainfall program	Rainfall time/d	24h rainfall/mm
Rainfall Program I	14	180
Rainfall Program II	14	260

4.1Effect of rainfall intensity on seepage field

Under the two rainfall intensity conditions of scenario I and scenario II, the pore water pressure changes within the distribution were recorded by setting different monitoring point maps on the slope. It can be clearly observed that as the rainfall intensity increases, the pore water pressure values at the preset monitoring points also increase. In the unsaturated soil, the soil gradually approaches the near-saturated state and even reaches the saturated state. Among them, the pore water pressure values at nodes a1, a2, and a3 at the top of the slope changed most dramatically ( Fig. 11a), followed by nodes b1, b2, and b3 in the middle of the slope (see Fig. 11b), while nodes c1, c2, and c3 at the foot of the slope (displayed in Fig. 11c), changed most slowly. In addition, the rate of change of pore water pressure values at the monitored nodes was also significantly faster under the condition of rainfall scenario 2 with greater rainfall intensity. The seepage field clouds calculated in the two rainfall scenarios show that under the rainfall duration of 14d ( Fig. 12a-b), the saturated zone performs a diffusion from the slope surface to the interior of the slope, and the area of the slope affected by rainfall gradually expands with the rainfall, and the positive pore water pressure first appears at the location of the slope surface, and the groundwater table surface at the foot of the slope has a obvious rise. And in the rainfall scenario II with greater rainfall intensity, the saturated area is significantly larger than that in scenario I.

4.2. Analysis of edge stability under rainfall conditions

The transient analysis method was adopted to analyze the change of slope stability coefficient with rainfall duration, of which the calculation results were saved every 1 day for 14 days for both Rainfall I and Rainfall II to figure out the stability coefficient. The change of slope stability coefficient with rainfall duration under rainfall scheme I and II is shown in Fig. The results show that: with the increase of rainfall duration, the stability of slope decreases continuously, and the impact of rainfall intensity on the safety coefficient is more significant, especially at the late stage of rainfall. What’s more, when the rainfall intensity increases from 180 mm/d to 260 mm/d, the slope safety coefficient under two different rainfall intensities decreases linearly with the rainfall duration from the early to the middle of the rainfall period. Under natural conditions, the slope safety coefficient is 1.184, which is in a stable state. as shown in Fig. 13, Under rainfall condition I, the slope safety coefficient decreases to 1.084, 1.029 and 1.012 on day 5, 10 and 14, respectively, while under rainfall condition II, the slope safety coefficient decreases to 1.078, 1.014 and 1.002 on day 5, 10 and 14, respectively. the change is 0.172 under rainfall condition I and 0.182 under rainfall condition II. The increase in rainfall intensity will lead to a decrease in the stability of the slope over time, and finally tends to 1 under rainfall condition II, which is likely to result in the destabilization of the whole slope.

From Fig. 14a, b, c, d, it can be seen that the maximum shear strain value increases continuously as the accumulation of rainfall time increases. A large amount of rainwater infiltrates into the slope surface. The plastic zone of the slope develops continuously, and a large number of them are distributed on the surface, foot of the slope and slip zone of the slope.

Comparing with Fig. 14a, b, c, d and e ,f, g ,h. with the increase of rainfall intensity, the maximum shear strain increases, and the distribution of shear strain increments is more dispersed than the sliding channel at the beginning and more coherent with the end of rainfall.

In summary, the maximum shear strain tends to increase continuously with the accumulation of rainfall and the increase of rainfall intensity. More specifically the higher the rainfall intensity, the faster the increase of the maximum shear strain value and the more concentrated the distribution of shear strain increments.

4.3. Analysis of slope displacement under rainfall conditions

4.3.1. Analysis of downslope displacement under rainfall conditions

On the first day of rainfall, the maximum horizontal displacement of the slope appeares on the central surface of the slope. With the extension of rainfall time, the maximum horizontal displacement of the slope gradually shifts to the foot of the slope. On the 5th day of rainfall, the maximum horizontal displacement appeares near basically follows the same pattern. After that, on the 10th and 14th day, the maximum horizontal displacement of the side slope gradually increases, but the distribution of horizontal displacement is more or less the same(Fig. 15a,b,c,d). On the whole, the horizontal displacement of the side slope mainly occurs in the upper part of the soil body, and increases gradually from the inside to the surface of that slope. At the beginning of the rainfall, the maximum horizontal displacement appears in the middle surface of the slope. Then, with the extension of rainfall time, the maximum horizontal displacement gradually increases. Finally, during the 14 days of rainfall, the horizontal displacement of the slope foot gradually increases, resulting in the gradual decrease of the slope stability.

4.3.2. Rainfall condition II downslope body displacement analysis

On the first day of rainfall, the maximum horizontal displacement of the slope appears at the central surface of the slope. The rainfall value becomes larger, and the maximum horizontal displacement of the slope is also larger compared with the rainfall conditions. After that, the maximum horizontal displacement of the slope increases gradually on the 5th, 10th and 14th days of rainfall, but the distribution of horizontal displacement basically follows the same pattern ((Fig. 16a,b,c,d). The maximum horizontal displacement is 0.16m on the 14th day.

4.4. Coupling analysis of rainfall landslides

With the purpose of doing simulation on the landslide sliding process, the rainfall results are saved to the internal cell through boundary wall that coupled to the continuous-discontinuous coupling model, and the rainfall loads are applied to the continuous and discontinuous models through equal effectiveness. It is that contact between the discrete particles and the cell grid is not close when the initial state gets generated leads to the uneven coupling, thus the contact forces are not uniform, and the maximum displacement is generated in the vicinity of the vertical coupling intersection, the displacement no longer develops until the model gradually equilibrates and transmits uniform contact forces, and the calculation results of the rainfall landslide model are shown in Fig. 17.

Under the influence of continuous rainfall and gravity, the landslide moves downward rapidly. The whole process lasts about 200 s.

It can be divided into 5 stages: (1) At the initial moment, the slope is damaged along the sliding surface and show a trend of downward displacement (Fig. 17a). (2) 0 ~ 25s, the head edge first separates from the slope surface and the sliding body moves rapidly to the lower ground level, and at the same time the landslide speed increases rapidly. The trailing edge part slides relatively slowly, with the maximum velocity exceeding 35 m/s, and the average velocity also increases rapidly to about 25 m/s. The remaining part of the sliding mass gathers and keeps moving downward (Fig. 17b). (3) During 25 ~ 50 s, the head edge portion separates from the slope and started to accumulate downstream,and its velocity also gradually decelerates to 7 ~ 10 m/s. The back edge of the slider forms a distinct steep slope. The remaining sliding blocks continue to move downstream until they reach the channel and gradually get accumulated (Fig. 17c). (4) During the propagation of the slide, the velocity gradually decreases due to the collision and friction between the particles. more and more sliding blocks get accumulated in the downstream channel (Fig. 17d,e). (5) After 200s, most of the sliding mass stops moving and get accumulated at the foot of the slope, and only a small amount of sliding mass continues to move downstream. Finally, the sliding mass is deposited along the channel (Fig. 17f). The average velocity and displacement time courses of the soil at different stages are shown in Fig. 18, and the velocity variation in each stage is consistent with Fig. 17. Throughout the process, the landslide body velocity peaks at the second stage at 4.8 m/s, and the average jump distance of the landslide was close to 500 m.

4.5 Impact loading studies

4.5.1.Motion process analysis

To analyze the interaction process between the debris flow and the barrier dam, Fig. 19 presents the detailed process of debris flow impacting the barrier dam. When t = 0.5s, the debris flow starts to move under the action of gravity, at this time the debris flow pattern has basically formed ( Fig. 19a). When t = 1.4s, the debris flow reaches the bottom of the barrier structure, at this moment the mudslide began to impact the blocking structure and began to climb higher (presented in Fig. 19b), the maximum velocity is 14.25m/s. At t = 1.8s, most of the debris flow has stopped, but the debris flow still has impact on the stopping structure ( Fig. 19c). When t = 2.2s, the debris flow partly overflows and partly flows back to siltation, and exerts load effect on the stopping structure ( Fig. 19d).

The process of climbing and desilting is shown in Fig. 20. At t = 1.4s, the debris flow reaches the bottom of the barrier structure and starts to impact the debris flow barrier structure and starts climbing, the climbing height is 0.7m (Fig. 20a); At t = 1.8s, the debris flow continues to climb with local desilting, at this time the SPH fluid particle splash and diffusion phenomenon can be observed, and the climbing height is 3.7m (see Fig. 20b);At t = 2.2s, the debris flow climbs to the highest with obvious desilting when it impacts the barrier structure, the climbing height is 3.7m; the debris flow climbs to the highest with obvious back-siltation and particle splash phenomenon when it impacts the barrier structure, and the climbing height reaches up to 5.7m (Fig. 20c).

4.5.2.Impact analysis

The displacement and time range of the middle of the blocking structure are monitored and shown in Fig. 21a. At about 2.2 s, the displacement of the middle of the blocking structure reaches a maximum value of 6.25 mm and rapidly decreases to 4.18 mm after reaching the sealed value, which is illustrated in Fig. 21a, and the subsequent impact force at different moments is shown in Fig. 21b. At t = 1.8s, the peak impact force is 44,500 KN, and then rapidly decreases to 28,900 KN (see Fig. 21b), at this moment the impact force distribution of the blocking structure is shown in Fig. 22.

The debris flow slides along the trench, and a certain amount of debris flow will overflow in the process of hitting the dam, but it will conduct a certain back siltation state due to the dam's blocking effect, the change of volume is shown in Fig. 22. When t = 1.8s, a certain amount of particles start to fly out, and at t = 2.8s, the overflow volume reaches the largest, with 46.73m³, and then slowly flows back to siltation. It makes some reference significance for the dam's design of debris flow interception.

This contribution, taking Nichanggou landslide as an example, makes further research on the sliding mechanism of the landslide using the results of numerical simulation, and discusses the influence of the landslide mudflow on the dam with the simplified terrain model, it is believed to provide significant reference for some practical designs.

In the research area, heavy rainfall occurred before the occurrence of Nichanggou landslide. When the rainfall infiltrates the soil, the water content of the soil increases rapidly, that is the cause of slope deformation, and it eventually results in the landslide under the action of gravity and rainfall. In this research, the major factors causing Nichanggou landslide are summarized as follows. (1) The average slope of the landslide area is 60°~70°, and the sliding direction is consistent with the dip of bedrock, which provides favorable conditions for landslide generation; (2) Under the condition of continuous rainfall, the soil body reaches saturation rapidly, making the safety coefficient of the slope decrease rapidly, and then triggering the breaking of the ring. (3) the inflow of rainwater into the landslide increases the mobility of the landslide body and accelerates the landslide movement.

The research area has also the condition that breeds the occurrence of debris flow, and the downstream area has residents, roads, and farmland. Therefore, the design of the check dam is also very important, which reproduces the formation process of "dragon's head, dragon's body and dragon's tail" of the debris flow, and provides significant reference for the check dam design.

In order to improve efficiency, the design of debris flow and dam is simplified, consequently, the reproduction process of the debris flow and dam may be inaccurate. To be specific, the simulation reproduce the slope instability caused by rainfall. The unstable slope slides along the broken surface, and the mixture of soil, rock and water slides down the trench to form part of the process of mudslide -mudslide impacting with the dam, but more efforts are needed to improve the performance of the model.

This paper takes the Nichanggou landslide mound as an example, of which a three-dimensional model considering rainfall is established to carry out a series of numerical studies, and the variation law of pore water pressure under the action of rainfall infiltration is analyzed using unsaturated seepage analysis method. Furthermore, research on the impact force problem, the dynamic response problem of debris flow, and its blocking structure has been conducted, and the following conclusions are drawn:

(1) Nichanggou landslide is located in a shallow surface stratum, also that landslide prone zone. Due to continuous heavy rainfall, precipitation infiltration causes slope damage and landslide. The slope slides downward along the steep cliff after the landslide, rushes into the river, and finally forms a mudslide, causing significant damage to the surrounding environment.

(2) Under the condition of long-term continuous rainfall, the maximum horizontal displacement of the slope grows continuously with the extension of rainfall time and the increase of rainfall intensity. At the beginning of rainfall, the maximum horizontal displacement of the slope appears in the middle of the slope. At the end of rainfall, the maximum horizontal displacement of the slope under rainfall condition II is 0.16 m.

(3) The saturated-unsaturated rainfall infiltration is analyzed with the continuous model, and the continuous-discontinuous coupling model is built through a boundary wall coupling method. Additionally, the seepage field is equivalently applied to the coupling model equivalently to realize the landslide coupling analysis under the influence of rainfall. The simulation results show that the whole process of Nichanggou landslide lasts about 200 s. It can be divided into five sub-stages: early accelerated deformation stage, slider trailing edge disintegration stage, jumping stage along the main slide surface, deceleration movement stage and final deposition stage. The average velocity of the landslide can reach 4.85 m/s and the average displacement is about 500 m. After the movement ceases, the landslide gets accumulated along the river basin hillside.

(4) In the simulation, the model involved with dam, debris flow, and channel is simplified through SPH-FEM coupling method, the interaction between particles, objects and structures is established, the dynamic response of debris flow impact force and barrier structure is analyzed, and the volume of debris flow overflowing from the dam is estimated, it is believed to provide significant reference for the ground disaster control energy.

Author Contributions: Conceptualization, Yu Li.; Methodology, Yu Li.; Software, Yu Li.; Writing-original draft, Yu Li.; Data curation, Xuezhi Yang and Xiao Hu.; Supervision, Yu Li and Xue zhi Yang.; Writing - review and editing, Yu Li.; Investigation,Liyan Wan and Erlong Ma. All authors have read and agreed to the published version of the manuscript.

Acknowledgments: We thank the anonymous reviewers for their careful review, contributions, and critiques,which led to the improvement of the article, and everyone who offered to help with this article.

Conflicts of Interest: The authors declare no conflict of interest.

Data availability statement: The datasets used and analysed during the current study available from the corresponding author on reasonable request.

Varnes, D.J. Slope movement types and processes. In Proceedings of the Landslides, Analysis and Control, Special Report 176:
Cruden, D.M.; Varnes, D.J. Landslide types and processes. In Proceedings of the Landslides, Investigation and Mitigation, Special.
Report 247; Turner, A.K., Schuster, R.L., Eds. Transportation Research Board: Washington, DC, USA, 1996; pp. 36–75.
Hou, Runing;Chen, Ningsheng; Hu, Guisheng; Characteristics, mechanisms, and post-disaster lessons of the delayed semi-diagenetic landslide in Hanyuan,Sichuan,China. Landslides. 2021, 19, 437-449.
Wu Yong; He Siming;Pei Xiangjun. Post-earthquake trench debris flow initiation conditions - Analysis of hydraulics mechanism of destabilization of loose accumulation in rain. Geotechnics. 2012, 10, 3043-3050.
Haruka Tsunetaka;Norifumi Hotta;Fumitoshi lmaizumi. Variation in rainfall patterns triggering debris flow in the initiation zoneof the Ichino-sawa torrent, Ohya landslide, Japan. Geomorphology.2021, 375, 107529.
Huilin Bai;Wenkai Feng;Xiaoyu Yi. Group‑occurring landslides and debris flows caused by the continuous heavy rainfall in June 2019 in Mibei Village, Longchuan County, Guangdong Province, China. Natural Hazards.2021, 108, 3181-3201.
Li, Yange; Tang. Numerical simulation of debris-flow behavior incorporating a dynamic method for estimating the entrainment. Engineering Geology. 2015.
Li Jing. Rheological properties and sliding mechanism of weak interlayer in high slope under the influence of rainfall and blasting. Ph.D.Thesis, The Wuhan University of Science and Technology, Wuhan, China, 2022.
Braun,Anika; Cuomo, Sabatino; Petrosino, Stefano. Numerical SPH analysis of debris flow run-out and related river damming scenarios for alocal case study in sw China. Landslides. 2018, 15, 535-550.
Dong Yaogang; Liu Xingrong; Fang Yuetao. Regulating performance of a debris flow check dam spillway culvert based on flume test. Science Technology and Engineering. 2022, 22, 1794-1802.
Li Bailong. A Study on lmpact Force and Dynamic Response of Debris Flow Retaining Structure Based on SPH-DEM-FEM Coupling. Ph.D.Thesis, The JiLin University, Changchun, China, 2022.
Wenbing Zhou,jinfeng Liu, Yong You. Experimental investigation of blocking and discharge regulation function of window-frame dam inviscous debris flow control. Natural Hazards, 2020,11,1505-1527.
Yuan Dong,Liu Jin feng,You Yong. Experimental study on the performance characteristics of viscous debris flows with a grid-type dam for debris flow hazards mitigation. Bulletin of Engineering Geology and the Environment. 2019, 78, 5763-5774.
Qu Shengjun; Zhao Jianjun; Ding Xiumei. Discrete element simulation of rainfall induced shelving bedding landslides. Hydrogeology & Engineering Geology. 2016, 43, 120-126.
Zhang Yan; Gao Shuzeng;Yan Chaoqun; Rainfall erosion on loess subgrade slopes. The Chinese Journal of Geological Hazard and Control. 2017, 28, 34-39.
Chang Xiaolin; Hu Chao; Ma gang. Continuous-Discontinuous deformable discrete elementmethod to simulate the whole failure process of rock mass and application. Chinese Journal of Rock Mechanics and Engineering. 2011, 30, 2004-2011.
Chao Shengtang;Xue Penggong;Zheng Taoshen; Soil micro-penetration resistance as anindex of its infiltration processes during Rainfall. Journal of Rock Mechanics and Geotechnical Engineering.2022, 14, 1580-1587.
Tan Xin;Chen Shanxiong; Yang Ming; Saturated-unsaturated seepage analysis of slope under rainfall. Rock and Soil Mechanics. 2003, 3, 381-384.
Van Genuchten, M T. A closed-form equation for predicting the hydraulic conductivity of unsaturated soils. Soil Science Society of America Journal. 1980, 44,892-898.
Wang Zhicao.TianJin University, TianJin,China. Research on droplet impact to discrete particle based on a SPH-DEM coupling method. 2015.
Xu Wenjie. Fluid-solid coupling method of landslide tsunamis and its application. Chinese Journal of Rock Mechanics and Engineering. 2020, 39, 1420-1433.
Xiao hui Lin.Gang Li;Fangchen Xu. Acoupled SPH-DEM approach for modeling of free-surface debris flows. AIP Advances. 2022, 12, 1-15.

No competing interests reported.

Download PDF

Journal Publication

published 18 Jan, 2024

Read the published version in Scientific Reports →

Editorial decision: Major revision
06 Aug, 2023
Reviews received at journal
24 Jul, 2023
Reviewers agreed at journal
18 Jul, 2023
Reviewers invited by journal
18 Jul, 2023
Editor assigned by journal
18 Jul, 2023
Editor invited by journal
13 Jul, 2023
Submission checks completed at journal
13 Jul, 2023
First submitted to journal
02 Jul, 2023

You are reading this latest preprint version

Mechanisms of rainfall-induced landslides and interception dynamic response: A Case Study of the Ni changgou Landslide in Shimian, China

Status:

Journal Publication

Version 1

Abstract

Figures

1. Introduction

2. Study Areas

2.1Overview of the landslide Area

2.2Geomorphology

2.3Landslide characteristic analysis

2.4 Rainfall Characteristics

3. Methodology

3.1. Study on the mechanism of landslide initiation by rainfall

3.2. Rainfall model

3.3. Continuous-Discontinuous Equivalence Method

3.4. SPH-FEM theoretical model (blocking dynamic response model)

3.5. Model Establishment

3.6. Calculated Solutions

4. Results

4.1Effect of rainfall intensity on seepage field

4.2. Analysis of edge stability under rainfall conditions

4.3. Analysis of slope displacement under rainfall conditions

4.3.1. Analysis of downslope displacement under rainfall conditions

4.3.2. Rainfall condition II downslope body displacement analysis

4.4. Coupling analysis of rainfall landslides

4.5 Impact loading studies

4.5.1.Motion process analysis

4.5.2.Impact analysis

4. Discussion

5. Conclusions

Declarations

References

Additional Declarations

Status:

Journal Publication

Version 1