PAPERS
Effect of Snow Cover on Slope Stability of Embankment during Rainfall and Snowmelt
2025 Volume 66 Issue 3 Pages 148-155
Details
2025 Volume 66 Issue 3 Pages 148-155
In this study, the effect of snow cover on slope stability was examined to more accurately evaluate the risk of snowmelt hazards. Firstly, laboratory tests were carried out to determine strength characteristics of snow. In addition to laboratory tests, sprinkling model experiments were conducted on a snow-covered embankment model to observe the moisture response and deformation of the embankment. In addition, slope stability analysis using a finite element method was carried out on the snow-covered embankment model. The result confirmed that snow cover could restrict the surface layer of an embankment and slightly improve the slope stability.
In snowy areas along railway lines in Hokkaido and on the Sea of Japan side of Honshu, there have been reports of cases of landslides triggered by the infiltration of melting snow, for example [1, 2]. In this paper “ground slope” is used to refer to embankment slopes, cut slopes and natural slopes; “snowmelt hazard” is defined as a landslide that occurs during the snow season due to the infiltration of melting snow or rain, and not due to external forces such as seismic activity. When a snowmelt hazards occurs, as shown in Fig. 1, there are many cases where snow cover remains on the slopes. In this study, slopes with residual snow cover are referred to as “snow-covered slopes.” Snow is composed of ice grains and the soil is composed of minerals, and each of these objects has its own weight, permeability, strength and deformation characteristics. It is therefore assumed that the presence or absence of snow has some effect on the stability of the ground slope in relation to issues such as permeability and stress deformation.
The objective of this research is to verify the effects of snow on slope stability where snowmelt and rainfall act on snow-covered slopes. In this research, we focus on snow-covered slopes in case of embankments. To collect information on the characteristics of the snow for the above verification, we conducted a laboratory test using a constant-pressure direct box shear test with snow specimens. In addition, after we conducted sprinkling experiments on embankment models of snow-covered slopes to observe the behavior of the models in rain, we evaluated the stability of snow-covered slopes based on slope stability analysis.
In this chapter, we report on the results of constant-pressure direct box shear test [3] (Fig. 2 and Fig. 3) which was conducted with snow and sand specimens under various conditions using a field direct box shear test machine to obtain the strength parameters. The conditions of each test case are summarized in Table 1. The purpose of changing the condition of specimens is to understand the effects of conditions such as “water content,” “snow particle size,” “undisturbed or reconstructed,” “shear rate,” and “snow/ground boundary” on the strength parameters.
| Case | Snow | Inagi sand | Test conditions | Results | Fig. 5 | |||||||||||||
| Initial wet density | Initial dry density | Initial weight moisture content | Saturated weight moisture content | Maximum load settlement | Sample composition | Snow type | Particle diameter | Hardness PR | Initial wet density | Initial dry density | Water content ratio | Specimen | Unsaturated or saturated | Shear rate | Friction angle | Cohesion | ||
| ρst | ρsd | θm | θmsat | δmax | ρt | ρd | w | vs | φd | cd | ||||||||
| (g/cm3) | (g/cm3) | (%) | (%) | (mm) | (mm) | (kPa) | (g/cm3) | (g/cm3) | (%) | (%/min) | (°) | (kPa) | ||||||
| S-1 | 0.536 | 0.471 | 12.1 | ― | 2.8 | Undisturbed | Compacted | 0.2～0.5 | 109.4 | ― | ― | ― | Snow | Unsaturated | 1.7 | 46.7 | 17.8 | |
| S-2 | 0.463 | 0.394 | 14.9 | ― | 8.3 | Undisturbed | Granulated | 1～2 | 26.2 | ― | ― | ― | Snow | Unsaturated | 1.7 | 38.5 | 7.5 | (a), (b) |
| S-3 | 0.563 | 0.394 | 30.0 | ― | 8.3 | Undisturbed | Granulated | 1～2 | ― | ― | ― | ― | Snow | Unsaturated | 1.7 | 37.1 | 7.6 | (b) |
| S-4 | 0.463 | 0.394 | 14.9 | 0.0 | 15.7 | Undisturbed | Granulated | 1～2 | ― | ― | ― | ― | Snow | Saturated | 1.7 | 50.7 | 4.5 | (b) |
| S-5 | 0.511 | 0.444 | 13.2 | ― | 7.8 | Reconst | Granulated | 2～5 | 13.9 | ― | ― | ― | Snow | Unsaturated | 11.1 | 37.4 | 5.8 | (c) |
| S-6 | 0.510 | 0.443 | 13.2 | ― | 12.5 | Reconst | Granulated | 2～5 | 17.5 | ― | ― | ― | Snow | Unsaturated | 1.7 | 25.2 | 5.5 | (c) |
| S-7 | 0.510 | 0.443 | 13.2 | 0.0 | 9.6 | Reconst | Granulated | 2～5 | ― | ― | ― | ― | Snow | Saturated | 11.1 | 47.5 | 6.8 | |
| S-8 | 0.510 | 0.443 | 13.2 | 0.0 | 12.0 | Reconst | Granulated | 2～5 | ― | ― | ― | ― | Snow | Saturated | 1.7 | 50.9 | 1.7 | |
| SG-1 | 0.463 | 0.394 | 14.9 | ― | 5.2 | Undisturbed | Granulated | 1～2 | ― | 1.47 | 1.41 | 3.98 | Snow & sand | Unsaturated | 1.7 | 41.4 | 1.1 | (a) |
| G-1 | ― | ― | ― | ― | ― | ― | ― | ― | ― | 1.51 | 1.41 | 5.30 | Sand | Unsaturated | 1.7 | 38.3 | 4.3 | (a) |
For the constant-pressure direct box shear test, a horizontal displacement meter and shear force meter were installed at the field direct box shear testing machine (Fig. 2). The test was conducted in a low-temperature chamber (room temperature was 2°C) with the aim of suppressing snow melting. The test machine has a rectangular shear box (cross-sectional area 100 mm × 100 mm, height 50 mm, upper box and lower box).
In preparation for the snow test specimens, outdoor snow accumulation was sampled at Shiozawa Snow Testing Station. Prior to sampling of snow specimens, the vertical excavation surface of the accumulated snow was observed. Small samples were taken at 10 cm interval depths from the snow surface to check the condition of the snow, including the wet density. In addition, the weight moisture content of the snow was measured with a dielectric moisture meter and the hardness PR [4] was measured using a push gauge at the same height at the vertical excavation surface of the snow. When collecting in-situ “undisturbed” snow specimens, the appropriate snow layer was selected based on the results of the above snow observations, and a shear box (cooled in advance in the snow) was pushed into the target snow layer to sample it. When using “reconstituted samples,” disturbed snow material was removed from the snow layer and placed in the shear box, and the snow specimens were produced by compaction.
In the test to determine the shear strength of the “snow-ground interface,” the ground was compacted in the lower shear box and then compacted snow specimens were placed in the upper shear box. Note that the name of ground material used in the test is Inagi sand.
The basic procedure for the constant-pressure direct box shear test is as follows. The snow specimen was placed in the testing machine and was saturated by filling the test box with cooling water for the saturation case. A loading plate was then placed on top of the specimen, and a weight was used to apply a vertical stress (three hypothetical stress conditions of snow accumulation with 2 m height: 0.6 kPa, 5.5 kPa, and 10.4 kPa) on the specimen at a constant pressure. For the shear test, the jack of the testing machine was manually operated at a constant speed to apply horizontal displacement to the shear box. The relationship between horizontal displacement D and shear stress τ was determined by shearing the specimen.
An example of the specimen after the test is shown in Fig. 3. Note that the shear stress was calculated by dividing the shear force T by the shear box cross-sectional area A. For the constant-pressure direct box shear test, the specimens were tested with the same snow condition for each of the three different normal stress patterns, and the relationship between normal stress and maximum shear stress was analyzed to determine the internal friction angle φ and cohesion c. The internal friction angle and cohesion for each test case are shown in Table 1.
Figure 4 shows the relationship between horizontal displacement and shear stress of the granular snow (case S-2) and Inagi sand (case G-1). The Inagi sand was a highly compacted test specimen, and the shear stress behavior shows a peak at a horizontal displacement of around 4 mm (around 4% horizontal strain). On the other hand, for the case of granular snow (case S-2), the shear stress increases gradually with horizontal displacement, even in the relatively large deformation region of approximately 10 mm horizontal displacement (10% horizontal strain). The cause of this is thought to be due to the sintering effect that occurs when the pressure between ice particles increases.
Figure 5 shows the relationship between the maximum shear stress and the normal stress. The legend in the figure corresponds to Table 1. To verify the effect of the water content of granular snow on the shear strength, three cases of S-2, S-3 and S-4 with different water contents are compared. In the case of S-4, a decrease in maximum shear stress is observed at low normal stress condition, but at other stress conditions, the maximum shear stress is almost the same in each case. From this, it can be concluded that the strength of granular snow, which is often observed during late season snowmelt, does not change significantly with the water content, except in saturated conditions and low stress conditions.
Next, we examined S-2, G-1, and SG-1 in order to compare the shear strength of unsaturated granular snow, Inagi sand, and the snow/ground interface. It is confirmed that snow (S-2) has a higher shear strength than sand (G-1). On the other hand, the shear strength of the snow/ground interface (SG-1) is the lowest. Therefore, in cases where there is no vegetation on the ground to increase the adhesion between the snow and the ground, the avalanche failure mode would be selected because the snow/ground interface has the lowest shear strength. On the other hand, in cases where the adhesion between the snow and the ground is high, the ground becomes the weakest, and it is assumed that the slope failure mode would be selected.
Next, we examine the shear rate dependence of snow cover strength by comparison of S-5 and S-6. Under the experimental conditions of the study, no clear velocity dependence is observed. In addition, the effect of snow type on shear strength is examined by comparing the internal friction angles and cohesion of S-1 and S-2 in Table 1, and it is confirmed that firmer snow had higher strength than granular snow in this study. This result confirms that changes in snow type are one of the main factors in changes in snow cover strength.
In terms of comparing undisturbed samples with reconstructed samples, the comparison of S-2 and S-6 shows that the reconstructed sample (S-6) has a slightly lower maximum shear stress. One reason for this is that the in-situ undisturbed sample has undergone consolidation between ice particles due to freezing. This is also confirmed from hardness PR (Table 1).
In this chapter, we have reported the results of the basic strength constants of the snowpack obtained from the constant-pressure direct box shear test. From these results, we have confirmed that even if the snow cover is granular snow, the maximum shear stress of snow is higher than that of Inagi sand.
In this chapter, we report on the results of a sprinkling model experiment that simulated the conditions where a snow-covered slope on an embankment is subjected to rainfall, in order to understand the mechanism of snow-covered slope instability. We focus on the deformation of the snow-covered slope. In addition to this, we also report on the measured results of water behavior in the embankment during sprinkling.
4.2 Experimental MethodThe geometric shape of the snow-covered slope model (experimental case with snow depth Hs=0.25 m) and the sensor installation location are shown in Fig. 6. The conditions for the experimental cases (snow, ground, and sprinkling conditions) are shown in Table 2. The main difference between each experimental case is the depth of snow cover on the embankment (Hs=0 m, 0.125 m, 0.25 m). In the sprinkling model experiment, the embankment was constructed twice in total. Case 1-1 was mainly conducted to observe the behavior of soil moisture during the sprinkling of the embankment without snow, and the sprinkling was stopped when deformation occurred at the toe of the slope. The damaged area of the embankment was repaired and reused to examine case 1-2. This case was used to examine the relationship between the occurrence of damage and the behavior of soil moisture in the embankment with snow cover. Case 2-1 was conducted with the aim of comparing it with case 1-1 (Fig.7). It was used to observe the behavior of soil moisture in the embankment with snow cover during water sprinkling. The embankment was then reused for case 2-2, which was to examine deformation and soil moisture behavior of the embankment without snow cover. In case 2-1, slope sprinkling was stopped before the embankment collapsed. In Case 2-3 however, sprinkling was continued to induce a clear collapse. However, due to the influence of the snow cover, no clear collapse occurred. Therefore, the snow cover was thinned, and the sprinkling experiment was repeated in Case 2-4.
| Test case | Focused Phenomena | Snow conditions | Ground conditions | Sprinkling | ||||||||||
| Snow depth | Initial wet density | Initial dry density | Initial moisture content | Saturated permeability | Model preparation | Initial wet density | Initial dry density | Water content ratio | Initial moisture content | Initial saturation | Saturated permeability | Rainfall intensity | ||
| Hs | ρst | ρsd | θ | ksat | ρt | ρd | w | θ | Sr | ksat | r | |||
| (m) | (g/cm3) | (g/cm3) | (%) | (m/s) | (g/cm3) | (g/cm3) | (%) | (%) | (%) | (m/s) | (mm/h) | |||
| Case1-1 | Moisture | 0 | ― | ― | ― | ― | New | 1.62 | 1.41 | 14.9 | 21.0 | 43.62 | 5.7×10-5 | 27 |
| Case1-2 | Deformation | 0.25 | 0.51 | 0.445 | 6.5 | 1.99×10-3 | Reuse | ― | ― | ― | ― | ― | 5.7×10-5 | 27 |
| Case2-1 | Moisture | 0.25 | 0.42 | 0.371 | 4.9 | 1.99×10-3 | New | 1.66 | 1.43 | 16.4 | 23.5 | 49.45 | 3.9×10-5 | 27 |
| Case2-2 | Deformation | 0 | ― | ― | ― | ― | Reuse | ― | ― | ― | ― | ― | 3.9×10-5 | 27 |
| Case2-3 | Deformation | 0.25 | 0.51 | 0.441 | 6.9 | 1.99×10-3 | Reuse | ― | ― | ― | ― | ― | 3.9×10-5 | 27 |
| Case2-4 | Deformation | 0.125 | 0.51 | 0.432 | 7.8 | 1.99×10-3 | Reuse | ― | ― | ― | ― | ― | 3.9×10-5 | 27 |
The method for constructing the embankment is explained below. The physical properties of the embankment ground material, “Inagi sand,” are shown in Table 3. This ground material was placed in a rigid soil tank after being adjusted with water, and the ground was built by evenly compacting it to a compaction ratio of 84%.
| Soil particle density | Average particle diameter | Fine particle content | Maximum dry density | Optimum water content |
| ρs(g/cm3) | D50(mm) | Fc(%) | ρdmax(g/cm3) | wopt(%) |
| 2.72 | 0.17 | 14.90 | 1.68 | 14.6 |
The ground was then shaped into an embankment with a height of 1 m and a slope gradient of 1:1.2. The slope gradient is steeper than the current design values for railway embankments (1:1.5, performance rank III, height less than 9 m) [5]. The strength characteristics of the Inagi sand embankment were obtained from a triaxial compression test (CD test), with an internal friction angle of φd = 35.0° and cohesion of cd = 6.3 kPa. For the snow build-up method, snow material (granular snow: grain diameter 2-5 mm) collected outdoors at Shiozawa Snow Testing Station was spread over the embankment surface by free fall (average fall height approx. 0.5 m) through a 1 cm mesh sieve and shaped to the planned thickness. In the sprinkling model experiment, the snow-covered slope was continuously sprayed with water at a rate of 27 mm/h indoors (air temperature was 4-10°C), based on the assumption that it would rain during the snow-melting season. In order to reduce the melting of the snow due to the warmth of the water, the water was supplied through a simple rapid cooler to lower the water temperature to 4°C. The main items measured during the sprinkling model experiment were the pore water pressure, volumetric water content and water head measured by manometers at the bottom of the rigid box. Note that the pore water pressure was obtained as additional information, so the measurement results are not shown in this report. In addition, slope deformation was observed from photographs of the side of the rigid box taken from fixed point.
4.3 Experimental resultsFocusing on the infiltration phenomenon on a snow-covered slope, we report on the water behavior in case 2-1 with snow cover (snow depth = 0.25 m). Specifically, Fig. 8 shows the spatial distribution (contour map) of the volumetric water content and the water head distribution at four stages during the 342 minutes from the start of the sprinkling in case 2-1. To compare the results shown in Fig. 8, the results for case 1-1 with no snow cover from the start of sprinkling to 342 minutes are shown in Fig. 9. Comparing Fig. 8 (d) and Fig. 9, it can be seen that the presence or absence of snow cover had no significant effect on the water head of the manometer under the conditions of this experiment. On the other hand, it should be noted that there is a possibility that measurement errors occurred due to air bubbles in the manometer M-3 in Fig. 9, so this should be considered when interpreting results. As mentioned above, the effect of snow cover on the infiltration phenomenon in the embankment was limited under the conditions of this experiment. The reason for this is that the snow used in this experiment was coarse granular snow (grain diameter 2-5 mm) with high permeability, and the snow cover thickness was 0.25 m, which is not thick. On the other hand, in nature there are various types of snow with finer grain size and lower permeability, such as “compacted snow” (grain size of about 0.5 mm). Therefore, rainwater and snowmelt water can be expected to take a long time to infiltrate through the snow cover and reach the ground surface where such snow types are interspersed with layers of snow. In order to verify the overall impact of the combined effects of snow depth, snow type and snowmelt on the moisture behavior of the soil in the embankment, more detailed validation using seepage flow analysis modelling a snow-covered slope is required.
Next, we consider the deformation issue of a snow-covered slope. Figure 10 shows the relationship between the horizontal displacement of the embankment at a height of 10 cm from the bottom (Fig. 7) and the manometer’s water head. The horizontal displacement was obtained by analyzing the side photographs of the model experiment. The water head of M-4 and M-5 manometer is adopted as representative value at the toe of the embankment. From these results, it can be concluded that the horizontal displacement of the embankment caused by the increase in the water head of the manometer was suppressed by the effect of the snow cover in the case with snow cover compared to the case without snow cover.
In order to understand the differences in the deformation of snow-covered slopes with different snow depths, the changes of the slope of the embankment after water sprinkling in case 1-2 (snow depth = 0.25 m) and case 2-4 (snow depth = 0.125 m) are shown as examples in Fig. 11. In case 1-2 where the snow cover is relatively thick, the snow accumulation has a restraining effect on the surface which acts as a slope protection. As a result, although the snow cover was subjected to earth pressure due to shallow sliding at the toe of the embankment and volume changes occur at the rear of the snow cover, the displacement on the snow cover surface was small. On the other hand, in Case 2-4 where the snow cover was relatively thin, the entire snow cover slid, dragged by the sliding of the embankment surface. Figure 12 shows photographs of the deformation of the embankment observed after the water sprinkling. The effects of the snow cover on the deformation of the embankment were then considered using Fig. 10 and Fig. 11. In both case 1-1 and case 2-2, where there is no snow cover, horizontal cracks were observed at a slope length of 45 cm from the toe of the slope. These cracks were observed to be progressing at the time of their occurrence, and it was assessed that if the water sprinkling had not been stopped, the embankment would have slipped and collapsed. On the other hand, in cases 1-2 and 2-3, where the slope of the embankment was constrained by the thick snow cover, no clear progressive deformation was visually observed during the experiment, even under conditions with higher water heads of manometer than in cases without snow cover. However, in case 1-2, when the snow cover was removed after water sprinkling, a horizontal crack was observed as a deformation at a slope length of 35 cm from the slope shoulder. We consider the possibility that the shallow deformation on the embankment did not progress because of the restraining effect of the snow cover, and that the granular snow has a higher cohesion c than the Inagi sand, which increases the apparent cohesion across the snow-covered slope, in case 1-2. Therefore, a deeper arc-slip mode would be selected in the embankment.
In case 2-4, where the snow cover was thin, the toe of the slope showed progressive deformation, the snow cover became extremely thin around areas of melting snow (the middle of the slope) and the eroded soil flowed down from the area, but there was no clear slip failure as shown in Fig. 1. These results indicate that snow cover has the effect of restraining the movement of small slides and muddy sediments on the slope surface of the embankment.
In this chapter, we report on the results of a verification of the stability of snow-covered slopes. Note that the result was obtained by performing stress deformation analysis and slope stability analysis using the finite element method analysis software “PLAXIS 2D.”
The geometric shape of the analysis model, the material conditions and the boundary conditions were based on the model experiment in chapter 4, and the other conditions are shown in Fig. 13. In addition, Table 4 shows the analysis parameters for the snow-covered slope obtained from the laboratory tests, and Table 5 shows the conditions for the slope stability analysis. In the slope stability analysis, the conditions are basically the same as in the water sprinkling experiment in Chapter 4. When the cohesion c of the embankment material Inagi sand (the internal friction angle φ and other parameters are fixed) and the groundwater level in the embankment (the water head hw at the M-4 location is used as the benchmark for the groundwater level in the embankment) are changed, the slope deformation mode and the slope safety factor Fs above conditions are confirmed.
| Material type | Inagi sand | Compacted snow | Granulated snow | Snow/Sand boundary | ||
| Material model | Mohr-Coulomb | |||||
| Drainage model | Drainage (without excess pore pressure) | |||||
| Wet unit weight | γunsat | kN/m3 | 16.2 | 4.2 | 4.2 | 4.2 |
| Saturated unit weight | γsat | kN/m3 | 18.9 | 9.7 | 9.7 | 9.7 |
| Elastic coefficient | E | kN/m2 | 6166 | 400 | 400 | 400 |
| Poisson's ratio | ν | 0.35 | 0.1 | 0.1 | 0.1 | |
| Shear rigidity | G | kN/m2 | 2284 | 181.8 | 181.8 | 181.8 |
| Dilatancy angle | ψ | ° | 5.0 | 0 | 0 | 0 |
| Cohesion | c | kN/m2 | 0～6 | 17.8 | 9.1 | 1.6 |
| Tensile strength | ct | kN/m2 | 0 | 0 | 0 | 0 |
| Friction angle | φ | ° | 35.0 | 46.7 | 46.3 | 47.0 |
| Water retention model | van Genuheten | |||||
| Residual saturation | Sres | 0.062 | 0.068 | 0.049 | 0.049 | |
| Submerged saturation | Ssat | 1.0 | 0.9 | 0.9 | 0.9 | |
| Model coefficient | gn | 1.38 | 3.17 | 3.99 | 3.99 | |
| Model coefficient | ga | 1/m | 3.83 | 10.8 | 37.6 | 37.6 |
| Model coefficient | gc | -0.420 | -0.684 | -0.749 | -0.749 | |
| Model coefficient | gl | 1.25 | 0 | 0 | 0 | |
| Saturated permeability | ksat | m/s | 3.90×10-5 | 1.00×10-3 | 1.99×10-3 | 1.99×10-3 |
The specific analysis procedure is described below. Since “PLAXIS 2D” can combine stress-deformation analysis and saturated-unsaturated seepage flow analysis, a steady seepage flow analysis was first performed for the entire embankment under rainfall. The rainfall conditions were varied by trial-and-error to reproduce the hydrostatic pressure distribution in the embankment that reproduces the water head hw (4 profiles) of M-4 in Table 5.
The information on the four hydrostatic pressure distribution profiles obtained in this way were used in the stress deformation analysis, and the stresses of the embankment under the conditions of the above steady pore water pressure distribution were calculated. The slope safety factor Fs is evaluated using the shear stress reduction method based on the above stresses and hydrostatic pressure distribution. Note that if the calculation does not converge and reaches a final state in the stress deformation analysis stage, the slope stability analysis using the shear stress reduction method will not be performed. In addition, we use the snow strength parameter of granular snow in these cases. Although the cohesion of Inagi sand in the triaxial compression test (CD test) was c=6.3 kPa, the slope safety factor did not fall below 1.0 under the hydrostatic pressure conditions under which cracks occurred at the toe of the embankment in the model experiments. On the other hand, the safety factor Fs falls below 1.0 when cohesion = 0.2 kPa. One of the reasons for this discrepancy is that the confining pressure conditions for the triaxial compression test of Inagi sand were carried out at 5 kPa to 20 kPa. On the other hand, the confining pressure at the surface of the small embankment model was lower, so it is assumed that the cohesion was actually lower on the embankment surface. Assuming that the actual cohesion of Inagi sand is 0.2 kPa, Fig. 14 shows the collapse modes of the ultimate state under the conditions of stability analysis A (without snow cover, water head at M-4 = 0.27 m) and stability analysis B (with snow cover, water head at M-4 = 0.27 m) based on an incremental displacement. The former shows the state that was reached the ultimate state as a result of the preliminary stress deformation analysis, and the latter shows the state that reached the ultimate state as a result of the shear strength reduction method. The deformation mode of stability analysis A (without snow cover) is similar to that observed in the sprinkling model experiment case 1-1 in chapter 5 (Fig. 12), with relatively large displacements occurring at the toe of the embankment. On the other hand, in stability analysis B (with snow cover), relatively large displacements occur in the area around the embankment shoulder. This deformation mode is consistent with the deformation tendency observed after the sprinkling model experiment case 1-2, when the snow was removed (Fig. 12).
Here, the results of the parameter study of the slope safety factor based on the conditions in Table 5 are shown in Fig. 15. These results confirm that the slope safety factor Fs in the stability analysis B (with snow cover) is higher than in the stability analysis A under the conditions where the water level in the embankment was formed by water sprinkling (water heads hw at M-4 =0.18 m, 0.24 m, and 0.27 m). This is because the snow cover restrains the shallow collapse of the toe of the embankment, as can be seen from the difference in deformation modes in Fig. 14. From these results, it is thought that the slope safety factor of snow-covered slope is higher than the slope without snow cover. On the other hand, as a peculiar phenomenon, the slope safety factor of the snow-covered slope was found to decrease under the condition where the groundwater level in the embankment was not formed (water head at M-4 = 0 m). In order to confirm the cause, Fig. 16 shows the distribution of incremental strain as the ultimate state after analysis using the shear strength reduction method under the condition that the water level in the embankment was not formed (water head at M-4 = 0 m). Here, in stability analysis B (with snow cover), a region of major shear strain increments Δγ is observed along the boundary between the snow cover and the embankment. In other words, under the condition where the groundwater level is not formed in the embankment in stability analysis B (with snow cover), it is assumed that the slip at the boundary between the snow cover and the ground was selected as the weakest failure mode. This can be called the all-layer avalanche mode.
In this study, we examined slope stability of snow-covered slope, and the results showed that snow generally has a positive effect on the stability of embankment slope, excluding the effects of snowmelt. However, depending on the slope gradient and ground surface conditions on natural slopes, for example, under conditions where snow accumulates thickly on the slope shoulders and snow loads are applied, it is assumed that snow accumulation may have a negative effect on the slope stability of snow-covered slopes. In the future, we would like to verify the stability of snow-covered slopes under various conditions.
In this study, we examined the effects of snow on slope stability from the perspective of evaluating the risk of snowmelt hazards caused by melting snow and rain in snowy regions. Specifically, the strength characteristics of snow were determined by laboratory test, and then a sprinkling model experiment was conducted on a 1m-high embankment model with snow cover, and the moisture response and deformation of the embankment were observed. Furthermore, slope stability analysis (shear stress reduction method) using the finite element method was carried out on this embankment model with snow. As a result, the following findings were obtained.
It was confirmed that the shear strength of the “granular snow” used in the experiment was higher than that of “Inagi sand” (sand ground) with a compaction degree of 84%. Note that while a peak in shear stress for the Inagi sand appeared at a horizontal strain of around 4%, that for the granular snow often appeared at a horizontal strain of 10% or more and in some cases no clear peak appeared. The reason for this is assumed to be caused by the sintering process that occurs when the pressure between ice particles increases as snow accumulates.
The result of a water sprinkling experiment on a model embankment with snow cover confirmed that in the case of a snow cover built on an embankment, the deformation around the toe of the embankment caused by the water sprinkling was suppressed compared to the case without snow cover.
The slope safety factor of the snow-covered slope (embankment) model was obtained using slope stability analysis (shear strength reduction method), which uses the strength, deformation and physical properties of the granular snow obtained in this study. As a result, it was found that although the melting water increases the water content of the embankment, the snow itself may slightly improve the slope safety factor of the embankment in the case of a snow-covered slope (embankment) affected by rain and melting snow.
The evaluation reported in this paper was carried out using a small embankment model, and we subsequently plan to apply the results of this research to evaluate the stability of various snow-covered slopes in their natural state.
|
Tsuyoshi TAKAYANAGI Senior Researcher, Geo-hazard & Risk Mitigation Laboratory, Disaster Prevention Technology Division Research Areas: Geotechnology, Slope Stability Assessment, Snowmelt Hazard. |
|
Shoma FUJIWARA Researcher, Geo-hazard & Risk Mitigation Laboratory, Disaster Prevention Technology Division Research Areas: Geotechnology, Slope Stability Assessment. |
|
Ryota SATO Assistant Senior Researcher, Meteorological Disaster Prevention Laboratory, Disaster Prevention Technology Division Research Areas: Glaciology |
