Keywords

1 Background

Causing billions of dollars of property damage and hundreds of human casualties each year, landslides are one of the worst natural disasters. Although there are several factors that can trigger landslides, the most common throughout the world is rainfall. Guzzetti et al. [1] identified rainfall intensity-duration thresholds required to trigger landslides using statistical data analyses in different regions of the world. The thresholds will depend on a number of parameters including the soil type, slope gradient, and the relative compaction of slopes. The rate of infiltration of the rainwater into the ground, or the seepage velocity, will be one of the most important factors controlling the stability of the partially saturated slopes. Larger seepage velocities will result in a quicker movements of the wetting front, which will reduce suction and increase the unit weight of the slope material. Both of these changes will increase the likelihood of the slope instability. Factors including the density of the slope material, the inclination of the slope as well as the intensity and duration of the rainfall will impact the seepage velocity [2,3,4,5,6,7]. As part of this study, small scale experimental models were prepared in a Plexiglas container to examine the influence of the density of the slope material and duration of rainfall on the seepage velocity through the slope material and settlement experienced as a result of the rainfall.

2 Materials and Methods

2.1 Slope Material

For the preparation of the model slopes tested in this study, soil collected from a new construction project at the Titan Student Union at California State University, Fullerton was used. The properties of the slope material are presented in Table 1. The material was air-dried and sieved through the U.S. #4 sieve (sieve opening of 4.75 mm) prior to its use in the construction of any model.

Table 1. Geotechnical properties of slope material.
Full size table

2.2 Preparation of Model Slopes

The experimental slopes were prepared in a Plexiglas container. The preparation of the slope model began with the installation of a 7 cm thick drainage layer. The drainage layer consisted of pea-sized gravel. It was overlain by geotextile that was 5 mm thick and acting as a filter between the slope material and the drainage layer. The slope was constructed on top of this geotextile in 5 cm thick lifts. The total dry weight of material required to obtain the desired relative compaction was determined and mixed with an initial moisture content of 7.2% (3% dry of optimum). This material was, then, compacted to the desired lift thickness. All of the slopes were constructed to an inclination of 40°. Nine different model slopes were prepared at different relative compaction levels. The relative compactions used in each of these models are summarized in Table 2. The slope models were constructed in a larger Plexiglas container with interior dimensions of 1.2 m by 2.4 m and a height of 1.8 m. Some pictures from the model preparation process in the Plexiglas container are presented in Fig. 1.

Table 2. Relative compaction in the models tested.
Full size table
Fig. 1.
figure 1

Model preparation process: (a) Plexiglas container with gravel drainage layer and geotextile, (b) compaction of soil lifts, and (c) model slope compacted to half height.

Full size image

After the slope was constructed to the desired height and required slope inclination, the slope was instrumented with eight miniature tensiometers and eight copper wires. The tensiometers had capacities to measure pore pressures from −180 kPa to 100 kPa. The copper wires were flexible and acted as inclinometers to determine the deformation in the slope. Figure 2 depicts the location of the instrumentation used in this study. The holes used to install the tensiometers were covered with a bentonite slurry in order to ensure that rainwater would not infiltrate rapidly through these holes. Three moisture sensors were also installed. The location of these moisture sensors are also included in Fig. 2. After the instrumentation was installed, the slopes were subjected to rainfall at an intensity of 30 mm/h using a rain simulator system. The location of the wetting front was recorded in 15 min intervals until the slope was completely saturated. Figure 3 contains some pictures of the instrumentation and the rain simulator system as well as the progression of the wetting front with time.

Fig. 2.
figure 2

Location of the instrumentation in the model slopes. Tensiometers are denoted with the letter T, while the copper wires are denoted with CW and soil moisture sensors are denoted with MS. All of the dimensions shown in centimeters.

Full size image
Fig. 3.
figure 3

(a) Instrumentation of the slope with tensiometers and copper wires, (b) completed slope model with installed rainfall simulator system, and (c) progression of wetting front with time during the application of rainfall.

Full size image

3 Results and Discussion

3.1 Wetting Front Locations

Presented in Fig. 4 is an example of the location of the wetting front with time recorded during the application of rainfall in Model 7. The movement of the wetting fronts were recorded on both sides as well as the back of the Plexiglas container. The seepage velocities were calculated at the tensiometer locations using the wetting front data as well as with the tensiometer data for comparison. The recordings in the tensiometers installed in Model 7 are shown in Fig. 5. These recordings give an indication of the movement of the wetting fronts inside of the soil mass and will approach a value of zero suction when the wetting front reaches the tensiometers.

Fig. 4.
figure 4

Movement of wetting front with time in model 6.

Full size image
Fig. 5.
figure 5

Tensiometer readings with time in model 6.

Full size image

Figure 6 contains a comparison of the time required for the tensiometers to obtain suction values of 0 kPa and the time required for the wetting front to reach the location of the tensiometers. Also included in Fig. 6 is a line representing that both times were equal. As it can be seen, the differences between the two measurements are similar in some of the cases, while in other cases time required to cease suction was slightly more than the time required for the wetting front to reach the tensiometer location. Such differences were expected as the location of the tensiometer and the, location where the weting front advancement was measured were slightly different.

Fig. 6.
figure 6

Comparison of the time required for the wetting front to reach the tensiometer location and the time required for the tensiometer reading to reach 0 kPa.

Full size image

The variation in the seepage velocity with relative compaction at the crest of the slopes is shown in Fig. 7. It can be seen from Fig. 7 that an increase in the relative compaction corresponded to a decrease in the seepage velocity when the relative compaction was greater than 63%. The seepage velocity in the model with a relative compaction of 60% appears to be lower than expected. However, this lower seepage velocity is attributed to the fact that the model slope experienced nearly 8% settlement in approximately the first hour after the application of rainfall. As a result, the slope material will densify impairing the infiltration of rainwater in the slope.

Fig. 7.
figure 7

Variation in the seepage velocity with relative compaction at the crest of the slopes.

Full size image

Figure 8 depicts the settlement and swelling experienced by the slopes as a result of the application of rainfall. In Fig. 8, the negative strain values represent settlement and positive strain values represent swelling of the slope. It can be seen that slopes with relative compactions initially less than 75% would experience settlement, while those with relative compactions initially greater than 75% would experience swelling with the application of rainfall.

Fig. 8.
figure 8

Settlement (negative strain values) and swelling (positive strain values) experienced by the model slopes when subjected to rainfall until complete saturation was achieved.

Full size image

4 Conclusions

Ten model slopes were prepared in Plexiglas container using soil obtained from a construction project at the Titan Student Union at California State University, Fullerton. The 40° inclined slopes were prepared to have different relative compactions ranging from 60% to 80% and were subjected to rainfall at an intensity of 30 mm/h. The location of the wetting front with duration of rainfall was recorded and used to determine the seepage velocity throughout the slope. The seepage velocity was found to decrease with an increase in the relative compaction. The model slopes were seen to experience settlement when the initial relative compaction was less than 75%, but would experience swelling when the initial relative compaction was greater than 75%.