The results of the analysis are divided into several discussions, namely the effect of the elevations of the existing groundwater level on slope stability, horizontal sub-drain on slope stability with the different water levels, vertical crack and weak layer on slope stability, and the effectiveness of horizontal drain with variations in soil conditions and groundwater level. This was carried out to determine the mechanism of the landslide at the study location, the effect of cracks and weak layers. It also makes horizontal sub-drain an effort to mitigate landslides by managing rainwater and understanding its effectiveness to increase the factor of safety.
In ES conditions, the elevation of the existing groundwater greatly affects the value of the safety factor on the slope. This indicates the lower the elevation of the existing groundwater level, the greater the value of the slope factor of safety. The daily factor of safety with real-time rain conditions has decreased (Fig. 12a) due to the absence of significant changes in the water level which can affect the factor of safety. Soil conditions that have a hydraulic conductivity value are much smaller than the intensity of the rain, causing most of the rainwater to become surface runoff, and not many seeps into the soil, which does not affect slope stability. This is in line with (Galeandro et al., 2013) which stated that when the rainfall intensity exceeds the infiltration capacity of the soil matrix, a certain amount of water is not able to infiltrate into the matrix. In this condition, the horizontal sub-drain (HD) does not have much effect to increase slope stability. At groundwater level conditions 1, 2, and 3, and in the installation of horizontal drains at sub-drain locations 1, 2, and 3, the factor of safety is the same as the slope without a horizontal drain (Figs. 12b and c). These results indicate that when the hydraulic conductivity is very small compared to the intensity, 30 days of rain does not significantly affect the stability of the slope. The selection of landslide handling with a horizontal drain in this condition is also very ineffective. in this condition, the installed horizontal sub-drain must be longer until it reaches the ground water level position.
Most of ES conditions obtained a factor of safety greater than 1.1 during the 30 days of rain in November, which had the highest intensity and is safe because it is above the critical slide limit (SF = 1). In this analysis, the soil parameters used are based on conventional tests, namely N-SPT and laboratory that do not capture the weathering conditions or the presence of other weaker layers below the soil surface. The conventional test was only carried out at several 3 points in a very large landslide area to avoid its representation of real conditions in the field. Therefore, the results of the analysis of slope stability gave a factor of safety that is safe against landslides when landslides occur. More in-depth observations were made on the geophysical data. The resistivity test also showed several zones which are crack soil and weak layers. The results of geophysical observations are interpreted under VC and VCWL conditions.
This is similar to the results of the analysis of ES conditions, where the elevation of the groundwater level in the VC condition also affected the factor of safety. This indicated that the higher the groundwater level, the lower the factor of safety, however, different results were obtained when a horizontal drain was used. Under the ES conditions, the horizontal drain did not effectively elevate slope stability but increased the factor of safety under VC conditions (Fig. 13), which tends to be higher after the 27th day of rain. Meanwhile, a significant decrease can also be caused by the rainfall that had accumulated from the previous day. According to (Suryo, 2013), a change in the factor of safety in vertical cracked conditions only began to occur after experiencing more than 70 days of rain. This decrease is influenced by the value of the hydraulic conductivity in the soil, the intensity of rain, number, position, and the dimension of cracks. The study also conducted a comparison of the factor of safety value with the presence of vertical cracks with variations in the position of the crack to decrease the safety value.
The location of the horizontal drain placement also affects its effectiveness. HD 1, which is in a lower position, improves the slope stability compared to HD 3, which is in the top position (Fig. 13a, b, and c). This result is only valid when the groundwater level is high (GWL 3). Meanwhile, different results were obtained in GWL 1 and 2, where horizontal drain did not increase slope stability (Fig. 13d). This occurred due to the presence of vertical surface cracks that did not increase the amount of rainwater that enters the soil, therefore, it does not affect the groundwater level at deep elevations. The results obtained will be different if there are more surface cracks and longer cracks on the slope. The effect of the horizontal position of the sub-drain, where HD is at a low elevation with a higher level of effectiveness, was also confirmed by (Rahardjo et al., 2003) and (Rahardjo, 2012) in their parametric study and field validation. (Ismail, Ng and Abustan, 2017) also obtained the same results on the relationship between the horizontal position of the sub-drain and its effectiveness.
VCWL obtained almost the same results as ES and VC in some conditions. Figure 15a shows that the higher the groundwater level, the lower the factor of safety. The same results were also obtained in the ES and VC conditions, which indicated that the elevation of the existing groundwater level is also very influential on the stability of the slope. The horizontal sub-drain did not increase the slope factor of safety at a low level. This is because the position of the weak layer is far above the level at low elevation and the horizontal sub-drain does not touch the groundwater. Meanwhile, the very small Ks condition of the soil restricted the rainwater from reaching the weak layer position and the horizontal drain to becomes ineffective. The increase in the factor of safety is observed only when it rains for 6 hours after experiencing 28 days of rain (Fig. 14c). This shows that the location of the horizontal sub-drain installation and the duration of one day of rain also affect the effectiveness of the sub-drain. From Fig. 13b, the HD 3 which is in the top position has the same factor of safety as the slope without horizontal drain. This occurred because the position of HD 3 is only in a small part of the weak layer. Therefore, the area and the location of the weak layer also affect the horizontal placement of the sub-drain when used in landslide mitigation. This result is almost the same as the study by (Martin, 2013), where the water from the Horizontal sub-drain is installed at the same elevation but various points have different discharges in a view location. (Ahmed et al., 2012a) also stated that horizontal sub-drain which emits the largest water discharge is only at certain points in this study area with a small resistivity value, this is also in line with the results of (Ismail, Ng and Abustan, 2017).
The effect of vertical crack and weak layer with no horizontal sub-drain condition was also observed in this study. Figure 15 shows the changes in the factor of safety in 30 days of rain with variations in groundwater level and soil parameters. In GWL 1 and 2 (Fig. 15a and b), the factor of safety under VCWL conditions is the highest, specifically at the beginning of the rain, but significantly decreases to the lowest for ES and VC conditions. However, in GWL 3 (Fig. 15c), which is the highest groundwater level, the factor of safety for VCWL is much lower compared to the ES and VC conditions. This indicates that at high groundwater levels, the presence of cracks and weak layers significantly reduces the stability of a slope. Meanwhile, at lower groundwater levels, the decrease in the factor of safety in VCWL conditions only occurred after several days of rain. These results showed that the stability of a slope that experiences rain with different intensities daily is influenced by soil conditions such as the presence or absence of cracks and weak layers as in this case, the elevation of the existing groundwater level, and contour topography.
Another experiment was carried out by adding the number of vertical cracks under VCWL conditions. The number of surface cracks previously was 6 pieces and was added to 18 pieces to determine the effect of rain when the ground water level is low. In Fig. 16a it can be seen that the more the number of cracks on the surface, the smaller the safety factor that occurs. In addition, the horizontal sub-drain in this condition also becomes effective during heavy rains on day 27 (Fig. 16b). This condition also shows that rain can affect the effectiveness of the horizontal drain. The presence of high intensity and duration of rain is proven to also affect the effectiveness of the horizontal drain. In Fig. 17a it can be seen that the horizontal sub-drain can be more effective in increasing the safety factor to reach 11.5% in rain with a higher intensity, which is between 1.41x10− 05 − 1.85x 10− 07 and occurs throughout the day for 14 days. At a lower rain intensity, namely 3.52x 16− 06-1.16x10− 08 and occurs throughout the day for 30 days, the horizontal sub-drain will be able to increase the safety factor when it rains with a higher intensity, namely on days 27–28 (Fig. 17b). These results indicate that the number of cracks, the intensity of the rain and the duration of the rain can affect the effectiveness of the horizontal sub-drain to increase the safety factor slope.
Furthermore, a slight oddity occurred in the VCWL conditions in GWL 2 and 3, where there was a slight increase in the factor of safety at the beginning of the rain. This occurred due to the changes in groundwater levels during rain. The changes in groundwater level in this condition are very different from the ES and VC because there is a weak zone with a larger Ks parameter. According to Vucovic (1997) and (Mattei et al., 2020), the differences in hydraulic conductivity parameters have a major effect on groundwater level fluctuations. This is because of an increase in the factor of safety caused by fluctuations in the groundwater level that occur. The increase in the value at the beginning of the rain does not occur when the existing groundwater level is very low.
Investigation on the effectiveness of horizontal drain during the rain was also carried out with a higher hydraulic conductivity value. (Tang, Tang and Wei Liu, 2011) discovered that the presence of a horizontal drain can increase the factor of safety by up to 25% when monitoring is carried out 48 hours after 60 mm/hour of rain for 9 hours. This study was conducted on soil with Ks = 9.2 x 10− 6 m/s. According to (Rahardjo, 2012), the presence of a horizontal drain increased the SF from 1,193 to 1,303 with Ks = 2.1 x 10 − 7 m/s. Meanwhile, the horizontal drains were installed at the foot of the slope with a length of 12–18 meters in Rahardjo's study and 25–30 meters in (Tang, Tang and Wei Liu, 2011). In the previous report by (Lin et al., 2016), increasing the factor of safety with horizontal drain combined with sub-surface drainage increased the value for slope from 1,098 to 1,228.
The position and length of the horizontal drain also affect its effectiveness. In this study, the installed horizontal drain for some conditions has not reached the location of the groundwater level. Based on (Cook, Santi and Higgins, 2008) summaries, drains need to extend far enough into the slope to achieve the desired water level drawdown throughout the slope. Since water needs to be removed from the slip zone, drains are installed to penetrate through this zone. Apart from the minimum length limit, the effectiveness of the horizontal drain can also be affected by its maximum length. (Royster, 1972) stated that drains must not extend more than 3–5 meters past the slip surface. Meanwhile, (Lau and Kenney, 1983) stated that no additional benefits can be achieved by installing drains that extend beyond where the critical slip surface intersects the top of the slope. This result was supported by (Nakamura, 1988) which showed that the maximum reduction in subsurface water is not affected by changes in drain length beyond a critical length. Furthermore, (Cai et al., 1998) stated that the increase in the factor of safety became smaller when the drains extend beyond a critical length. This showed that installing drains that significantly exceed the slip surface is uneconomical and can cause more water to be conveyed into the failure zone (Royster, 1972).
In this landslide case, the soil at the study location has the dominant parameters of hard clay and rock with high shear strength, but there are weathered layers in several zones below the soil surface. This showed that there is a slight decrease in the factor of safety due to rainfall with and without horizontal drain installation. From the results of numerical analysis, the decrease in the value of the factor of safety under ES conditions is between 0.01–0.03 without and with horizontal drain, while VC is between 0.02–0.035. However, the decrease in the safety value in the VCWL condition is slightly higher, namely 0.11. The presence of a horizontal drain also increases the slope factor of safety. Based on the results, the increase due to horizontal sub-drain installation ranges from 0.02 to 0.083 in 30 days of rain. This increase in the factor of safety is slightly smaller than the study by (Rahardjo, 2012). Meanwhile, the result is due to the Ks parameter on the slope, which is much smaller than the rainfall intensity causing the value of the increase in the factor of safety. The soil shear parameters and the horizontal sub-drain length also affect the effectiveness of the horizontal drain.
Gambar 17. Changes in safety factor during 14 and 30 days of rain; a) Effect of horizontal sub-drain due to rain for 14 days and occurs throughout the day with high intensity; a) Effect of horizontal sub-drain due to 30 days of rain and occurs throughout the day with lower intensity
(Ismail, Ng and Abustan, 2017) conducted a study to determine the effectiveness of horizontal drain on cutting slopes with high GWL by varying the length. The result showed that a horizontal drain with a minimum length of 22.5 meters is the most effective way of increasing safety. In this case, the value of the factor of safety increased from 1,021 without a horizontal drain to 1,456 with a horizontal drain of 25 meters. This investigation was conducted on a cutting slope with lower shear strength and Ks parameters compared to this case study. The increase in factor of safety after a horizontal drain was installed in the report by Ismail was greater than the results of this study, (Rahardjo, 2012), and (Lin et al., 2016) but almost the same as (Tang, Tang and Wei Liu, 2011). Therefore, the effectiveness of the horizontal drain is very sensitive and is influenced by many factors, namely soil parameters, topography slope, groundwater level conditions, rainfall intensity and duration, length, and horizontal position of the drain. When the horizontal drain is installed in a suitable location, its effective length can significantly increase the factor of safety by lowering the groundwater level and reducing the effect of rain on slope stability.
Based on the landslide case, the results showed that the condition of the subsurface soil is very heterogeneous and affects its stability. Soil parameters obtained from conventional testing in form of boring, CPTU, N-SPT, and laboratory tests are not sufficient to explain the real conditions for all layers below the soil surface. Therefore, the soil parameters from the conventional test are stable, making it impossible for landslides to occur. There are weathered soil and surface cracks that are not known from the test. A geophysical test that obtains a resistivity value can help identify subsurface zones with a higher water content, which is believed to be a weak layer due to soil weathering. This causes the differences in soil parameters that occur where the shear parameter becomes lower (Zhang, Tao and Morvant, 2005) and hydraulic conductivity is greater ((Gofar et al., 2006), (Wang and Li, 2011) and (Suryo, 2013)).
In this study, the landslide occurred due to heavy rain and seeped into the ground through vertical cracks. The continuous ingress of rainwater through vertical cracks is suspected to have caused two things, namely, first, crack propagation until it reaches the weathered deep layers. The second is weathering of hard soils, a decrease in soil shear parameters, and an increase in hydraulic conductivity parameters. When this condition is accompanied by continuous rain, it causes a decrease in soil parameters leading to slope instability and landslides. The horizontal drain will be effective in increasing slope stability when it is installed in a good position in the weak layer zone and locations with high rainfall intensity as well as a fairly long duration of rain such as in tropical countries. This will lead to an increase in the groundwater level due to rainwater infiltration which can be removed through the horizontal drain.