Landslide Potential Analysis on Hilly Terrain Based on the CPT Data along the Malang-Kediri Road, Indonesia

Natural disasters such as landslides can generate a wide range of adverse impacts, including structural damage, economic disruption, agricultural losses, and even fatalities [1, 2]. Landslides occur worldwide, particularly in mountainous regions [3]. Their potential increases with high rainfall intensity and steep slope conditions [4]. During the rainy season, elevated water discharge increases the load acting on slopes, degrades soil structure, and reduces soil bearing capacity, ultimately triggering slope failures [5]. Slope stability is directly related to the occurrence of landslides and mass movements [6], defined as the natural displacement of soil material from a higher to a lower elevation. Several factors may induce ground movement, including steep topography, slope modification for roadway construction, and weak soil layers [7].

Batu City, Indonesia, is widely known as a tourism destination; however, one of its main access routes, the Malang-Kediri road, is highly prone to landslides. According to the Batu City Disaster Management Agency (BPBD), 78 landslide events were recorded in 2022 alone [8]. The most frequently affected area is the Malang-Kediri road segment, which lies within a hilly terrain with slope gradients ranging from 40º to 50º [9, 10]. This region also experiences high rainfall intensity. Moreover, the roadway serves as a primary transportation corridor between the two cities, resulting in high traffic density. Consequently, landslides in this area cause severe disruption, as road repairs require considerable time, leading to complete isolation of transportation access [11].

Although landslides cannot be entirely prevented, various measures as mitigation strategies can be implemented to minimize risks in landslide-prone areas [12, 13]. One such approach is conducting landslide potential assessments as part of disaster mitigation strategies. Investigating landslide susceptibility requires an understanding of subsurface soil stratigraphy at the study location. Traditionally, landslide boundary identification and characterization rely on borehole investigations (direct methods) [14]. Several geotechnical in-situ tests commonly used to determine soil layer properties include the Pressuremeter Test (PMT), Standard Penetration Test (SPT), Cone Penetration Test (CPT), and others [15]. This study utilizes CPT because it offers several advantages over alternative in-situ tests, including high accuracy in characterizing soft soils, minimal disturbance to soil structure, and the ability to evaluate soft soils that are difficult to sample using conventional methods [16].

Previous research by Potvin et al. [17] used CPT to assess reworked soil formation following a landslide event at Mud Creek, Ottawa. Their findings demonstrated that slip surfaces can be identified through changes in cone resistance, marking the transition between reworked material above the slip surface and intact soil below it. Ryan et al. [18] used CPT to test the hypothesis that water-saturated clay-rich basal layers act as slip planes. The results confirmed that smectite-rich clay layers served as the slip surface, and the CPT profile also showed strength reduction at depths corresponding to the smectite horizon. Champi et al. [19] conducted a slope stability assessment in the Palapye region, Botswana, using CPTu data coupled with limit equilibrium analysis. Their use of random variable parameters and probability density functions showed that factors of safety yielded by limit equilibrium and finite element methods were consistent, with an average difference of ±5.28%. Building on these findings, the present study provides a slope stability analysis using CPT data complemented by factor-of-safety calculations for multiple conditions, enabling the identification of the most critical scenarios that trigger landslide potential along the Malang-Kediri road.

CPT results can be utilized to determine the physical and mechanical properties of soils at various depths, providing essential input for soil structure modelling. Therefore, this study aims to analyze the slope factor of safety (FS) to evaluate landslide potential along the Malang-Kediri (Trunojoyo road) corridor based on CPT data. The outcomes of this research are expected to serve as a reference for relevant stakeholders—such as road authorities, planners, engineers, and local communities—in future development and maintenance of this critical transportation route [20].

3.1 CPT data processing

The CPT test data (Figure 3) is presented graphically to identify landslide potential, with the blue line showing the q_c fluctuation. It is concluded that the change in cone price is due to changes in soil layer type; therefore, the bearing capacity of each soil layer will differ. The preparation of parameters at this depth uses additional existing data (SPT and laboratory data). CPT testing has the advantage of providing accurate data on the penetration of soft (shallow) soils. Still, it is not effective for determining the bearing capacity of deep soils due to tool limitations, and the test tool will stop if small rocks hold the steel cone, even though the test tool has not reached the depth of hard soil [40]. Figure 3 shows the differences in soil layers from S-1 to S-3 and the plotting results used to determine soil classification.

tu_pian_3.png

(a) S-1 CPT test result

tu_pian_4.png

(b) S-2 CPT test result

tu_pian_1.png

(c) S-3 CPT test result

Figure 3. CPT test data

Generally, soil classification (Figure 4) consists of the topsoil from organic soil to the deeper layers, from non-cohesive, coarse-grained to cohesive, fine-grained [41]. These layers are susceptible to sliding during the rainy season. The non-cohesive coarse-grained material that exists between the cohesive fine-grained material can trigger movement of the soil towards the foot of the slope. Meanwhile, fine-grained soils are classified as high to moderate landslide susceptibility [42]. Large external loads acting on the pavement layer further accelerate ground movement. The change in water content is possible as part of the alternate shear strength of soil layers.

tu_pian_6.png

Figure 4. Soil classification using the CPT data [43]

There were some correlations in determining physical and mechanical properties from CPT data (q_c; f_s; R_f). The geotechnical parameters used to analyze the slope stability of each slope are shown in Table 2. Sometimes, the physical and mechanical parameters must be defined by the soil layer type data only from (q_c) and (R_f) parameters because the cohesion (c) and internal shear angle (ϕ) could not be determined from (q_c) and or (f_s) only or (R_f) parameters as the baseline, soil layer from classification from prediction data graphically classified as silty sand and sandy silt layers could be used the other geotechnical parameters, such as saturated weight volume (γ_sat), unsaturated weight volume (γ_unsat), elastic modulus (E_s), etc. In one case shown in Figure 2, the correlation analysis was used to determine weight volume (γ/γ_w) and soil classification using (q_t or q_c) and (R_f).

Table 2. Soil parameter correlation result

Results of the correlation analysis between CPT data, unit weight, and soil classification. Based on the graph above, in addition to obtaining the soil classification, (γ/γ_w) is also obtained (γ/γ_w)_, where the soil content weight is determined by pulling the intersection point of (q_t) and (R_f) down to get (γ/γ_w)_. The weight of the soil content is obtained from (γ/γ_w)multiplied by (γ_w). The average slope gradient is larger than 60%, meaning a steep slope. Based on the results of the correlation of CPT data and geotechnical parameters, the unit weight soil (γ), internal friction angle (ϕ), and cohesion (c) for each soil layer were obtained.

The sounding tool is handy for predicting landslides because it provides detailed soil layer profiles, estimates shear strength parameters, and detects weak zones and pore water pressures, which play significant roles in slope failure. In practical analysis some pay attention should be given in determining soil bulk density or weight volume (γ); overburden stress (σ_v0) and (σ’_v0); undrained cohesion (c_u) and internal shear angle (f); and deformation modulus (E_s) especially for cohesive fine grained; and finally for non-cohesive in determining density relative (D_r) during simulation using the finite element analysis or software for genotechnical analysis (geostudio or slope W).

3.2 Discussion

Landslide potential analysis was conducted for 3 scenarios on each slope to determine the lowest FS. Scenario 1 simulates the slope not receiving any load. The study of scenario 1 was intended to determine the primary FS of the hill. Scenario 2 slopes received dead and live loads. Scenario 2 slopes were simulated under normal traffic conditions and in dry conditions. Scenario 3 slope receives dead load, live load, and pore water pressure—scenario 3 slope condition with normal traffic during rainy conditions. Scenario 3 was chosen because landslides in research locations occur more frequently during the rainy season. Scenario 2: to determine the slope that is already in critical condition and even unstable during dry conditions. Scenario 1 is to determine whether the foundation soil layer on the road body is stable or unstable. The analysis with 3 slope scenarios aims to simulate the actual conditions as accurately as possible. Rainfall conditions were added to assess landslide potential.

tu_pian_7.png

Figure 5. Slope analysis result

The slope analysis found that 4 of 9 scenarios are unstable: Slope S-1 in scenarios 2 and 3, and Slopes S-2 and S-3 in scenario 3. Based on Figure 5, the unstable slope at the point of S-1 was more caused by the composition of the organic layer failure at the top layer. This layer was very soft and could not withstand traffic loads. The FS value at the S-1 point decreased drastically from 1.744 to 1.049 between scenario 1 and scenario 2. Then, scenario 3 at the S-1 point reached the lowest FS (around 0.55) during rainy conditions. Some references stated that an FS value lower than 1.07 would be classified as an unstable slope.

According to some references, if the FS < 1.07, the slope is unstable. The slip surface obtained from the analysis can be used to simulate the slope collapse. Slope loading scenarios can weaken the slope’s strength and cause the slope’s condition to become unstable. Further research is required to design and improve the topsoil on this scene based on the slip surface. Mainly, the S-2 and S-3 points consisted of sandy silt and silty sand. Based on scenarios 1 and 2, S-2 and S-3 were stable with an FS > 1.07. However, S-2 and S-3 points became unstable or FS < 1.07 during scenario 3 (rainy session). The unstable slope was caused by the accumulation of water from the crest to the toe, and all soil layers near the phreatic line became saturated during the rainy season. Finally, the pore water pressure would increase due to water infiltration, thereby influencing the total stress. Furthermore, the presence of water in rocks reduces the coefficient of friction and cohesion between mineral particles [44]. Then, the effective stress would decrease, and soil layers could move to the toe of the slope because the effective stress would not be able to retain the soil grains around the slope.

Slopes with natural soil conditions without loads are stable with FS > 1.25; when they receive traffic loads, they tend to remain stable even though the FS has decreased. The determination of this value was based on considerations of the external loads (in this case, the traffic load) acting on the slope area, and the slope remains stable along the FS for a few moments. However, in scenario 3, when the slope is subjected to traffic loads and pore water pressure, the slope becomes unstable and has the potential for landslides when FS < 1.07. This study found the FS equals 0.55, 0.59, and 0.74, respectively, according to S-1, S-2, and S-3. Furthermore, the depth of the slip surface can be determined from changes in the q_c value during the penetration process. According to some references, the slip plane location can be inferred from the (q_c) value, which is higher in the layer above the slip plane than in the layer below it [17]. More in-depth analysis is needed to determine a more accurate slip surface; one approach is to use Geostudio or slope-W software.

Generally, the soil layers in the study area are dominated by sandy silt and silty sand layers. From some references, both soil types have low cohesion (c) and high internal friction angles [45]. The type of soil layers and parameters result in different types of landslides. An unstable slope with toppling failure and a rotational slip surface is more likely to occur because cohesion (c) is very small and cannot resist the movement of soil grains when subjected to external loads around the slope. On the contrary, a translational landslide occurs when the internal shear angle is minimal, resulting in soil movement. Slopes with organic material can be weak, susceptible to landslides, or have a low FS. Soil replacement and ground improvement systems would be needed to stabilize these slopes. Pore water pressure increases during rainy periods and triggers soil movement from the crest to the toe of the slope. Slope instability can be caused by water infiltration from the slope crest due to rainfall intensity and duration. Pore water pressure can be reduced by installing drainage channels along the slope and/or replanting vegetation to prevent soil movement from the slope crest.

The CPT can effectively predict slope stability in hilly terrain and determine FS, as it is the easiest tool to mobilize at the study site. However, sometimes the CPT cannot penetrate the hard layers because soil layers contain some gravel fractions between the grains of the soft soil. Advanced CPT tools, such as CPTU, can measure pore water pressure directly during penetration into soft soil and evaluate soil parameters, providing more accurate estimates of the FS value.

CPT equipment can be used to assess pore pressure and shear strength, two critical factors that regulate slope stability. CPT devices were deployed in undisturbed soil layers of the body slope. CPT devices were placed in the main body of the slide, across the toe of the slope, and in undisturbed slope soil layers. Cone resistance of the body slope, which ranges between 10 and 1765 kPa, will correspond to undrained shear strengths from 5.0 to 25 kPa. 

Despite the evidence for landslide features in the study area, the gravity core description alone cannot unambiguously distinguish between the undisturbed and soil-layer deposits. The deposit represents primary sedimentary rock and, therefore, the Landslide is either older and was not reached by the CPT data. Since we were unable to perform long-term pore-pressure dissipation experiments, we cannot safely propose the physical trigger mechanism for the landslide types [46]. Scientific assessment methods for evaluating soil strength are still lacking. Thus, future research should focus on predicting the spatial distribution of the physical and mechanical parameters of slope soils from the top to the toe.

This work is supported by the Directorate of Research, Technology, and Community Service (DRTPM) through funding for doctoral dissertation research (Grant No.: 045/E5/PG.02.00.PL/2024 and Derivative Grant No.: 00309.9/UN10.A0501/B/PT.01.03.2/2024).