Identification of Landslides Controlling and Triggering Factors using the Analytical Hierarchy Process (AHP) in Dermasuci Village, Pangkah District, Tegal Regency, Central Java Province - Indonesia

The National Disaster Management Agency (BNPB) stated that in 2022 around 634 landslides occurred in Indonesia. One of the regions in Indonesia that experienced landslide disasters is Dermasuci Village, Pangkah District, Tegal Regency, Central Java Province. The first landslide disaster occurred on February 5, 2022, and the second landslide occurred on February 10, 2022. The landslide disasters caused damage to 61 residents’ houses and four road sections. The objective of this study is to identify the controlling and triggering factors that mainly contributed to the landslide event. These two factors are determined based on the order of their static and dynamic properties using the analytical hierarchy process (AHP) method. Controlling and triggering factors of the landslide in the study area are determined based on geological and engineering geological conditions which were obtained from surface mapping results. The controlling factors obtained are the rock bedding inclination, geological structure in the form of faults, slope inclination, engineering properties of soil and rock, land use, and groundwater table. The triggering factors that are known and commonly occur are rainfall in the form of antecedents and intensity, and earthquakes that can induce landslides.


Introduction
Indonesia is one of the countries that have a high level of vulnerability to natural disasters in the world.Indonesia ranks 38 th out of 181 countries with the highest natural disaster risk index 21 .As of December 31, 2022, there were 3,544 natural disaster events occurred in Indonesia; where the landslide disaster which is the main topic of this study, ranks third with a total of 634 incidents 11 .One of the areas in Indonesia that experienced landslides throughout 2022 is Dermasuci Village, Pangkah District, Tegal Regency, Central Java Province.The first landslide disaster in Dermasuci Village occurred on February 5, 2022, and the second landslide occurred on February 10, 2022.The landslide disaster caused damage to 61 residents' houses and four road sections 17 .The administrative map for the Dermasuci Village and the location points where the landslide occurred is shown in Figure 1.According to the geological literature study, there are two formation units lied in the research area 5 , namely: the Rambatan Formation (Tmr) and the Halang Formation (Tmph).The Rambatan Formation (Tmr) is composed of shale, marlstone, and limed sandstones; and the Halang Formation (Tmph) is composed of andesitic sandstones, tuffaceous conglomerates, and marlstone.Based on regional physiographic conditions, the research area is situated in the North Serayu Anticlinorium zone 2 , and based on regional structural geology is located in the Tegal High 16 .These two conditions cause the Dermasuci Village area to have a complex rock bedding inclination and massive geological structure, which can potentially be the factors causing a landslide.
A landslide is the downslope movement of rock or soil or both, that occurs at the ruptured surface.There are two types of rupture surfaces, namely curved rupture/rotational slide, or planar rupture/translational slide, in which much of the material often moves as a coherent or semicoherent mass with little internal deformation 19 .A landslide can occur because it is influenced by two main factors, namely: controlling factors and triggering factors.The controlling factors are factors that make the condition of a slope vulnerable and ready to move, and the triggering factors are factors that change the condition of a slope from a prone condition or ready to move to a critical condition and finally move 7 .
This research was carried out with the objective of identifying the controlling and triggering factors that caused landslides in Dermasuci Village.Controlling and triggering factors are determined based on geological and engineering geological conditions obtained from surface mapping results.The controlling factors obtained from surface mapping results include the rock bedding inclination, geological structure in the form of faults, slope inclination, engineering properties of soil and rock, land use, and groundwater level.The triggering factors obtained from secondary data include rainfall in the form of antecedents and intensity, and earthquakes.All these factors will be sorted from the most static to the most dynamic characteristics using the analytical hierarchy process (AHP) method.

Data and method
Identification of controlling and triggering factors of landslides in Dermasuci Village was carried out using primary and secondary data.Primary data is obtained from the results of surface geological and engineering geological mapping.The results of surface geological mapping obtained data as follows: strike and dip of rock, distribution of rock types, and distribution of geological structure.Based on the results of the mapping that has been carried out, a total of 134 observation points were obtained, of which 48 observation points obtained strike and dip of rock data.There are four types of rock identified and distributed in the study area, namely marlstone (Tmr), limed sandstone (Tmr), marlstone (Tmph), and andesitic sandstone (Tmph).Geological structures that can be identified include strike-slip faults, normal faults, reverse faults, and indications of anticlines and synclines.The surface geological mapping results are presented in the form of a geological map, as shown in Figure 2.

Figure 2. Geological Map of Research Area
Surface engineering geological mapping obtained data in the form of landslide distribution and direction, slope, and soil and rock engineering properties.According to the results of surface engineering geological mapping, there are 15 landslide points spread across Dermasuci Village (Figure 1).Slope data is used to determine the angle of the slope that has the potential to experience landslides, as well as those that have experienced landslides.Data on the engineering properties of soil and rock presented in this paper are limited only to those based on physical tests carried out in the field with reference to the degree of weathering of the rock mass 6 .Based on the results of the degree of rock weathering identification, there are three types of weathering degrees obtained for each rock.Limed sandstone (Tmr) is classified as fresh rock, marlstone (Tmr) and marlstone (Tmph) are classified as highly weathered rocks, and andesitic sandstone (Tmph) is classified as completely weathered rock.The distribution of rocks based on the degree of weathering and the landslide directions are presented in the form of an engineering geological map, as shown in Figure 3.

Figure 3. Engineering Geological Map of Research Area
The result of surface geological and engineering geological mapping data in the form of land use distribution combined with secondary data in the form of spatial data on the distribution of residents' houses 4 and satellite image data in the form of Landsat 8-9 OLI/TIRS X2 L2 20 , it can be used to identify four types of land use, namely forest, fields, settlements, and vacant land.The land use distribution is presented in the form of a land use map, as shown in Figure 4.

Figure 4. Land Use Map of Research Area
The next set of data is secondary data which includes rainfall data in the form of antecedents and intensities, seismicity data and earthquake potential.Based on rainfall intensity from January 25 to February 14, 2022 8 , combined with information on the landslides' occurrence time and groundwater level, an analysis was carried out to identify the highest rainfall intensity before the landslide occurred in Dermasuci Village.The rainfall intensity data is presented in graphical form, as shown in Figure 5.The rainfall intensity data used cannot fully represent the conditions at the study site because the data was obtained outside the boundaries of the research area.The rainfall intensity measurement data is located at the coordinates UTM 49 S 292372 E 9240417 N. Based on historical seismic data and earthquake potential 4 , it is known that there are six points of seismicity spread across Tegal Regency and its environs from different time periods, and very low earthquake potential.The distribution of seismic points and earthquake potential in Tegal Regency and its environs is presented in the form of a seismic and earthquake potential map as shown in Figure 6.

Figure 6. Seismicity and Earthquake Potential Map
Based on the data presented above, the controlling and triggering factors that cause landslides in Dermasuci Village can be determined.The landslide-controlling factors include the rock bedding inclination, geological structure in the form of faults, slope inclination, engineering properties of soil and rock, land use, and groundwater level.The landslide-triggering factors include rainfall in the form of antecedents and intensity, and earthquakes.These factors will then be sorted from the most static to the most dynamic by using the analytical hierarchy process (AHP) method.
The analytical hierarchy process (AHP) is a general theory of measurement used to derive ratio scales from both discrete and continuous paired comparisons.These comparisons may be taken from actual measurements or from a fundamental scale that reflects the relative strength of preferences and feelings.The comparisons are arranged in a pairwise comparison matrix to form the reciprocal matrix needed to get the eigenvalue.Furthermore, the eigenvalue is used to calculate the level of consistency of the comparison results in order to obtain a consistency index and consistency ratio 13 .
An assumption is needed to compile a pairwise comparison matrix table before the weighting process begins.The set assumption is the dynamic levels of the factors that cause landslides to begin when it rains with low, moderate, or high intensity, and there is a potential for an earthquake to occur.Weighting is done by comparing a parameter with other parameters.The weighting value given ranges from 1 to 9. A value of 1 is interpreted to have the same level of dynamics.The higher the value given to a factor parameter, the more the factor parameter is considered to have more dynamic properties compared to the comparison factor parameters. Pairwise comparison matrices of the predetermined parameters of the controlling factors and landslide triggers are shown in Table 1.

Result and discussion
The weighting that has been carried out using the analytical hierarchy process method in Table 1 obtained a principle Eigenvalue (λmax) of 8.181, a ratio index (RI) value of 1.41, a consistency index (CI) value of 0.030, and a consistency ratio (CR) value of 1.84%.Based on the provisions in the analytical hierarchy process calculation, if the weighting results in a CR value of less than 10%, the CR value is consistent and acceptable 13 , so the CR value resulting from the weighting of the parameters of the factors that cause and trigger landslides in Dermasuci Village is acceptable.The weighting value results are then sorted from the smallest value to the largest value, where the value represents the level from the most static to the most dynamic factors causing landslides, as shown in  1, it is known that the controlling factors in the form of the rock bedding inclination, geological structure in the form of faults, slope inclination, and land use have the same dynamic factor value of 0.033 or have the same static properties.This can be interpreted to mean that when the rain conditions start to fall, the four factors do not experience any changes until landslides occur.The next controlling factor in the form of engineering properties (soil and rock) and groundwater level has a dynamic value change that is increasingly dynamic, namely 0.084 and 0.161, which means that when the rain conditions start to fall, these two factors begin to experience a change in conditions.The triggering factor in the form of rainfall intensity has a dynamic factor value of 0.262, and earthquakes have the highest dynamic factor value of 0.361.Each of the factors listed in Table 1 that cause landslides will then be discussed in terms of how they influence the causes of landslides in Dermasuci Village, as follows:

Rock bedding inclination and geological structure (fault)
The rock bedding inclination and geological structures in the form of faults which are identified from the results of surface geological mapping (Figure 2), can be one of the controlling factors that cause landslides.One example of a landslide case due to the rock bedding inclination that caused a landslide in one of the buildings is located at observation location 134 (Figure 1), as shown in Figure 7. Fault zone processes control the rainfall-induced landslides in three ways 10 : (1) increasing fracture density near major faults that produce more debris and reduce rock strength, (2) weak minerals growing in fault zones in which clay minerals grow, and (3) topographic features produced by active faulting.
Because fault planes and rock bedding planes are both types of discontinuity planes, the processes of rainfall-induced landslides at the fault zone can also be applied to rock bedding.An example of a landslide due to a geological structure (fault) that was identified as causing damage to several buildings on the surface is located at observation point 59 (Figure 1), as shown in Figure 8.The geological cross-section in this case cannot be used to show how the influence of geological controlling factors (rock inclination, geological structure in the form of fault, and engineering properties of soil and rock) on the triggering factors for landslides.This is due to several reasons, namely; First, geological cross-sections cannot provide detailed rock layering data such as cross-sections made from the correlation results of core drilling.Second, the condition of the rock layers in the study area is composed of rocks with a relatively thin average layer thickness of 3 -12 cm (Figure 9), so it will be very difficult to draw the detailed rock layers on a geological cross-section with a large-scale.

Slope inclination
The slope inclination which is identified from the results of surface engineering geological mapping (Figure 3), can be one of the controlling factors that cause landslides.The slope determines how influential that strength is in ensuring the movement of the object; the steeper the slope, the stronger the gravitational strength that ensures the shear of the slope and allows the object to slide 3 .This statement is in accordance with the results of the mapping of points that experienced landslides, where there were three landslide points (38, 41, and 124) that occurred due to the steep slope of the slope, namely 70 o , 71 o , and 57 o .An example of a landslide due to a slope inclination that was identified as causing damage to several buildings is located at observation point 124 (Figure 1), as shown in Figure 10.

Engineering properties (soil and rock)
Weathering is an essential process that affects the mechanical properties of rock material and mass at shallow depths and on the surface through chemical and physical weathering 1 .The sliding occurs because of strength reduction caused by external influences, such as the propagation of weathering from the surface towards the slope inside, and the progress of weathering through bedding joints and other joints 9 .Those statements are consistent with the results of surface engineering geological mapping (Figure 3), where 11 landslide points are located in rock areas with completely weathered conditions, two landslide points are located in rock areas with highly weathered conditions, and two landslide points with special mechanisms are located in rock areas with fresh conditions.

Land use
Recent studies focusing on the effect of human-induced land use changes on slope stability have shown that in populated regions, the impact of humans on the environment contributes significantly to the initiation and reactivation of landslides 12 .This statement is consistent with the results of geological and engineering geological surface mapping (Figure 4), where 13 out of 15 landslide points in Dermasuci Village are located in areas with land use in the form of settlements.

Groundwater level and rainfall intensity
Groundwater level and rainfall intensity, respectively, are controlling and triggering factors that are interrelated to the causes of landslides, as shown in Figure 5, and are followed by their influence on controlling factors in the form of engineering properties (soil and rock).Based on the time sequence of events, it can be known that landslides did not occur immediately at the time of the highest rainfall, but some time after the occurrence of rain with sufficient and/or prolonged intensity until causing the soil condition to become saturated.This is in accordance with the results of a literature study 18 , where the intensity of rainfall will take some time to influence the magnitude of the increase in the groundwater level, followed by an increase in pore water pressure and a decrease in shear stress on the soil.When the soil shear stress reaches its critical point, landslides will occur.The length of time required for landslides to occur is interpreted to be highly influenced by the amount of rainfall intensity, and the physical and mechanical conditions of the soil or rocks.

Earthquakes
Earthquakes are a triggering factor for landslides in the most dynamic research locations because when and where an earthquake will occur cannot be predicted.In addition, the earthquake potential at the study site, as shown in Figure 6, is classified as very low.Based on a literature study from 14,15 , in a simple explanation, during the main shock of an earthquake loading will cause a rapid increase in pore water pressure followed by a decrease in soil shear strength in progressive failure motions.

Conclusions
Based on the results of the research that has been carried out, it can be concluded as follows: 1) Controlling factors that cause landslides in Dermasuci Village include the rock bedding inclination, geological structure in the form of faults, slope inclination, engineering properties of soil and rock, land use, and groundwater level.Triggering factors that cause landslides in Dermasuci Village include rainfall intensity, and earthquakes.2) Static properties are factors that do not experience any changes in conditions before the occurrence of landslides.Meanwhile, dynamic properties are factors that experience changes in conditions either individually or as a result of the influence of other factors before the occurrence of landslides.
3) The factors that cause landslides from the most static to the most dynamic based on the results of the analysis using the analytical hierarchy process (AHP) method include the rock bedding inclination, geological structure in the form of faults, slope inclination, and land use that have the same dynamic factor value (0.033), engineering properties of soil and rock (0,084), groundwater level (0.161), rainfall intensity (0.262), and earthquakes (0.361).
The results of this research have limitations and suggestions for further research as follows: 1) This research does not discuss and compare which factors have more influence on the causes of landslides.Further research can be carried out to further explore which factors have the greatest influence on the causes of landslides disaster in Dermasuci Village.2) This research does not explain in detail how the mechanism of the influence of triggering factors on each of the controlling factors that cause landslides occurs.Further research can be carried out using more detailed data (such as laboratory test data or in-situ testing) to explain the effect of triggering factors on controlling factors in the mechanical analysis.

Figure 1 .
Figure 1.Administrative Map of Research Area (Source: Geospatial Information Agency/BIG, accessed in 2022)

Figure 7 .
Figure 7. Influence of Rock Bedding Inclination toward a Landslide (Point 134 in Figure 1)

Figure 8 .
Figure 8. Influence of the Geological Structure Toward a Landslide (Point 59 in Figure 1)

Figure 9 .
Figure 9. Relatively Thin Rock Layers Condition on Halang Formation (Point 37, Not a Landslide Point Area)

Figure 10 .
Figure 10.Influence of Slope Inclination toward a Landslide (Point 124 in Figure 1)

Table 2 .
Controlling and triggering factors of landslide.
According to the weighting results of the dynamic factor values in Table The first highest intensity of rainfall (39.7 mm) on 3 February 2022, is interpreted to be the cause of landslides at point 124 which occurred on 5 February 2022.As a result of this highest rainfall, the average groundwater level has increased from -10 m in the day without rain (1 February 2022) to -0 m.The second landslide at point to occur due to prolonged rainfall that occurred from 3 February 2022 until the day the second landslide occurred on 10 February 2022.The second-highest rainfall (7 February 2022) and the third-highest rainfall (9 February 2022) are interpreted to have a significant influence against landslides at point 67, and causing the groundwater level to reach -0 m.