Case analysis on the characteristics of the landcreep area: focusing on Seo-myeon in Suncheon, South Korea

This study was conducted to prevent landcreep damages by examining various features such as geology, topography, and physical characteristics. The first landcreep occurred in Seo-myeon in Suncheon, South Korea, in 2000. The second occurred in 2020, the damaged surface area of the second event was approximately 21.7 times that of the first, and it was discovered that landcreep in Sunchoen was contributed to forest road construction. The study area was a typical landcreep area caused by weathered rocks, and this area belonged to a type of convex terrain ground form. As a result of analyzing the landslide hazard grade, the landcreep is found to be different from general landslides because it includes most areas of landslide hazard grade 3–5 and out of the grade. Results show that the landcreep has proceeded in the directions of S15°E, S20°E, S25°E, and S30°E, which is consistent with the direction following the flow direction of groundwater determined using the electrical resistivity survey method.


Introduction
The term "landcreep" used in this paper is expressed as a phenomenon in South Korea, where a part of the soil layer continues to gradually move while maintaining some extent landslide toward a lower area along a part in weak shear resistance or adhesion in the ground.
It is caused by groundwater on a mountain slope or slope face composed mainly of soil or rock layers. Distinguishing it from general landslides caused by heavy rains is often difficult. Landcreep tends to occur intensively in a zone with specific geology or geological structure. The most vulnerable type is the presence of an impermeable layer on the sliding surface, which often occurs when the soil layer acts as an active lubricant due to the presence of groundwater; in such a zone composed of shale, mudstone, tuff, etc., and in the 1 3 case of Japan, this phenomenon is observed (Takaya 2017). The "landcreep" refers to the slow descent of rocks or dirt from downhill in other countries. It is explained as soil deformation due to long-term pressure. The investigators may consider it a continuous phenomenon, but it is defined as a minute and discontinuous movement caused by gravity.
In South Korea, the landcreep has existed for a long time, but there has been no exposure before looking for large-scale mountainous areas. But currently, large-scale landcreep is exposed due to tensile cracks and collapses resulting from slope cutting or natural phenomena that are caused by various development activities (Park 2018).
Until the 1970s, numerous shallow landslides had occurred in South Korea because of erosion and loss of soil due to mountain degradation (Choi 2018). Currently, studies regarding the landcreep in South Korea are being conducted for finding the landcreep area by analyzing topographic maps and aerial photographs caused by the deformation of microtopography due to long-term ground movement (Park et al. 2021) and the geological and topographical characteristics of the landcreep area (Davis 1899;Takaya 2017), along with the research regarding restoration measures (National Institute of Forest Science 2017). The landcreep occurs in the relationship between clay stratum in the soil and rainfalls (Culling 1963;Jau et al. 2000) and is also caused due to the collapse of discontinuity surface and fragmental zone of fault (Tomio et al. 1990;Park et al. 2003;2005;Kim 2004). It is caused by freezing and thawing (Auzet and Ambroise 1996) or an increase in the groundwater level (Matsukura et al., 1983;Anderson and Richards 1987;Montgomery et al. 1998), as well as artificial development, such as forest road opening (Park et al. 2003; National Institute of Forest Science 2017). Therefore, this study has been conducted to provide basic data for preventing landcreep damages by analyzing the characteristics of the landcreep area at San 41, Pangyo-ri, Seo-myeon, Suncheon, South Korea.

Material and methods
The forest type of this survey area is a coniferous forest where the Pinus thunbergii is the dominant tree. The geology of this site is composed of the Precambrian Nangrim group, gneiss, and mixed gneiss, with a lot of weathering. There is a concern of additional landcreep. The soil type of the landcreep area was slightly dry brown forest soil (B2), and the soil depth was 30-60 cm or less, which was the same as the domestic average soil depth (30-60 cm, Jeong et al. 2002). From the particle size analysis, the content (%) in sequence appeared (sand: grit sand: clay = 45.2: 41.0: 13.8), which was a loam, and among the slopes of the target area, the steep slopes (58% or more and 30° or more) was accounted for the most at 78.8%, followed by 3.0% of the steep slopes (47-58% and 25°-30°), and 3.0% of steep slopes (37-47 and 20°-25°), 1.9% of the sloped area (27-37% and 15°-20°), and 3.8% of the flat area (< 9% and < 5°) including gently sloped area (9-27% and 5°-15°) (Fig. 1).
Along forest roads, there is a tensile crack in the middle of the slope on the left, and there is a valley on the right as steep slope. The elevation distribution has been shown at 200-300 m with 86.1% the most, at 100-200 m with 5.2%, and at 300-400 m with 8.7%. The geology of the target area was classified based on the geological map. During an onthe-spot investigation, the size of the investigated area (length × width × depth of collapse) and the location of the tensile crack were identified using GPS (Fig. 2). This study used clinometer (Suunto) measured tensile crack and cut slope of the landcreep area. Site characteristics, such as the parent rock, bearings, shape, and mountain slope of the land crept area were investigated and analyzed using GIS (Arcmap 10.8; ESRI, Redland, CA, USA). And this study identified the damaged condition of the slopes caused by the landcreep, and the joint slope, direction, and spacing of the joints were investigated. The strike and dip of the joint were investigated on-site using a clinometer.
Investigations have been conducted as follows: (1) a strike direction by the direction of the horizontal line existing on the joint surface, (2) the direction of a dip by the inclination of the steepest line existing on the joint surface sloped from the horizontal, and (3) the inclination direction measured by clockwise from true north. To identify the topographical characteristics and signs of landcreep occurrence, the characteristics and microtopography of the location where landcreep occurred were investigated using a topographic map (1:25,000 and 1:50,000) (Varnes 1978;Jau et al. 2000). The landcreep areas were classified based on the parameters affecting landcreep, including the parent rock, dip, strike, soil type, soil physics, and the type of landcreep occurrence shown in Park et al. (2005), Park (2015), and Takaya (2017), and the slopes of the landcreep area were investigated by dividing them into straight, arc, and horseshoe (Fig. 3).
A 500 g of soil was collected and analyzed at the Korea Forestry Promotion Institute (KOFPI) for the surface geological investigation. And to determine the strata of the slope, the researchers conducted electric resistivity survey, drilling survey on three points. To identify the traces of landcreep occurrences in the past, aerial photos (1:5,000) for the target area (scale; 1:5,000) in the years 1993, 2000, 2010, and 2019 were acquired using the land information platform. The soil hardness at each survey site soil layer level was measured thrice using a penetration-type soil hardness tester (Daiki, DIK-5520) that can measure up to a soil depth of 90 cm to identify the soil bulk density differential between the upper and lower regions where the tensile cracking occurred. The permeability rate for 100 cc of soil collected using a soil sampler was determined using a water permeability measuring device with the constant head permeability test (Fig. 4) (Daiki, DIK-4000). The calculation formula was then used to calculate the soil's permeability coefficient.

Overview of the target area and results of aerial photo analysis
The tensile cracks and collapse had occurred in the target area due to landcreep after the forest road construction. Pinus thunbergii has been the most spread tree in the landcreep area. Pinus densiflora and Quercus variabilis grow in the upper layer, and juniperus, Camellia japonica, and Picea jezoensis grow in the middle layer. There are spreads of shrubs, such as Quercus serrata and Hosta longipes, and Festuca ovina is spread in lower vegetation layers. Pinus densiflora and Pinus rigida are the dominant species, indicating that these have been considerably reforested by erosion projects in the past (Choi 2018). It is assumed that the wasteland being greened and granitic gneiss beneath the fallen leaves and branches being weathered through the soiling process played a role in accelerating the subterranean landcreep (Choi 2018;Park et al. 2019).
The first landcreep occurred in 2000. Distinguishing a landslide occurred area (Debris flow) and soil runoff area (Soil runoff) is possible by visually reading out a high-resolution satellite image in a target area of natural changes 2011); no traces of landcreep were found in 1993, as a result of analyzing the high-resolution aerial photos of the 51-m-class (portal map service site; Daum) and the land information platform in 1987 and the area damaged by the primary landcreep was 1990.2 m 2 as a result of analyzing the aerial photos in 2000. As a result of the on-site investigation of this area, the traces of collapse caused by tensile cracking were clear, and the surface area was similar to the result calculated using the aerial photograph. There were tensile cracks at the upper end of the area where tensile cracking and tensile cracks with a width of 3.5 m × length of 7.5 m and a depth of 1.2 m had occurred by collapsing in March 2020. After that, the second landcreep occurred on March 31, 2020. As a result of the investigation, the damaged surface area by the landcreep was found to be 43,176.0 m 2 , and the volume caused by the landcreep was 107,940.0 m 3 . The damaged surface area of the second landcreep was approximately 21.7 times that of the first one.
When analyzing the areas cut while opening the forest road in the target area to be investigated, it was found that the joints being developed were nearly vertical, and the weathering had occurred deeply along the joint surface so that the weathered formation developed about 30 cm between the horizontal joints. Landslides are widespread when worn soils are Q : seepage quantity of water A : surface area of sample t : permeable time L : length of sample soaked with water because they shape soils with high clay content. The joint surfaces of landcreep in this area classify them as wedge-shaped landslides. The landslide in this location occurred due to the deep geological and topographical causes of weathering formation and the cutting of the slope due to the construction of the forest road (Figs. 5, 6 and 7).

Topographical analysis of target area using GIS
The type was a relative ground topography of convex, which belonged to a weathered rock landcreep, and the landcreep had occurred in the convex area. In the target area to be investigated, the ratio L/W of the width (W) to the length (L) of the landcreep area was 1.5 or less and approximately 0.8. The thickness of the soil block was lower than that of both sides along the centerline of the horseshoe-shaped from the head at the beginning of the occurrence, and it corresponds to the horseshoe type, which occurs in thick cases . Furthermore, the convex-shaped area on the slopes affects landcreep and is continuously pushed in the form of a base. The all stone retaining wall functions as a raised end section at the lower part (National Institute of Forest Science 2017). Landcreep had occurred at the inquiry target region in the relatively high convex topography. It is considered a result of elements affecting landcreep, such as talus at the upper part of the slope .
As a result of conducting the on-site GPS survey of the area where the collapse had occurred (inner boundary line) in the investigation target area and the area at risk of additional landcreep (outer boundary line) in the event of a local downpour in the future (Fig. 8), the area of the outer boundary was analyzed to be approximately 3.6 ha. The landslide hazard grade, direction, slope, elevation, and geology were analyzed using the digital topographic map of the area (Fig. 9) with GIS.
The Korea Forest Service assigns weights to topographic factors (slope, orientation, length, curvature, and topographic wetness index), stock map, diameter class, soil depth, and parent rock. It categorizes them into five different landslide grades (Grade 1-5). As a result of analyzing the landslide hazard grade in the area based on the seven criteria specified by the Korea Forest Service, it was found to be 0.5% for grade 1, 26.6% for grade 2, 48.4% for grade 3, 20.1% for grade 4, and 4.3% for grade 5. Thus, according to the National Institute of Forest Science (2017), it was found that the landcreep area is different from general landslides as most of the landslide hazard grades are 3-5 grades, and non-grade areas are included.
As a result of analyzing the direction distribution of the target area, 63.0% has been identified to be in the east direction, 23.9% in the southeast direction, 5.2% in the south direction, 4.3% in the northeast direction, 2.4% in the southwest direction, and 1.1%  in the west direction. In other words, when compared with the results of analyzing the direction of growing in distortion of trees in the past due to the landcreep, the landcreep direction is found to be growing toward the east and southeast. Most of the results are consistent with the direction of the outflow of groundwater that has the most significant effect on the landcreep. These findings are consistent with the conclusion that during long-term terrain changes microtopography considerably affects landcreep (Takaya 2017;Park et al. 2019). Furthermore, a phenomenon has been discovered in which trees are grown bent in the direction of the contour line in areas where tensile cracks occurred in the target area (Park et al. 2021). As a result of analyzing the elevation distribution, the result has been consistent with the study result (National Institute of Forest Science 2017) that 200-300 m were the most at 86.1%.

Geology of the target area
In the investigation area, there was an approximate tensile crack of 1.2 m caused by the first landcreep that occurred 20-30 years ago (elevation 266 m, N35° 1′ 59.00″, E127° 30′ 49.80″) due to the recent cutoff of the slopes following the forest road opening, the second landcreep has been progressed, and an additional tensile crack (approximate length of 80 m and height of 1.5 m) has occurred. There was aberrant tree growth surrounding the first region of occurrence, and groundwater was drained from the cutoff section of the forest road. The weathering of the strata was exceptionally well developed, and the weathered layer of the collapsed area was clayed and vulnerable to landcreep. Concern exists regarding additional landcreep as the bedrock is floating due to the distribution of convex-shaped talus in the lower part of the concave area. Park (2018) and the National Institute of Forest Science (2017) have reported that slope cutting in talus areas is vulnerable to landcreep. As shown in Mari (Park et al. 2005), Sangdong-myeon, Gimhae, South Korea, the cutoff of the mountain angle resulting from the opening of the forest road seemed to be the cause of the reoccurrence of the landcreep. As a result of analyzing the geological map (1:50,000) and the strata and main constituent rocks of the target area, it was identified to be in the Precambrian period with low integrity and weathering, and the supergroup was the Nangrim stratum, which was vulnerable to the landcreep (Takaya 2017). The main constituent rocks were metamorphic rocks in which weathering was relatively severe; and thus, landcreep quickly occurred (Park 2018). The landslide scrap of the target area was clear, there were no other strata or intrusive rocks, and the rocks were investigated in a state in which weathering had progressed significantly (Tomio et al. 1990;Kim et al. 2015;Choi 2018). The faults and folds appeared, and the strike was in the direction of N13°SE. The discontinuity surface promoted landcreep in the same direction as the slope ). The number of joints was discovered to be three or more, and the joint interval was thick (6-20 mm) and responsive to external variables (National Institute of Forest Science 2017; Park 2018). According to the direction and spacing of the joints, as the number of directions of joints increases, it is discovered that landcreep occurs more quickly, which is consistent with the findings of the previous studies (Kim et al. 2015;Takaya 2017;Park et al. 2019), which show that with denser discontinuities and more number of directions, the more vulnerable it is to landcreep, and the easier it is to detect (Kim et al. 2015;Takaya 2017;Park et al. 2019).
Meanwhile, the strike direction of the rock joint in the forest road adjacent to the landcreep area was E70°-77°. The rock formation on the opposite strike was N67°-80°E, which appeared opposite the landcreep area. The lower rock of the groundwater effluent point in the area where the collapsing progressed because the landcreep was progressing in the S17°W direction. The joint of the collapsed rock at the top was N65°W, indicating that the collapse proceeded in the valley direction.

Moisture measurement by each soil depth
Soil moisture content (gravitational moisture content, macroporosity), maximum waterholding capacity, weight water content, and permeability coefficient have been measured up to a soil thickness of 250 cm in the area collapsed due to landcreep. After collecting the soil thrice in 100 cc cans, soil collection was analyzed using a soil sampler for collapsed areas in the target area. Topsoil exhibited the highest gravitational moisture content among the soil moisture content at 32.0%. As for the microporosity content, the topsoil was 34.9% as the topsoil was deeper, and the pore volume decreased as the topsoil became deeper. At a depth of 250 cm, it was 34.1%, which was lower than that of the topsoil but exhibited a similar trend. It was considered the result of the condition of the collapsed soil being composed very inhomogeneously. The maximum water-holding capacity was 48.9-54.1% from the topsoil to the soil depth of 250 cm, and the weight water content was in the range of 0.40-0.50%. The permeability coefficient of the soil was the highest (0.0173 cm/s) at a depth of 200 cm, and at a depth of 20 and 100 cm, low values of 0.0052 and 0.0035 cm/s were identified, respectively. And appeared at 0.0148 cm/s in the topsoil but again showed a low value at 250 cm to 0.0080 cm/s. Similarly, the trend of permeability was dependent on the depth of the soil; therefore, the rainwater deposited on the soil during accumulating rainfall in a place with a low permeability coefficient, and it was identified that the landcreep occurred while the groundwater was being pushed out at once to a place with a high permeability coefficient (Park 2018). As shown in the study results of Woo (1992) and Choi (2018), the soil structure is vulnerable to the landcreep.

Particle size analysis of target area
The soil type of the landcreep area was slightly dry brown forest soil (B2). The soil depth was relatively shallow, with < 30-60 cm. It is a result of the distribution of the talus zone, and the soil depth was relatively more in the left and right parts of the area where the collapse occurred due to the landcreep. It was found that the moving zone (remnant lamination, depth, cm, A + B + C layer) corresponded to 100-200 cm. After collecting 500 g of soil from the area that collapsed due to the landcreep and the control area where landcreep did not occur, the average soil pH was 5.95 (4.90-6.60) as a result of particle size analysis and soil pH analysis at the Korea Forestry Promotion Agency. It was higher than the average pH of forest soil in South Korea (Jeong et al., 2002). The content (%) of sand, silt, and clay of the topsoil in the landcreep area was 45.2, 41.0, and 13.8, which was determined to be loam, and the natural mountain area was 81.7, 16.1, and 2.2, which was determined to be loamy sand, based on particle size analysis. And collapsed soil by landcreep was 52.9, 30.3, and 16.9, which was determined to be loamy sand; lastly, the soil at the groundwater outlet result was 64.2, 23.8, and 12.0 analyzed loamy sand, so the soil content of landcreep and soil content of each soil were found to differ based on the location of the natural mountain area used as a control and the area where landcreep occurred. And the heterogeneous soils (Tomio et al. 1990;Kim et al. 2015) exhibited soil characteristics vulnerable to the landcreep.

Soil hardness of the target area
The soil hardness for each soil depth was measured thrice using a penetration-type soil hardness meter (Daiki, DIK-3300) to measure up to a depth of 90 cm (Table 3). Soil hardness (kgf/cm 2 ) was measured for the areas at risk of landcreep, the boundaries of the areas at risk of landcreep (the boundary between the collapsed areas and the areas at risk of landcreep), the areas collapsed due to landcreep, and the control group in which no landcreep occurred. For the soil depth, 80 cm of the area was at risk of landcreep, 20 cm at the boundary of the landcreep area, 40 cm at the collapsed area, and 25 cm in the control area.

3
The control area in the areas at risk of landcreep and the boundary area was harder than natural soil because the soil hardness was assumed to be due to talus and weathered rocks underground.
Furthermore, the landslide collapsed the area's soil hardness and quickly dropped below 30 cm of soil depth. It is considered the case because the soil hardness was quickly broken as the higher soil mass was pushed down, causing the collapse below. As shown in Table 3, the upper end part of the tensile crack caused by landcreep exhibits a relatively less disturbed state. In contrast, the lower part of the tensile crack caused by landcreep exhibits lower soil hardness due to disturbance and disintegration. It was discovered that the soil disturbance caused by landcreep affected the soil hardness (Tables 1, 2 and 3).

Analysis of the flow direction of groundwater
The landcreep is closely related to groundwater (Woo 1992;Park 2015); the investigation target area has a stream flowing downstream; when it rains, groundwater flows out from the forest slope coincidentally with the direction of the landcreep and enters into the stream. The groundwater is assumed to be easily flowing in the lower part of the landcreep area and the lower part of the tensile cracked, affecting the landcreep (Park 2018;Choi 2018). In other words, the electrical resistivity value distribution of the lateral line L-1 is 632-8,735 Ωm, indicating a very high distribution. This was assumed to be affected by gravel, gravel-cobble, and megalith mixed in the soil layer. There is a clear difference in electrical resistivity at the 65 m point in which the slope activity is confirmed due to landcreep; this is the effect of the increase in porosity resulting from the relaxation of the soil by the slope activity or the effect of the groundwater in the soil. The strange seepage zone is assumed to be distributed up to 10 m below the surface, and it is affected by the relaxation of the soil layer and the groundwater up to 10 m below the surface. As a result of the electrical resistivity survey of the L-2 sidelines, the distribution of electrical resistivity values is 609-20,667 Ωm, indicating a very high distribution, which is affected by gravel, gravel-cobble, and megalith mixed in the soil layer. In particular, the electrical resistivity cross-sectional diagram shows a low electrical resistivity distribution at starting and ending points. The leakage water observed at the downstream tip is assumed to be flowing and leaking to the downstream slope along the topographical gradient centered on the low anomalous seepage zone caused by the effect of groundwater. The landcreep proceeds in the directions of S15°E, S20°E, S25°E, and S30°E, and this result was consistent with the direction along with the flow direction of groundwater by the electrical resistivity survey, which was consistent with the results of Lee et al. (2019) and Park et al. (2019) (Figs. 10,  11).

Conclusion
In Korea, most landcreep is caused by artificial development (park et al. 2003); this landcreep occurred due to the construction of a forest road in areas at risk of landcreep, such as cracks. So before artificial development on the mountain slope, if it is determined as a place of concern for landcreep through a close investigation, it should be carried out after countermeasures are taken, and policy issues should be prepared. The risk rating of landslides developed by the Korea Forest Service accounted for most of this study's 3-5 grades. Most of the landcreeps in Korea are not included in the danger zone , so there is a difference in the causes of landcreep and landslides; it is necessary to produce a landcreep risk map by through a nationwide landcreep survey.
In the case of Korea, the study of landcreep is now in the primary stage, and prior research is needed. And also, as in this study, many studies are necessary to understand the characteristics of the landcreep that occurred in a specific area. Supposed the characteristics of landcreeps are collected I big data. In that case, it will be possible to predict areas of concern for landcreep and to protect the property and life of Korean people by developing a method to take pre-emptive measures before landcreep occurs.
Funding This study was conducted with the support of the "R and D Program for Forest Science Technology (Project No. 2021347B10-2123-CD01)" provided by the Korea Forest Service (Korea Forestry Promotion Institute).