1. Overview of the target area and results of aerial photo analysis
The tensile cracking, gaps, and collapsing had occurred in the target area due to land creeping after the construction of the forest road, and as a result of three times repeating the vegetation survey on a quadrat with the size of 20 m × 20 m in the crept land area, the Pinus thunbergii has been found to be the most spread vegetation in the crept land area, pine trees, and oyster oak trees in the upper layer, and juniper and camellia trees are growing in the middle layer. There are spreads of shrubs, such as serrata, spruce, and red spruce, and bibichu and festuca are spread in lower vegetation layers. Pine trees and Rigida pines 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 land creeping (Choi, 2018; Park et al., 2018).
The first land creeping 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 (Lee and Kim; 2011); no traces of land creeping 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 (Fig. 2), and the area damaged by the primary land creeping was 1,990.2 m2 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 gaps and tensile cracks at the upper end of the area where tensile cracking and gaps 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 land creeping occurred on March 31, 2020. As a result of the investigation, the damaged surface area by the land creeping was found to be 43,176.0 m2, and the volume caused by the land creeping was 107,940.0 m3, and the damaged surface area of the second land creeping 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 soaked with water because they shape soils with high clay content. The joint surfaces of land creeping landslides 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.
2. Topographical analysis of target area using GIS
The target area was a typical land creeping-type landslide area formed by groundwater, and the type
was a relative ground topography of convex (凸), which belonged to a weathered rock land creeping-type landslide and the land creeping had occurred in the convex area (凸). In the target area to be investigated, the ratio L/W of the width (W) to the length (L) of the land crept area was 1.5 or less and approximately 0.8, and 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. It corresponds to the horseshoe type, which occurs in thick cases (Park et al., 2018). Furthermore, the convex-shaped area (凸) on the slopes affects land creeping and is continuously pushed in the form of a base, and the all-stone retaining wall functions as a raised end section at the lower part (National Institute of Forest Science, 2017). Land creeping had occurred at the inquiry target region in the relatively high convex topography (凸). It is considered a result of elements affecting land creeping, such as talus at the upper part of the slope (Park et al., 2019).
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 land creeping (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, and the landslide hazard grade, direction, slope, elevation, and geology were analyzed using the digital topographic map of the area (Figs. 9–12) with GIS (Arc map Ver. 10.8).
The Korea Forest Service assigns weights to topographic factors (slope, orientation, length, curvature, topographic wetness index), stock map, diameter class, soil depth, and parent rock and divides them into landslide hazard Grades 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 20.1% for grade 5. 4.3%. Thus, according to the National Institute of Forest Science (2017), it was found that the creeping land area is different from common landslides as most of the landslide hazard grades are 3–5 grades and nongrade 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 land creeping, the land creeping 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 land creeping. These findings are consistent with the conclusion that microtopography during long-term terrain changes considerably affects land creeping (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 cracking and gaps occurred in the target area (Park et al., 2021), and as a result of analyzing the elevation distribution, result has been consistent with the study result (National Institute of Forest Science, 2017) that 200–300 m were the most at 86.1%.
3. Geology of the target area
In the investigation area, there was an approximate gap of 1.2 m caused by the first land creeping that occurred 20–30 years ago (elevation 266 m, N35°1′59″, E127°30′49.80″) due to the recent cutoff of the slopes following the forest road opening; the second land creeping has been progressed, and an additional gap (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 land creeping. Concern exists regarding the occurrence of additional land creeping as the bedrock is floating due to the distribution of convex-shaped talus (凸) in the lower part of the concave area (凹). Park (2018) and the National Institute of Forest Science (2017) have reported that slope cutting in talus areas are vulnerable to land creeping, and as shown in Mari (Park et al., 2005), Sangdong-myeon, Gimhae-si, 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 land creeping.
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 (Takaya, 2017), and the supergroup was the Nangrim stratum, which was vulnerable to the land creeping. The main constituent rocks were metamorphic rocks in which weathering was relatively severe, and thus, land creeping 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 land creeping in the same direction as the slope (Park et al., 2019). 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 land creeping occurs more quickly, which is consistent with the findings of 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 land creeping, 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 land crept area was E70°~77°. The rock formation on the opposite strike was N67°~80°E, which appeared opposite to the land crept area. The lower rock of the groundwater effluent point in the area where the collapsing progressed because the land creeping 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.
4. Soil environmental parameter analysis
4.1 Soil dispersion
Checking the possibility of soil loss between the soil at a depth of 20, 100, 200, and 250 cm from the surface soil and the collapsed weathered rocks is necessary to observe if the clay and groundwater between the bedrock and groundwater are leaching out by groundwater or rainfall infiltration The Crumb Test (ASTM D 6572-06) was conducted to check the soil dispersion by collecting 100 g of soil, such as clay soil and groundwater runoff. The soil’s dispersion indicates the extent to which the aggregates of the soil are decomposed by water, so that the possibility of soil erosion and loss caused by the flow of groundwater can be confirmed. Soils with a high degree of dispersion are very vulnerable to groundwater’s surface and internal erosion (McCook, 1991; Marcus H, 2009).
The soil dispersion degree was measured at depths of 20, 100, 200, and 250 cm from the surface soil and after 1 hour, it was found to be mild at a soil depth of 20 cm but severe at soil depths of 100, 200, and 250 cm. In addition, the clay soil between the bedrock was mild to severe. It was also discovered to be a severe problem in the area in which the groundwater was leaking. The degree of dispersion in this area was similar after 1 h and 2 h, and erosion was discovered to proceed swiftly within 1 h. In particular, it was analyzed that erosion occurred fast when groundwater flowed out, increasing the risk of land creeping in the direction of groundwater outflow.
4.2 Moisture measurement by each soil depth
Soil moisture content (gravitational moisture content, porous volume), maximum water capacity, weight water content, and permeability coefficient have been measured up to a soil thickness of 250 cm in the area collapsed due to land creeping. After collecting the soil thrice in 100 cc cans, soil collection was analyzed using a soil sampler for collapsed areas in the target area. Among the soil moisture content, topsoil exhibited the highest gravitational moisture content at 32.0%. As for the porous volume 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. This was considered to be the result of the condition of the collapsed soil being composed very inhomogeneously. The maximum water 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 of rainfall in a place with a low permeability coefficient, and it was identified that the land creeping occurred while the groundwater was being pushed out at once to a place with a high permeability coefficient (Park, 2018), and as shown in the study results of Woo (1992) and Choi (2018), the soil structure is vulnerable to the land creeping.
Table 1
Analysis by each depth of the soil of the investigation target area
Soil depth (cm)
|
Soil moisture content
|
Maximum water holding capacity (%)
|
Weight moisture content (%)
|
Volume moisture content (%)
|
Coefficient of permeability (cm/sec)
|
Gravitational water content (%)
|
Porous volume (%)
|
Surface soil
|
32.0
|
34.9
|
54.1
|
0.50
|
0.54
|
0.0148
|
20 cm
|
27.2
|
30.3
|
53.0
|
0.50
|
0.53
|
0.0052
|
100 cm
|
26.4
|
29.8
|
51.2
|
0.43
|
0.51
|
0.0035
|
200 cm
|
27.7
|
30.0
|
48.9
|
0.40
|
0.49
|
0.0173
|
250 cm
|
30.6
|
34.1
|
52.9
|
0.48
|
0.53
|
0.0080
|
4.3 Particle size analysis of target area
The soil type of the land crept area was slightly dry brown forest soil (B2), and the soil depth was relatively shallow with <30–60 cm. This 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 land creeping. 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 collapsed due to the land creeping and the control area where land creeping did not occur, the average soil pH was 5.95 (4.90–6.60) as a result of particle size analysis and soil and soil pH analysis at the Korea Forestry Promotion Agency. It was higher than the average pH of forest soil in Korea (Jeong et al., 2002). The content (%) of sand, silt, and clay of the topsoil in the creeping land area were 45.2: 41.0: 13.8, which was determined to be loam, and the natural mountain area was 81.7: 16.1: 2.2, which was determined to be loamy sand, based on particle size analysis. and collapsed soil by landcreep was 52.9: 30.3: 16.9, which was determined to be loamy sand, lastly soil at the groundwater outlet result was 64.2: 23.8: 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 land creeping occurred. And the heterogeneous soils (Tomio et al., 1990; Kim et al., 2015) exhibited soil characteristics vulnerable to the land creeping.
Table 2
Soil particle size analysis in areas where land creeping occurred in natural mountains
Site
|
Grading analysis(%)
|
Soil texture
|
Soil ph
|
sand
|
silt
|
clay
|
Land creep
Surface soil
|
45.2
|
41.0
|
13.8
|
Loam
|
4.9
|
Land creep
Sedentary
Deposit
|
52.9
|
30.3
|
16.9
|
Sandy Loam
|
5.7
|
Landcreep
Groundwater
Elution point
|
64.2
|
23.8
|
12.0
|
Sandy Loam
|
6.6
|
Normal site
(control site)
|
81.7
|
16.1
|
2.2
|
Loamy
Sand
|
6.6
|
4.4 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/cm3) was measured for the areas at risk of land creeping, the boundaries of the areas at risk of land creeping (the boundary between the collapsed areas and the areas at risk of land creeping), the areas collapsed due to land creeping, and the control group in which no land creeping occurred. For the soil depth, 80 cm of the area was at risk of land creeping, 20 cm at the boundary of the land crept area, 40 cm at the collapsed area, and 25 cm in the control area The control area in the areas at risk of land creeping and the boundary area was hardness than nature soil. Because The soil hardness which was assumed to be due to the presence of talus and weathered rocks underground. Furthermore, the landslide collapsed area’s soil hardness quickly dropped below 30 cm of soil depth. This is considered to be 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 land creeping exhibits a relatively less disturbed state, whereas the lower part of the tensile crack caused by land creeping exhibits lower soil hardness due to disturbance and disintegration, and it was discovered that the soil disturbance caused by land creeping affected the soil hardness.
Table 3
Soil hardness (kgf/cm3) in areas at risk of land creeping, the boundary of the land crept area, collapsed areas by land creeping, and control areas (natural mountain areas)
Soil depth (cm)/Soil hardness (kgf/cm3)
|
Landcreep
dangerous
area
|
Landcreep boundary
|
Landcreep area
|
Normal site (control site)
|
0
|
2.2
|
7.2
|
4.0
|
1.5
|
5
|
5.5
|
12.0
|
9.0
|
3.0
|
10
|
10.1
|
2.0
|
11.0
|
4.0
|
15
|
15.0
|
16.1
|
8.5
|
5.5
|
20
|
15.2
|
17.5
|
10.0
|
8.5
|
25
|
15.2
|
-
|
10.5
|
13.5
|
30
|
15.2
|
-
|
10.0
|
-
|
40
|
15.5
|
-
|
2.0
|
-
|
50
|
19.0
|
-
|
-
|
-
|
60
|
21.0
|
-
|
-
|
-
|
70
|
21.0
|
-
|
-
|
-
|
80
|
22.0
|
-
|
-
|
-
|
90
|
-
|
-
|
-
|
-
|
5. Analysis of the flow duration of groundwater
The land creeping is closely related with 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 land creeping and enters into the stream. The groundwater is assumed to be easily flowing in the lower part of the land crept area and the lower part of the tensile cracked, affecting land creeping (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 land creeping; 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, and 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 sideline, 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 land creeping proceeds in the directions of S15°E, S20°E, S25°E, and S30°E, and this result was consistent with the direction along the flow duration of groundwater by the electrical resistivity survey, which was consistent with the results of Lee, et al. (2019), and Park et al. (2019).