• Journal of the Korean Association of Geographic Information Studies
• /
• v.23 no.4
• /
• pp.253-265
• /
• 2020

As the risk of landslide is recently increasing due to the typhoons and localized heavy rains, effective techniques for the landslide damage detection are required to support the establishment of the recovery planning. This study describes the analysis of landslide-damaged areas using ISODATA(Iterative Self-Organizing Data Analysis Technique Algorithm) with Sentinel-2 image, regarding the case of Gokseong in August 7, 2020. A total of 4.75 ha of landslide-damaged areas was detected from the Sentinel-2 image using spectral characteristics of red, NIR(Near Infrared), and SWIR(Shortwave Infrared) bands. We made sure that the satellite remote sensing is an effective method to detect the landslide-damaged areas and support the establishment of the recovery planning, followed by the field surveys that require a lot of manpower and time. Also, this study can be used as a reference for the landslide management for the CAS500-1/2(Compact Advanced Satellite) scheduled to launch in 2021 and the Korean Medium Satellite for Agriculture and Forestry scheduled to launch in 2024.

• Journal of the Korean Society of Hazard Mitigation
• /
• v.10 no.3
• /
• pp.133-145
• /
• 2010

This study analyzed the characteristics of sediment produce by landslide triggered by rainfall. One-dimensional unsaturated groundwater model and infinite slope stability analysis were used to estimate the behavior of soil moisture and slope stability according to rainfall, respectively. Slope stability analysis was performed considering on soil depth and characteristics of trees. As the results considering on recovery of the failed slopes, much amount of sediment was produced in 1963, 1970, and 2002. As the results of verification of simulation results using Landsat 5 TM images, we can find differences of landslide location between the results from model and satellite images. These differences can be caused by uncertainties of the rough parameters in the model. However, in the case that Obong-dam basin was divided into two subbasin, Wangsan-chun and Doma-chun basin, the results of each subbasin show errors around 20%. And only 4% of error occurred in the case of comparing landslide area on the entire Obong-dam basin. These errors seem insignificant considering on the errors which can be caused from the analyses in this study such as estimation of sediment produce, soil cover classification, and estimation of landslide area.

• Journal of the Korean GEO-environmental Society
• /
• v.17 no.6
• /
• pp.13-21
• /
• 2016

Since the landslide hazard areas prediction was analyzed by slope-angle and soil properties, regional characteristics is not taken. Therefore, in order to make more rational prediction, it is necessary to consider the characteristics of the region. Tree roots have been known to increase soil cohesion in landslide hazard areas and to vary the degrees depending on the tree type. In addition, a reasonable prediction of landslide hazard areas can be made by considering crown density based on crown distribution patterns of the area of interest. In this study, using the roots cohesion considering the crown density of the trees, which is in the landslides risk areas around Mt. Gwehwa in Sejong City, the landslides risk areas were predicted and compared with predicted results obtained by not considering root cohesion.

• Journal of the Korean Professional Engineers Association
• /
• v.25 no.4
• /
• pp.102-108
• /
• 1992

This study was carried to find out the soil characteristics of landslide site and to develope landslide prediction method by seismic refraction prospecting. For these aims, landslide condition and travel time were investigated at 68 Landslide sites over the country during 1990 to 1991. The results were as follows. 1. The surface of rupture was included mainly in C layer. Its Hardness was less than 3kg / $\textrm{cm}^2$ at the upper pare of landslide. 2. When the profile line length was 20m, the range of travel time was 40 to 90 msec. The travel time did not differ between bedrocks. 3. Refraction distance ranged from 1 to 7m and mean of that was 2.5m. Travel time was increased according to receiving distance without large variance in the refraction distance but that was appeared large variance out of the refraction distance on slope that has shallow soil depth and discontinuous ground surface. Therefore, the spread distance must be shorten to 10-l5m. 4. The seismic velocity at the first layer(layer of rupture) was less than 500m1sec by degree of weathering and the velocity at the second layer decreased in order of Granite> Granitic gneiss >Sedimentary rock. 5. The first layer observed by seismic refraction was contained C layer that has parent material and weathered rocks of hardness 10-20kg/$\textrm{cm}^2$. 6. Among the range of seismic velocity was less than 200m/sec in 63% of the total plots, 200-300m/sec in 34% and 300-500m /sec in 3%. 7. There was a proportional relationship between seismic prospecting soil depth and executive soil depth, and seismic propection soil depth was about 10 to 20cm deeper than the order.

• Proceedings of the Korean Geotechical Society Conference
• /
• 2006.03a
• /
• pp.12-19
• /
• 2006

Landslide model experiments considering hourly rainfall were performed to investigate and predict the run out distance induced by landslides. The model flume and the rainfall simulator were designed and produced. The model flume was designed in consideration of the landslide characteristics of Korea. The landslides in Korea were mainly occurred in the interface between soil layer and rock layer. The rainfall simulator was produced for controlling hourly rainfall ranged from 100mm/hr to 1,000mm/hr. Jumnunjin standard sand as slope soils was placed on the model flume. The model experiments were performed with changing the hourly rainfall ranged from 150mm/hr to 250mm/hr. In this experiments, the inclination of slope was 25o and the relative density of slope soils was 35%. As a result of experiments, the pore water pressure is rapidly increased at landslide occurring time, and the scale of landslide is increased with increasing in hourly rainfall. The spreading range of run out distance is occurred with pan type, and the spreading width and length are rapidly increased in its early stage and slowly increased after early stage. Also, The increasing velocity of run out distance of debris is influenced by hourly rainfall.

• Journal of Soil and Groundwater Environment
• /
• v.26 no.6
• /
• pp.106-117
• /
• 2021

The purpose of this study is prediction of landslide occurrence reflecting the subsurface flow characteristics within the soil layer in the future due to climate change in a large scale watershed. To do this, we considered the infinite slope stability theory to evaluate the landslide occurrence with predicted soil moisture content by SWAT model based on monitored data (rainfall-soil moisture-discharge). The correlation between the SWAT model and the monitoring data was performed using the coefficient of determination (R²) and the model's efficiency index (Nash and Sutcliffe model efficiency; NSE) and, an accuracy analysis of landslide prediction was performed using auROC (area under Receiver Operating Curve) analysis. In results comparing with the calculated discharge-soil moisture content by SWAT model vs. actual observation data, R² was 0.9 and NSE was 0.91 in discharge and, R² was 0.7 and NSE was 0.79 in soil moisture, respectively. As a result of performing infinite slope stability analysis in the area where landslides occurred in the past based on simulated data (SWAT analysis result of 0.7~0.8), AuROC showed 0.98, indicating that the suggested prediction method was resonable. Based on this, as a result of predicting the characteristics of landslide occurrence by 2050 using climate change scenario (RCP 8.5) data, it was calculated that four landslides could occur with a soil moisture content of more than 75% and rainfall over 250 mm/day during simulation. Although this study needs to be evaluated in various regions because of a case study, it was possible to determine the possibility of prediction through modeling of subsurface flow mechanism, one of the most important attributes in landslide occurrence.

• Journal of Korean Society of Forest Science
• /
• v.113 no.3
• /
• pp.349-360
• /
• 2024

This study aimed to analyze the main forest environmental factors affecting the discernment of slow-moving landslide-prone areas in the Republic of Korea, based on data from a detailed landslide survey conducted from 2019 to 2021. Field survey data from 256 sites were collected covering 29 forest environmental factors in seven categories, including geology, soil, and topography. The analysis was conducted using the Random Forest model (AUC = 0.910) and XGBoost model (Accuracy = 0.808, Kappa = 0.594, F1 - measure = 0.494), which were evaluated as having high classification accuracy during the machine learning model development process. Consequently, factors with a high mean decrease Gini (MDG), representing classification importance, were identified as the presence of cracks (average MDG of both models: 22.1), peak elevation (14.8), and the presence of steps (7.0), indicating that these were significant factors in determining slow-moving landslide-prone areas. The presence of cracks and steps aligned well with the characteristics of slow-moving landslides, suggesting that their importance should be emphasized in future detailed landslide surveys. However, the influence of the peak elevation was considered somewhat overestimated due to the characteristics of the input data used in the analysis. These findings are expected to further improve the accuracy and efficiency of final judgments in detailed landslide surveys.

• The Journal of Engineering Geology
• /
• v.27 no.4
• /
• pp.475-487
• /
• 2017

We performed landslide flume tests to analyze characteristics of landslide occurrence and change in the ground materials due to rainfall infiltration. The test apparatus is composed of flume, rainfall simulator, and measurement sensors and landslides were triggered by heavy rainfall (Intensity=200 mm/hr) sprinkled at the above of an artificial slope. The measurement sensors for matric suction and volumetric water content were installed with 3 sets at shallow (GL-0.2 m), middle (GL-0.4 m), and deep depth (GL-0.6 m) in the slope and the tests were performed with in-situ, loose, and dense condition of each weathered soils of granite, gneiss, and mudstone. The analyses show that surface erosion was dominant in initial time of the test due to heavy rainfall and then landslides occur following locally happened transverse tension cracks. The characteristics of landslide were both shallow failure because of a spread of wetting front induced by the rainfall infiltration and retrogressive failure. While the matric suction was decreased rapidly without any precursor in the soil saturation, the volumetric water content was increased gradually, reached its maximum value, and then decreased rapidly with landslide.

• Proceedings of the Korean Geotechical Society Conference
• /
• 2010.09b
• /
• pp.161-170
• /
• 2010

This paper is results of analyzing causes of damage and rainfall characteristics to investigate characteristics of landslide occurrence and its causes at Inje area in Gangwondo around July in 2006 through collection of related data, literature review and field reconnaissance. From results of analyzing the geometrical characteristics of landslide slope, the slope width of showing the most probable frequency were in the range of 10~50m and the most frequent slope angle was in the range of $30{\sim}40^{\circ}$. The most probable elevation of slope was 200~300m. For the slope direction of landslides, the most frequent directions were NW and SE.

• Journal of Forest and Environmental Science
• /
• v.33 no.2
• /
• pp.130-135
• /
• 2017

In temperate monsoon regions, extensive shallow landslides triggered by heavy rainfall are recurrent phenomena in mountainous areas. 1,357 landslides over Jinbu area, Korea that totaled 127 km2 were identified from aerial photographs and field survey. We examined characteristics of rainfall-induced shallow landslides and casual factors affecting landslide distribution with respect to topographic and forest settings, and land use. Most landslides occurred in the study area were the results of a complex combination of precondition, preparatory factors and triggering factors. Cumulative rainfall and high intensity rainfall during short period of time made the study area very sensitive to landslides and played as catalysts to enable other factors including topographic and forest settings, and land use to act more effectively. In addition, some landslides at lower elevation involved channel incision or bank erosion influenced by land use changes such as deforestation and intensification of agriculture surrounding riparian forests or hillslopes. The results suggest that most of landslide were triggered by heavy rainstorms while topographic, forest settings, and land use affected landslide distribution occurred in the study area.