• Economic and Environmental Geology
• /
• v.52 no.2
• /
• pp.199-212
• /
• 2019

In the machine learning techniques, the sampling strategy of the training data affects a performance of the prediction model such as generalizing ability as well as prediction accuracy. Especially, in landslide susceptibility analysis, the data sampling procedure is the essential step for setting the training data because the number of non-landslide points is much bigger than the number of landslide points. However, the previous researches did not consider the various sampling methods for the training data. That is, the previous studies selected the training data randomly. Therefore, in this study the authors proposed several different sampling methods and assessed the effect of the sampling strategies of the training data in landslide susceptibility analysis. For that, total six different scenarios were set up based on the sampling strategies of landslide points and non-landslide points. Then Random Forest technique was trained on the basis of six different scenarios and the attribute importance for each input variable was evaluated. Subsequently, the landslide susceptibility maps were produced using the input variables and their attribute importances. In the analysis results, the AUC values of the landslide susceptibility maps, obtained from six different sampling strategies, showed high prediction rates, ranges from 70 % to 80 %. It means that the Random Forest technique shows appropriate predictive performance and the attribute importance for the input variables obtained from Random Forest can be used as the weight of landslide conditioning factors in the susceptibility analysis. In addition, the analysis results obtained using specific sampling strategies for training data show higher prediction accuracy than the analysis results using the previous random sampling method.

• Journal of The Geomorphological Association of Korea
• /
• v.27 no.1
• /
• pp.61-89
• /
• 2020

The objective of this study is to characterize landslide susceptibility depending on various geo-environmental variables as well as to compare the Frequency Ratio (FR) and Evidential Belief Function (EBF) methods for landslide susceptibility analysis of rainfall-induced landslides. In 2013, a total of 259 landslides occurred in Chuncheon, Gangwon Province, South Korea, due to heavy rainfall events with a total cumulative rainfall of 296~721mm in 106~231 hours duration. Landslides data were mapped with better accuracy using the geographic information system (ArcGIS 10.6 version) based on the historic landslide records in Chuncheon from the National Disaster Management System (NDMS), the 2013 landslide investigation report, orthographic images, and aerial photographs. Then the landslides were randomly split into a testing dataset (70%; 181 landslides) and validation dataset (30%; 78 landslides). First, geo-environmental variables were analyzed by using FR and EBF functions for the full data. The most significant factors related to landslides were altitude (100~200m), slope (15~25°), concave plan curvature, high SPI, young timber age, loose timber density, small timber diameter, artificial forests, coniferous forests, soil depth (50~100cm), very well-drained area, sandy loam soil and so on. Second, the landslide susceptibility index was calculated by using selected geo-environmental variables. The model fit and prediction performance were evaluated using the Receiver Operating Characteristic (ROC) curve and the Area Under Curve (AUC) methods. The AUC values of both model fit and prediction performance were 80.5% and 76.3% for FR and 76.6% and 74.9% for EBF respectively. However, the landslide susceptibility index, with classes of 'very high' and 'high', was detected by 73.1% of landslides in the EBF model rather than the FR model (66.7%). Therefore, the EBF can be a promising method for spatial prediction of landslide occurrence, while the FR is still a powerful method for the landslide susceptibility mapping.

• Proceedings of the KSRS Conference
• /
• 2004.10a
• /
• pp.314-317
• /
• 2004

The aim of this study was to evaluate the susceptibility from landslides in the Lai Chau region of Vietnam, using Geographic Information System (GIS) and remote sensing data, focusing on the relationship between tectonic fractures and landslides. Landslide locations were identified from an interpretation of aerial photographs and field surveys. Topographic and geological data and satellite images were collected, processed, and constructed into a spatial database using GIS data and image processing techniques, and a scheme of the tectonic fracturing of the crust in the Lai Chau region was established. In this scheme, Lai Chau was identified as a region with low crustal fractures, with the grade of tectonic fracture having a close relationship with landslide occurrence. The factors found to influence landslide occurrence were: topographic slope, topographic aspect, topographic curvature, distance from drainage, lithology, distance from a tectonic fracture and land cover. Landslide prone areas were analyzed and mapped using the landslide occurrence factors employing the probability-likelihood ratio method. The results of the analysis were verified using landslide location data, and these showed a satisfactory agreement between the hazard map and existing landslide location data.

• Journal of The Geomorphological Association of Korea
• /
• v.24 no.2
• /
• pp.91-101
• /
• 2017

Quantitative forecasting methods based on spatial data and geographic information system have been used in predicting the landslide location. This study compared the simulated results of logistic, Bayesian, and maximum entropy models to understand the uncertainties of each model and identify the main factors that influence landslide. The study area is Boeun gun where 388 landslides occurred in the year of 1998. The verification results showed that the AUC of the three models was 0.84. However, the landslide susceptibility distribution of Maxent model was different from those of the other two models. With the same landslide occurrence data, the result of high susceptible area in Maxent model is smaller than Logistic or Bayesian. Maxent model, however, proved to be more efficient in predicting landslide than the other two models. In Maxent's simulations, the responsible factors for landslide susceptibility are timber age class, land cover, timber diameter, crown closure, and soil drainage. The results suggest that it is necessary to consider the possibility of overestimation when using Logistic or Bayesian model, and forest management around the study area can be an effective way to minimize landslide possibility.

• Journal of Forest and Environmental Science
• /
• v.39 no.4
• /
• pp.213-224
• /
• 2023

As abnormal weather events due to climate change continue to rise, landslide damage is also increasing. Given the substantial time and financial resources required for post-landslide recovery, it becomes imperative to formulate a proactive response plan. In this regard, landslide susceptibility analysis has emerged as a valuable tool for establishing preemptive measures against landslides. Accordingly, this study conducted an annual landslide susceptibility analysis using the history of landslides that occurred over many years in the Jeolla region, and analyzed areas with a high potential for landslides in the Jeolla region. The analysis employed an ensemble model that amalgamated 10 data-based models, aiming to mitigate uncertainties associated with a single-model approach. Furthermore, based on the cumulative data regarding landslide susceptible areas, this research identified regions vulnerable to recurring landslide damage in Jeolla region and proposed specific strategies for utilizing this information at various levels, including local government initiatives, adaptation plan development, and development approval processes. In particular, this study outlined approaches for local government utilization, the determination of adaptation plan types, and considerations for development permits. It is anticipated that this research will serve as a valuable opportunity to underscore the significance of information concerning regions vulnerable to recurring landslide damage.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.34 no.4
• /
• pp.443-451
• /
• 2016

Globally, landslides triggered by natural or human activities have resulted in enormous damage to both property and life. Recent climatic changes and anthropogenic activities have increased the number of occurrence of these disasters. Despite many researches, there is no standard method that can produce reliable prediction. This article discusses the process of landslide susceptibility mapping using various methods in current literatures and applies the FR (Frequency Ratio) method to develop a susceptibility map for the 2015 earthquake region of Sindhupalchowk, Nepal. The complete mapping process describes importance of selection of area, and controlling factors, widespread techniques of modelling and accuracy assessment tools. The FR derived for various controlling factors available were calculated using pre- and post- earthquake landslide events in the study area and the ratio was used to develop susceptibility map. Understanding the process could help in better future application process and producing better accuracy results. And the resulting map is valuable for the local general and authorities for prevention and decision making tasks for landslide disasters.

• Geomechanics and Engineering
• /
• v.37 no.2
• /
• pp.135-147
• /
• 2024

A GIS-based landslide susceptibility mapping (LSM) was carried out using frequency ratio (FR), modified frequency ratio (M-FR), analytic hierarchy process (AHP), and modified analytic hierarchy process (M-AHP) methods to identify and delineate the potential failure zones along the Halong - Vandon expressway. The thematic layers of various landslide causative factors were generated for modeling in GIS, including geology, rainfall, distance to fault, distance to road, slope, aspect, landuse, density of landslide, vertical relief, and horizontal relief. In addition, a landslide inventory along the road network was prepared using data provided by the management department during the course of construction and operation from 2017 to 2019, when many landslides were documented. The validation results showed that the M-FR method had the highest AUC value (AUC = 0.971), which was followed by the FR method with AUC = 0.961. The AUC values were 0.939 and 0.892 for the M-AHP and AHP methods, respectively. The generated LSM obtained from M-FR method classified the study area into five susceptibility classes: very low (0), low (0-1), moderate (1-2), high (2-3), and very high (3-4) classes, which could be useful for various stakeholders like planners, engineers, designers, and local public for future construction and maintenance in the study area.

• Journal of Environmental Impact Assessment
• /
• v.15 no.6
• /
• pp.373-383
• /
• 2006

The spatial mapping of risk is very useful data in planning for disaster preparedness. This research presents a methodology for making the landslide life risk map in the Boeun area which had considerable landslide damage following heavy rain in August, 1998. We have developed a three-stage procedure in spatial data analysis not only to estimate the probability of the occurrence of the natural hazardous events but also to evaluate the uncertainty of the estimators of that probability. The three-stage procedure consists of: (i)construction of a hazard prediction map of "future" hazardous events; (ii) validation of prediction results and estimation of the probability of occurrence for each predicted hazard level; and (iii) generation of risk maps with the introduction of human life factors representing assumed or established vulnerability levels by combining the prediction map in the first stage and the estimated probabilities in the second stage with human life data. The significance of the landslide susceptibility map was evaluated by computing a prediction rate curve. It is used that the Bayesian prediction model and the case study results (the landslide susceptibility map and prediction rate curve) can be prepared for prevention of future landslide life risk map. Data from the Bayesian model-based landslide susceptibility map and prediction ratio curves were used together with human rife data to draft future landslide life risk maps. Results reveal that individual pixels had low risks, but the total risk death toll was estimated at 3.14 people. In particular, the dangerous areas involving an estimated 1/100 people were shown to have the highest risk among all research-target areas. Three people were killed in this area when landslides occurred in 1998. Thus, this risk map can deliver factual damage situation prediction to policy decision-makers, and subsequently can be used as useful data in preventing disasters. In particular, drafting of maps on landslide risk in various steps will enable one to forecast the occurrence of disasters.

• The Journal of Engineering Geology
• /
• v.33 no.4
• /
• pp.673-687
• /
• 2023

Exploratory data analysis is the process of observing and understanding data collected from various sources to identify their distributions and correlations through their structures and characterization. This process can be used to identify correlations among conditioning factors and select the most effective factors for analysis. This can help the assessment of landslide susceptibility, because landslides are usually triggered by multiple factors, and the impacts of these factors vary by region. This study compared two stages of exploratory data analysis to examine the impact of the data exploration procedure on the landslide prediction model's performance with respect to factor selection. Deep-learning-based landslide susceptibility analysis used either a combinations of selected factors or all 23 factors. During the data exploration phase, we used a Pearson correlation coefficient heat map and a histogram of random forest feature importance. We then assessed the accuracy of our deep-learning-based analysis of landslide susceptibility using a confusion matrix. Finally, a landslide susceptibility map was generated using the landslide susceptibility index derived from the proposed analysis. The analysis revealed that using all 23 factors resulted in low accuracy (55.90%), but using the 13 factors selected in one step of exploration improved the accuracy to 81.25%. This was further improved to 92.80% using only the nine conditioning factors selected during both steps of the data exploration. Therefore, exploratory data analysis selected the conditioning factors most suitable for landslide susceptibility analysis and thereby improving the performance of the analysis.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.8 no.1
• /
• pp.75-87
• /
• 2005

This study aims at predicting and mapping of the landslide susceptibility in the Geumho river watershed using GIS and Remote Sensing techniques. We constructed the spatial database of affecting factors such as slope angle, slope aspect, lithology, landuse, and vegetation index (NDVI) at a $30m{\times}30m$ resolution. The landslide susceptibility of the study area was predicted through overlay analysis and adding up estimation matrix, and the predicted map of landslide susceptibility with six categories (stable, very low, low, moderate, high, very high) was constructed. As the results, it showed that the very high susceptibility zones made up approximately 0.3% of the total study area, and these zones were mainly distributed in the forest area with the high slope angle and low vegetation index.