Landslide Conditioning Factors Research Articles

GIS-based landslide susceptibility modeling: A comparison between fuzzy multi-criteria and machine learning algorithms

Hazards and disasters have always negative impacts on the way of life. Landslide is an overwhelming natural as well as man-made disaster that causes loss of natural resources and human properties throughout the world. The present study aimed to assess and compare the prediction efficiency of different models in landslide susceptibility in the Kysuca river basin, Slovakia. In this regard, the fuzzy decision-making trial and evaluation laboratory combining with the analytic network process (FDEMATEL-ANP), Naïve Bayes (NB) classifier, and random forest (RF) classifier were considered. Initially, a landslide inventory map was produced with 2000 landslide and non-landslide points by randomly divided with a ratio of 70%:30% for training and testing, respectively. The geospatial database for assessing the landslide susceptibility was generated with the help of 16 landslide conditioning factors by allowing for topographical, hydrological, lithological, and land cover factors. The ReliefF method was considered for determining the significance of selected conditioning factors and inclusion in the model building. Consequently, the landslide susceptibility maps (LSMs) were generated using the FDEMATEL-ANP, Naïve Bayes (NB) classifier, and random forest (RF) classifier models. Finally, the area under curve (AUC) and different arithmetic evaluation were used for validating and comparing the results and models. The results revealed that random forest (RF) classifier is a promising and optimum model for landslide susceptibility in the study area with a very high value of area under curve (AUC = 0.954), lower value of mean absolute error (MAE = 0.1238) and root mean square error (RMSE = 0.2555), and higher value of Kappa index (K = 0.8435) and overall accuracy (OAC = 92.2%).

Landslides susceptibility assessment by analytical network process: a case study for Kuttiyadi river basin (Western Ghats, southern India)

In the case of Western Ghats of Kerala located in southern India, landslides become the most devastating hazard during the monsoon season is mainly because of intense rainfall. The recent landslides struck in the months of June and August 2018 were the worst in the history of the state as the damages to life and property happened were very dangerous. From this evidences, being a landslide-prone area a hazard zonation attempted using the terrain susceptibility concept for highlands of Kuttiyadi river basin is a part of Western Ghats in Kozhikode district with a maximum elevation of 1640 m above MSL. Through landslide susceptibility mapping, the assessment of areas likely to have or probable to landslides in the future identified using geospatial and multi-criteria decision-making (MCDM) tool mix. This work explores the usage of the MCDM tool—analytical network process (ANP)-based decision-making approach to determine the relationship between each landslide conditioning factors and to assign weight effectively. Factor maps of various landslide triggering parameters such as slope, soil, landuse, drainage density, stage of landform, lineament density, geology, and geomorphology were prepared and integrated through geospatial analyses in the GIS platform, and the study area (318 km2) is categorized into four classes of unstable (121 km2), moderately unstable (76 km2), stable to moderately stable (124 km2), high vulnerability (67 km2), and very high vulnerability (74 km2). The result of this work concluded that the application of ANP is an effective tool to weight the factors causing landslides and to generate landslide susceptibility zonation map with maximum accuracy using geospatial technology. The validation of output data with landslide inventory data is satisfying it.

Landslide susceptibility modelling using integrated evidential belief function based logistic regression method: A study from Southern Western Ghats, India

The present study aims to develop a spatially integrated evidential belief function based logistic regression model (EBF-LR) for landslide susceptibility mapping in a naturally sloping terrain of Southern Western Ghats in Kerala, India. For this, a landslide inventory map of 83 previous landslides was prepared using satellite imageries and verified in field check. Thereafter the landslide inventory was randomly divided into 70%–30% basis for model training and testing. Twelve landslide conditioning factors viz., lithology, land use/land cover, NDVI, slope angle, slope aspect, profile curvature, distance to stream, distance to roads, distance to lineaments, soil texture, topographic wetness index and average annual rainfall were considered for landslide susceptibility modelling. The resultant susceptibility maps were validated using receiver operating characteristics curve (ROC) with area under the curve (AUC) value, sensitivity, specificity, kappa index, mean absolute error (MAE) and root mean square error (RMSE). Analysis shows that the integrated EBF-LR model outweigh other conventional bivariate and multivariate approaches with a ROC-AUC value of 0.935, Kappa score 0.719, sensitivity 0.885, specificity 0.833 and least RMSE of 0.456 in the validation stage. The study also reveals that anthropogenic disturbances have a significant role on landslide initiation in the study area. Considering the ROC- AUC, MAE, RMSE and other validation measures, the accuracy of the proposed landslide susceptibility model is satisfactory and can be used for future land use planning and landslide mitigation in the study area.

Random Forest-Based Landslide Susceptibility Mapping in Coastal Regions of Artvin, Turkey

Natural disasters such as landslides often occur in the Eastern Black Sea region of Turkey owing to its geological, topographical, and climatic characteristics. Landslide events occur nearly every year in the Arhavi, Hopa, and Kemalpaşa districts located on the Black Sea coast in the Artvin province. In this study, the landslide susceptibility map of the Arhavi, Hopa, and Kemalpaşa districts was produced using the random forest (RF) model, which is widely used in the literature and yields more accurate results compared with other machine learning techniques. A total of 10 landslide-conditioning factors were considered for the susceptibility analysis, i.e., lithology, land cover, slope, aspect, elevation, curvature, topographic wetness index, and distances from faults, drainage networks, and roads. Furthermore, 70% of the landslides on the landslide inventory map were used for training, and the remaining 30% were used for validation. The RF-based model was validated using the area under the receiver operating characteristic (ROC) curve. Evaluation results indicated that the success and prediction rates of the model were 98.3% and 97.7%, respectively. Moreover, it was determined that incorrect land-use decisions, such as transforming forest areas into tea and hazelnut cultivation areas, induce the occurrence of landslides.

Investigating the Effect of Cross-Modeling in Landslide Susceptibility Mapping

To mitigate the negative effects of landslide occurrence, there is a need for effective landslide susceptibility mapping (LSM). The fundamental source for LSM is landslide inventory. Unfortunately, there are still areas where landslide inventories are not generated due to financial or reachability constraints. Considering this led to the following research question: can we model landslide susceptibility in an area for which landslide inventory is not available but where such is available for surrounding areas? To answer this question, we performed cross-modeling by using various strategies for landslide susceptibility. Namely, landslide susceptibility was cross-modeled by using two adjacent regions (“Łososina” and “Gródek”) separated by the Rożnów Lake and Dunajec River. Thus, 46% and 54% of the total detected landslides were used for the LSM in “Łososina” and “Gródek” model, respectively. Various topographical, geological, hydrological and environmental landslide-conditioning factors (LCFs) were created. These LCFs were generated on the basis of the Digital Elevation Model (DEM), Sentinel-2A data, a digitized geological and soil suitability map, precipitation, the road network and the Różnów lake shapefile. For LSM, we applied the Frequency Ratio (FR) and Landslide Susceptibility Index (LSI) methods. Five zones showing various landslide susceptibilities were generated via Natural Jenks. The Seed Cell Area Index (SCAI) and Relative Landslide Density Index were used for model validation. Even when the SCAI indicated extremely high values for “very low” susceptibility classes and very small values for “very high” susceptibility classes in the training and validation areas, the accuracy of the LSM in the validation areas was significantly lower. In the “Łososina” model, 90% and 57% of the landslides fell into the “high” and “very high” susceptibility zones in the training and validation areas, respectively. In the “Gródek” model, 86% and 46% of the landslides fell into the “high” and “very high” susceptibility zones in the training and validation areas, respectively. Moreover, the comparison between these two models was performed. Discrepancies between these two models exist in the areas of critical geological structures (thrust and fault proximity), and the reliability for such susceptibility zones can be low (2–3 susceptibility zone difference). However, such areas cover only 11% of the analyzed area; thus, we can conclude that in remaining regions (89%), LSM generated by the inventory for the surrounding area can be useful. Therefore, the low reliability of such a map in areas of critical geological structures should be borne in mind.

GIS-based landslide susceptibility assessment using optimized hybrid machine learning methods

Globally, but especially in China, landslides are considered to be one of the most severe and significant natural hazards. In this study, bivariate statistical-based kernel logistic regression (KLR) models with different kernel functions (Polynomial, PUK, and Radial Basis Function), named the PLKLR, PUKLR, and RBFKLR models, were proposed for landslide susceptibility evaluation in Zichang City, China. Meanwhile, the present study aims to build landslide susceptibility maps based on bivariate statistical correlation analysis, optimization of different kernel functions, comparison of three landslide susceptibility maps and systematic analysis of spatial patterns. The steps of this article are organized as follows: Firstly, a landslide inventory containing 263 historical landslide locations was constructed. For the purpose of training and validation of models, 263 landslide locations were randomly divided into two parts with a ratio of 70/30. Secondly, 14 landslide conditioning factors were extracted from the spatial database. Subsequently, correlation analysis between the conditioning factors and the occurrence of landslides was conducted using frequency ratios. Then, the conditioning factors with normalized frequency ratios values were used as inputs to build the landslide susceptibility maps using the three models. A multicollinearity analysis was performed using collinearity statistics. Finally, the area under the receiver operating characteristic curve (AUC) was used for comparison and validation of models for recognizing the prediction capability. By further quantitative comparing mapped susceptibility values on a pixel-by-pixel basis, which can acquire underestimations and overestimations of factors (distance to river and slope) and susceptibility area. The results indicated that the PUKLR model had superior performance in landslide susceptibility assessment, with the highest AUC values of 0.884 and 0.766 for training and validation datasets, respectively. This model was followed by the RBFKLR model and the PLKLR model for the training datasets (AUC values of 0.879 and 0.797, respectively), and the PLKLR model and the RBFKLR model for the validation datasets (AUC values of 0.758 and 0.752, respectively). The landslide susceptibility map could help government agencies and decision-makers make wise decisions for future natural hazards prevention in Zichang region.

Susceptibility mapping of damming landslide based on slope unit using frequency ratio model

In the mountainous area of southwest China, there is a large number of large hydropower stations that have already been built or under construction. The construction and the safety of the operation of hydropower stations directly affect the residents’ safety of life and property on both sides of the downstream river. Landslide hazards are one of the main problems affecting the construction and operation of hydropower stations. Therefore, it is currently more urgent to evaluate landslide susceptibility within the basin. The goal of this paper is to carry out landslide susceptibility mapping that affects the construction and operation of hydropower stations. In the case of a large research area and lack of landslide hazard data, this paper selects the sample of blocking river landslide in a larger area with similar topography and geological conditions and applies the quantitative method to carry out landslide sensitivity mapping. In this paper, the slope unit works as the evaluation unit, 191 landslide sample data were obtained, and nine landslide conditioning factors (slope, aspect, curvature, lithology, distance to fault, distance to the river, land use, soil, and the topographic wetness index (TWI)) were obtained. Combined with the quantitative frequency model (FR model), the landslide susceptibility map was obtained. The research process and results show that (1) when the study area is large, the landslide data is scarce, and it is difficult to get the landslide inventory map; the landslide samples can be selected in a larger area with similar topography and geological conditions. It is feasible to use the analysis data of landslide samples to evaluate the landslide susceptibility; (2) for specific industry applications such as hydropower station operation safety, the landslide evaluation range can be reduced by the spatial correlation between landslide aspect and river flow direction, which is beneficial to the management department to reduce the scope of attention; and (3) the verification results show that the slope unit can be used for quantitative evaluation of landslide sensitivity based on the FR model and can obtain the highest evaluation accuracy. The main advantage of this method is its capability to carry out landslide susceptibility mapping with quantitative methods and to ensure that higher mapping accuracy is obtained in areas where landslide hazard data is lacking.

Landslide susceptibility mapping and rock slope stability assessment using frequency ratio and kinematic analysis in the mountains of Mgulat area, Northern Ethiopia

Landslides, mainly triggered by rainfall, are one of the most common natural hazards in the northern and central highlands of Ethiopia, resulting in the loss of human life and property. As part of the Ethiopian highlands, Mgulat Mountain (88.6 km2) is a typical example for the occurrence of landslides and associated damages. The main aim of this research was to generate landslide susceptibility map (LSM) for the first time by considering various causative factors typical for the Mgulat Mountain area. The methods used include frequency ratio (FR) and kinematic analysis (KA) supported by GIS tools. The research involved conventional fieldwork (e.g., geological mapping, landslides inventory) and analysis and interpretation of images. From the geological mapping, major rock units were identified: quaternary deposit, basalt, red sandstone, tillite, and white sandstone. A total of 220 landslides (15.45 km2) were identified in the study area from field survey, Google Earth and aerial photo interpretation. Five causative factors (lithology, slope steepness, distance from drainage, land use-land cover (LULC), and slope aspect) were selected based on their contribution to landslide initiation as the landslide conditioning factors for generation of the LSM. From the five causative factors, lithology and slope steepness were found to be influential as landslide conditioning factors. Specifically, lithology (red sandstone and tillite), distance to stream (0 to 300 m), LULC (bare land and settlement), and slope steepness (> 25°) have FR > 1, indicating a high relationship with landslide occurrence. Based on the generated LSM, five susceptibility classes or zones were identified (very low, low, moderate, high, and very high susceptibility). The area of each class as percentage of total area was found to be 19.53, 24.15, 22.57, 18.74, and 15.01%, for the very low to the very high class, respectively. The results of KA show potentially and marginally planar, topple, and wedge failures. The generated LSM can serve as an important tool or guide for further land-use planning and to prevent and mitigate landslide-related risks. Consequently, to consider cost-effective and technically sound mitigation measures, detailed and site-specific multidisciplinary investigations are recommended.

Comparison of Support Vector Machine, Bayesian Logistic Regression, and Alternating Decision Tree Algorithms for Shallow Landslide Susceptibility Mapping along a Mountainous Road in the West of Iran

This paper aims to apply and compare the performance of the three machine learning algorithms–support vector machine (SVM), bayesian logistic regression (BLR), and alternating decision tree (ADTree)–to map landslide susceptibility along the mountainous road of the Salavat Abad saddle, Kurdistan province, Iran. We identified 66 shallow landslide locations, based on field surveys, by recording the locations of the landslides by a global position System (GPS), Google Earth imagery and black-and-white aerial photographs (scale 1: 20,000) and 19 landslide conditioning factors, then tested these factors using the information gain ratio (IGR) technique. We checked the validity of the models using statistical metrics, including sensitivity, specificity, accuracy, kappa, root mean square error (RMSE), and area under the receiver operating characteristic curve (AUC). We found that, although all three machine learning algorithms yielded excellent performance, the SVM algorithm (AUC = 0.984) slightly outperformed the BLR (AUC = 0.980), and ADTree (AUC = 0.977) algorithms. We observed that not only all three algorithms are useful and effective tools for identifying shallow landslide-prone areas but also the BLR algorithm can be used such as the SVM algorithm as a soft computing benchmark algorithm to check the performance of the models in future.

Comparative landslide susceptibility assessment using statistical information value and index of entropy model in Bhanupali-Beri region, Himachal Pradesh, India

Landslide is a complex natural hazard that sometimes causes disaster resulting in loss of life, assets and infrastructure, especially in the Himalayas. Recent studies suggest that for effective mitigation and resilience through proper planning and policymaking, it is equally important to justify and select a suitable scientific technique that most appropriately addresses the salient causes of a landslide in any area. The principal objective of this study is to carry out a comparative assessment between two contemporary statistical techniques, i.e., the statistical information value (SIV) and index of entropy (IOE), to find out the effectiveness of the two said methods in landslide susceptibility mapping in Bhanupali-Beri region. During the analysis, the higher-resolution satellite images, i.e., World view-2 image of 2017 and Landsat-8 OLI image of 2018, have been used for delineation of various triggering parameters used for landslide susceptibility. The contemporary GIS technique integrated with the remote sensing applications was distinct in preparing the prominent landslide conditioning factor layers such as slope, slope aspect, thrust and fault proximity, geomorphology, landuse–landcover, stream power index, topographic wetness index, geology, roads proximity, lineament density and past landslide inventory. The final assessment was performed using GIS software through raster re-sampling, and the values derived for each conditioning factors were combined using defined SIV and IOE equations. The study area was categorized into five distinct landslide susceptible zones (very low, low, moderate, high and very high) using the Jenk’s Natural Breaks algorithm. Index of entropy model has given better results compared to SIV. The utmost vital factors triggering landslide (estimated for entropy values) in the area are landuse–landcover with barren land and sparse vegetation followed by TWI, lineament density, geomorphology, and slope.

Seismic landslide susceptibility assessment based on ADASYN-LDA model

Seismic landslide susceptibility (SLS) assessment can be used to estimate the susceptibility of landslides induced by an earthquake, which has great significance for emergency measure making and land use planning to mitigate the landslide hazard and risk. At present, one of the key problems in SLS evaluation is that the sample is extremely unbalanced, that is, the number of landslide units is usually far less than that of non-landslide units, which easily leads to poor prediction performance of SLS model. In addition, when using high-resolution mapping units, the sample size greatly increases, and thus the training time cost is considerable especially for complicated nonlinear artificial intelligence such as the Support Vector Machine and the Neural Networks. In view of this, this paper proposes an SLS model (called ADASYN-LDA) combining ADASYN sample method and linear discriminant analysis (LDA), and evaluates it using 2014 Ludian earthquake event. The results show that the two ADASYN-LDA models are highly efficient in training, and have good prediction performance with the high AUC of 0.8737 and 0.8535 respectively, and with the balanced accuracy of 0.7950 and 0.6174 respectively. Compared with the traditional LDA model without ADASYN method, ADASYN-LDA model improves the results of the SLS assessment affected by the sample imbalance. The further evaluation is needed in other earthquake events, and the further improvement can be made with more landslide conditioning factors and seismic landslide data set.

Landslide susceptibility mapping along the upper Jinsha River, south-western China: a comparison of hydrological and curvature watershed methods for slope unit classification

The mapping unit is the smallest indivisible unit of landslide susceptibility mapping. The mapping unit considerably affects the effect of landslide susceptibility mapping. Among many mapping units, the slope unit is one of the most suitable units for landslide susceptibility mapping. The methods available for classifying slope unit are not uniform, and the dividing effects differ. In this study, the hydrological method and curvature watershed method were used to classify the slope units in the study area on the upper Jinsha River, south-western China. Lithology, slope angle, slope aspect, normalized difference vegetation index (NDVI), land cover, rainfall, curvature, distance-to-river, distance-to-fault, were selected as the landslide conditioning factors. Support vector machine were applied to the landslide susceptibility modeling. Statistical indexes, 5-flod cross validation, Kappa coefficient, and AUC values were introduced to validate the prediction accuracy of the landslide susceptibility model. By comparing the shape characteristics of slope units classified by the two methods, we found that the slope unit classified based on curvature watershed method has a uniform size, shape between circle and equilateral triangle, and small internal terrain variation. As for the mean prediction accuracy, Kappa coefficient and AUC values of landslide susceptibility model, in the training stage, were 81.62%, 63.23%, and 89.72%, respectively, for the hydrological method, while 84.26%, 68.51%, and 90.80%, respectively, for the curvature watershed method; in the testing stage, were 80.70%, 61.40%, and 88.08%, respectively, for the hydrological method, while 83.24%, 66.49%, and 88.96%, respectively, for the curvature watershed method, which means that the curvature watershed method is more effective to produce landslide susceptibility map of the study area.

Evaluating the Performance of Individual and Novel Ensemble of Machine Learning and Statistical Models for Landslide Susceptibility Assessment at Rudraprayag District of Garhwal Himalaya

Landslides are known as the world’s most dangerous threat in mountainous regions and pose a critical obstacle for both economic and infrastructural progress. It is, therefore, quite relevant to discuss the pattern of spatial incidence of this phenomenon. The current research manifests a set of individual and ensemble of machine learning and probabilistic approaches like an artificial neural network (ANN), support vector machine (SVM), random forest (RF), logistic regression (LR), and their ensembles such as ANN-RF, ANN-SVM, SVM-RF, SVM-LR, LR-RF, LR-ANN, ANN-LR-RF, ANN-RF-SVM, ANN-SVM-LR, RF-SVM-LR, and ANN-RF-SVM-LR for mapping landslide susceptibility in Rudraprayag district of Garhwal Himalaya, India. A landslide inventory map along with sixteen landslide conditioning factors (LCFs) was used. Randomly partitioned sets of 70%:30% were used to ascertain the goodness of fit and predictive ability of the models. The contribution of LCFs was analyzed using the RF model. The altitude and drainage density were found to be the responsible factors in causing the landslide in the study area according to the RF model. The robustness of models was assessed through three threshold dependent measures, i.e., receiver operating characteristic (ROC), precision and accuracy, and two threshold independent measures, i.e., mean-absolute-error (MAE) and root-mean-square-error (RMSE). Finally, using the compound factor (CF) method, the models were prioritized based on the results of the validation methods to choose best model. Results show that ANN-RF-LR indicated a realistic finding, concentrating only on 17.74% of the study area as highly susceptible to landslide. The ANN-RF-LR ensemble demonstrated the highest goodness of fit and predictive capacity with respective values of 87.83% (area under the success rate curve) and 93.98% (area under prediction rate curve), and the highest robustness correspondingly. These attempts will play a significant role in ensemble modeling, in building reliable and comprehensive models. The proposed ANN-RF-LR ensemble model may be used in the other geographic areas having similar geo-environmental conditions. It may also be used in other types of geo-hazard modeling.

Application of statistical probabilistic methods in landslide susceptibility assessment in Kurseong and its surrounding area of Darjeeling Himalayan, India: RS-GIS approach

Kurseong and its surrounding areas are frequently affected by the landslide which causes huge loss of properties and lives. The landslide susceptibility maps can play an important role in human development and sustainable environment management in Darjeeling Himalayan. The present research focused on preparing the effective landslide susceptible models using different statistical probabilistic methods, namely landslide nominal susceptibility factor (LNSF), information value (InfoVal) and certainty factor (CF) models. Experiments have been carried out in Kurseong region, a part of Darjeeling Himalaya as a field of this research. Landslide sites were identified from previous records through extensive field survey and Google pro satellite imagery. Totally, 273 landslide sites were compiled and prepared the inventory map. Out of 273 landslides, 70% were used for training and 30% were used for validating the models. Seventeen landslide conditioning factors were selected for modeling landslide susceptibility, i.e., elevation, aspect, slope degree, rainfall, geological structure, geomorphologic division, lineament, land use/land cover, distance to roads, earthquake zone, soil texture, soil depth, normalized difference vegetation index, drainage density, stream power index and topographic wetness index. The produced susceptibility maps were validated using the receiver???s operating characteristic (ROC) curves. The prediction accuracy of LNSF, InfoVal and CF models as per ROC are 80.78%, 82.91% and 86.13%, respectively. The CF achieved the highest accuracy (86.13%), while the LNSF produced the lowest ROC value (80.78%). However, the comparison of the produced landslide maps revealed that all the applied models have good precision for studying susceptibility of landslide in Kurseong and its surrounding of Darjeeling Himalaya, India. The findings of current study can be supportive for the mitigation of landslide risk in the Kurseong range as well as the surrounding comparable areas having same geoenvironmental conditions.

Comparing the efficiency of weight of evidence, support vector machine and their ensemble approaches in landslide susceptibility modelling: A study on Kurseong region of Darjeeling Himalaya, India

The main objectives of the current study is to generate and compare the landslide susceptibility maps (LSMs) using a probabilistic model i.e. weight of evidence (WoE), a machine learning technique i.e. support vector machine (SVM) and ensemble approach i.e. WoE-SVM. Experiments were conducted in the Kurseong region, a part of Darjeeling Himalaya as a study area. Previous landslide sites were identified through field survey and previous records. Total 273 landslide sites were compiled and considered for the landslide inventory map to calibrate as well as validate the models. Seventeen landslide conditioning factors (LCFs) were used for landslide susceptibility modelling, i.e. slope degree, elevation, aspect, curvature, rainfall, geological structure, lineament, distance to main roads, soil texture, soil depth, earthquake zone, land-use/land-cover, drainage density, stream power index, sediment transport index, and topographic wetness index. The generated landslide susceptibility maps were validated applying the receiver operating characteristic curves (ROC), Positive Predictive Value (PPV), Proportion Correctly Classified (PCC) and the seed cell area index (SCAI). The results revealed that prediction rates are 89.47%, 91.81%, and 94.45% with standard error 0.0359, 0.0289 and 0.0183 respectively for WoE, SVM and ensemble WoE-SVM models. The study shows that the fellow land, plantation area, roads side areas and denudation structural hills having elevation more than 1500 m and 24° to 69° slopes are extremely susceptible to landslides. The comparison of the produced landslide maps revealed that all the applicable models have committed accuracy for studying sensitivity in Kurseong region of Darjeeling Himalaya. The ensemble model has better capability than the other two models. The results may be helpful for landslide risk mitigation in the Kurseong and its surrounding region having similar terrain region and geological conditions.

Spatial modelling of shallow landslide susceptibility: a study from the southern Western Ghats region of Kerala, India.

ABSTRACT The study was undertaken to produce the landslide susceptibility maps by using Dempster–Shafer, Bayesian probability and logistic regression methods for the southern Western Ghats, Kerala, India. A landslide inventory database of 82 landslides is prepared and used for landslide susceptibility modelling. Twelve landslide conditioning factors including lithology, geomorphological features, slope angle, soil texture, distance from stream, distance from road, distance from lineaments, land use/land cover, slope curvature, rainfall, topographic wetness index and relative relief are extracted from the spatial database and used for modelling. Multi-collinearity among the independent variables were tested and landslide susceptibility maps are constructed. The constructed models were validated with sensitivity, specificity, classification accuracy, ROC-AUC, root mean square error (RMSE) and kappa index. The Bayesian probability model obtained highest ROC-AUC (0.833), sensitivity (0.870), specificity (0.800) and kappa index (0.667) with least RMSE (0.4550) in validation phase. In addition, the study reveals that the agricultural areas have 10°–40° slopes falling on the denudational structural hills are extremely susceptible to landslide occurrence with extended influence from distance from roads, distance from streams and soil texture. The predicted model is trustworthy for future land use planning in the southern Western Ghats to mitigate the risk from landslide hazard.

Comparisons of heuristic, general statistical and machine learning models for landslide susceptibility prediction and mapping

Commonly used data-driven models for landslide susceptibility prediction (LSP) can be mainly classified as heuristic, general statistical or machine learning models. This study plans to compare the prediction performance of these data-driven models on the landslide susceptibility mapping, thus further to explore the inherently features of these data-driven models. As a result, a more accurate and reliable LSP can be realized through choosing an optimal data-based model. A heuristic model represented by the analytic hierarchy process (AHP), a general statistical model represented by the general linear model (GLM) and information value (IV) model, and machine learning models represented by binary logistic regression (BLR), Multilayer Perceptron (MLP), back-propagation neural network (BPNN), support vector machine (SVM) and C5.0 decision tree (C5.0 DT) are adopted in this study. Shicheng County in China is used as the study area. In total, 369 landslides identified through field investigation are classified as training (70%) and testing datasets (30%). Next, 13 landslide conditioning factors (elevation, slope, aspect, plan curvature, profile curvature, relief amplitude, total surface radiation, population density, Normalized difference vegetation index, distance to river, topographic wetness index and rock types) are acquired from data sources of the free remote sensing images, Digital Elevation Model, field investigation and government reports. The correlations between these conditioning factors and the landslide locations are determined by frequency ratio analysis. Then, the landslide susceptibility indexes (LSIs) calculated by the eight trained models are imported into GIS software to produce landslide susceptibility maps of Shicheng County. Finally, the area under receiver operating characteristic curve (AUC), the calculated LSIs are applied to assess the LSP performance of the present eight models. The testing results show that these eight models generate reasonable LSP results as a whole, further showing that the C5.0 DT is of the highest prediction accuracy with an AUC value of 0.868, followed by the SVM (0.813), BPNN (0.803), MLP (0.792), BLR (0.784), GLM (0.779), IV (0.774) and AHP (0.773). It can be inferred that the machine learning models have higher LSP performance than general statistical and heuristic models due to its high AUC accuracy and reasonable LSIs distribution features, while general statistical model is limited by its linear analysis and heuristic model is limited by subjective weighting process.

A Comparative Study of the Frequency Ratio, Analytical Hierarchy Process, Artificial Neural Networks and Fuzzy Logic Methods for Landslide Susceptibility Mapping: Taşkent (Konya), Turkey

In this study, the four landslide susceptibility (LS) mapping methods, frequency ratio (FR), analytic hierarchy process (AHP), artificial neural networks (ANN) and fuzzy logic (FL) method, are compared. The study has been conducted in Taskent (Konya, Turkey) Basin which is located between 36.88 N to 36.95 N latitudes and 32.35 E to 32.53 E longitudes. The survey area is approximately 80 km2. The FR, AHP, ANN and FL methods are used to map LS. Thematic layers of fourteen landslide conditioning factors including landslide inventory, elevation, slope, slope aspect, plan, and profile curvature, sediment loading factor, stream power, and wetness index, drainage, and fault density, distance to drainage, and fault, geological units, and land use-land cover are used for preparing the LS maps. Estimation power of models has been evaluated by the relative operating characteristic curve method. The areas under the curve for FR, AHP, ANN and FL method have been computed as 0.926, 0.899, 0.916 and 0.842, respectively. These results showed that FR method is relatively good, whereas FL method is a relatively poor estimator for susceptibility. The validity of the LS maps was evaluated by test landslides. The 58 test landslides (76 pixels), 43 training landslides (200 pixels), and 101 total landslides (276 pixels) have been put onto the LS maps prepared by the various methods. The percentages of the existing landslide pixels within the different landslide occurrence potential classes were determined. It is determined that a significant portion of all landslides (76% in the ANN, 83% in the FR, 87% in the AHP and 89% in the FL method) belong to the high and very high LS class. The produced four susceptibility maps were also compared using cross-correlation methods. The cross-correlation coefficients were found to be 0.82, 0.70, 0.63, 0.54, 0.48, and 0.45 for AHP versus FR, FR versus FL, AHP versus FL, AHP versus ANN, FR versus ANN, and FL versus ANN maps, respectively. Here, the confidence level is 0.95. The FR and AHP methods have been assessed to be more suitable methods among other used methods.

Optimizing an Adaptive Neuro-Fuzzy Inference System for Spatial Prediction of Landslide Susceptibility Using Four State-of-the-art Metaheuristic Techniques.

Four state-of-the-art metaheuristic algorithms including the genetic algorithm (GA), particle swarm optimization (PSO), differential evolutionary (DE), and ant colony optimization (ACO) are applied to an adaptive neuro-fuzzy inference system (ANFIS) for spatial prediction of landslide susceptibility in Qazvin Province (Iran). To this end, the landslide inventory map, composed of 199 identified landslides, is divided into training and testing landslides with a 70:30 ratio. To create the spatial database, thirteen landslide conditioning factors are considered within the geographic information system (GIS). Notably, the spatial interaction between the landslides and mentioned conditioning factors is analyzed by means of frequency ratio (FR) theory. After the optimization process, it was shown that the DE-based model reaches the best response more quickly than other ensembles. The landslide susceptibility maps were developed, and the accuracy of the models was evaluated by a ranking system, based on the calculated area under the receiving operating characteristic curve (AUROC), mean absolute error, and mean square error (MSE) accuracy indices. According to the results, the GA-ANFIS with a total ranking score (TRS) = 24 presented the most accurate prediction, followed by PSO-ANFIS (TRS = 17), DE-ANFIS (TRS = 13), and ACO-ANFIS (TRS = 6). Due to the excellent results of this research, the developed landslide susceptibility maps can be applied for future planning and decision making of the related area.

Ensemble modeling of landslide susceptibility using random subspace learner and different decision tree classifiers

In this study, we have developed five spatially explicit ensemble predictive machine learning models for the landslide susceptibility mapping of the Van Chan district of the Yen Bai Province, Vietnam. In the model studies, Random Subspace (RSS) was used as the ensemble learner with Best First Decision Tree (BFT), Functional Tree (FT), J48 Decision Tree (J48DT), Naïve Bayes Tree (NBT) and Reduced Error Pruning Trees (REPT) as the base classifiers. Data of 167 past and present landslides and various landslide conditioning factors were used for generation of the datasets. The results showed that the RSSFT model achieved the highest performance in terms of Fgiurepredicting future landslides, followed by RSSREPT, RSSBFT, RSSJ48, and RSSNBT, respectively. Therefore, the RSSFT model was found to be more robust model than the other studied models, which can be used in other areas of landslide susceptibility mapping for proper landuse planning and management.