Landslide Conditioning Factors Research Articles

Landslide susceptibility mapping through continuous fuzzification and geometric average multi-criteria decision-making approaches

Landslide is a type of natural hazards causing many casualties in mountainous and rainy areas. Therefore, recognizing areas those that have potentials for happening such type of hazards is an important task. For this, methods of landslide susceptibility mapping, categorized mainly into two general data- and knowledge-driven approaches, have been widely developed and applied. In this regard, stochastic and systemic errors, respectively, associated with adequacy in the number of known landslide locations and subjectivity of expert judgment applied to assign weights of landslide conditioning factors are two main issues affecting the data- and knowledge-driven approaches. These issues are, in fact, types of bias and uncertainties that adversely affect landslide susceptibility mapping practices. This paper aims to adapt continuous fuzzification and geometric average multi-criteria decision-making approaches to overcome the aforementioned disadvantages of the existing landslide susceptibility mapping methods. In the method proposed weights of landslide conditioning factors are continuously assigned without using known landslide locations as training points, and without using expert opinion in categorization of values of landslide conditioning factors into arbitrary classes and assigning subjective weights. We applied the procedure proposed on a dataset of Oshvand watershed, Hamadan Province, Iran, to demonstrate its effectiveness. The results demonstrated that the continuous weighting method applied is more reliable than the existing methods those which apply classified values of landslide conditioning factors.

DInSAR and statistical modeling to assess landslides: The case study of Sierras Chicas (central Argentina)

The main goal of this investigation was to assess landslides in the steep West-facing slope of the Sierras Chicas mountains in central Argentina using two complementary approaches: Differential SAR interferometry (DinSAR) and statistical modeling. The combination of the geological characteristics of the area, wildfires, intense rainfall events and human disturbances have made the area highly susceptible to landslides. This condition has been recently aggravated with the construction of a roadway. In this work, we applied DInSAR methods to explore the suitability of COSMO-SkyMed (CSK) acquisitions to measure displacements and we used Sentinel-1 C-band images for accuracy comparison. In addition, Generalized Linear Models were fitted to identify landslide-conditioning factors. Based on the results obtained, Sentinel-1 proved to be the most adequate source of images to perform interferometry, in contrast to Cosmo Skymed imagery that showed very poor coherence in the study area. The statistical modeling identified slope degree, distance to roads and fire frequency as the main variables explaining landslide occurrence on the west escarpment of the Sierras Chicas. Results will hopefully contribute to improving future applications of SAR interferometry both in the study area and others with similar characteristics, as well as the tools and knowledge for decision making in engineering and urban development projects in order to minimize risk.

A hybrid optimization method of factor screening predicated on GeoDetector and Random Forest for Landslide Susceptibility Mapping

The aim of this study was to develop a hybrid model (Geo-RFE-RF) for Landslide Susceptibility Mapping (LSM) predicated on GeoDetector and Random Forest (RF) using the Recursive Feature Elimination (RFE-RF) method for eliminating redundant and noise factors. At the outset, for the sample of 1522 investigated landslides and 1522 non-landslides, twenty-two factors were chosen as the initial landslide-conditioning factors to construct a spatial database. Subsequently, the GeoDetector and RFE-RF methods were adopted to eliminate the least effective factors, respectively, with the Geo-RFE-RF model being formulated with the combination factors of these two methods. Finally, the performance of the Geo-RFE-RF and RF model with twenty-two initial factors (In-RF) were compared and assessed, and the higher accuracy model was employed to generate a LSM in a case study area, Fengjie County (China). The results indicate that, the Area Under Curve, Accuracy, Precision, and F1 Score of the test dataset is increased by 0.9%, 0.4%, 1.5%, and 0.3%, respectively, under the Geo-RFE-RF model, as compared to the In-RF model. The conditioning factors used to construct the model have been reduced from twenty-two to thirteen, but the predictive ability of the Geo-RFE-RF model performs better, proving the effectiveness of the hybrid model that combines the factors from GeoDetector and RFE-RF methods. This hybrid model not only considers the spatial pattern characteristics of spatial data for screening factors, but the selected factors are also in good agreement with the adopted machine learning method, offering potential use as a general framework for eliminating redundant and noise factors in LSM across the globe.

A Comparative Study of Landslide Susceptibility Mapping Using SVM and PSO-SVM Models Based on Grid and Slope Units

The main purpose of this study aims to apply and compare the rationality of landslide susceptibility maps using support vector machine (SVM) and particle swarm optimization coupled with support vector machine (PSO-SVM) models in Lueyang County, China, enhance the connection with the natural terrain, and analyze the application of grid units and slope units. A total of 186 landslide locations were identified by earlier reports and field surveys. The landslide inventory was randomly divided into two parts: 70% for training dataset and 30% for validation dataset. Based on the multisource data and geological environment, 16 landslide conditioning factors were selected, including control factors and triggering factors (i.e., altitude, slope angle, slope aspect, plan curvature, profile curvature, SPI, TPI, TRI, lithology, distance to faults, TWI, distance to rivers, NDVI, distance to roads, land use, and rainfall). The susceptibility between each conditioning factor and landslide was deduced using a certainty factor model. Subsequently, combined with grid units and slope units, the landslide susceptibility models were carried out by using SVM and PSO-SVM methods. The precision capability of the landslide susceptibility mapping produced by different models and units was verified through a receiver operating characteristic (ROC) curve. The results showed that the PSO-SVM model based on slope units had the best performance in landslide susceptibility mapping, and the area under the curve (AUC) values of training and validation datasets are 0.945 and 0.9245, respectively. Hence, the machine learning algorithm coupled with slope units can be considered a reliable and effective technique in landslide susceptibility mapping.

Landslide Susceptibility Mapping Using Rotation Forest Ensemble Technique with Different Decision Trees in the Three Gorges Reservoir Area, China

This study presents a new ensemble framework to predict landslide susceptibility by integrating decision trees (DTs) with the rotation forest (RF) ensemble technique. The proposed framework mainly includes four steps. First, training and validation sets are randomly selected according to historical landslide locations. Then, landslide conditioning factors are selected and screened by the gain ratio method. Next, several training subsets are produced from the training set and a series of trained DTs are obtained by using a DT as a base classifier couple with different training subsets. Finally, the resultant landslide susceptibility map is produced by combining all the DT classification results using the RF ensemble technique. Experimental results demonstrate that the performance of all the DTs can be effectively improved by integrating them with the RF ensemble technique. Specifically, the proposed ensemble methods achieved the predictive values of 0.012–0.121 higher than the DTs in terms of area under the curve (AUC). Furthermore, the proposed ensemble methods are better than the most popular ensemble methods with the predictive values of 0.005–0.083 in terms of AUC. Therefore, the proposed ensemble framework is effective to further improve the spatial prediction of landslides.

Advanced data mining techniques for landslide susceptibility mapping

This paper describes the development and validation of landslides susceptibility models for mountainous regions using advanced data mining techniques. The investigation was carried out to ascertain the effectiveness of Naïve Bayes Multinomial (NBM) and Random Trees (RT) in landslide susceptibility mapping. The NBM is an advancement of the frequently used Naïve Bayes classifiers, while the RT was built to overcome the limitations of the traditional forest classifiers. A geospatial database for this investigation comprises 148 landslide locations influenced by ten (10) landslide conditioning factors. The factors (Slope Angle, Slopes Elevation, Slope Aspect, Plan curvature, Profile Curvature, Lithology, Soil type, Stream power index (SPI), Sediment transport index (STI), and Rainfall precipitation) were drawn using a Multi Collinearity Decision Making (MCDM) technique. A Frequency Ratio (FR) analysis was used to obtain the relative significance of the factors in the slides. Predictive models were also developed by quantifying these models using data mining techniques. A section of the entire geospatial data (70%) was used as training datasets, while the remaining part of the data (30%) was used to validate the trained datasets. SVM, RT, and NBM algorithms were used to produce predicted datasets from the training datasets. These predicted datasets were used to develop the Landslides Susceptibility Models. A comparative assessment between the two classifiers against the famous traditional learning algorithm, the Support vector machines (SVM), was conducted. Model performance evaluators such as the AUROC, RSME, F-measure, MAE, and ACC were employed to check the predictive capabilities and accuracies of the models. The indices indicated that the SVM model performed better than the other two algorithms in both training and validation datasets. Further analysis and comparison of the models reveal that the new data mining techniques are reliable for landslide susceptibility. Simultaneously, the traditional algorithm is also useful and remains relevant, especially with similar site conditions. This study has provided insights on better planning and development and provision of mitigation strategies and further analysis on landslides in the study area, particularly in cases of limited data availability.

Spatial prediction of shallow landslide: application of novel rotational forest-based reduced error pruning tree

Landslides are a form of soil erosion threatening the sustainability of some areas of the world. There is, therefore, a need to investigate landslide rates and behaviour. In this research, we introduced a novel hybrid artificial intelligence approach of rotation forest (RF) as a meta classifier based on reduced error pruning tree (REPTree) as a base classifier called RF-REPTree, for landslide susceptibility mapping (LSM) in the Kalaleh watershed, Golestan Province, Iran. Some benchmark models, including the open-source Java decision tree (J48), naive Bayes tree (NBTree), and REPTree were used to compare the designed model. A total of 249 landslide locations were identified and mapped. The group was split into training (70%) and testing (30%) data for modelling and reliability analysis. Based on a literature review and multi-collinearity tests, 16 landslide conditioning factors (LCFs) were selected. Of the LCFs, the topographical position index (TPI) had the highest correlation with landslide occurrence. The LSM produced by RF-REPTree revealed that nearly 29% of the study areas have high to very high landslide susceptibility (LS). Statistical analysis of the model results included the receiver operating characteristic curve (ROC), the efficiency test, the true skill statistic (TSS), and the kappa index. ROC demonstrated that the AUC values of RF-REPTree, REPTree, J48, and NBTree models were 0.832, 0.700, 0.695, and 0.759 for succession rate curves and 0.794, 0.740, 0.788, and 0.728 for prediction rate curves, respectively. Therefore, all models were judged to be acceptably accurate for LSM. Among the LS models, the RF-REPTree model achieved the highest accuracy, followed by REPTree, J48, and NBTree. The results of LSM can be used to target the mitigation of landslide hazards and provide a foundation for sustainable environmental planning.

Effect of spatial resolution and data splitting on landslide susceptibility mapping using different machine learning algorithms

With the increasing computational facilities and data availability, machine learning (ML) models are gaining wide attention in landslide modeling. This study evaluates the effect of spatial resolution and data splitting, using five different ML algorithms (naïve bayes (NB), K nearest neighbors (KNN), logistic regression (LR), random forest (RF) and support vector machines (SVM)). The maps were developed using twelve landslide conditioning factors at two different resolutions, 12.5 m and 30 m. To identify the effect of data splitting on model performance, 2162 landslide points and an equal number of non-landslide points were used for training and testing the models using k-fold cross-validation, by varying the number of folds from two to ten. Results indicated that the spatial resolution of the dataset affects the performance of all the algorithms considered, while the effect of data splitting is significant in KNN and RF algorithms. All the algorithms yielded better performance while using the dataset with 12.5 m resolution for the same number of folds. It was also observed that the accuracy and area-under-the-curve values of 7, 8, 9, and 10-fold cross-validations with 30 m resolution was better than 2 and 3-fold cross-validations using 12.5 m resolution, in the case of RF algorithm.

A comparison among fuzzy multi-criteria decision making, bivariate, multivariate and machine learning models in landslide susceptibility mapping

Landslides are dangerous events which threaten both human life and property. The study aims to analyze the landslide susceptibility (LS) in the Kysuca river basin, Slovakia. For this reason, previous landslide events were analyzed with 16 landslide conditioning factors. Landslide inventory was divided into training (70% of landslide locations) and validating dataset (30% of landslide locations). The heuristic approach of Fuzzy Decision Making Trial and Evaluation Laboratory (FDEMATEL)-Analytic Network Process (ANP) was applied first, followed by bivariate Frequency Ratio (FR), multivariate Logistic Regression (LR), Random Forest Classifier (RFC), Naïve Bayes Classifier (NBC) and Extreme Gradient Boosting (XGBoost), respectively. The results showed that 52.2%, 36.5%, 40.7%, 50.6%, 43.6% and 40.3% of the total basin area had very high to high LS corresponding to FDEMATEL-ANP, FR, LR, RFC, NBC and XGBoost model, respectively. The analysis revealed that RFC was the most accurate model (overall accuracy of 98.3% and AUC of 97.0%). Besides, the heuristic approach of FDEMATEL-ANP model (overall accuracy of 93.8% and AUC of 92.4%) had better prediction capability than bivariate FR (overall accuracy of 86.9% and AUC of 86.1%), multivariate LR (overall accuracy of 90.5% and AUC of 91.2%), machine learning NBC (overall accuracy of 76.3% and AUC of 90.9%) and even deep learning XGBoost (overall accuracy of 92.3% and AUC of 87.1%) models. The study revealed that the FDEMATEL-ANP outweighed the NBC and XGBoost machine learning models, which suggests that heuristic methods should be tested out before directly applying machine learning models.

Mapping Landform and Landslide Susceptibility Using Remote Sensing, GIS and Field Observation in the Southern Cross Road, Malang Regency, East Java, Indonesia

There has been an increasing trend of land area being brought under human’s use over time. This situation has led the community to carry out land-use development activities in landslide hazard-prone areas. The use of land can have a positive impact by increasing economic conditions, but it can have negative impacts on the environment. Therefore, this study aimed to identify the landslide hazard, focusing on the development of a landform map to reduce the risk of landslide disaster in JLS, Malang Regency. The integration of remote sensing and geographic information systems, as well as field observation, were used to create a landform map and a landslide susceptibility map. Using the geomorphological approach as a basic concept in landform mapping, the morphology, morphogenesis, and morphoarrangement conditions were obtained from the remote sensing data, GIS, and field observation, while morphochronological information was obtained from a geological map. The landslide susceptibility map was prepared using 11 landslide conditioning factors by employing the index of entropy method. Thirty-nine landform units were successfully mapped into four landslide susceptibility classes. The results showed that the study area is dominated by a high level of landslide susceptibility with a majority of moderate to strongly eroded hill morphology. It also reaffirms that landform mapping is a reliable method by which to investigate landslide susceptibility in JLS, Malang Regency.

Landslide hazard assessment of the Neom promising city, northwestern Saudi Arabia: An integrated approach

Landslides hazard as one of the major environmental threats of the Neom promising city in northwestern Saudi Arabia has been assessed through data integration of several sources including topographic, geologic, and seismologic data sets. Moreover, Landsat imagery and historical landslide data are introduced in assessing landslide susceptibility of the study area. All these data oriented to generate eight thematic layers representing eight landslide conditioning factors. These factors arranged according to their relative importance for landslide susceptibility mapping as follows: slope angle, elevation, curvature, aspect, earthquake density, fault density, drainage density, and lineaments density. These factors were assessed using GIS through a numerical rating scheme for thematic layers. The ranks and weights were given in association with intensive field studies and authors’ experiences. The estimated values of Landslide Potential Index (LPI) varies from 32 up to 107. Then, the resultant landslide susceptibility map classifies the landslide hazards into three relative classes of susceptibility: high, moderate, and low. Furthermore, this map was tested through several field checks and the extracted historical landslides within Neom area as well. This landslide susceptibility map will boost decision-making authority while executing the developing projects throughout the Neom promising area either avoiding strategic facilities or applying corrective geotechnical procedures to minimize the landslide hazardous impact for Neom city.

Landslide susceptibility index based on the integration of logistic regression and weights of evidence: A case study in Popayan, Colombia

In this paper, we present a suitable integration of discrete and continuous data in a unique methodology based on systematically collected landslide inventory data. Eleven landslide conditioning factors were analyzed and used, where eight correspond to DEM–derived variables, and three to thematic polygon–type variables (shallow geology, geomorphology and soil land–use). Principal Component Analysis (PCA) was used to avoid the effect of multicollinearity. Additionally, 3 proposals were developed using Logistic Regression (LR) and Weights of Evidence (WoE) methods that use the continuous and discrete variables efficiently, respectively. The performance of each proposal was evaluated by the Area Under the Curve (AUC) of the Receiver Operating Characteristic (ROC) curves. The analysis indicated that Proposal 1 with AUC = 0.8578 and Proposal 2 with AUC = 0.8459 have the best LSI assessment, while the performance of Proposal 3 with AUC = 0.8054 shows the lowest prediction approaches. In comparison with the WoE method, our proposal shows an increase in high and very high susceptibility in areas with complex topography, which is consistent with the reported landslides.

An Optimized Random Forest Model and Its Generalization Ability in Landslide Susceptibility Mapping: Application in Two Areas of Three Gorges Reservoir, China

Numerous researches have been published on the application of landslide susceptibility assessment models; however,they were only applied in the same areas as the models were originated,the effect of applying the models to other areas than the origin of the models has not been explored. This study is purposed to develop an optimized random forest (RF) model with best ratios of positive-to-negative cells and 10-fold cross-validation for landslide susceptibility mapping (LSM),and then explore its generalization ability not only in the area where the model is originated but also in area other than the origin of the model. Two typical counties (Fengjie County and Wushan County) in the Three Gorges Reservoir area,China,which have the same terrain and geological conditions,were selected as an example. To begin with,landslide inventory was prepared based on field investigations,satellite images,and historical records,and 1 522 landslides were then identified in Fengjie County. 22 landslide-conditioning factors under the influence of topography,geology,environmental conditions,and human activities were prepared. Then,combined with 10-fold cross-validation,three typical ratios of positive-to-negative cells,i.e.,1:1,1:5,and 1:10,were adopted for comparative analyses. An optimized RF model (Fengjie-based model) with the best ratios of positive-to-negative cells and 10-fold cross-validation was constructed. Finally,the Fengjie-based model was applied to Fengjie County and Wushan County,and the confusion matrix and area under the receiver operating characteristic (ROC) curve value (AUC) were used to estimate the accuracy. The Fengjie-based model delivered high stability and predictive capability in Fengjie County,indicating a great generalization ability of the model to the area where the model is originated. The LSM in Wushan County generated by the Fengjie-based model had a reasonable reference value,indicating the Fengjie-based model had a great generalization ability in area other than the origin of the model. The Fengjie-based model in this study could be applied in other similar areas/countries with the same terrain and geological conditions,and a LSM may be generated without collecting landslide information for modeling,so as to reduce workload and improve efficiency in practice.

Performance Evaluation and Comparison of Bivariate Statistical-Based Artificial Intelligence Algorithms for Spatial Prediction of Landslides

The purpose of this study is to compare nine models, composed of certainty factors (CFs), weights of evidence (WoE), evidential belief function (EBF) and two machine learning models, namely random forest (RF) and support vector machine (SVM). In the first step, fifteen landslide conditioning factors were selected to prepare thematic maps, including slope aspect, slope angle, elevation, stream power index (SPI), sediment transport index (STI), topographic wetness index (TWI), plan curvature, profile curvature, land use, normalized difference vegetation index (NDVI), soil, lithology, rainfall, distance to rivers and distance to roads. In the second step, 152 landslides were randomly divided into two groups at a ratio of 70/30 as the training and validation datasets. In the third step, the weights of the CF, WoE and EBF models for conditioning factor were calculated separately, and the weights were used to generate the landslide susceptibility maps. The weights of each bivariate model were substituted into the RF and SVM models, respectively, and six integrated models and landslide susceptibility maps were obtained. In the fourth step, the receiver operating characteristic (ROC) curve and related parameters were used for verification and comparison, and then the success rate curve and the prediction rate curves were used for re-analysis. The comprehensive results showed that the hybrid model is superior to the bivariate model, and all nine models have excellent performance. The WoE–RF model has the highest predictive ability (AUC_T: 0.9993, AUC_P: 0.8968). The landslide susceptibility maps produced in this study can be used to manage landslide hazard and risk in Linyou County and other similar areas.

Factors Affecting Occurrence of Landslides Induced by the M7.8 April 2015, Nepal Earthquake

The M7.8 Nepal earthquake on 25 April 2015 and its aftershocks triggered thousands of landslides. Understanding correlations between the landslide occurrences and landslide conditioning factors is essential to generate a landslide susceptibility map. This study considers the two most reliable inventory data: landslide inventory 1 (LI-1) and landslide inventory 2 (LI-2). The LI-1 contains 1,352 landslide polygons, and the LI-2 contains 4,000 landslide points. The centroid locations of LI-1 were also considered as additional point inventory data for the statistical analyses. The landslide area percentage and landslide number density are computed for important factors such as slope angle, slope roughness, aspect, peak ground acceleration (PGA), distance to epicenter, distance to rivers and roads, Normalized Difference Vegetation Index (NDVI), land use and land cover, and mean annual precipitation. The cumulative percentage curves and the areas under the curve (AUC) are also compared for all the considered factors. PGAs of the main shock and two major aftershocks, distance to epicenter, slope angle, distance to river, and NDVI have relatively stronger correlations with the landslide occurrence than other factors. It is found that the LI-1 polygons have weaker correlations with the factors than LI-1 points and LI-2 points do because the LI-1 does not distinguish debris from scarps.

Rainfall induced landslide susceptibility mapping using novel hybrid soft computing methods based on multi-layer perceptron neural network classifier

In this study, we have investigated rainfall induced landslide susceptibility of the Uttarkashi district of India through the developmentof different novel GIS based soft computing approaches namely Bagging-MLPC, Dagging-MLPC, Decorate-MLPC which are a combination Multi-layer Perceptron Neural Network Classifier (MLPC) and Bagging, Dagging, and Decorate ensemble methods, respectively. The proposed models were trained and validated with the help of 103 historical landslide events (divided into 2 samples: training (70%) and validation (30%)) and 12 landslide conditioning factors. The accuracy of the models was evaluated using different statistical methods including Area Under Curve (AUC) of Receiver Operating Characteristic (ROC). The results show that though performance of all the studied models is good (AUC > 0.80) but of the hybrid Bagging-MLPC model is the best (AUC:0.965). Therefore, this newly hybrid model (Bagging-MLPC) can be used for the accurate landslide susceptibility mapping and assessment of landslide prone areas for landslide prevention and management.

Assessing and mapping landslide susceptibility using different machine learning methods

The main aim of the present study was to produce and compare landslide susceptibility maps by using five machine learning techniques, namely, artificial neural network (ANN), logistic regression (LR), support vector machine (SVM), random forest (RF) and, classification and regression tree (CART). The study area was determined as the Arhavi-Kabisre river basin, a region in which the most landslide incidents occur in Turkey. Firstly, a landslide inventory was produced by identifying a total of 252 landslides. Secondly, a total of 11 landslide conditioning factors were considered for the landslide susceptibility mapping. Subsequently, the five machine learning techniques were constructed with the help of the training dataset for the landslide susceptibility maps. Finally, the receiver operating characteristic (ROC), sensitivity, specificity, F-measure, accuracy and kappa index were applied to compare and validate the performance of the five machine learning techniques.

Prediction of landslide susceptibility in Rudraprayag, India using novel ensemble of conditional probability and boosted regression tree-based on cross-validation method

The present research examines the landslide susceptibility in Rudraprayag district of Uttarakhand, India using the conditional probability (CP) statistical technique, the boost regression tree (BRT) machine learning algorithm, and the CP-BRT ensemble approach to improve the accuracy of the BRT model. Using the four fold of data, the models' outcomes were cross-checked. The locations of existing landslides were detected by general field surveys and relevant records. 220 previous landslide locations were obtained, presented as an inventory map, and divided into four folds to calibrate and authenticate the models. For modelling the landslide susceptibility, twelve LCFs (landslide conditioning factors) were used. Two statistical methods, i.e. the mean absolute error (MAE) and the root mean square error (RMSE), one statistical test, i.e. the Freidman rank test, as well as the receiver operating characteristic (ROC), efficiency and precision were used for authenticating the produced landslide models. The results of the accuracy measures revealed that all models have good potential to recognize the landslide susceptibility in the Garhwal Himalayan region. Among these models, the ensemble model achieved a higher accuracy (precision: 0.829, efficiency: 0.833, AUC: 89.460, RMSE: 0.069 and MAE: 0.141) than the individual models. According to the outcome of the ensemble simulations, the BRT model's predictive accuracy was enhanced by integrating it with the statistical model (CP). The study showed that the areas of fallow land, plantation fields, and roadsides with elevations of more than 1500 m. with steep slopes of 24° to 87° and eroding hills are highly susceptible to landslides. The findings of this work could help in minimizing the landslides' risk in the Western Himalaya and its adjoining areas with similar landscapes and geological characteristics.

Comparative analysis of gradient boosting algorithms for landslide susceptibility mapping

The aim of the study is to compare four recent gradient boosting algorithms named as Gradient Boosting Machine (GBM), Categorical Boosting (CatBoost), Extreme Gradient Boosting (XGBoost), and Light Gradient Boosting Machine (LightGBM) for modelling landslide susceptibility (LS). In the first step of the study, the geodatabase including landslide inventory map and landslide conditioning factors was constructed. In the second step, chi-square (CHI) statistic-based feature selection (FS) technique was utilized to compute the importance of the landslide causative factors. In the third step, tree-based ensemble learning algorithms were applied to predict the potential distribution of landslide susceptibility. Also, the prediction performance of ensemble methods was compared to that of Random Forest (RF) ensemble method. Finally, the prediction capabilities of the methods were assessed using overall accuracy (Acc), area under the receiver operating characteristic curve (AUC), kappa index, root mean square error (RMSE), and F score measures. In order to further evaluation, the McNemar's test was utilized to assess statistical significance in the differences between the four gradient boosting models. The accuracy results indicated that the CatBoost model had the highest prediction capability (Acc= 0.8503 and AUC= 0.8975), followed by the XGBoost (Acc= 0.8336 and AUC= 0.8860), the LightGBM (Acc= 0.8244 and AUC= 0.8796) and the GBM (Acc= 0.8080 and AUC= 0.8685). On the other hand, the estimated accuracy measures considered in this study showed that the RF method had the lowest prediction capability of compared the others. Although the individual performances of the methods were found to be acceptable level, the CatBoost method showed the superior performance compared to others with respect to the AUC and Acc values estimated in this study. The results of the study confirmed that the relatively new ensemble learning techniques were efficient and robust for producing LS maps and furthermore, it is probably that these algorithms will be preferred more often in the future studies due to their robustness.

Improving Spatial Agreement in Machine Learning-Based Landslide Susceptibility Mapping

Despite yielding considerable degrees of accuracy in landslide predictions, the outcomes of different landslide susceptibility models are prone to spatial disagreement; and therefore, uncertainties. Uncertainties in the results of various landslide susceptibility models create challenges in selecting the most suitable method to manage this complex natural phenomenon. This study aimed to propose an approach to reduce uncertainties in landslide prediction, diagnosing spatial agreement in machine learning-based landslide susceptibility maps. It first developed landslide susceptibility maps of Cox’s Bazar district of Bangladesh, applying four machine learning algorithms: K-Nearest Neighbor (KNN), Multi-Layer Perceptron (MLP), Random Forest (RF), and Support Vector Machine (SVM), featuring hyperparameter optimization of 12 landslide conditioning factors. The results of all the four models yielded very high prediction accuracy, with the area under the curve (AUC) values range between 0.93 to 0.96. The assessment of spatial agreement of landslide predictions showed that the pixel-wise correlation coefficients of landslide probability between various models range from 0.69 to 0.85, indicating the uncertainty in predicted landslides by various models, despite their considerable prediction accuracy. The uncertainty was addressed by establishing a Logistic Regression (LR) model, incorporating the binary landslide inventory data as the dependent variable and the results of the four landslide susceptibility models as independent variables. The outcomes indicated that the RF model had the highest influence in predicting the observed landslide locations, followed by the MLP, SVM, and KNN models. Finally, a combined landslide susceptibility map was developed by integrating the results of the four machine learning-based landslide predictions. The combined map resulted in better spatial agreement (correlation coefficients range between 0.88 and 0.92) and greater prediction accuracy (0.97) compared to the individual models. The modelling approach followed in this study would be useful in minimizing uncertainties of various methods and improving landslide predictions.