Landslide Susceptibility Research Articles

Environmental critical thresholds based on statistical analysis for modelling landslide susceptibility in Continental Basaltic Provinces

Abstract. The study aims to estimate the environmental critical thresholds using statistical approaches to understand the landslide conditioning factors that can trigger landslides in the Continental Basaltic Provinces a landslide-prone area, using as reference the landslides that occurred in an extreme rainfall event. The study area is a region that was the scene of an extreme hydrological event in January 2017, with an accumulated volume of rain of 163.9 mm in 8 hours, causing a widespread event of shallow planar landslides with more than 400 scars detected. Hydrological, anthropic, geological, geomorphological, and topographical features of this region were analyzed considering landslides and non-landslides samples set, and their influence in the event was carried out using the Frequency Ratio method, followed by Pearson's Linear Correlation Coefficient and Linear Regression. The results showed that this process helped us to understand environmental critical thresholds based on classes of conditioning factors that have a greater influence on rainfall-triggered landslide occurrences and, consequently, higher predictive capacity in the landslide susceptibility models with the same geoenvironmental parameters which is a valuable insight for risk management.

A Spatial Landslide Risk Assessment Based on Hazard, Vulnerability, Exposure, and Adaptive Capacity

Landslides threaten human life, property, and vital infrastructure in most mountainous regions. As climate change intensifies extreme weather patterns, the landslide risk is likely to increase, resulting in challenges for disaster management, sustainability development, and community resilience. This study presents a comprehensive framework for assessing landslide risk, integrating advanced machine learning models with the Iyengar–Sudarshan method. Our case study is Son La province, the Northwest region of Vietnam, with data collected from 1771 historical landslide occurrences and fifteen influencing factors for developing landslide susceptibility maps using advanced ensemble machine learning models. The Iyengar–Sudarshan method was applied to determine the weights for landslide exposure, vulnerability, and adaptive capacity indicators. The resulting landslide risk map shows that the highest-risk districts in Son La province are located in the central and northeastern regions, including Mai Son, Phu Yen, Thuan Chau, Yen Chau, Song Ma, and Bac Yen. These districts experience high landslide hazards, exposure, and vulnerability, often affecting densely populated urban and village areas with vulnerable populations, such as young children, the elderly, and working-age women. In contrast, due to minimal exposure, Quynh Nhai and Muong La districts have lower landslide risks. Despite having high exposure and vulnerability, Son La City is situated in a low-susceptibility zone with high adaptive capacity, resulting in a low landslide risk for this region. The proposed framework provides a reference tool for mitigating risk and enhancing strategic decision making in areas susceptible to landslides while advancing our understanding of landslide dynamics and fostering community resilience and long-term disaster prevention.

Incorporating modelling uncertainty and prior knowledge into landslide susceptibility mapping using Bayesian neural networks

ABSTRACT In landslide susceptibility (LS) mapping, the reliance on deep learning (DL) models within the frequentist paradigm often leads to an oversight of modelling uncertainties and priors, undermining reliability. This study applied Bayesian neural network (BNN) using variational inference and Bayes backpropagation to LS mapping in Lin’an, China. We developed two BNN configurations: one employing a zero-mean Gaussian distribution as priors (BNN-normal) and another grounded in domain knowledge (BNN-knowledge), and benchmarked them against deep neural networks (DNNs). The study further investigates the models’ performance against random noise impacts on conditional factors and challenges posed by incomplete landslide inventories. Results indicated although BNNs were comparable to the DNNs in generalisation, they demonstrated superior calibration accuracy, resistance to noise and overfitting. Notably, BNN-knowledge effectively mitigated the adverse effects of incomplete data. A landslide confidence map was constructed by integrating the LS and epistemic entropy. Compared to LS maps without uncertainty analysis, the landslide densities in very high and very low LS area with low uncertainty increased by 8.34 and decreased by 0.1, respectively, providing more information for scientific and effective decision-making. Our results further confirmed that sparse vegetation, anthropogenic disturbances and economic forest expansion were main factors of high LS in the region.

Stacking Ensemble Technique Using Optimized Machine Learning Models with Boruta–XGBoost Feature Selection for Landslide Susceptibility Mapping: A Case of Kermanshah Province, Iran

Landslides cause significant human and financial losses in different regions of the world. A high-accuracy landslide susceptibility map (LSM) is required to reduce the adverse effects of landslides. Machine learning (ML) is a robust tool for LSM creation. ML models require large amounts of data to predict landslides accurately. This study has developed a stacking ensemble technique based on ML and optimization to enhance the accuracy of an LSM while considering small datasets. The Boruta–XGBoost feature selection was used to determine the optimal combination of features. Then, an intelligent and accurate analysis was performed to prepare the LSM using a dynamic and hybrid approach based on the Adaptive Fuzzy Inference System (ANFIS), Extreme Learning Machine (ELM), Support Vector Regression (SVR), and new optimization algorithms (Ladybug Beetle Optimization [LBO] and Electric Eel Foraging Optimization [EEFO]). After model optimization, a stacking ensemble learning technique was used to weight the models and combine the model outputs to increase the accuracy and reliability of the LSM. The weight combinations of the models were optimized using LBO and EEFO. The Root Mean Square Error (RMSE) and Area Under the Receiver Operating Characteristic Curve (AUC-ROC) parameters were used to assess the performance of these models. A landslide dataset from Kermanshah province, Iran, and 17 influencing factors were used to evaluate the proposed approach. Landslide inventory was 116 points, and the combined Voronoi and entropy method was applied for non-landslide point sampling. The results showed higher accuracy from the stacking ensemble technique with EEFO and LBO algorithms with AUC-ROC values of 94.81% and 94.84% and RMSE values of 0.3146 and 0.3142, respectively. The proposed approach can help managers and planners prepare accurate and reliable LSMs and, as a result, reduce the human and financial losses associated with landslide events.

Enhanced landslide susceptibility mapping in data-scarce regions via unsupervised few-shot learning

Given the critical need to assess landslide hazards, producing landslide susceptibility map (LSM) in regions with scarce historical landslide inventories poses significant challenges. This study introduces a novel landslide susceptibility assessment framework that combines unsupervised learning strategies with few-shot learning methods to increase the accuracy of LSM in these areas. The framework has been practically validated in a representative geological disaster-prone area along the West-East Gas Pipeline in Shaanxi Province, China. We employed three advanced few-shot learning models: a support vector machine, meta-learning, and transfer learning. These models implement feature representation learning for weakly correlated influencing factors through an unsupervised approach, thereby constructing an effective landslide susceptibility assessment model. We compared traditional learning methods and used the receiver operating characteristic (ROC) curve and SHAP values to quantify the effectiveness of the models. The results indicate that the meta-learning algorithm outperforms both the SVM and transfer learning in areas with limited landslide data. The integration of unsupervised strategies significantly improves performance, achieving area under the curve (AUC) values of 0.9385 and 0.9861, respectively. Compared with using meta-learning alone, incorporating unsupervised learning strategies increased the AUC by 4.76%, enhancing both the predictive power of the model and the interpretability of the features. Meta-learning under unsupervised conditions effectively mitigates the evaluation difficulties caused by insufficient landslide records, providing a viable path and empirical evidence for performance improvement in similar data- scarce regions worldwide.

Review on the Artificial Intelligence-based methods in Landslide Detection and Susceptibility Assessment: Current Progress and Future Directions

Review on the Artificial Intelligence-based methods in Landslide Detection and Susceptibility Assessment: Current Progress and Future Directions

Space-time modeling of cascading hazards: Chaining wildfires, rainfall and landslide events through machine learning

The current study sets out to explore yearly landslide susceptibility dynamics on slopes regularly affected by fires. To do so, two yearly inventories have been generated, one for the landslides and one for the wildfires, for an area of approximately 2 km2 and for a period of 24 years. It is important to stress that space–time data-driven models employed for susceptibility assessment are relatively new, and their application so far has mostly linked landslide occurrences to the precipitation trigger and the standard morphometric characteristics of the landscape at hand. Here we also consider an additional element of disturbance to the slope equilibrium, in the form of burnt areas tested from one up to three years priors to the reference landslide occurrence time. The relevance of the wildfire spatiotemporal signal is tested as part of a multi-variate modeling procedure. The results highlight at least a 10 % performance increase when these wildfire-related predictors are featured (from an average AUC of 0.75 to 0.85 in a random forest modeling framework). The associated yearly variations in the landslide occurrence probability are translated into individual maps, stressing the extent to which the standard and static definition of susceptibility does not hold especially in the context of multiple hazards and in an urban setting such as the Camaldoli hills (Naples, Italy) we chose for our test site.

Landslide Susceptibility Assessment Using a CNN–BiLSTM-AM Model

Landslide Susceptibility Assessment Using a CNN–BiLSTM-AM Model

Evaluating the geological and soil characteristics influencing landslide susceptibility in Anambra state, Nigeria

Evaluating the geological and soil characteristics influencing landslide susceptibility in Anambra state, Nigeria

A-posteriori analysis of the performance of a rockfall susceptibility map

BackgroundRockfalls pose a serious threat along the main road network, representing a major hazard in mountainous territory and causing damage and victims. Currently, susceptibility mapping represents a starting point to identify areas more susceptible to rockfall occurrence, a key approach in land use planning and risk management. Despite the extensive use of these maps by decision makers and administrators, the usability of such maps over time and their reliability represent a poorly discussed and examined feature.MethodsHere, we proposed a-posteriori analysis of a three-year-old rockfall susceptibility map, generated along the main road network of the Aosta Valley, an alpine region of north-western Italy. To verify map consistency over time, we implemented a dual-analysis in GIS-environment and by text mining, to respectively analyse the geocoded data and textual information derivable from the regional landslide inventory. The first one allowed us to extract rockfall events occurred after the susceptibility map generation. By this way, we operated to spatially and temporally verify the map consistency. Jointly, the textual information reported in the Event Description Form, linked to each geocoded event, are being exploited. This allowed us to derive relevant information about occurred damage and their degree, presence of protective measures or secondary roads, i.e. involvement of, farm or forestry, road.ResultsThe implemented approach allowed us to prove the quality of the previous map in terms of reliability, robustness and degree of fitting respect to the succession of rockfall occurrence over time. After only three years as many as 198 rockfall events have been occurred and collected since the map was generated. Particularly, 80% of rockfall fit with “high” and “very high” susceptibility classes, and pointed out large involvement of main roads in rockfall occurrence, representing the most affected target with damage to road pavement and vehicles, as well as a relevant involvement of existing rockfall barriers and of the dense network of forestry roads and footpaths that characterize this alpine region.ResultsThe proposed approach representing a starting point in landslide susceptibility map verification and usability as valid instrument for a reliable land planning.

Generating a Landslide Susceptibility Map Using Integrated Meta-Heuristic Optimization and Machine Learning Models

A landslide susceptibility assessment is one of the critical steps in planning for landslide disaster prevention. Advanced machine learning methods can be used as data-driven approaches for landslide susceptibility zonation with several landslide conditioning factors. Despite there being a number of studies on landslide susceptibility assessment, the literature is limited in several contexts, such as parameter optimization, an examination of the factors in detail, and study area. This study addresses these lacks in the literature and aims to develop a landslide susceptibility map of Kentucky, US. Four machine learning methods, namely artificial neural network (ANN), k-nearest neighbor (KNN), support vector machine (SVM), and stochastic gradient boosting (SGB), were used to train the dataset comprising sixteen landslide conditioning factors after pre-processing the data in terms of data encoding, data scaling, and dimension reduction. The hyperparameters of the machine learning methods were optimized using a state-of-the-art artificial bee colony (ABC) algorithm. The permutation importance and Shapley additive explanations (SHAP) methods were employed to reduce the dimension of the dataset and examine the contributions of each landslide conditioning factor to the output variable, respectively. The findings show that the ABC-SGB hybrid model achieved the highest prediction performance. The SHAP summary plot developed using the ABC-SGB model shows that intense precipitation, distance to faults, and slope were the most significant factors affecting landslide susceptibility. The SHAP analysis further underlines that increases in intense precipitation, distance to faults, and slope are associated with an increase in the probability of landslide incidents. The findings attained in this study can be used by decision makers to develop the most effective resource allocation plan for preventing landslides and minimizing related damages.

Exploring Topological Information Beyond Persistent Homology to Detect Geospatial Objects

Accurate detection of geospatial objects, particularly landslides, is a critical challenge in geospatial data analysis due to the complex nature of the data and the significant consequences of these events. This paper introduces an innovative topological knowledge-based (Topological KB) method that leverages the integration of topological, geometrical, and contextual information to enhance the precision of landslide detection. Topology, a fundamental branch of mathematics, explores the properties of space that are preserved under continuous transformations and focuses on the qualitative aspects of space, studying features like connectivity and exitance of loops/holes. We employed persistent homology (PH) to derive candidate polygons and applied three distinct strategies for landslide detection: without any filters, with geometrical and contextual filters, and a combination of topological with geometrical and contextual filters. Our method was rigorously tested across five different study areas. The experimental results revealed that geometrical and contextual filters significantly improved detection accuracy, with the highest F1 scores achieved when employing these filters on candidate polygons derived from PH. Contrary to our initial hypothesis, the addition of topological information to the detection process did not yield a notable increase in accuracy, suggesting that the initial topological features extracted through PH suffices for accurate landslide characterization. This study advances the field of geospatial object detection by demonstrating the effectiveness of combining geometrical and contextual information and provides a robust framework for accurately mapping landslide susceptibility.

Machine learning-based assessment of regional-scale variation of landslide susceptibility in central Vietnam.

Recurrent landslide events triggered by typhoons and tropical storms over Vietnam pose a longstanding threat to the nation's population and infrastructure. Changes in hydroclimatic conditions, especially the growing intensity and frequency of storms, have elevated landslide susceptibility in many parts of the country. This research examines the spatio-temporal variations in landslide susceptibility across central Vietnam over several years, using multi-temporal landslide inventories from Typhoon Ketsana (2009), Tropical Storm Podul (2013), and Typhoon Molave (2020). Additionally, the research explores the impact of individual landslide causative factors on the probabilistic occurrences of landslides. The post-event landslide susceptibility models of these three climate extreme events were developed using nine causative factors and a Random Forest machine learning algorithm. The results indicate a notable areal expansion of high to very high landslide susceptibility in the northern and eastern regions and a moderate reduction in the central and southern areas during the post-Molave period compared to the post-Ketsana period. These changes may be early indicators of increasing landslide susceptibility in response to changing hydro-climatic conditions. The research found that annual average rainfall and topographic elevation are the two most important variables influencing landslide prediction, showing a nonlinear relationship with landslide probability. The landslide susceptibility models achieved high Area Under the Receiver Operating Characteristic Curve (AUC) (>95%), accuracy (>89%), and sensitivity (>90%) scores, signifying the robustness of the models. Additionally, the uncertainty of the models was quantified and spatially mapped. This multi-temporal analysis of landslide susceptibility is crucial for understanding the regional susceptibility trends and identifying areas with increasing, decreasing, and consistently high susceptibility to landslides. These insights are invaluable for prioritizing mitigation and risk reduction strategies in landslide-prone regions and guiding appropriate land use planning.

Impact of sampling for landslide susceptibility assessment using interpretable machine learning models

Impact of sampling for landslide susceptibility assessment using interpretable machine learning models

Analyzing the posterior predictive capability and usability of landslide susceptibility maps: a case of Kerala, India

AbstractLandslide susceptibility maps serve as the basis for hazard and risk assessment, as well as risk-informed land use planning at various spatial scales. Researchers create these maps aiming to fulfil a variety of purposes, including infrastructure planning and restrictive land use zoning. These applications require accurate and specific information to fulfil these purposes, as decisions based on these maps have the potential to cost lives and cause infrastructure damage. The usability of the maps depends on whether they provide the required information and their accuracy to be utilized for the intended purpose. Therefore, assessing the usability and predictive accuracy of landslide susceptibility maps is of paramount importance. Typically, the accuracy of the maps is evaluated using the same landslide inventory that was used to create the map, which does not actually test the predictive ability of the maps in future situations. In this study, we briefly reviewed the purposes of the map creation using literature and stakeholder interviews and assessed the accuracy of three landslide susceptibility maps in a posterior manner. We generated a multi-temporal landslide event inventory after the creation dates of these landslide susceptibility maps. We devised a method to evaluate classified maps by making use of Unique Conditions Units (UCUs) to compare the posteriorly predicted susceptibility classes and the new landslide occurrences. Interviews with stakeholders revealed a disconnection between the aims set forth by map producers and the specific needs of the end users. Our posterior assessment shows that overall predictions of the maps provide plausible results; however, their interpretations for different use cases make them less likely to be used. When comparing the maps using UCUs, landslide densities overlap between the different susceptibility classes, indicating low predictive performance of the maps. Direct comparison of all maps shows a low agreement between susceptibility classes, which pinpoints the uncertainties in data and methods used to create different maps. This study highlights the need for purpose-oriented landslide susceptibility mapping and posterior assessment of the predictive capabilities of these maps aiming to fulfil respective purposes.

Spatiotemporal dynamics of landslide susceptibility under future climate change and land use scenarios

Abstract Mountainous landslides are expected to worsen due to environmental changes, yet few studies have quantified their future risks. To address this gap, we conducted a comprehensive analysis of the eastern Hindukush region of Pakistan. A geospatial database was developed, and logistic regression was employed to evaluate baseline landslide susceptibility for 2020. Using the latest CMIP6 models under three Shared Socioeconomic Pathways (SSPs) and the CA-Markov model, we projected future rainfall and land use/land cover patterns for 2040, 2070, and 2100, respectively. Our results reveal significant changes in future rainfall and land use patterns, particularly in the long-term future (2070 and 2100). Future landslide susceptibility was then predicted based on these projections. By 2100, high-risk landslide areas are expected to increase substantially under all SSP scenarios, with the largest increases observed under SSP5-8.5 (56.52%), SSP2-4.5 (53.55%), and SSP1-2.6 (22.45%). By 2070, high-risk areas will rise by 43.08% (SSP1-2.6), 40.88% (SSP2-4.5), and 12.60% (SSP5-8.5). However, by 2040, the changes in high-risk areas are minimal compared to the baseline, with increases of 9.45% (SSP1-2.6), 1.69% (SSP2-4.5), and 7.63% (SSP5-8.5). These findings provide crucial insights into the relationship between environmental changes and landslide risks and support the development of climate risk mitigation, land use planning, and disaster management strategies for mountainous regions.

Multi-hazard modeling of erosion and landslide susceptibility at the national scale in the example of North Macedonia

Abstract Due to favorable natural conditions and human impact, the territory of North Macedonia is very susceptible to natural hazards. Steep hillslopes combined with soft rocks (schists on the mountains; sands and sandstones in depressions), erodible soils, semiarid continental climate, and sparse vegetation cover give a high potential for soil erosion and landslides. For this reason, this study presents a multi-hazard approach to geohazard modeling on the national extent in the example of North Macedonia. Utilizing Geographic Information Systems, relevant data about the entire research area were employed to analyze and assess soil erosion and susceptibility to landslides and identify areas prone to both hazards. Using the Gavrilović Erosion Potential Method (EPM), an average value of 0.36 was obtained for the erosion coefficient Z, indicating low to moderate susceptibility to erosion. However, a significant area of the country (9.6%) is susceptible to high and excess erosion rates. For the landslide susceptibility assessment (LSA), the Analytical hierarchy process approach is combined with the statistical method (frequency ratio), showing that 29.3% of the territory belongs to the zone of high and very high landslide susceptibility. Then, the accuracy assessment is performed for both procedures (EPM and LSA), showing acceptable reliability. By overlapping both models, a multi-hazard map is prepared, indicating that 22.3% of North Macedonia territory is highly susceptible to erosion and landslides. The primary objective of multi-hazard modeling is to identify and delineate hazardous areas, thereby aiding in activities to reduce the hazards and mitigate future damage. This becomes particularly significant when considering the impact of climate change, which is associated with increased landslide and erosion susceptibility. The approach based on a national level presented in this work can provide valuable information for regional planning and decision-making processes.

Advancements in Technologies and Methodologies of Machine Learning in Landslide Susceptibility Research: Current Trends and Future Directions

Landslides are pervasive geological hazards that pose significant risks to human life, property, and the environment. Understanding landslide susceptibility is crucial for predicting and mitigating these disasters. This article advocates for a comprehensive review by systematically compiling and analyzing 146 relevant studies up to 2024. It assesses current progress and limitations and offers guidance for future research. This paper provides a comprehensive overview of the diverse challenges encountered by machine learning models in landslide susceptibility assessment, encompassing aspects such as model selection, the formulation of evaluation index systems, model interpretability, and spatial heterogeneity. The construction of an evaluation index system, which serves as the foundational data for the model, profoundly influences its accuracy. This study extensively investigates the selection of evaluation factors and the identification of positive and negative samples, proposing valuable methodologies. Furthermore, this paper briefly deliberates and compares classical machine learning models, offering valuable insights for model selection. Additionally, it delves into discussions concerning model interpretability and spatial heterogeneity issues. These research findings promise to enhance the precision of landslide susceptibility assessments and furnish effective strategies for risk management.

Optimization of SVR and CatBoost models using metaheuristic algorithms to assess landslide susceptibility

In this study, a landslide susceptibility assessment is performed by combining two machine learning regression algorithms (MLRA), such as support vector regression (SVR) and categorical boosting (CatBoost), with two population-based optimization algorithms, such as grey wolf optimizer (GWO) and particle swarm optimization (PSO), to evaluate the potential of a relatively new algorithm and the impact that optimization algorithms can have on the performance of regression models. The Kerala state in India has been chosen as the test site due to the large number of recorded incidents in the recent past. The study started with 18 potential predisposing factors, which were reduced to 14 after a multi-approach feature selection technique. Six susceptibility models were implemented and compared using the machine learning algorithms alone and combining each of them with the two optimization algorithms: SVR, CatBoost, SVR-PSO, CatBoost-PSO, SVR-GWO, and CatBoost-GWO. The resulting maps were validated with an independent dataset. The performance rankings, based on the area under the receiver operating characteristic curve (AUC) metric, are as follows: CatBoost-GWO (AUC = 0.910) had the highest performance, followed by CatBoost-PSO (AUC = 0.909), CatBoost (AUC = 0.899), SVR-GWO (AUC = 0.868), SVR-PSO (AUC = 0.858), and SVR (AUC = 0.840). Other validation statistics corroborated these outcomes, and the Friedman and Wilcoxon-signed rank tests verified the statistical significance of the models. Our case study showed that CatBoost outperformed SVR both in case the models were optimized or not; the introduction of optimization algorithms significantly improves the results of machine learning models, with GWO being slightly more effective than PSO. However, optimization cannot drastically alter the results of the model, highlighting the importance of setting up of a rigorous susceptibility model since the early steps of any research.

Mapping Landslide Risk in the United States and Puerto Rico

A new method provides highly accurate continental-scale landslide susceptibility maps that are being used in the aftermath of Hurricane Helene.