Forest Cover and Landslide Susceptibility Assessment Using a Machine Learning Approach in Northern Midland and Mountainous Region of Vietnam

As a critical predisaster warning tool, landslide susceptibility assessment is crucial in disaster prevention and mitigation efforts. However, earlier methods for assessing landslide susceptibility have often ignored the impact of similarities in geographical attributes, restricting their feasibility in regions with diverse characteristics. The geographical-optimal-similarity (GOS) model effectively captures similarity relations within geospatial data and can isolate region-specific landslide features, thus overcoming this challenge. Consequently, a landslide susceptibility assessment method was developed by integrating the information value (IV) model with the GOS model. Huangshan City in Anhui Province, China, was selected as the study region. This research used 11 remote sensing feature factors and 657 historical landslide points, combined with the IV model, to construct a dataset for landslide prediction and susceptibility assessment using the GOS model. The findings indicate that, compared to conventional methods such as random forest, logistic regression, and radial basis function classifier, the GOS model enhances the area under the curve (AUC) value by 2.81% to 8.92%, reaching 0.846. This demonstrates superior performance and confirms the effectiveness and accuracy of the method in landslide susceptibility assessment. Furthermore, compared to the basic-configuration-similarity (BCS) model, the GOS model increases the AUC value by 9.64%, achieving 0.846. This approach substantially diminishes the effects of historical data accuracy, revealing upgraded applicability in landslide susceptibility evaluations. Landslides in Huangshan City are primarily influenced by rainfall and vegetation cover. High-susceptibility zones are predominantly located in areas with high precipitation and low vegetation cover. In contrast, low-susceptible and non-susceptible zones are primarily found in flat areas with high vegetation cover and farther from fault lines. The majority of the study region lies within landslide-prone zones, with non-susceptible areas comprising only 12.43% of the total area. Historical landslides are largely concentrated in moderate- to high-susceptibility zones, accounting for 92.24% of all landslide occurrences. Landslide density increases with the susceptibility level, with a density of 0.15 landslides per square kilometre in high-susceptibility zones. This study brings forward a reliable strategy for establishing the spatial relationship between geographical attribute similarity and landslide susceptibility, bolstering the method’s adaptability across various regions.

Fast physically-based probabilistic modelling of rainfall-induced shallow landslide susceptibility at the regional scale considering geotechnical uncertainties and different hydrological conditions

The recent census in India revealed that India is now housing 17% of the world’s population, and India is on the way to become the most populated country. Landslides are an increasing concern in India due to the rapid population expansion in hilly and mountainous terrain. Landslides affect vast areas within India, in particular in the Himalayan chain in the North and Eastern part of the country and the Western Ghats in the Southwest. The Geological Survey of India (GSI) has been designated as the nodal agency for landslides by the Indian government, and they are responsible for landslide inventory, susceptibility and hazard assessment. Until recently their landslide susceptibility assessment was based on a heuristic approach using fixed weights or ranking of geofactors, based on guidelines of the Bureau of Indian Standards (BIS). However, this method is disputed as it doesn’t provide accurate results. This paper gives an overview of recent research on how the existing methods for landslide inventory, susceptibility and hazard assessment in India could be improved, and how these could be used in (semi)quantitative risk assessment. Due to the unavailability of airphotos in large parts of India, satellite remote sensing data has become the standard data input for landslide inventory mapping. The National Remote Sensing Center (NRSC) has developed an approach using semi-automatic image analysis algorithms that combine spectral, shape, texture, morphometric and contextual information derived from high resolution satellite data and DTMs for the preparation of new as well as historical landslide inventories. Also the use of existing information in the form of maintenance records, and other information to generate event-based landslide inventories is presented. Event-based landslide inventories are used to estimate the relation between temporal probability, landslide density and landslide size distribution. Landslide susceptibility methods can be subdivided in heuristic, statistical and deterministic methods. Examples are given on the use of these methods for different scales of analysis. For medium scales a method is presented to analyze the spatial association between landslides and causal factors, including those related to structural geology, to select the most appropriate spatial factors for different landslide types, and combine them using the multivariate methods. For transportation corridors a method is presented for quantitative hazard and risk assessment based on a landslide database. Deterministic methods using several dynamic slope-hydrology and slope stability models have been applied to evaluate the relation between land use changes and slope stability in a steep watershed. The paper ends with an overview how the susceptibility maps can be combined with the landslide databases to convert them into hazard maps which are subsequently used in (semi) quantitative risk assessment at different scales of analysis, and how the results can be used in risk reduction planning.

Nepal is facing the threat of landslides each year causing huge loss of lives and properties. Landslide prediction and susceptibility assessments help in identifying the potential zones of landslide occurrences and provide opportunities to treat them prior to their occurrence. Among different methods of landslide susceptibility mapping, the InfoVal method is one of the simple and useful methods. In this study, this method is used to study the landslide susceptibility in the Marin Khola Watershed within the Siwaliks of central Nepal as this area comprises of the weak geological formations that contribute to high potentialities of landslides, yet there are no studies for predicting landslides. A total of 217 landslides were taken for the study and they were divided into two groups: working landslides and validating landslides. 75% of these total landslides were selected as working landslides and the remaining 25% were selected for validating landslides. Spatial relationships of the landslide distribution with different causative factors including topographic factors, hydrologic factors, geological factors and landuse factors were employed and analyzed. The results depict that very high, high, moderate, low and very low susceptibility classes cover 1.15%, 49.93%, 30.17%, 11.48%, and 11.28% area, respectively. The Middle Siwaliks are most susceptible to landslides compared to the Upper Siwaliks and Lower Siwaliks. The accuracy values are found to be affected by the difference in the landslide characteristics and types occurring in the study area. The model accuracy remains at 66% and predictive accuracy at 75%.

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.

Abstract: The assessment of landslide susceptibility holds significant importance in disaster risk reduction. This study comprehensively examines the current research on landslide susceptibility from two aspects: the steps involved in landslide susceptibility assessment and modeling methods. Initially, we retrieved pertinent research articles, published between 2014 and 2023, and focused on “Landslide SensitivityAssessment” from the Web of Science database. Subsequently, we identified frequently occurring keywords in landslide susceptibility assessment studies employing ensemble learning methods during the past decade and created analytical charts. The standard methods for landslide inventory, evaluation indicators, and validation techniques were introduced along with their advantages and limitations. The shortcomings of eachmethod were identified, and potential future research directions were outlined. Finally, a detailed analysis of the use of ensemble methods in landslide susceptibility assessment was conducted; this is presented in several sections. The findings indicate that the advancement of ensemble learning methods has facilitated the development of landslide susceptibility assessment, rendering the landslide modeling process more efficient and accurate. In turn, this has enhanced the intelligence of models in landslide susceptibility research. The results of this study can help researchers understand the current conditions of landslide susceptibility research and provide a reference for subsequent research in this field.

Instability events are recurring phenomena in Southern Italy due to its geological history and tectonic-geomorphological evolution leading to the occurrence of several formations identified as Structurally Complex Formations (SCFs; Esu, 1977) in a territory mainly composed of densely populated areas also in mountainous and hilly regions. SCFs are clay-dominant terrains that, usually, give origin from very-slow to extremely-slow phenomena (Cruden and Varnes, 1996) with a long evolutionary history made up of multiple reactivations that makes difficult their identification, monitoring and susceptibility evaluation. The study has been carried out from point-wise (Bisaccia, Costa della Gaveta and Nerano cases) to wide areas (Palermo province case) where crops out SCFs as the Termini sandstones Formation (CARG, 2011), the Varicoloured Clays of Calaggio Formation (Ciaranfi et al., 1973), the Varicoloured Clays Unit (Mattioni et al., 2006) the Sicilide Unit (Vitale and Ciarcia, 2013 and references therein), the Numidian Flysch (Johansson et al., 1998) and the Corleone Calcarenites (Catalano R. et al., 2002). The aim of this thesis is to produce updated Landslide Inventory Maps and, whenever possible, Landslide Susceptibility Maps following a new approach during the landslide mapping and landslide monitoring stages. The Landslide Inventory Maps have taken into account the combination of geological, geomorphological, and stereoscopic surveys, as well as engineering geological investigations, namely conventional techniques. In addition innovative Advanced-Differential Interferometry Synthetic Aperture Radar (A-DInSAR) techniques have been used: the Coherent Pixels Technique – CPT (Mora et al., 2003; Blanco et al., 2008), the Intermittent Small BAseline Subset – ISBAS (Sowter et al., 2013) and the Ground-Based Synthetic Aperture Radar. Finally, the Weight of Evidence method (van Westen, 1993) has been chosen to generate the Landslide Susceptibility Maps only for the point-wise studies. In the case of Nerano (Province of Naples), the ISBAS analysis on ENVISAT images (for the period 2003-2010) has been carried out and compared with inclinometric and rainfall data. These have revealed several reactivations of a rotational slide + earth flow (Cruden and Varnes, 1996) that involves reworked clay olistostromes and limestone olistoliths inside the Termini sandstones Formation; even in recent years the landslide, despite many engineering works, has given evidence of a continuing activity. The results highlight a very slow movement in the detachment zone (<1 mm/yr), which assumes slightly higher values in the accumulation area (5 mm/yr). The Landslide Susceptibility Map confirms the high levels in the flow track and the accumulation area. In Bisaccia (Province of Avellino), a conglomeratic slab undergoes a Deep Seated Gravitational Slope Deformation (DSGSD; Pasuto and Soldati, 2013 and references therein) due to the bedrock consolidation, made of the Varicoloured Clays of Calaggio Formation. Here the CPT processing on ENVISAT images (covering the period between 2002 and 2010), displays a vertical displacement for the town center, suffering a progressively increasing velocity from the southern (4.2 mm/yr) to the northern (15.5 mm/yr) portion of the slab that localizes four different sectors. The pattern is confirmed from the building damage map. The landslides susceptibility reaches the highest values in the adjacent valleys and at the edges of each sector. Multiple datasets have been employed for the Costa della Gaveta case-study (Province of Potenza), these encompass: ENVISAT, TerraSAR-X and COSMO-SkyMed constellations together with Ground Based Synthetic Aperture Radar (GBSAR). The A-DInSAR data have been compared with stereoscopic analysis and the available rainfall and inclinometric data. The analysis allows for the identification of 16 landslides (complexes and earth flows; Cruden and Varnes, 1996) developed in the Varicoloured Clays Unit that show, according to all the existing instruments, velocities between 1.5 and 30 mm/yr. The western side of Costa della Gaveta slope is the portion which suffers the highest landslides susceptibility levels. In the Province of Palermo (northwestern Sicily) information deriving from A-DInSAR processing, specifically the ISBAS technique, have been focused on three subareas (Piana degli Albanesi, Marineo and Ventimiglia di Sicilia) for a total extension of 182 Km2 where standard A-DInSAR algorithms showed limitations due to the widespread presence of densely vegetated areas. The radar-detected landslides have been validated through field geomorphological mapping and stereoscopic analysis proving to be highly consistent especially with slow phenomena. The outcome has allowed to confirm 152 preexisting landslides, to detect 81 new events and to change 133 previously mapped landslides, modifying their typology, boundary and/or state of activity. The study demonstrates how a better knowledge of landslide development and their cause-effect mechanisms provided by new Earth Observation techniques is useful for Landslide Inventory and Susceptibility Maps. The research project has been carried out at the University of Naples "Federico II", including nine months (September 2013 – May 2014) spent in the United Kingdom, at the British Geological Survey under the supervision of Dr. Francesca Cigna and Dr. Jordan Colm and at the University of Nottingham (Department of Civil Engineering), under the supervision of Dr. Andrew Sowter where the ISBAS technique has been recently developed.

The accuracy of landslide susceptibility mapping is influenced by the quality of sample data, factor systems, and assessment methods. This study aims to enhance the representativeness and overall quality of the sample dataset through an effective sample expansion strategy, achieving greater precision and reliability in the landslide susceptibility model. An integrated interpretative framework for landslide susceptibility assessment is developed using the XGBoost-SHAP-PDP algorithm to deeply investigate the key contributing factors of landslides in karst areas. Firstly, 17 conditioning factors (e.g., surface deformation rate, land surface temperature, slope, lithology, and NDVI) were introduced based on field surveys, satellite imagery, and literature reviews, to construct a landslide susceptibility conditioning factor system in line with karst geomorphology characteristics. Secondly, a sample expansion strategy combining the frequency ratio (FR) with SBAS-InSAR interpretation results was proposed to optimize the landslide susceptibility assessment dataset. The XGBoost algorithm was then utilized to build the assessment model. Finally, the SHAP and PDP algorithms were applied to interpret the model, examining the primary contributing factors and their influence on landslides in karst areas from both global and single-factor perspectives. Results showed a significant improvement in model accuracy after sample expansion, with AUC values of 0.9579 and 0.9790 for the training and testing sets, respectively. The top three important factors were distance from mining sites, lithology, and NDVI, while land surface temperature, soil erosion modulus, and surface deformation rate also significantly contributed to landslide susceptibility. In summary, this paper provides an in-depth discussion of the effectiveness of LSM in predicting landslide occurrence in complex terrain environments. The reliability and accuracy of the landslide susceptibility assessment model were significantly improved by optimizing the sample dataset within the karst landscape region. In addition, the research results not only provide an essential reference for landslide prevention and control in the karst region of Southwest China and regional central engineering construction planning but also provide a scientific basis for the prevention and control of geologic hazards globally, showing a wide range of application prospects and practical significance.

Maize is an important crop in the northern midland and mountainous region of Vietnam with the area of 406.1 thousand hectares, accounting for 45.8% of the whole country. In order to exploit the potential and advantages for developing sustainable and highly valued maize production, the program on development of maize hybrids for production and economic growth in the northern midland and mountainous region of Vietnam in accordance with the Government policies and the Ministry of Agriculture and Rural Development’s directions on sustainable agricultural development strategy has been successfully implemented in which newly released maize hybrids developed by Maize Research Institute (Vietnam Academy of Agricultural Sciences) with high yield, good quality, good resistance and over-all advanced technologies were accordingly introduced in rainfed based maize areas of larger scale. The program partly contributed to improving the grain yield from 1.5 tons/ha in the early 1990s to 4.99 tons/ha in this region that played an important role not only in maize production development but also in national food security and environmental protection as well.

A multi-annual landslide inventory for the assessment of shallow landslide susceptibility – Two test cases in Vorarlberg, Austria

China-Pakistan Economic Corridor (CPEC) is a framework of regional connectivity, which will not only benefit China and Pakistan but will have positive impact on Iran, Afghanistan, India, Central Asian Republic, and the region. The surrounding area in CPEC is prone to frequent disruption by geological hazards mainly landslides in northern Pakistan. Comprehensive landslide inventory and susceptibility assessment are rarely available to utilize for landslide mitigation strategies. This study aims to utilize the high-resolution satellite images to develop a comprehensive landslide inventory and subsequently develop landslide susceptibility maps using multiple techniques. The very high-resolution (VHR) satellite images are utilized to develop a landslide inventory using the visual image classification techniques, historic records and field observations. A total of 1632 landslides are mapped in the area. Four statistical models i.e., frequency ratio, artificial neural network, weights of evidence and logistic regression were used for landslide susceptibility modeling by comparing the landslide inventory with the topographic parameters, geological features, drainage and road network. The developed landslides susceptibility maps were verified using the area under curve (AUC) method. The prediction power of the model was assessed by the prediction rate curve. The success rate curves show 93%, 92.8%, 92.7% and 87.4% accuracy of susceptibility maps for frequency ratio, artificial neural network, weights of evidence and logistic regression, respectively. The developed landslide inventory and susceptibility maps can be used for land use planning and landslide mitigation strategies.

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.

Introduction: Landslide is one of the most widespread geohazards around the world. Therefore, it is necessary and meaningful to map regional landslide susceptibility for landslide mitigation. In this research, landslide susceptibility maps were produced by four models, namely, certainty factors (CF), naive Bayes (NB), J48 decision tree (J48), and multilayer perceptron (MLP) models.Methods: In the first step, 328 landslides were identified via historical data, interpretation of remote sensing images, and field investigation, and they were divided into two subsets that were assigned different uses: 70% subset for training and 30% subset for validating. Then, twelve conditioning factors were employed, namely, altitude, slope angle, slope aspect, plan curvature, profile curvature, TWI, NDVI, distance to rivers, distance to roads, land use, soil, and lithology. Later, the importance of each conditioning factor was analyzed by average merit (AM) values, and the relationship between landslide occurrence and various factors was evaluated using the certainty factor (CF) approach. In the next step, the landslide susceptibility maps were produced based on four models, and the effect of the four models were quantitatively compared by receiver operating characteristic (ROC) curves, area under curve (AUC) values, and non-parametric tests.Results: The results demonstrated that all the four models can reasonably assess landslide susceptibility. Of these four models, the CF model has the best predictive performance for the training (AUC=0.901) and validating data (AUC=0.892).Discussion: The proposed approach is an innovative method that may also help other scientists to develop landslide susceptibility maps in other areas and that could be used for geo-environmental problems besides natural hazard assessments.

Landslides pose a significant threat to the safety of reservoirs, particularly those situated in canyon terrains. This study aims to enhance the safety and security of reservoir areas by proposing an integrated method for the automatic identification and assessment of landslides. By combining the SBAS-InSAR technique with spatial clustering analysis, we successfully delineated landslide areas and developed a new landslide susceptibility assessment model. This model operates independently of historical landslide inventory data. Based on the delineated landslide areas, we enhanced the information value model using the inverse tangent function, which was then integrated with Random Forest and Extreme Gradient Boosting methods for landslide susceptibility assessment. The identified landslides were validated through field tests, demonstrating a high degree of consistency with actual conditions. The results indicated that, in canyon-type reservoirs, aspect was a critical factor influencing landslide occurrence, with susceptibility being greater near water bodies. In model comparisons, the RF-NIV model outperformed, providing a more realistic representation of landslide susceptibility distribution. These findings offer valuable insights for landslide safety management in canyon-type reservoirs, such as those in Hekou Village and Baihetan.

The present study evaluated the landslide (LS) susceptibility using RBF net, Naïve-Bayes Tree, Random subspace, and Rotational forest advanced hybrid machine learning (HML) algorithm in landslide hazard-prone area of Kurseong, Darjeeling Himalaya, India. The locations of landslides were detected by field surveys. 352 LS coordinates have been derived, displayed as an LS inventory map to calibrate LS susceptibility models, and used to authenticate the models. 16 LCFs (landslide conditioning factors) were utilized to prepare LS susceptibility maps. The developed landslide models were validated using two statistical methods, i.e., the mean absolute error (MSE) and the root mean square error (RMSE) as well as the receiver operating characteristics (ROC), efficiency, and accuracy. The results of the accuracy measures (area under curve for RBF net = 85.76%, NB tree = 86.54%, Random sub = 87.29%; and Rotational for = 84.81%) revealed that all models have good potentiality to forecast the landslide susceptibility in the Kurseong region of Darjeeling Himalaya. The Random subset model achieved higher accuracy (ROC = 87.29%; MSE = 0.145; and RMSE = 0.118) than other used models. The research revealed that in the fellow land, plantation areas, sides of highways, and structural hills where elevation is above 1150 m and slope ranges from 26° to 69° the susceptibility to landslide is very high. The prepared landslide susceptibility maps can be helpful in introducing location-specific proper management strategies for reducing landslide hazards in Kurseong region of Darjeeling Himalaya.KeywordsHybrid machine learning modelsRBF netNaïve-Bayes TreeRandom subspaceRotational forestKurseong