Current Landslide Research Articles

Synthetic libraries of urban landslide simulations to identify slope failure hotspots and drivers across spatial scales and landscapes

AbstractRainfall-triggered landslides are most deadly in developing countries, and future urban sprawl and climate change could intensify existing risks. In these regions, enhancing efforts in urban landslide risk mitigation and climate change adaptation is crucial. Current landslide probability assessment methodologies struggle to support effective mitigation because they fail to represent local anthropogenic factors (e.g. informal housing) across space and time scales. To meet this challenge, we demonstrated in previous work that hillslope-scale mechanistic models representing such localised changes can be used to create synthetic libraries of urban landslides that account for both data and future scenario uncertainty. Here, we show how these libraries can become an explorative tool for researchers and stakeholders, allowing them to investigate slope stability variations across spatial scales and landscapes. Results highlight, for example, how the main slope instability drivers change according to the location (e.g., upper vs lower catchment), the landcover (e.g. forest vs urban) and the spatial scale analysed (e.g. at hillslope scale slope stability was mostly controlled by water table height, whereas at regional scale by slope geometry). Ultimately, we demonstrate that stochastic analyses can lead to a greater understanding of the system interactions and they can support the identification of mitigation strategies that perform well across spatial scales and uncertain scenarios. These strategies should be prioritised even if future conditions are unknown. This reasoning is shown on a data-scarce region with expanding informal housing. However, the same methodology can be applied to any urban context and with any mechanistic-based model.

Enhancing landslide inventory mapping through Multi-SAR image analysis: a comprehensive examination of current landslide zones

Enhancing landslide inventory mapping through Multi-SAR image analysis: a comprehensive examination of current landslide zones

Incorporating mitigation strategies in machine learning for landslide susceptibility prediction

This study proposes an approach that considers mitigation strategies in predicting landslide susceptibility through machine learning (ML) and geographic information system (GIS) techniques. ML models, such as random forest (RF), logistic regression (LR), and support vector classification (SVC) are incorporated into GIS to predict landslide susceptibilities in Hong Kong. To consider the effect of mitigation strategies on landslide susceptibility, non-landslide samples were produced in the upgraded area and added to randomly created samples to serve as ML models in training datasets. Two scenarios were created to compare and demonstrate the efficiency of the proposed approach; Scenario I does not considering landslide control while Scenario II considers mitigation strategies for landslide control. The largest landslide susceptibilities are 0.967 (from RF), followed by 0.936 (from LR) and 0.902 (from SVC) in Scenario II; in Scenario I, they are 0.986 (from RF), 0.955 (from LR) and 0.947 (from SVC). This proves that the ML models considering mitigation strategies can decrease the current landslide susceptibilities. The comparison between the different ML models shows that RF performed better than LR and SVC, and provides the best prediction of the spatial distribution of landslide susceptibilities.

Remote Sensing and GIS-Based Accuracy Assessment of LULC Map and Landslide Susceptibility Prediction for Meghalaya, India

ABSTRACT Through this study, a Landslide Susceptibility Map (LSM) has been developed for the Meghalaya state, India using an Analytical Hierarchy Process (AHP). According to a 2012 Geological Survey of India report, the annual average number of landslides in Meghalaya is nearly 30, which is due to a combination of mountains, steep slopes, and excessive rainfall, leading the state to suffer a huge loss of life and property from landslides. For effective management of the current landslide situation, information about prior landslides is needed. Therefore, the landslide inventory map is prepared with 380 previously occurred events. The Landslide inventory records were separated into training samples (70%) and testing samples (30%) for the purpose of validation. In this regard, the present study has 15 conditioning factors, i.e., slope, rainfall, elevation, relative relief, aspect, distance from the road, curvature, distance from the stream, LULC, lineament density, geomorphology, geology, NDVI, MSAVI, NDWI, which are used to develop susceptibility map. Classification and accuracy assessment of LULC is carried out with segregation as 77% vegetation, 16.4% range land, 3.1% built area, 2.8% crops, 0.4% waterbodies, and 0.3% others (bare land, flooded vegetation, etc.). The Kappa for LULC categorization is 0.92, which is quite satisfactory and suggests that the LULC categorization is reliable. The developed susceptibility map is classified into four different classes, low susceptibility (35%), moderate susceptibility (41%), high susceptibility (20%), and very high susceptibility (4%), and has been verified using physical and Receiver Operating Characteristics (ROC) techniques. Results show that anticipated susceptibility classes are in good match with previous landslide events. The prepared map is reliable and can be used for land-use planning of the state in the future.

Active landslide detection using integrated remote sensing technologies for a wide region and multiple stages: A case study in southwestern China

While significant progress has been achieved in utilizing remote sensing technologies for landslide investigation in China, there remains a notable gap in consolidating information on applicable conditions, application stages, and workflows across various remote sensing methodologies. This paper proposes a comprehensive framework for active landslide detection, incorporating multiple stages and data sources, successfully implemented in a vast region of southwestern China. Furthermore, detailed discussions are provided on the effects of the geometric distortion, land cover type, and various InSAR methods on the accuracy of active landslide identification results. Additionally, the paper delves into the advantages of integrated remote sensing technology in active landslide investigation, encompassing the assessment of current landslide activity status, precise delineation of boundaries, identification of different deformation stages, and determination of damage patterns. Through comprehensive analysis of multisource data, it enhances understanding of the active landslide process, ultimately contributing to the mitigation of casualties and property damage.

Evaluation of U-Net transfer learning model for semantic segmentation of landslides in the Colombian tropical mountain region

Landslides in tropical regions, like the Colombian Andean region, pose unique challenges due to factors such as intense rainfall, steep slopes, and complex terrains. Mapping historical and current landslide activity through inventory maps is essential in tropical mountainous regions. While satellite data is commonly used for mapping, it can be time-consuming and manual-intensive, limiting inventory availability. Deep Learning (DL) models, especially Convolutional Neural Networks (CNNs), have shown promise in remote sensing applications with High Resolution (HR) imagery, including landslide detection. Despite advancements, their use in this field is still relatively limited. This study assesses the effectiveness of U-Net model, for automated landslide detection using spectral data from optical satellite imagery (RGB bands), two DEM-derived geo-indices (slope and curvature), and two Synthetic Aperture Radar (SAR) layers (VV amplitude pre- and post-landslide event in May 2015) across three image models (3, 5, and 7 bands). Initially, data is combined into multiband images, and the model is trained in the “La Argelia” river basin in Colombia’s Pacific region. Subsequently, the model is tested in the “La Liboriana” river basin in the western Andean region. The landslide detection results within the inference area are validated by comparing them with the landslide inventory and segmentation results. The U-Net model demonstrates good performance (F1-score around 0.70) for landslide detection, as confirmed in various geographical settings. By utilizing DL models and combining high-resolution satellite imagery, topographical, and SAR data, a comprehensive space-time mapping of landslides can be achieved. This approach has the potential to greatly improve the accuracy and effectiveness of landslide mapping, offering a more holistic view of the temporal dynamics related to these natural hazards.

Mapping landslide susceptibility in the Debretabor-Alember road sector, Northwestern Ethiopia through geospatial tools and statistical approaches

This study aimed to locate areas along the Debretabor-Alember route segment in northern Ethiopia that are susceptible to landslides. Geospatial tools, specifically frequency ratios (FR) and information values (IV), were used to develop landslide susceptibility maps (LSMs). A comprehensive on-site investigation and analysis of Google Earth imagery were conducted, resulting in the detection and analysis of 89 landslides, including current and historical events. The dataset used for validation comprised 78% of the previously documented landslides, whereas the remaining 22% was used for training. Several factors were considered in this study to determine landslide susceptibility, including "slope, aspect, curvature, elevation, lithology, distance from streams, land use and cover, precipitation, normalized difference vegetation index (NDVI)", and the FR and IV models. Based on the results obtained using the FR approach, specific areas exhibited different levels of susceptibility, ranging from very low to moderately high, medium, high, and very high. These areas covered a total of 18.4 km2 (19.9%), 18.9 km2 (20.5%), 19.7 km2 (20.3%), 17.7 km2 (20%), and 17.7 km2 (19%), respectively. The LSMs generated by the IV model indicated multiple susceptibility classes in the study area, varying from very low to very high. These maps revealed that 18.4 km2 (19.8%), 18.8 km2 (20%), 18.9 km2 (19.5%), 18.8 km2 (20.5%), and 18.3 km2 (19.8%) of the area fell into these susceptibility classes. The landslide density indicator method was employed to validate the LSMs. The FR and IV models demonstrated that a significant proportion of confirmed past and current landslide records (72.16% and 73.86%, respectively) occurred in regions with a high or very high susceptibility to landslides. Overall, the IV model, which utilized latent variable structural modeling (LSM) in the independent variable model, outperformed the fixed effects regression model (FR).

Developing a probability-based technique to improve the measurement of landslide vulnerability on regional roads

Considering the uncertainty in vulnerability curves, this study employs a probabilistic technique to improve the current landslide vulnerability on roads. The uncertainty in current vulnerability was first elaborated, and slope failure and the damage caused to roads were investigated. To minimize uncertainty, subsequently, the probabilistic landslide intensity was proposed, and hired to calibrate the empirical landslide intensity. Finally, an improved vulnerability curve was developed and compared with the empirical vulnerability, in terms of the fluctuation band and residual value. Results show that the uncertainty of the improved vulnerability is significantly reduced compared with the empirical vulnerability model. In detail, the uncertainty reduction in landslide vulnerability is reflected not only in the decrease of the residual range, from 0.6 to 0.3, but in the narrowing of the fluctuation band. It provides a route for enhancing the accuracy of landslide vulnerability from a probabilistic view.

Investigation of landslide dam life span using prediction models based on multiple machine learning algorithms

A rapid and accurate prediction of a landslide dam’s life span is of significant importance for emergency geological treatment. However, current prediction models for the state of a landslide dam are based solely on geomorphological indexes, and do not take into consideration attribute properties such as landslide types, trigger factors, and dam types. This study investigates the relationships between a landslide dam’s geometry and the capacity of the barrier lake and proposes fitting models, which supplement the current landslide dam database. Subsequently, six predictive models for landslide dam life span are established, utilizing machine learning algorithms such as logistic regression, k-nearest neighbors, support vector machine, Naïve Bayes, decision tree, and random forest, which consider five factors, including geometry parameters and attribute properties. The performances of these six models are analyzed and compared to a typical prediction model, the dimensionless blockage index (DBI). The results suggest that the models established in this study not only have a consistent absolute accuracy as the DBI model, but also overcome the disadvantage that a large number of cases cannot be judged by the DBI model. Among the formulated machine learning models, the random forest model exhibits the highest absolute accuracy (89%), lowest error rate (7%), lowest false alarm rate (15%), and no uncertainty rate. Additionally, three renowned landslide dams, namely the Costantino, Hsiaolin, and Baige landslide dams, are analyzed to illustrate the applicability of the established machine learning models. The study results provide essential guidance for the predictions and emergency geological treatments of landslide dam disasters.

Machine learning for high-resolution landslide susceptibility mapping: case study in Inje County, South Korea

Landslides are a major natural hazard that can significantly damage infrastructure and cause loss of life. In South Korea, the current landslide susceptibility mapping (LSM) approach is mainly based on statistical techniques (logistic regression (LR) analysis). According to previous studies, this method has achieved an accuracy of approximately 75.2%. In this paper, we expand upon this traditional approach by comparing the performance of six machine learning (ML) algorithms for LSM in Inje County, South Korea. The study employed a combination of geographical data gathered from 2005 to 2019 to train and evaluate six algorithms, including LR, Gaussian Naive Bayes (GNB), Support Vector Machine (SVM), K-Nearest Neighbors (KNN), Random Forest (RF), and Extreme Gradient Boosting (XGB). The effectiveness of these models was measured by various criteria, such as the percentage of correct classification (PCC) score, F1 score, and Kappa score. The results demonstrated that the PCC and F1 scores of the six models fell between [0.869–0.941] and [0.857–0.940], respectively. RF and XGB had the highest PCC and F1 scores of 0.939 and 0.941, respectively. This study indicates that ML can be a valuable technique for high-resolution LSM in South Korea instead of the current approach.

An Overview of Landslide Management: Scope, Difficulties, Limitations with Future Directions and Opportunities

This study explores the scope, difficulties, limitations, and future directions of landslide management. Landslides pose significant threats to human lives, infrastructure, and the environment, necessitating effective management strategies. The study begins by introducing the topic of landslide management and its importance in mitigating landslide risks. It then delves into the scope of landslide management, including activities such as early warning systems, hazard mapping, and slope stabilization. The study highlights the various challenges faced in landslide management, such as limited resources, technical expertise, and data availability. Additionally, it discusses the limitations of current landslide management practices, emphasizing their inability to completely prevent landslides and the reliance on available technologies. Furthermore, the study presents case studies to provide real-world examples and insights into landslide management approaches. Future directions and opportunities in landslide management are explored, including the integration of advanced technologies, community-based approaches, and improved risk assessment techniques. The study concludes by emphasizing the need for a holistic approach that combines scientific expertise, community engagement, and policy support to achieve effective landslide management and reduce the impacts of these natural hazards.

Prompt Quantitative Risk Assessment for Rain-Induced Landslides

An extreme rainstorm can cause thousands of landslides and kill hundreds of people. In the changing climate, fatal rainstorms become more frequent and intense. The current landslide emergency management evaluates hazard intensities but lacks key information on likely consequences. This study presents a novel prompt quantitative risk assessment method for rain-induced landslides. The proposed method automatically generates a one-page risk assessment report within minutes to support effective risk communication, resource allocation, and emergency response. The propagation of uncertainties is quantified in a scientific probabilistic framework. The proposed method is tested using 83 major rainstorms during 1995–2016 in Hong Kong. The method accurately predicts the number of affected buildings and the number of potential fatalities and identifies rainstorms that can trigger fatal landslides. The proposed method contributes to the advancement of landslide emergency management from hazard-informed to risk-informed, which will significantly enhance societal resilience and facilitate climate change adaptation.

Research on prediction of slope displacement based on a weighted combination forecasting model

With the continuous development of slope engineering, people have put forward new requirements for slope safety. Slopes widely exist both above-ground and underground engineering. Landslide prevention is crucial to building and personal safety, and predicting the deformation and evolution trend of landslides is an important part of it. Aiming at the limitation that most of the current landslide displacement prediction models belong to a single model, this paper proposes a weighted combination prediction model to predict the displacement of the slope. Taking the measured displacement of the Three Gorges slope as an example, this paper firstly uses Grey-Marcof, GA-BP neural network and ARIMA model for prediction, and then combines the prediction results of the above three models by the reciprocal variance weighted method, and finally the linear combined prediction model of the slope displacement is established to complete the final prediction. The prediction results show that, compared with the minimum RMSE value of 0.3080 and the maximum RMSE value of 2.1379 predicted by the three methods alone, the combined prediction with RMSE value of 0.1249 has the best prediction accuracy. Moreover, compared with the average relative error of 1.46%, 1.49%, 4.32% and 1.24% of the prediction results by the single method of the authors who provided the data, the prediction results of the combined prediction model also showed a significant improvement in accuracy with the average relative error of 0.51%. Therefore, the weighted combination prediction model can further improve the slope displacement prediction accuracy, and has good scientificity and feasibility.

Performance evaluation of a real-time high-precision landslide displacement detection algorithm based on GNSS virtual reference station technology

Global navigation satellite system (GNSS) technology is one of the most effective landslide displacement monitoring methods, but most of the current landslide research based on GNSS is focused on long-term landslide monitoring, but short-term landslide monitoring is particularly important for the early warning of landslides. The main purpose of this paper is to extend the existing GNSS positioning model and design a set of different landslide displacement detection algorithms, which can guarantee high-precision displacement detection and give consideration to real-time, mainly for short-term landslide displacement, and can also give consideration to long-term landslide monitoring. In addition, the traditional layout mode of “one reference station corresponds to one landslide” has many problems, such as high cost, low utilization rate of reference station and excessive reliance on single reference station, which restrict the large-scale promotion and application of GNSS in landslide monitoring. Based on the existing VRS processing method and the displacement detection algorithm presented in this paper, the service effect of virtual reference station and physical reference station is compared and analyzed in detail, and the feasibility of virtual reference station replacing physical reference station is demonstrated. Experimental results show that the positioning accuracy of real time kinematic (RTK) is slightly worse than that of physical reference station due to the existence of atmospheric residual errors in the regional reference station network when VRS replaces physical reference station for displacement detection. However, after processing the real-time sliding window model, it significantly suppressed the influence of all kinds of observation noise and gross error, and the solution accuracy of the output sliding window period was not much different from that of the physical reference station, basically at the sub-centimeter level, which could fully meet the needs of centimeter and sub-centimeter level landslide displacement detection. In addition, the “short and long period mutual checking method” proposed in this paper can realize the detection and early warning of the rapid displacement type in a short period by reasonably adjusting the window length of the short and long period. The “ epoch difference method” can realize the detection and early warning of the sudden displacement in a few epoch. In practical application, the combined detection of the two methods can satisfy the real-time detection and early warning of short-term landslide displacement. In addition, by storing the solution positions of each long period sliding window period, the periodic position comparison can be carried out to judge the displacement changes of the monitoring body, which is equivalent to the usual long-term landslide monitoring strategy.

Spatiotemporal Evolution Pattern and Driving Mechanisms of Landslides in the Wenchuan Earthquake-Affected Region: A Case Study in the Bailong River Basin, China

Understanding the spatiotemporal evolution and driving mechanisms of landslides following a mega-earthquake at the catchment scale can lead to improved landslide hazard assessment and reduced related risk. However, little effort has been made to undertake such research in the Wenchuan earthquake-affected region, outside Sichuan Province, China. In this study, we used the Goulinping valley in the Bailong River basin in southern Gansu Province, China, as an example. By examining the multitemporal inventory, we revealed various characteristics of the spatiotemporal evolution of landslides over the past 13 years (2007–2020). We evaluated the activity of landslides using multisource remote-sensing technology, analyzed the driving mechanisms of landslides, and further quantified the contribution of landslide evolution to debris flow in the catchment. Our results indicate that the number of landslides increased by nearly six times from 2007 to 2020, and the total volume of landslides approximately doubled. The evolution of landslides in the catchment can be divided into three stages: the earthquake driving stage (2008), the coupled driving stage of earthquake and rainfall (2008–2017), and the rainfall driving stage (2017–present). Landslides in the upstream limestone area were responsive to earthquakes, while the middle–lower loess–phyllite-dominated reaches were mainly controlled by rainfall. Thus, the current landslides in the upstream region remain stable, and those in the mid-downstream are vigorous. Small landslides and mid-downstream slope erosion can rapidly provide abundant debris flow and reduce its threshold, leading to an increase in the frequency and scale of debris flow. This study lays the foundation for studying landslide mechanisms in the Bailong River basin or similar regions. It also aids in engineering management and landslide risk mitigation under seismic activity and climate change conditions.

SAR data and field surveys combination to update rainfall-induced shallow landslide inventory

The Campania region has been recurrently hit by severe landslides in volcanoclastic deposits. The city of Naples, and in particular the Camaldoli and Agnano hills (Phlegraean Fields), also suffered several landslide crises in weathered volcanoclastic rocks as a consequence of intense rainfalls or wildfires. To identify slope failures phenomena occurred in the winter season 2019–2020 an innovative procedure has been proposed. The purpose of this procedure is to highlight areas where major land cover changes occurred within our area of study, which can be potentially related to mass movements. The amplitude of spaceborne SAR images has been exploited for the change detection analysis and the output derived from the segmentation procedure has been compared with field observations. The amplitude-based method has been already applied in the detection of landslides, but never on the event with limited extensions, such as for this application. The achieved outcomes allowed the mapping of 62 new landslides that have been used to update the current landslide inventory database. This type of information is expected to help decision-makers with land planning and risk assessment.

Three-Dimensional Measuring Device and Method of Underground Displacement Based on Double Mutual Inductance Voltage Contour Method.

Landslide is a very common and destructive geo-hazard, and displacement monitoring of it is integral for risk assessment and engineering prevention. Given the shortcomings of current landslide displacement monitor technologies, a new three-dimensional underground displacement monitoring technology is proposed based on the double mutual inductance voltage contour method. The underground displacement measuring device mainly consists of an information processing unit and sensing array, connected by power and RS-485 communication lines. An underground displacement measurement model to convert the double mutual inductance voltages and the inter-axis angle into the relative displacement between adjacent sensing units is established based on the interval-interpolation and contour-modeling. Under the control of the information processing unit, the relative displacement between any two adjacent sensing units can be calculated through the underground displacement measurement model, so as to obtain the total displacement from underground depth to surface, and the measurement data can be further sent to the Internet of things cloud platform through the 4G module; thus the remote real-time monitoring of underground displacement three-dimensional measurement for the rock and soil mass from underground depth to the surface is realized. The measurement model is verified by building an experimental platform to simulate the underground displacement of rock and soil mass. The experimental results show that for each measuring unit, when the horizontal displacement and vertical displacement are within the measurement range of 0–50 mm, the maximum measurement error will not exceed 1 mm, which can meet the accuracy requirements of underground displacement monitoring of landslide.

Modelling reconstruction and boulder size-frequency distribution of a young (

We focus on a young (~ 4.5 Ma), 3.4 km long landslide located in the floor of Simud Vallis, Oxia Palus Quadrangle of Mars. By making use of a 2 m-scale HiRISE DEM we reconstruct the terrain surface before the landslide and in doing so we estimate the release and deposition heights and volumes related to the different stages of the landslide. Using the r.avaflow software we simulate the mass movement as a multi-stage event, and obtain simulated deposits that are both spatially and longitudinally comparable to the current landslide deposits. Through two 0.25 m-scale HiRISE images we identify and manually count >130,000 boulders that are located along the landslide, deriving their size-frequency distribution and spatial density per unit area for boulders with an equivalent diameter ≥1.75 m. Our analyses reveal that the distribution is of a Weibull-type, suggesting that the rocky constituents fractured and fragmented progressively during the course of the mass movement, consistent with our proposed two-stage model of landslide motion.

Historical dynamics of landslide risk from population and forest-cover changes in the Kivu Rift

Human activity influences both the occurrence and impact of landslides in mountainous environments. Population pressure and the associated land-use changes are assumed to exacerbate landslide risk, yet there is a lack of statistical evidence to support this claim, especially in the Global South where historical records are scarce. In this work, we explore the interactions between population, deforestation and landslide risk in the Kivu Rift in Africa. To do so, we develop a holistic landslide risk model that evaluates 58 years of population and forest-cover trends. We show that the current landslide risk in the eastern Democratic Republic of the Congo (DRC) is twice as high as in neighbouring Rwanda and Burundi. Congolese households, on average, populate more hazardous terrain, probably as a result of conflicts and economic pull factors such as mining. Moreover, the recent large-scale deforestation of primary rainforest in the DRC has considerably exacerbated the landslide risk. Our analysis demonstrates how the legacy of deforestation, conflicts and population dynamics is reflected in the landslide risk in the Kivu Rift. A mapping study covering over 50 years finds that landslide risk is much higher in the Democratic Republic of the Congo, due to more recent deforestation and more people living in more susceptible areas, than in similar landscapes in neighbouring countries.

Assessment of landslide susceptibility zonation using frequency ratio and statistical index: a case study of Al-Fawar basin, Tartous, Syria

Landslide is one of the most morpho-dynamic patterns of slopes in the coastal region of Syria. The assessment of landslide susceptibility is a complex procedure because of limited data and lack of related literature in Syria. Therefore, the current paper explored the landslide susceptibility in Al-Fawar basin, based on remote sensing data in the GIS environment. Inventorization of the current landslide was carried out as points utilizing satellite images, fieldwork, and photographs. Thirteen dominant factors of landslide including slope angle, slope aspect, curvature, plan curvature, profile curvature, elevation, proximity to faults, proximity to streams, proximity to roads, NDVI, rainfall, lithology, and TWI were utilized for landslide susceptibility mapping. In this study, two bivariate statistical models, i.e., frequency ratio (FR) and statistical index (SI), were used for spatial calibration between the resulting inventory map and the geofactors. The quality of models performance in mapping the landslide susceptibility was assessed utilizing receiver operating characteristic (ROC) curves. The prediction accuracy curves showed that the landslide susceptibility map using FR model has the greatest prediction accuracy (AUC = 0.824) compared with SI model that showed 0.801 of AUC. Identically, the success rate curve graph showed that the AUC for FR and SI models was 0.799 and 0.778, respectively. Although the susceptibility assessment utilizing the FR model produced higher accurate output, the validation results indicate that the two models used in this study can be encouraging approaches for the assessment of landslide susceptibility hazard in the Syrian coastal basin, thus fruitfully creating spatial mitigation and conservation measures.