Search

Due to their practicability, simple and data-driven empirical models have been extensively developed and applied in practical engineering to predict the run-out distance of landslides. However, the definition of the most appropriate empirical model for specific landslide data is rarely discussed. Moreover, the empirical model is subjected to the high variability of landslide data, which should be quantified into the model to provide more reliable predictions. As such, we propose in this paper, a simple, practical, and probabilistic run-out distance prediction method based on an empirical model and Bayesian method. This method was implemented with a regional landslide inventory compiled from 34 loess landslides in Heifangtai Terrace, Gansu Province, China. In this method, we performed a Bayesian model selection to determine the most appropriate empirical model for the compiled database among the possible candidate models adapted from previous literature. Considering the high variability of data, unknown parameters of the empirical model are regarded as random variables, and their posterior distributions are obtained by Bayesian updating with the compiled database. Then, we developed the probabilistic run-out prediction model to evaluate the run-out distance exceedance probability of landslides based on the most appropriate model and its associated posterior random variable information. We utilized data from two recent landslides that occurred in Heifangtai Terrace to validate the performance of the proposed model. In addition, we produced a run-out distance exceedance probability curve using the proposed method for a potential landslide in Heifangtai Terrace, in which the sliding volume interval is estimated using the sloping local base level (SLBL) method. In general, this study presents a practical method for landslide run-out distance analyses within a probabilistic framework, aiming to provide support for risk-based decisions. [ABSTRACT FROM AUTHOR]

As the global climate warms, the frequency of soil slope failures induced by snowmelt is gradually increasing, especially loess landslide disasters in seasonally frozen areas. Landslide disasters in seasonally frozen areas pose a serious threat to human lives and engineering constructions and are gradually drawing greater public attention around the world. However, the mechanism that regulates deformation and failure processes, resulting in snowmelt-driven landslides, remains elusive. The Yili River Valley in Xinjiang, China, is an ideal location for exploring the mechanism of snowmelt-driven landslides. Based on detailed field surveys, remote sensing image identification, meteorological data analysis and loess characteristic tests, this paper takes a representative high-level loess landslide in north-west China as a case in order to explore its evolution history, movement process and resurrection mechanism. It was found that this loess slope had suffered two large-scale sliding failures and different degrees of inherited slope deformation had been found between the two slope failures. The 32 surface cracks identified have a total length of 3,505.10 m, and the total area of the landslide was found to be 135,462m2. The average thickness of the sliding body was about 30 m, and the volume was approximately 504,000 m3. The actual movement time lasted for 32 s with an average moving speed of about 15 m/s. Human grazing activities dominate the formation and development of slope surface cracks in the early stage. The strong water sensitivity of distinctive loess controls soil strength deterioration in the slip zone. Rapid snow removal and infiltration, driven by an abnormal temperature rise in Spring, is the most important triggering factor for slope deformation, evolution and failure. As regional grazing activities increase and global warming intensifies, the potential for resurrection landslides will increase. The results provide essential information for a comprehensive understanding of early warning systems and risk assessment for snowmelt-triggered landslides in cold areas. [ABSTRACT FROM AUTHOR]

Due to the huge losses caused by landslides, landslide deformation monitoring has become essential. The change of the tilting angle of the slope surface can reflect the instability process of the landslide. However, there are few studies on the relationship between the instability process of landslides and the tilting angle of the slope surface on the time scale. Therefore, there are even fewer cases in using the inclination angle of the landslide as an early warning method in practice. On August 24, 2018, a wireless tilt sensor network was deployed in the Heifangtai area of Gansu Province in China, to monitor the inclination state of loess landslides to study the relationship between the tilting angle of the slope surface and the landslide instability process. At 4 o'clock on October 5, 2019, the wireless tilt sensor network mentioned above detected a new loess fall landslide. Through the analysis of the monitoring data from the tilting sensor nodes named T3 and T4, we found that the tilting angle of the landslide surface in the pre-sliding stage has significant change compared with the stable stage, which directly verifies that the inclination data can be used for landslide warning. At the same time, we inputted the monitored inclination data into the inclination warning model proposed by Xie et al. (Landslides 17:301–312, 2020) for verification and found that the trend of inclination changes matches the model well. However, the forecast time error is large. Therefore, this paper optimizes the inclination early warning model proposed by Xie. Using the optimized early warning model, the predicted landslide occurrence time based on the data of the T3 node was 0.517 days, which was about 8 h later than the actual landslide that occurred on October 4, 2019, which is an improvement of forecasting. This case serves as an on-site direct evidence to further verify the feasibility of using the landslide body inclination state as a landslide monitoring and early warning parameter which can be promoted and used in practice. [ABSTRACT FROM AUTHOR]

Landslide disasters are associated with severe losses on the Loess Plateau of China. Although early warning systems and susceptibility mapping have mitigated this issue to some extent, most methods are qualitative or semi-quantitative in the site-specific range. In this paper, a quantitative spatial distribution model is presented for site-specific loess landslide hazard assessment. Coupled with multi-temporal remote sensing images and high-precision UAV cloud point data, a total of 98 loess landslides that have occurred since 2004 on the Heifangtai terrace were collected to establish a landslide volume-date and retreating distance database. Eleven loess landslides are selected to construct a numerical model for parameter back analysis, and the accuracy of the simulation results is quantitatively evaluated by the centroid distance and overlapping area. Different volumes and receding distance rates of landslides are fitted to determine the relationship between cracks and potential volume, and different volumes and parameters are combined to simulate the spatial distribution of potential loess landslides. The results of this study reveal that landslide volumes mainly range between 1 × 103 and 5 × 105 m3, and the historical occurrence probability reaches 0.551. The optimal parameters are estimated by the maximum likelihood method to obtain a uniform distribution parameter value probability model, and the results show that the error of the estimated length within a range of 0.05 from the optimal parameter does not exceed 15%. In the selected slope slide case, farmland near the toe of the slope primarily includes exposed hazards with probabilities greater than 0.7. This work provides a useful reference for local disaster reduction and a theoretical methodology for hazard assessments. [ABSTRACT FROM AUTHOR]

In March 15, 2019, a small-scale loess landslide occurred at Zaoling, Shanxi Province, China, destroying several buildings and killing 20 people. Based on a digital orthophoto map (DOM) and digital elevation model (DEM) derived from Dajiang Pro4 UAV aerial images, coupled with field investigations, this paper describes the geometric features of this slope failure and analyzes its possible inducing factors. Results show that this landslide has a circular chair shape with a sliding direction of 355°. The total projected area of the landslide is 11,800 m2, in which the source is 9880 m2, and the main scarp is 1995 m2. The volume of the source material and loose deposit debris are about 10 × 104 m3 and 12 × 104 m3, respectively. Our research suggests that the inducing factors of this event may include the low strength and structure of the loess under a wet condition, special landforms of the slope, and human activities such as water discharge from a nearby bath center building and coal mining. [ABSTRACT FROM AUTHOR]

Currently, the theory and methodology of digital terrain analysis (DTA) has been well developed. However, this technique has not been widely applied in the research of loess landslides in China. This study investigated the application of DTA on loess landslides with the high-resolution terrain data obtained from low-cost unmanned aerial vehicles (UAVs). Taking a high-speed and long-runout landslide occurring on the Bailu Loess Tableland, a typical landform type on the Loess Plateau, as an example, we illustrated the fundamental characteristics and spatial patterns of the landslide from various perspectives and performed hydrology analysis, geomorphic change detection, hypsometric integral (HI) and stability analysis, morphology analysis, and structure analysis. The results prove that the DTA methodology cannot only advance understanding of the geomorphology and structure of landslides and detect geomorphic change but also reveal the evolution principles of landforms and demonstrate unique advantages in the prediction of the internal stability of landslides. In conclusion, the DTA methods adopted in this paper are useful to better understand loess landslide and its relationship with geomorphologic evolution. [ABSTRACT FROM AUTHOR]

Loess landslides pose significant physical and environmental threats to communities of the Heifangtai Terrace in Gansu Province, China. Estimating potentially affected areas using dynamic numerical simulations provides a foundation for risk assessment and management. However, current studies often use rheological input parameters derived from the back analysis of an individual historical landslide, which requires that the potential landslide is quite similar to the historical landslide to guarantee the accuracy of predictions. We collected data from 20 loess flowslides that occurred from 2015 to 2019 in the Heifangtai Terrace. A friction model embedded in MassFlow was used to simulate run-out behavior. Accumulation depths at multiple control points were used as observation information. A multi-objective optimization back analysis was used to calibrate optimal rheological parameters for each flowslide, yielding a region-specific database of 20 datasets for statistical analyses. Finally, two flowslides that occurred in 2019 were employed as potential landslides to conduct probabilistic run-out predictions based on the database. These predictions agreed well with the observed accumulation areas in both cases, thus validating our method. This database will be useful for future landslide run-out predictions in this area. [ABSTRACT FROM AUTHOR]

Loess landslides are widely distributed in Northern China and pose a significant threat to human life, natural resources, and infrastructure in mountainous regions. Accurate segmentation and measurement of loess landslides is crucial to documenting their occurrence and extent and investigating the distribution, types, and patterns of slope failures. The measurement of landslides also assists in assessing their susceptibility and risk. Herein, a novel loess landslide segmentation and measurement framework based on deep learning is proposed. Multiple experts label the ground-truth landslide regions, and simultaneous truth and performance level estimation (STAPLE) masks are generated. The U-Net segmentation algorithm is trained using a supervised approach to segment the loess landslide region. The fully connected conditional random field is integrated into the U-Net to further optimize the segmentation quality. In the final step, the predicted landslide boundaries are visualized, and the diameters (e.g., length and width) of the segmentation outcome are computed simultaneously. Four state-of-the-art segmentation algorithms are selected for the comparative analysis. The computational results demonstrate that the proposed framework outperforms all the other algorithms tested in terms of segmentation accuracy and boundary errors. The results verify the advantages of using STAPLE and U-Net integrated with a conditional random field. [ABSTRACT FROM AUTHOR]

On October 5, 2021, a small-scale catastrophic loess landslide occurred in Niangniangba town, Tianshui City, Gansu Province in China. The landslide with a volume of approximately 15,360 m3 buried 2 houses and damaged another 2 houses. On the basis of field investigation, unmanned aerial vehicle photography, and numerical simulation, the characteristics and dynamic processes of the Zhongzhai landslide are analyzed, and the potential destructive zone of the unstable loess mass is preliminarily assessed. The Zhongzhai landslide is a typical rainfall–reactivated flow-type loess landslide, and its failure mode is creep-tensile type. The upper sliding mass is the typical retrogressive failure, and the lower sliding mass shows thrust load caused failure. The runout simulation results show that the entire sliding process of Zhongzhai landslide lasted for approximately 35 s, with a maximum velocity of 12 m/s. The maximum travel distance of the sliding mass is 130 m. In addition, the presence of masonry buildings has a great influence on the sliding distance. Simulation results of potential destructive zone of the unstable loess mass show that the travel distance of the unstable loess mass is approximately 250–295 m, the thickness of the deposited mass is approximately 2.7–3.3 m in the potential destructive zone, and the maximum velocity is 28 m/s. This work contributes to improving the understanding of evolution law, dynamic processes, and risk prevention of this kind of landslide in loess plateau areas. [ABSTRACT FROM AUTHOR]

According to local villagers, the main Beiguo landslide on the Heyang Loess Tableland in China occurred several decades ago (the specific time is unknown). After the initial occurrence of the main landslide, there was no major activity for several decades until December 17–22, 2017 (2 months after the rainy season), when the main landslide was completely reactivated. One of the interesting findings of our field survey is that there are two erosion gullies and numerous loess caves (groups) developed at the top of the Beiguo landslide, which are generally considered to be the dominant channels by which rainwater or surface runoff enters the deep underground space. In order to better understand the process and triggering factors of the main landslide's reactivation, a multidisciplinary investigation including satellite remote sensing, interferometric synthetic aperture radar (InSAR), an unmanned aerial vehicle (UAV) survey, a terrestrial laser scanning (TLS) survey, an electrical resistivity tomography (ERT) survey, laboratory experiments, real-time monitoring using an inclinometer, and field investigations of the Beiguo landslide was conducted. We believe that the reactivation of the main Beiguo landslide may have been occurring for more than 10 years, or perhaps even longer. Since 2011, there have been several signs of local activity in the main landslide body, and the last accelerated stage of the entire deformation process occurred in 2017. The development and existence of loess caves directly destroyed the stability of the slope, and the rainfall and surface runoff quickly entered the interior of the main landslide. The influence of the loess caves on the landslide's reactivation was long-term and continuous. The results of this study suggest that the several days of abundant and continuous rainfall in the rainy season in 2017 were the last direct trigger of the final reactivation of the main landslide. The secondary landslide significantly changed the original slope, including the aspect of the slope profile's morphology, the slope gradient, relief, erosion, and accumulation, which accelerated the evolution of the tableland landform. This study enhances our understanding of the reactivation behavior of loess landslides, and also provides important support for the monitoring and prevention of reactivated landslides in the future. [ABSTRACT FROM AUTHOR]

The Heifangtai terrace, in Northwest China, is a typical area where loess landslides have been induced by agricultural irrigation, and many of the landslides are prone to reactivation. However, the spatiotemporal evolution and hydrological-triggering mechanisms of loess landslide reactivation are not well understood. In this research, multiple remote sensing (SBAS-InSAR, TLS, and optical remote sensing), integrated with time-lapse ERT (tl-ERT) imaging, was used to monitor the post-failure evolution of the Luojiapo landslide in Heifangtai during the period of May 2015 to Nov. 2020. Pronounced temporal and spatial differences in the deformation and hydrological evolution of landslides after sliding were observed. The largest displacement rates occurred in the landslide source area, and the lateral extension of the landslide source area caused by spatial differences in reactivation is an important feature of landslide evolution. In the landslide area, the groundwater table (GWT) decreased at first ascribed to the spring hole caused by the exposure of the GWT after sliding and then increased due to the subsequent continuous irrigation, and the lag time of the GWT response to irrigation decreased significantly. Spatial differences in GWT evolution are one of the main causes of spatial differences in landslide reactivation, and reactivation was more likely to occur where the GWT fluctuated at a high level. The GWT also fell with local reactivation. Our findings highlight the potential for obtaining internal and external spatiotemporal information of loess landslide evolution using multiple remote sensing integrated with tl-ERT. Our results also help to understand the reactivation process of irrigated loess landslides and provide a reference for the monitoring and early warning of such landslides. [ABSTRACT FROM AUTHOR]

Frequent occurrence of landslides in the Chinese Loess Plateau has seriously threatened the urban and rural development. Heifangtai loess platform was selected as a case study for the current research. The platform surface is being largely irrigated since 1968. About 70 landslides of different types have occurred in the region including loess slides, loess flows, combined loess, and bedrock landslides along/across bedding planes. The loess slides occurred in the unsaturated soil and have a shallow failure depth. The loess flows initiated at the contact between loess and underlying clay and can travel a distance of more than 300 m at a speed of up to 5 m/s. Controlled by the topography of the underlying clay layer, the loess flows mainly occurred along the eastern margins of Heifangtai. The combined loess and bedrock landslides along bedding planes typically travel a distance of 30-60 m at a speed of 0.002-0.005 m/s with a volume of up to several million cubic meters. This type of landslide commonly occurred along the southern margins of Heifangtai, where the slope aspects are close to the inclination direction of the bedding planes of underlying bedrocks. Monitoring of perched water table in loess layer at the northern margin of Heifangtai shows the water table is rising at a speed of about 18 cm/year. [ABSTRACT FROM AUTHOR]

Landslides are catastrophic natural hazards that often lead to loss of life, property damage, and economic disruption. Image-based landslide investigations are crucial for determining landslide susceptibility and risk. In practice, satellite images have been widely utilized for such investigations; however, they still require significant labor and time resources. In this study, we propose an image-based two-phase data-driven framework for detecting and segmenting landslide regions using satellite images. In phase I, an object detection algorithm, Faster-RCNN, is trained to detect the landslide location within the large-scale satellite images. The bounding boxes of each landslide location are proposed and visualized. In phase II, we crop the satellite images into small images using the location information of the bounding boxes. Next, we use a boundary detection algorithm to identify the boundary information of each detected loess landslide to strengthen the segmentation performance. Finally, we improve the architecture of the segmentation U-Net by integrating additional inception blocks with dilation to enhance the landslide segmentation performance. A total of 150 local loess landslide occurrences in northern China are selected as our case study to validate the effectiveness, efficiency, and universality of the proposed two-phase framework. Segmentation of loess landslides is considered a challenging task due to the intrinsic nature of vague boundary information. The proposed framework is compared with the conventional U-Net and other recent benchmarking landslide segmentation algorithms. Computational results indicate that the proposed framework produces more accurate segmentation of loess landslides compared with the other tested benchmarking algorithms. [ABSTRACT FROM AUTHOR]

On January 27, 2021, a small-scale loess landslide occurred in Dangchuan Village, Yongjing County, Gansu Province, China. Timely monitoring and warning helped to prevent any casualties. The sliding mass of the Dangchuan landslide fell in two stages with an interval of 24 s and lasted for a total of 31 s. Eventually, the Dangchuan landslide had a descent height of approximately 120 m and a sliding portion stretching approximately 350 m. The landslide volume was approximately 3 × 104 m3. Based on the investigation of different stages of the Dangchuan landslide and the characteristics of the deposition area, the induction of this landslide was attributed to the effect of frozen stagnant water. Following field investigation of the landslide source and crack areas, a potential sliding mass model with a volume of 5 × 104 m3 was obtained via calculation. To predict the disaster scope of potential landslides, we first used DAN3D software to invert this Dangchuan landslide. Then, based on the numerical model parameters obtained from the inversion, the sliding accumulation process of the potential slide mass was simulated. The results showed that the maximum disaster scope of the potential slide mass would be larger than that of the Dangchuan landslide. The new deposit would cover an area of approximately 3.5 × 104 m2, with a thickness in the range of 0–6 m. The new deposition area would expand by approximately 50 m in the long-axis direction compared to the Dangchuan landslide, impacting the houses and roads. The DAN3D prediction results of the disaster scope will help residents in the potential danger zone evacuate in advance, reducing unnecessary damage. [ABSTRACT FROM AUTHOR]

Information about the long-term spatiotemporal evolution of landslides can improve our understanding of the landslide development process and can help prevent landslide disasters. However, few studies have been devoted to the pre- and post-failure spatiotemporal evolution process and pattern of landslides. Therefore, we studied the pre- and post-failure geomorphic changes and spatiotemporal evolution of the 2019 Jiangou landslide based on field investigations, interferometric synthetic aperture radar (InSAR), unmanned aerial vehicle (UAV) observations, and remote sensing techniques. The results show that the volume of the deposition of the Jiangou landslide is less than the depletion volume, which means the remaining landslide materials were washed away when the dammed lake collapsed. Moreover, the InSAR technique has an advantage in terms of the retrieval of pre- and post-failure creep deformation. Our analysis suggests that the Jiangou landslide has experienced long-term creep. Potential landslide risks still exist after the previous failure event. Furthermore, we found that the pre- and post-failure spatiotemporal deformation processes and evolutionary patterns of the landslide are different. The pre-failure evolutionary pattern of the landslide is a progressive failure mode, while the post-failure evolutionary pattern is a retrogressive failure mode. This evolution provides a reference for local governments to further monitor or take effective prevention measures against future landslide failures. [ABSTRACT FROM AUTHOR]

Loess landslides are a severe engineering geological problem on the Loess Plateau of China. This study relies on interferometric synthetic aperture radar (InSAR) and unmanned aerial vehicle (UAV) photogrammetry techniques to characterise a large-scale loess landslide deformation process in Northwest China. First, a total of 85 Sentinel-1 SAR images acquired from March 2015 to March 2019 were used to detect the spatial displacement and characterise the pattern of deformation from presliding to postsliding. Second, three UAV flight surveys were conducted to reconstruct landslide morphology, identify scarps, cracks and fissures; and generate elevation differences. By comparing these two technologies, we found that InSAR has more advantages in the retrieval of creeping deformation, while UAV data is valuable for detecting large sudden slides. Dynamic deformation zonation maps were generated based on comprehensive analysis, and the deformation pattern of the large loess landslide was deduced. The Hongheyan landslide was induced by heavy rainfall and then gradually became suspended owing to the topography effect. Recently, the western part of the slope showed a clear accelerating trend. This evolution provides comprehensive kinematic information for local governments to make further intensive observations or take effective precautions. [ABSTRACT FROM AUTHOR]