Earthquake-induced Landslides Research Articles

Comparative Study of Artificial Neural Network and Random Forest Model for Susceptibility Assessment of Landslides Induced by Earthquake in the Western Sichuan Plateau, China

Earthquake-induced landslides are one of the most dangerous secondary disasters in mountainous areas throughout the world. The nowcasting of coseismic landslides is crucial for planning land management, development, and urbanization in mountainous areas. Taking Wenchuan County in Western Sichuan Plateau (WPS) as the study area, a landslide inventory was built using historical records. Herein, eight causative factors were selected for a library of factors, and then a landslide susceptibility assessment (LSA) was performed based on the machine learning techniques of Random Forest (RF) and Artificial Neural Network (ANN) models, respectively. The prediction abilities of the above two LSM models were assessed using the area under curve (AUC) value of the receiver operating characteristics (ROC) curve, precision, recall ratio, accuracy, and specificity. The performances of both machine learning techniques were found to be excellent, but RF outperformed in accuracy. There were still some differences between the models’ performances shown by the results: RF (AUC = 0.966) outperformed ANN (AUC = 0.914). The RF model demonstrated a higher degree of correlation between the areas classified as very low and high susceptibility in comparison to the ANN model. The results provided a theoretical framework upon which machine learning applications could be applied (e.g., RF and ANN), a reliable and low-cost tool to assess landslide susceptibility. This comparative study will provide a useful description of earthquake-induced landslides in the study area, which can be used to anticipate the features of landslides in the future, and have played a very important role in proper anthropogenic activities, resource management, and infrastructural development of the mountainous areas.

Estimating the debris-flow magnitude using landslide sediment connectivity, Qipan catchment, Wenchuan County, China

In an earthquake disturbed basin, the earthquake-induced landslides’ connectivity with debris-flow channels determines the landslide sediment volume that is transported into debris flows, and the debris-flow magnitude as well. In this study, the index of sediment connectivity (IC) was used to analyze the evolution of landslide sediment transport capacity. A normalized IC was developed to quantify the movable volume of the landslide sediment available for debris flows. Combined with a landslide empirical relationship, sediment connectivity calculation, particle model, and numerical algorithm, two debris-flow events after the 2008 Wenchuan earthquake were simulated for the Qipan catchment. The results show good agreement between the simulated and observed debris-flow magnitudes of peak discharges, inundated areas, and depths with approximately 9% average error, which indicates that the normalized landslide connectivity index is suitable for estimating the movable volume. Moreover, the IC and landslide characteristics (area and location) showed significant negative correlations, indicating that the spatial connectivity of landslide sediment generally decayed over time, with a corresponding reduction in debris-flow magnitude.

Earthquake-induced landslides in Haiti: analysis of seismotectonic and possible climatic influences

Abstract. First analyses of landslide distribution and triggering factors are presented for the region affected by the 14 August 2021 earthquake (Mw=7.2) in the Nippes Department, Haiti. Landslide mapping was mainly carried out by comparing pre- and post-event remote imagery (∼0.5–1 m resolution) available on Google Earth Pro® and Sentinel-2 (10 m resolution) satellite images. The first covered about 50 % of the affected region (for post-event imagery and before completion of the map in January 2022), and the latter were selected to cover the entire potentially affected zone. On the basis of the completed landslide inventory, comparisons are made with catalogs compiled by others both for the August 2021 and the January 2010 seismic events, including one open inventory (by the United States Geological Survey) that was also used for further statistical analyses. Additionally, we studied the pre-2021 earthquake slope stability conditions. These comparisons show that the total number of landslides mapped for the 2021 earthquake (7091) is larger than the one recently published by another research team for the same event but slightly smaller than the number of landslides mapped by a third research team. It is also clearly smaller than the one observed by two other research teams for the 2010 earthquake (e.g., 23 567, for the open inventory). However, these apparently fewer landslides triggered in 2021 cover much wider areas of slopes (>80 km2) than those induced by the 2010 event (∼25 km2 – considering the open inventory). A simple statistical analysis indicates that the lower number of 2021 landslides can be explained by the missing detection of the smallest landslides triggered in 2021, partly due to the lower-resolution imagery available for most of the areas affected by the recent earthquake; this is also confirmed by an inventory completeness analysis based on size–frequency statistics. The much larger total area of landslides triggered in 2021, compared to the 2010 earthquake, can be related to different physical reasons: (a) the larger earthquake magnitude in 2021, (b) the more central location of the fault segment that ruptured in 2021 with respect to coastal zones, (c) and possible climatic preconditioning of slope instability in the 2021 affected area. These observations are supported by (1) a new pre-2021 earthquake landslide map; (2) rainfall distribution maps presented for different periods (including October 2016 – when Hurricane Matthew had crossed the western part of Haiti), covering both the 2010 and 2021 affected zones; and (3) shaking intensity prediction maps.

Characteristics of physical properties of the sliding and its surrounding layers in landslides caused by the 2018 Hokkaido Eastern Iburi Earthquake

A 6.6-Mw earthquake struck the Iburi region of Hokkaido, Japan, in 2018, triggering massive landslides. Most of these landslides were shallow and occurred mostly in the Atsuma and Abira towns. Ta-c and Ta-d tephra layers have been found in the Towa landslide at Atsuma from the Tarumai volcano, while Ta-d, En-a, and Spfa-1 tephra layers have been found in the Mizuho landslide at Abira from the Tarumai and Eniwa volcanos, as well as the Shikotsu caldera. Field observations from previous studies revealed that the sliding layers were located in the Ta-d and En-a layers at the Towa and Mizuho landslides, respectively. Unlike previous research on earthquake-induced landslides, which were investigated using mechanical properties, this study investigates the characteristics of physical properties, saturated permeability properties, and content of clay minerals on sliding and surrounding tephra layers. Results from this study reveal that the physical properties of sliding layers from two landslides demonstrated the same characteristics: non-plastic soil with a low density of soil particles, void ratio, and dry density; these characteristics could influence earthquake-induced landslides. It also reveals a relationship between the plasticity chart and the age of tephra materials, including the relationship between the weathering process and density of soil particles and the dissimilarity in characteristics of saturated permeability properties in tephra materials.

Distribution of large- and medium-scale loess landslides induced by the Haiyuan Earthquake in 1920 based on field investigation and interpretation of satellite images

Abstract Studying the distribution law and influencing factors of coseismic landslides has important scientific significance and engineering value for understanding the mechanism of seismic landslides and predicting the occurrence of seismic landslides. After a hundred years, these large- and medium-scale landslides induced by the 1920 Haiyuan earthquake are still well-preserved and have extremely high academic research value. About 620 loess seismic landslides induced by the Haiyuan earthquake in 1920 were investigated on site. On this basis, the shape differences between seismic landslides and gravity landslides were summarized; 605 landslides were identified by satellite images, and the Haiyuan earthquake-induced loess landslide database containing seismic information and landslide information was established. The distribution law and morphological characteristics of large- and medium-sized landslides induced by the Haiyuan earthquake were systematically counted according to the conditions of the landslide-intensive area, intensity, and fault upper and lower plates. The influencing factors of loess earthquake landslides were summarized, and the following conclusions were obtained: (1) 1,225 large- and medium-sized landslides were induced by the Haiyuan earthquake in 1920. These landslides have the characteristics of long sliding distance, large single scale, and strong disaster-causing. They are mainly distributed in three concentrated areas of Xiji, Haiyuan, and Pengyang. The landslide morphological characteristics of the three landslide-intensive areas are different because the landslide sliding mechanism caused by topography and lithology is different. (2) The landslide distribution has obvious clustering, zonation, and directivity, and has an obvious river distance effect and fault hanging wall effect. (3) The internal influencing factors such as stratum lithology, topography, fault location, and direction, and the role of water control the occurrence location and scale of landslide. The external factor of an earthquake is an important incentive and control factor for landslide occurrence.

A Numerical Study of Wave Propagation and Cracking Processes in Rock-Like Material under Seismic Loading Based on the Bonded-Particle Model Approach

An earthquake is usually followed by a considerable number of aftershocks that play a significant role in earthquake-induced landslides. During the aftershock, the cracking process in rocks becomes more complex because of the formation of faults. In order to investigate the effects of seismic loading on the cracking processes in a specimen containing a single flaw, a numerical approach based on the bonded-particle model (BPM) was adopted to study the seismic loading applied in two orthogonal directions. The results reveal that no transmission and reflection phenomena were observable in the small specimens (76 mm × 152 mm) because they were considerably smaller than the wavelength of the P-wave. Furthermore, under seismic loading, the induced crack was solely tensile in nature. Repeated axial seismic loading did not induce crack propagation after the first axial seismic loading. Cracks began to propagate only when the seismic loading direction was changed from axial to lateral, and then back to axial, ultimately resulting in the failure of the specimen.

Experimental study on seismic response and progressive failure characteristics of bedding rock slopes

Bedding rock slopes are common geological features in nature that are prone to failure under strong earthquakes. Their failures induce catastrophic landslides and form barrier lakes, posing severe threats to people's lives and property. Based on the similarity criteria, a bedding rock slope model with a length of 3 m, a width of 0.8 m, and a height of 1.6 m was constructed to facilitate large-scale shaking table tests. The results showed that with the increase of vibration time, the natural frequency of the model slope decreased, but the damping ratio increased. Damage to the rock mass structure altered the dynamic characteristics of the slope; therefore, amplification of the acceleration was found to be nonlinear and uneven. Furthermore, the acceleration was amplified nonlinearly with the increase of slope elevation along the slope surface and the vertical section, and the maximum acceleration amplification factor (AAF) occurred at the slope crest. Before visible deformation, the AAF increased with increasing shaking intensity; however, it decreased with increasing shaking intensity after obvious deformation. The slope was likely to slide along the bedding planes at a shallow depth below the slope surface. The upper part of the slope mainly experienced a tensile-shear effect, whereas the lower part suffered a compressive-shear force. The progressive failure process of the model slope can be divided into four stages, and the dislocated rock mass can be summarized into three zones. The testing data provide a good explanation of the dynamic behavior of the rock slope when subjected to an earthquake and may serve as a helpful reference in implementing antiseismic measures for earthquake-induced landslides.

The Mobility of Landslides in Pumice: Insights from a Flume Experiment

Risk of landslide hazards strongly depends on how far landslide sediment travels, known as landslide mobility. Previous studies mentioned enhanced mobility of earthquake-induced landslides in volcanic deposits compared to those from other geologic/soil settings. A flume apparatus constructed at a 1:300 scale was used to examine the mobility of landslides with pumice. Four pumice samples were collected from landslides induced by the 2018 Eastern Iburi earthquake, Hokkaido, Japan. Laboratory tests confirmed the unique low specific gravity of the pumice (1.29–1.33), indicating numerous voids within pumice particles. These voids allowed pumice to absorb a substantial amount of water (95–143%), about 9–15 times higher than other coarse-grained soils. Our flume experiments using various saturation levels (0–1) confirmed the influence of this inner-particle water absorption on pumice mobility. Because a low value of specific gravity indicates a low strength of soil, grain crushing may occur on the pumice layer, causing water from the internal voids to discharge and fluidize the transported landslide mass. Our findings indicate that such earthquake-induced landslides can be as mobile as those induced by rainfall, depending on the initial water content of the pumice layers. These conditions might be associated with water accumulation from previous rainfall events and the water-holding capability on pumice layers.

A new permanent displacement model considering pulse-like ground motions and its application in landslide hazard assessment

Permanent displacement predictive modeling can be used with geotechnical and seismological parameters to assess hazard assessment of earthquake-induced landslides (ELHA). Pulse-like ground motions (PLGMs) have a significant influence on landslide initiation, especially in near-fault areas. In this study, several permanent displacement predictive models were developed by analyzing PLGMs and horizontal-and-vertical combined shaking based on a modified permanent displacement model. Five hundred and fifty-five horizontal-and-vertical groups of near-fault PLGMs with distances less than 80 km from the Next Generation Attenuation (NGA) database were selected and used as the excitation of the permanent displacement calculation model. A model utilizing peak ground acceleration (PGA), peak ground velocity (PGV), and Arias intensity (Ia) parameters, was selected owing to its excellent performance in the comparative analysis of the correlation coefficient and standard deviation. The characteristics of near-fault PLGMs with azimuth and vertical component considered in the developed model can be applied to ELHA of the 1994 Northridge earthquake. The predictive capability of the presented model was evaluated from the area under the receiver operating characteristic (ROC) curve, with a value of 0.84. The results show that the model performs well in the ELHA in the near-fault areas and provides a basis for disaster prevention and mitigation.

A mixed-mode data collection approach for building inventory development: Application to school buildings in Central Sulawesi, Indonesia

Urban disaster risk management and reduction requires the development and periodic updating of regional building inventories. However, the development of such inventories can be very cost-intensive and time-consuming, making this a challenging task, particularly for low- and middle-income countries. This article discusses a mixed-mode building inventory data collection framework using a rapid and cost-effective remote survey technique that can be deployed in various geographic contexts. A key component of the proposed approach is an inter-rater reliability analysis of data collected from traditional sidewalk surveys and remote surveys for a small subset of buildings in the considered building portfolio, which is used to assess the suitability of the remote survey for the location(s) considered. The framework is demonstrated by developing a regional database of school buildings in the Central Sulawesi region of Indonesia. The database consists of 2536 school buildings from 454 elementary and high schools in the Palu, Sigi, and Donggala regions, susceptible to earthquake-induced ground shaking, tsunami, liquefaction, and landslides. The developed database can be used in pre-event/long-term risk analysis and management, post-event/near-real-time loss estimation, and regional-level decision-making on school assets and related policies. The database has been made available for public use and can be readily harmonized with similar databases for other regions.

Co-seismic landslide hazard assessment of Uttarakhand state (India) based on the modified Newmark model

Co-seismic landslides are the most dangerous geological hazards in seismically active mountainous regions. These landslides cause damage to roads, drainage pipelines, buildings, agriculture, and loss of human lives. It is crucial but difficult to precisely map the hazards resulting from earthquake-induced landslides. Newmark's approach is widely used to determine the permanent displacement of a probable sliding mass and simulate the mechanism of seismically-induced slope failure. The present study assesses co-seismic landslide hazards for Uttarakhand state (India), which falls under the highest seismically active zones as per the seismic code of India. It proposes an improved Newmark’s approach aby incorporating rock joint shear strength parameters and the size effect of possible sliding surfaces. Static factor of safety and critical acceleration maps were prepared for the study area by combining various geological, geotechnical, and topographical parameters. The ground motion parameters were determined in terms of surface-level Arias Intensity. Newmark permanent displacement map was prepared by taking critical acceleration and Arias Intensity as input parameters. Finally, the modified Newmark’s model results were compared with the conventional Newmark’s model using the area under curve (AUC) analysis. This assessment approach can be used effectively to forecast the area affected by co-seismic landslide hazards and prepare guidelines for major infrastructure projects and post-disaster damage studies.

Mechanisms of rock slope failures triggered by the 2016 Mw 7.8 Kaikōura earthquake and implications for landslide susceptibility

Landslide failure mechanisms are influenced by topography, lithology, structure and rock mass damage – factors that also control landslide susceptibility. Failure mechanisms, however, are rarely considered in regional-scale coseismic landslide susceptibility analyses. In this study, we use 3D pixel tracking in pre- and post-earthquake aerial imagery, geomorphic mapping, rock mass characterisation, and geophysical ground investigations to develop conceptual models for three earthquake-induced landslides triggered by the 2016 Mw 7.8 Kaikōura earthquake on New Zealand's South Island. Analysis of two incipient landslides in Cretaceous greywacke illustrates the failure stages of a rock mass comprising multiple sets of closely-spaced and low persistence discontinuities. Conversely, analysis of a landslide in Neogene massive siltstones illustrates the role of high-persistence bedding planes in generating large translational rockslides. These two distinct mechanisms typify failures in highly deformed basement and overlying massive sedimentary rocks respectively, and highlight the link between rock mass damage, failure mechanism, and initiation. In greywacke failures, the rupture plane initiated close to the ridgetop and propagated as a joint-step-path failure along pre-existing, but low-persistence joints whereas the landslide in Neogene siltstone initiated by sliding along weak, high persistence, bedding planes near the base of the slope where topographic stresses are highest. Analysis of landslide displacements furthermore points out that the evolution of landslides is closely related to rock mass characteristics. In highly jointed Cretaceous greywacke, relatively little displacement is needed for rock mass disintegration and avalanching to occur whereas in Neogene siltstone, the landslide displaces as a coherent body - even at large displacements.Our results suggest that (1) failure mechanisms and, as a result, coseismic landslide susceptibility factors, may vary fundamentally in different geological settings; (2) characteristic spatial landslide displacement patterns may be used to remotely identify relevant failure mechanisms; and (3) the amount of displacement that can be accommodated before a failure transitions from sliding to avalanching, is highly dependent on material characteristics and topographic constraints.

Arias intensity attenuation relationship in Sichuan–Yunnan region, China

Arias intensity is an essential ground motion measure correlating with the potential for earthquake-induced landslides. The Sichuan–Yunnan region, which is primarily mountainous, is a high incidence region of earthquake-induced landslides in China. However, there is no available attenuation relationship for this intensity measure due to the backward construction of the stations. In this study, we developed a region-specific Arias intensity attenuation relationship using the China Strong-Motion Networks Center (CSMNC) database which was established in 2008. We recommend this relationship be applied in the Sichuan–Yunnan region for moment magnitudes ranging between 4.2 and 7.0, distances ranging between 0 and 300 km and with Vs30 (the average shear-wave velocity in the upper 30 m of a soil profile) ranging between 128 and 760 m/s. The current study finds that this relationship’s intra-event, inter-event, and total standard deviations are greater than for other regions. This is likely caused by the complicated seismotectonic activities, nonlinear site effects, error from inferring Vs30, basin effects, etc. However, this relationship has the best performance in fitting and predicting the data from the Sichuan–Yunnan region.

The effect of surficial soil on the seismic response characteristics and failure pattern of step-like slopes

Earthquake-induced shallow landslides are widely distributed in earthquake-affected areas. These hazards mainly originate from fragmented surface rock masses or surficial loose deposits overlaid on bedrock structure slopes. The seismic response of the surficial soil-bedrock slope is the key to revealing the dynamic failure mechanism of earthquake-induced landslides. This study investigates the seismic response of a surficial soil-bedrock step-like slope to evaluate the acceleration ground motion amplification along the slope. Parametric analysis focusing on the geometry of the surficial soil, thickness of the surficial soil, surficial soil-bedrock impedance contrast and geometry of the slope is conducted using the finite difference modeling code FLAC3D. The results show that both the surficial soil characteristics and geometry of the slope have great influences on the acceleration amplification factor in the horizontal direction (AAF). A maximum AAF zone is always formed near the interface between the surficial soil and bedrock, and the AAF always abruptly amplifies near the slope surface, especially at the crest of the slope. The potential dynamic failure surfaces under earthquakes are formed near the interface between the surficial soil and bedrock, which is in accordance with the AAF amplification zone of the slope, and the depth of the dynamic failure of the surficial bedrock slope is deeper than the static failure surface. It can be concluded that the topographic effect on the seismic response of the slope is strongly dependent on the surficial soil material characteristics and geometry. The findings of this study can contribute to a better understanding of the mechanism of shallow earthquake-induced landslides and have a number of important implications for disaster prevention and mitigation.

Study on the Uncertainty of Machine Learning Model for Earthquake-Induced Landslide Susceptibility Assessment

The landslide susceptibility assessment based on machine learning can accurately predict the probability of landslides happening in the region. However, there are uncertainties in machine learning applications. In this paper, Artificial Neural Network (ANN), Random Forest (RF), Support Vector Machine (SVM), and Logistic Regression (LR) are used to assess the landslide susceptibility in order to discuss the model uncertainty. The model uncertainty is explained in three ways: landslide susceptibility zoning result, risk area (high and extremely high) statistics, and the area under Receiver Operating Characteristic Curve (ROC). The findings indicate that: (1) Landslides are restricted by influence factors and have the distribution law of relatively concentrated and strip-shaped distribution in space. (2) The percentage of real landslide in risk area is 86%, 87%, 82%, and 61% in SVM, RF, LR, and ANN, respectively. The area under ROC of RF, SVM, LR, and ANN, respectively, is 90.92%, 80.45%, 73.75%, and 71.95%. (3) Compared with the prediction accuracy of the training set and test set from the same earthquake, the accuracy of landslide prediction in the different earthquakes is reduced.

Use of scenario earthquakes for seismic hazard assessment in low-seismicity, stable continental regions: A case study from Indiana, USA

While large earthquakes have been documented in stable continental regions, such as the central United States, they occur too infrequently to provide reliable observations of earthquake-related impacts. Using the state of Indiana as a case study, we investigate the impacts of five deterministic scenarios—three occurring outside the borders of the state and two within the state. They include a M7.3 Wabash Valley event in southern Illinois, a M7.6 New Madrid event in southeastern Missouri, a M6.2 event in west central Ohio, a M6.2 event near Evansville, Indiana, and a M5.8 Indianapolis event. The locations and magnitudes are based on known fault locations and credible interpretations of the earthquake history in the region. We use a combination of the US Geological Survey’s ShakeMap and the Federal Emergency Management Agency’s (FEMA) Hazus software packages to assess earthquake-triggered ground shaking and its effects on the built environment. We also use the United States Geological Survey (USGS) Ground Failure estimation tool to examine the spatial distribution of anticipated earthquake-induced landslides and liquefaction. The five main deterministic scenario models indicate that moderate-sized urban earthquakes may represent a greater threat to a state like Indiana than a large-magnitude event from the New Madrid seismic zone. To better understand the influence of earthquake source parameters on impacts, we conducted a sensitivity analysis for earthquakes near Indianapolis and Evansville, where we reviewed losses due to differences in magnitude, depth, strike, and dip. Based on the model projections, magnitude and depth have first-order and relatively predictable influence on losses. The orientation and dip of the causative fault in relation to populated areas can alter economic losses for an event of the same magnitude by 13%–32%. Deterministic scenario analyses, such as those presented here, can potentially be of widespread utility for understanding earthquake impacts in the central United States and other low-seismicity, stable continental regions.

Earthquake-induced landslide monitoring and survey by means of InSAR

Abstract. This study uses interferometric synthetic aperture radar (SAR) techniques to identify and track earthquake-induced landslides as well as lands prone to landslides, by detecting deformations in areas struck by earthquakes. The pilot study area investigates the Mila region in Algeria, which suffered significant landslides and structural damage (earthquake: Mw 5, 7 August 2020). DInSAR analysis shows normal interferograms with small fringes. The coherence change detection (CCD) and DInSAR analysis were able to identify many landslides and ground deformations also confirmed by Sentinel-2 optical images and field inspection. The most important displacement (2.5 m), located in the Kherba neighborhood, caused severe damage to dwellings. It is worth notice that CCD and DInSAR are very useful since they were also able to identify ground cracks surrounding a large zone (3.94 km2 area) in Grarem City, whereas the Sentinel-2 optical images could not detect them. Although displacement time-series analysis of 224 interferograms (April 2015 to September 2020) performed using LiCSBAS did not detect any pre-event geotechnical precursors, the post-event analysis shows a 110 mm yr−1 subsidence velocity in the back hillside of Kherba.

Earthquake-induced landslide erosion coupled to tectonics and river incision, and effects of ground motion on coupled patterns

Landslide erosion and long-term denudation for the natural surface erosion processes in tectonically active mountain ranges are coupled to tectonics and river incision. However, such relationships have rarely been quantified in earthquake-induced landslides. Here we assess seismic landslide erosion caused by the Wenchuan earthquake in the Longmen Shan and explore relationships between tectonic uplift, river incision, and earthquake-induced landslide erosion yields and thus evaluate the effects of ground motion on coupled patterns. We show that nonlinear correlations exist between seismic landslide erosion rates and topographic metrics (mean slope gradient and normalized channel steepness) under various shaking conditions (peak ground acceleration PGA > 0.4 g). The areas with stronger shaking cause greater landslide erosion yields, higher erosion efficiency, weaker nonlinearity, and lower mean landslide slope although they have a similar threshold hillslope. Taken together, these analyses support an emerging view that the stronger ground motion reduces the equivalent rock mass strength on the hillslope and breaks the topographic threshold, which consequently produces a systematic increase in landslide erosion below or above the threshold state and results in nonlinearity variation. Our finding demonstrates that earthquake-induced landslide erosion is coupled to tectonics and river incision under various shaking conditions and presents a quantitative relationship between seismic landslide erosion rates and topographic metrics, providing potentially important implications for mountain belt evolution.

Application of novel deep boosting framework-based earthquake induced landslide hazards prediction approach in Sikkim Himalaya

A major earthquake (6.9 Moment magnitude) occurred in the Sikkim and Darjeeling areas of the Indian Himalaya as well as in the adjacent Nepal on 18th September 2011, triggering a large number of landslides. A total of 188 landslide locations were extracted in order to create the landslide inventory map (LIM). The earthquake-induced landslide susceptibility maps (LSMs) were created using an Artificial Neural Network (ANN) model and three novel deep learning approaches (DLAs), namely Deep Boosting (DB), Deep Learning Neural Network (DLNN), and Deep Learning Tree (DLT), as well as training points and 22 conditioning factors. The earthquake-induced LSMs validated using several statistical indices and the results showed optimal accuracy for all models, where DB yielding the highest prediction rate curve (PRC) of 98.5%. This is followed by DLT (97%), DLNN (96%), and ANN (91%). The results demonstrate maximum efficacy of the proposed earthquake-induced LSM.

A Web-GIS hazards information system of the 2008 Wenchuan Earthquake in China

The typical earthquake hazards chain includes main shock, surface ruptures, aftershocks, and earthquake-induced landslides, leading to structural damage and casualties. Lots of investigation and research were published. But there is no Web-GIS platform to integrate earthquake hazards chain data and compare relative research results. This paper studies compiling such a hazards chain database and building a cloud-based Web-GIS system of the 2008 earthquake. The system introduced by this paper provides buttons and tools to browse, query and analyze map data. In addition, this system is open to anyone to add data for viewing and comparing in the web map. The result showcases that the cloud-based system runs efficiently and flexibly. The multi-scale and multi-source data can be integrated into a “one-map” system by well-design map services and geodatabase.