Rapid Factor Screening for Landslide Susceptibility Mapping of Linear Engineering Slopes Using a Reduced-Factor Information Value Model: A Case Study of the Jing-Zhang Railway, China

<p><strong>Abstract</strong><strong>:</strong> Landslides are one of the most common and devastating natural hazards worldwide, which cause injuries to life and damage to properties, infrastructures leading to high-cost maintenance. In this study frequency ratio, information value and fuzzy logic models were used for landslide susceptibility mapping of an area of 356km<sup>2</sup> in and around Dharamshala, Himachal Pradesh, using earth observation data. Dharamshala, a part of North-western Himalaya, is one of the fastest-growing tourism hubs with a total population of 30,764 according to the 2011 census and is amongst one of the hundred Indian cities to be developed as a smart city under PM’s Smart Cities Mission. The thrust for infrastructure development has led to a need for prior planning to minimize the consequences of landslide hazards. The final produced landslide susceptibility zonation maps with better accuracy could be used for land-use planning to prevent future losses. A landslide inventory for the study area was prepared through visual interpretation of high-resolution satellite imagery and available inventory report. Remote sensing data and other ancillary data like geological data were collected and processed in the GIS environment to generate thematic maps of parameters influencing landslide occurrence. The landslide causative parameters used in the study are slope angle, slope aspect, elevation, curvature, topographic wetness index, relative relief, distance from lineaments, land use land cover, and geology. Using these parameters and landslide inventory weight and membership value was calculated for the Frequency ratio, information value and Fuzzy logic model, respectively. In the frequency ratio and information value model, all the landslide causative parameters were arithmetically overlaid using calculated weights for landslide susceptibility mapping. In the fuzzy logic model, different fuzzy operators were applied to the calculated fuzzy membership values. Unlike the normalization process for membership calculation present study used the cosine amplitude method, which will give more reliable results. A total of ten landslide susceptibility maps (LSM) were produced using two models, 9 from fuzzy logic and 1 from frequency ratio. All the results were verified spatially and statistically using landslide locations and ROC curves. Further, the performance and significance of different outputs were compared to select the most suitable LSM for the study area. Among all fuzzy operators, “gamma” with λ = 0.9 showed the best accuracy (84.3%) and operator “and” has the worst accuracy (77.6%). But among all 9 output maps of fuzzy logic except the output of gamma (λ = 0.9) gives satisfactory LSM rest all show the unacceptable result as the maximum number of pixels is either in very low or high susceptible zone. The validation and comparison result exhibited that the fuzzy logic (accuracy=84.3%) is better than the information value (83.46) and the frequency ratio method (accuracy=83.43%).</p><p><strong>Keywords</strong>: Bivariate Statistical Techniques, Information Value, Frequency Ratio, Fuzzy Logic, ROC</p>

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

Geological hazards present one of the most important constraints for the development of the Arzew sector (Oran province), North Western of Algeria. Landslides are considered us one of the most common phenomena in the study area and especially in the hilly area. For minimizing and reducing the consequences of this problem, it is necessary to carry out preliminary studies on the cartography of the different zones exposed to the slope instability phenomena. The main objective of this study is to perform the landslide susceptibility mapping by statistical models and GIS techniques for the Arzew area. To achieve this goal, an analytical approach was carried out. Firstly, a landslide inventory map was prepared using previous inventory maps, satellite images, aerial photos and field surveys. Secondly seven conditioning factors such as slope degree, aspect, lithology, land use, distance to the streams, distance to the road and altitude were exploited to assess landslide susceptibility. Thirdly, the weight value for each class of the conditioning factors was determined using Frequency Ratio (FR) and Information Value (IV) models based in GIS functionalities. Consequently, Landslide Susceptibility Maps (LSMs) were produced by the classification process of the global Landslide Susceptibility Indexes (LSIs) into five classes. Finally, for experiment verification, the LSMs obtained with the FR and IV models were confirmed comparing LSMs with landslide inventory map using both the Receiver Operating Characteristics (ROC) and the Seed Cell Area Index (SCAI) models. The area under curve (AUC) results, demonstrate that the IV method more performance (89.03%) for LSM than FR method (85.57%). Furthermore, the validation results using SCAI also confirmed that the IV model was more accurate than FR model. The models employed in this study are capable to resolve the issue of the landslide susceptibility of the study area. The produced susceptibility maps can be used for future land use planning and can be considered as a powerful tool to resolve the spatial distribution of the risk associated to landslides.

Goncha Siso Eneses area of East Gojam Zone in northwestern Ethiopia is one of the most landslide-prone regions, which is characterized by frequent landslide occurrences causing fatalities and damages in cultivated and non-cultivated lands, infrastructure and properties. Hence, preparing a landslide susceptibility map is very helpful in reducing the damages in infrastructure and properties and loss of animal and human lives. In this study, GIS-based information value and logistic regression models were applied. A reliable and detailed landslide inventory with 894 landslides was prepared through detailed fieldwork and Google Earth image interpretation. These landslides were randomly divided into training data set for model development and testing data set for model validation. Nine landslide causative factors like slope, curvature, aspect, lithology, distance to stream, distance to lineament, distance to spring, rainfall and land use/cover were integrated with training landslides to determine the weight(s) of each landslide factor and factor classes using Information Value and Logistic Regression models, respectively. The landslide susceptibility index map was then produced by summing the weights of all the landslide factors using raster calculator of the spatial analyst tool in GIS. To evaluate the performance of the information value and logistic regression models for landslide susceptibility modeling, the relative landslide density index and area under the curve (AUC) of the receiver operating characteristic curves were performed on both the training and testing landslide data sets. The model has an AUC accuracy of 88.9% success rate and 85.9% prediction rate for information value model whereas 81.8% success rate and 80.2% predictive rate for logistic regression model.

The main objective of this study was to apply a statistical (information value) model using geographic information system (GIS) to the Chencang District of Baoji, China. Landslide locations within the study area were identified using reports and aerial photographs, and a field survey. A total of 120 landslides were mapped, of which 84 (70 %) were randomly selected for building the landslide susceptibility model. The remaining 36 (30 %) were used for model validation. We considered a total of 10 potential factors that predispose an area to a landslide for the landslide susceptibility mapping. These included slope degree, altitude, slope aspect, plan curvature, geomorphology, distance from faults, lithology, land use, mean annual rainfall, and peak ground acceleration. Following an analysis of these factors, a landslide susceptibility map was produced using the information value model with GIS. The resulting landslide susceptibility index was divided into five classes (very high, high, moderate, low, and very low) using the natural breaks method. The corresponding distribution area percentages were 29.22, 25.14, 15.66, 15.60, and 14.38 %, respectively. Finally, landslide locations were used to validate the results of the landslide susceptibility map using areas under the curve (AUC). The AUC plot showed that the susceptibility map had a success rate of 81.79 % and a prediction accuracy of 82.95 %. Based on the results of the AUC evaluation, the landslide susceptibility map produced using the information value model exhibited good performance.

The information value (IV) model is a conventional method for landslide susceptibility prediction (LSP). However, it is inconsistent with the actual situation to regard all conditioning factors as equally weighted in the modeling process. In view of this, this paper studied the optimization effect of different weight calculation methods for IV model. Xingshan County, a typical landslide-prone area located in Hubei Province, China, was taken as a case study. The procedure was as follows: First, six conditioning factors, including elevation, slope angle, aspect, curvature, distance to river, and distance to road, were selected to form an evaluation factor library for analyzing the landslide susceptibility. Then, the weight of factors was calculated by fuzzy analytical hierarchy process (FAHP) and principal component analysis (PCA). On this basis, combined with the IV model, two weighted IV models (FAHP-IV model and PCA-IV model) were formed for LSP. The results shows that the optimization effect of PCA was the best. Moreover, compared with the IV-only model (AUC = 0.71), the FAHP-IV model (AUC = 0.76) and PCA-IV model (AUC = 0.79) performed better. The outcome also provided a feasible way for the study of regional LSP.

Landslides are usually initiated under complex geological conditions. It is of great significance to find out the optimal combination of predisposing factors and create an accurate landslide susceptibility map based on them. In this paper, the Information Value Model was modified to make the Modified Information Value (MIV) Model, and together with GIS (Geographical Information System) and AUC (Area Under Receiver Operating Characteristic Curve) test, 32 factor combinations were evaluated separately, and factor combination group with members Slope, Lithology, Drainage network, Annual precipitation, Faults, Road and Vegetation was selected as the optimal combination group with an accuracy of 95.0%. Based on this group, a landslide susceptibility zonation map was drawn, where the study area was reclassified into five classes, presenting an accurate description of different levels of landslide susceptibility, with 79.41% and 13.67% of the validating field survey landslides falling in the Very High and High zones, respectively, mainly distributed in the south and southeast of the catchment. It showed that MIV model can tackle the problem of “no data in subclass” well, generate the true information value and show real running trend, which performs well in showing the relationship between predisposing factors and landslide occurrence and can be used for preliminary landslide susceptibility assessment in the study area.

At present, landslide susceptibility assessment (LSA) based on the characteristics of landslides in different areas is an effective prevention measure for landslide management. In Enshi County, China, the landslides are mainly triggered by high-intensity rainfall, which causes a large number of casualties and economic losses every year. In order to effectively control the landslide occurrence in Enshi County and mitigate the damages caused by the landslide. In this study, eight indicators were selected as assessment indicators for LSA in Enshi County. The analytic hierarchy process (AHP) model, information value (IV) model and analytic hierarchy process-information value (AHP-IV) model were, respectively, applied to assess the landslide distribution of landslides in the rainy season (RS) and non-rainy season (NRS). Based on the three models, the study area was classified into five levels of landslide susceptibility, including very high susceptibility, high susceptibility, medium susceptibility, low susceptibility, and very low susceptibility. The receiver operating characteristic (ROC) curve was applied to verify the model accuracy. The results showed that the AHP-IV model (ROC = 0.7716) was more suitable in RS, and the IV model (ROC = 0.8237) was the most appropriate model in NRS. Finally, combined with the results of landslide susceptibility in RS and NRS, an integrated landslide susceptibility map was proposed, involving year-round high susceptibility, RS high susceptibility, NRS high susceptibility and year-round low susceptibility. The integrated landslide susceptibility results provide a more detailed division in terms of the different time periods in a year, which is beneficial for the government to efficiently allocate landslide management funds and propose effective landslide management strategies. Additionally, the focused arrangement of monitoring works in landslide-prone areas enable collect landslide information efficiently, which is helpful for the subsequent landslide preventive management.

This work was to generate landslide susceptibility maps for the Three Gorges Reservoir (TGR) area, China by using different machine learning models. Three advanced machine learning methods, namely, gradient boosting decision tree (GBDT), random forest (RF) and information value (InV) models, were used, and the performances were assessed and compared. In total, 202 landslides were mapped by using a series of field surveys, aerial photographs, and reviews of historical and bibliographical data. Nine causative factors were then considered in landslide susceptibility map generation by using the GBDT, RF and InV models. All of the maps of the causative factors were resampled to a resolution of 28.5 m. Of the 486289 pixels in the area, 28526 pixels were landslide pixels, and 457763 pixels were non-landslide pixels. Finally, landslide susceptibility maps were generated by using the three machine learning models, and their performances were assessed through receiver operating characteristic (ROC) curves, the sensitivity, specificity, overall accuracy (OA), and kappa coefficient (KAPPA). The results showed that the GBDT, RF and InV models in overall produced reasonable accurate landslide susceptibility maps. Among these three methods, the GBDT method outperforms the other two machine learning methods, which can provide strong technical support for producing landslide susceptibility maps in TGR.

The landslide susceptibility prediction (LSP) is generally implemented using certain types of single models; however, certain drawbacks exist in the single models; e.g., it is difficult to clearly reflect the weights of landslide-related environmental factors if only the information value (IV) model is adopted. To overcome these limitations, this study proposes an IV-logistic regression (IV-LR) model for LSP. The landslides that occurred in the southern part of Chongyi County, China, are used as study cases. Nine environmental factors—elevation, slope, plane curvature, profile curvature, relief amplitude, distance to river, lithology, normalized difference vegetable index, and normalized difference built-up index—are adopted based on remote sensing and geographic information system. Certain landslide grid units and the same number of non-landslide grid units are used as the output variables of these models. The IV, LR, and IV-LR models are used to implement the LSP in the southern part of Chongyi County. The predicted landslides susceptibility in Chongyi County mostly occurred in areas with low elevations, close distance to rivers, carbonate lithology, low vegetation coverage rate, and densely populated areas. The results show that the prediction rate of the IV-LR model (80.4%) is higher than that of the LR model (76.8%), followed by the IV model (72.8%); they further demonstrate that the IV-LR model has its unique superiority and rationality compared with the IV and LR models.

Landslides have frequently occurred in deeply incised valleys in the upper reaches of the Minjiang River. Long-term interactions between rock uplift and river undercutting developed widely distributed landslides in this catchment, which recorded the typical tectonic geomorphology in the eastern margin of the Tibetan Plateau. In this study, we examined the landslides in the Minjiang catchment and aimed to compare the prediction ability of the statistical and machine learning (ML) models in landslide susceptibility assessment. We adopted the statistical models of the frequency ratio (FR) and information value (IV) models, and the ML models represented by a logistic model tree (LMT) and radial basis function classifier (RBFC) for landslide prediction. An inventory map of 668 landslides was compiled, and the landslides were randomly divided into training (80%) and validation (20%) datasets. Furthermore, 11 control factors of landslides based on topography, geology, hydrology, and other environments were applied for the analysis. The comprehensive performance of the four models was validated and compared using accuracy and area under the receiver operating characteristic curve (AUC). The results indicated that both sides of the valley along the Mingjiang and Heishuihe Rivers are in the high and very high susceptibility zones; in particular, the river segment from Wenchuan to Maoxian County has the highest susceptibility. The AUC values of the FR, IV, LMT, and RBFC models with the training data were 0.842, 0.862, 0.898, and 0.894, respectively, while the validation dataset illustrated the highest AUC value of 0.879 in the LMT model, followed by the RBFC (0.871), IV (0.869), and FR (0.839) models. Moreover, the LMT and RBFC models had higher accuracy values than the FR and IV models. This suggests that the ML models are superior to the statistical models in generating adequate landslide susceptibility maps, and the LMT model is the most efficient one for landslide prediction in the study region. This study provides a typical case in a landslide-prone region in the plateau margin to advance the understanding of landslide susceptibility assessment.

This paper aims to identify potential areas of landslides in the Amzaz watershed in northern Morocco with its precarious environmental balance using the Information Value (IV) Model. Van Westen (1994) defines bivariate methods as a modified form of the quantitative map combination with the exception that weightings are assigned based upon the statistical relationship between past landslides and various factor maps, individual factor maps (independent variable). A set of factor maps were overlaid with a landslide map (dependent variable) to create cross-tabulations for each one and class. The landslide inventory is used to result in the susceptibility maps for better mitigation of the risks and losses related to this phenomenon. The results demonstrated that the percentage of rotational landslides varies between 8.79 and 30.08%, and between 9.79 and 23.36% for translational slides susceptibility in the Amzaz watershed.

This study investigates the application of information value (InV) and logistic regression (LR) models for producing landslide susceptibility maps (LSMs) of the Zigui–Badong area near the Three Gorges Reservoir in China. This area is subject to anthropogenic influences because the reservoir’s water level cyclically fluctuates between 145 and 175 m. In addition, the area suffers from extreme rainfall events due to the local climate and has experienced significant and widespread landslide events in recent years. In this study, a landslide inventory map was initially constructed using field surveys, aerial photographs, and a literature search of historical landslide records. Eight causative factors, including lithology, bedding structure, slope, aspect, elevation, profile curvature, plane curvature, and fractional vegetation cover, were then considered in the generation of LSMs by using the InV and LR models. Finally, the prediction performances of these maps were assessed through receiver operating characteristics (ROC) that utilized both success-rate and prediction-rate curves. The validation results showed that the area under the ROC curve for the InV model was 0.859 for the success-rate curve and 0.865 for prediction-rate curve; these results indicate the InV model surpassed the LR model (0.742 for success-rate curve and 0.740 for prediction-rate curve). Overall, the two models provided nearly similar results. The results of this study show that landslide susceptibility mapping in the Zigui–Badong area is viable with both approaches.

Landslides, as geological hazards, cause significant casualties and economic losses. Therefore, it is necessary to identify areas prone to landslides for prevention work. This paper proposes an improved information value model based on gray clustering (IVM-GC) for landslide susceptibility mapping. This method uses the information value derived from an information value model to achieve susceptibility classification and weight determination of landslide predisposing factors and, hence, obtain the landslide susceptibility of each study unit based on the clustering analysis. Using a landslide inventory of Chongqing, China, which contains 8435 landslides, three landslide susceptibility maps were generated based on the common information value model (IVM), an information value model improved by an analytic hierarchy process (IVM-AHP) and our new improved model. Approximately 70% (5905) of the inventory landslides were used to generate the susceptibility maps, while the remaining 30% (2530) were used to validate the results. The training accuracies of the IVM, IVM-AHP and IVM-GC were 81.8%, 78.7% and 85.2%, respectively, and the prediction accuracies were 82.0%, 78.7% and 85.4%, respectively. The results demonstrate that all three methods perform well in evaluating landslide susceptibility. Among them, IVM-GC has the best performance.

The completeness of landslide inventories and the selection of evaluation models significantly impact the accuracy of landslide susceptibility assessments. Conventional field geological survey methods and single remote-sensing technology struggle to reliably identify landslides under complex environmental conditions. Moreover, prevalent landslide susceptibility evaluation models are often plagued by issues such as subjectivity and overfitting. Therefore, we investigated the uncertainty in susceptibility modeling from the aspects of landslide inventory quality and model selection. The study focused on Luquan County in Yunnan Province, China. Leveraging multisource remote-sensing technologies, particularly emphasizing optical remote sensing and InSAR time-series deformation detection, the existing historical landslide inventory was refined and updated. This updated inventory was subsequently used to serve as samples. Nine evaluation indicators, encompassing factors such as distance to faults and tributaries, lithology, distance to roads, elevation, slope, terrain undulation, distance to the main streams, and average annual precipitation, were selected on the basis of the collation and organization of regional geological data. The information value and two coupled machine-learning models were formulated to evaluate landslide susceptibility. The evaluation results indicate that the two coupled models are more appropriate for susceptibility modeling than the single information value (IV) model, with the random forest model optimized by genetic algorithm in Group I2 exhibiting higher predictive accuracy (AUC = 0.796). Furthermore, comparative evaluation results reveal that, under equivalent model conditions, the incorporation of a remote-sensing landslide inventory significantly enhances the accuracy of landslide susceptibility assessment results. This study not only investigates the impact of landslide inventories and models on susceptibility outcomes but also validates the feasibility and scientific validity of employing multisource remote-sensing technologies in landslide susceptibility assessment.