ArticlePublisher preview available

Landslide susceptibility assessment for the Darjeeling Toy Train route: a GIS and machine learning approach

Authors:
Prasanya Sarkar at COOCH BEHAR PANCHANAN BARMA UNIVERSITY
Prasanya Sarkar
  • COOCH BEHAR PANCHANAN BARMA UNIVERSITY
Madhumita Mondal
Madhumita Mondal
Madhumita Mondal
  • This person is not on ResearchGate, or hasn't claimed this research yet.
Alok Sarkar
Alok Sarkar
Shasanka Kumar Gayen at COOCH BEHAR PANCHANAN BARMA UNIVERSITY
Shasanka Kumar Gayen
  • COOCH BEHAR PANCHANAN BARMA UNIVERSITY
Read publisher preview
Download citation
To read the full-text of this research, you can request a copy directly from the authors.

Abstract and Figures

Landslide susceptibility mapping is crucial for reducing risks in culturally and historically significant areas like the Darjeeling Toy Train route, a UNESCO World Heritage site. In this study, the risk of landslides along this road is evaluated using Geographic Information System (GIS) tools and advanced machine learning models, such as Support Vector Machine (SVM), Gradient Boosting Machine (GBM), Logistic Regression, and Classification and Regression Trees (CART). It uses a set of 512 landslide and non-landslide sites, with a 70:30 split between training and testing. Within the research area, thirteen topographical, hydrological, and geological factors linked to landslides are shown as GIS layers to make maps of landslide susceptibility (LSM). The study area particularly vulnerable to various types of landslides, including debris slides, rock falls, and soil slips. ROC–AUC results show that the SVM model did the best (0.813), followed by GBM (0.807), Logistic Regression (0.797), and CART (0.781). SVM had the highest accuracy rate at 83.2%, followed by GBM at 81.5% and LR at 80.3%. CART had the lowest overall accuracy rate at 78.6%. Furthermore, confusion matrix analysis showed that SVM and Logistic Regression were better at finding actual landslide-prone areas, with 84.6% and 82.1% recall rates, respectively. This made them more accurate in predicting high-risk areas. Susceptibility levels were categorized, revealing high-risk areas like Darjeeling and Rishihat and safer areas like Kurseong and Mohanbari. For lowering the risk of landslides and protecting this historic route, these results are very useful for land management and disaster preparation.
Figures - available from: Stochastic Environmental Research and Risk Assessment
This content is subject to copyright. Terms and conditions apply.
A preview of this full-text is provided by Springer Nature.
Learn more
Content available from Stochastic Environmental Research and Risk Assessment
This content is subject to copyright. Terms and conditions apply.
ORIGINAL PAPER
Stochastic Environmental Research and Risk Assessment (2025) 39:613–637
https://doi.org/10.1007/s00477-024-02885-y
1 Introduction
In recent years, there has been increased attention to land-
slides and related slope movements due to their uncontrol-
lable nature (Khalil et al. 2022). These natural events, which
can cause signicant damage to lives and property, occur
frequently worldwide, especially in mountainous regions
(Sarkar and Kanungo 2004; Bera et al. 2019). Landslides,
characterized by the sudden collapse of rock or soil on a
slope (Varnes 1978; Goetz et al. 2011; Chakraborty and
Pradhan 2012), are often triggered by factors such as heavy
rainfall, earthquakes, rapid snowmelt, and human activities
(Mersha and Meten 2020). Despite their smaller scale than
other natural disasters, landslides are common and can be
highly destructive (Trigila et al. 2015; Nohani et al. 2019).
In India, the region most vulnerable to landslides include
the Himalayan mountain ranges, the Meghalaya plateau,
and the Western Ghats (Mondal and Mandal 2019).
Prasanya Sarkar
sarkarprasanya@gmail.com
Madhumita Mondal
mgayen79@gmail.com
Alok Sarkar
aloksarkar162@gmail.com
Shasanka Kumar Gayen
gshasanka@gmail.com
1 Department of Geography, Cooch Behar Panchanan Barma
University, Vivekananda Street, Cooch Behar,
West Bengal 736101, India
2 Department of Geography, Bhairab Ganguli College,
Kolkata, West Bengal 700056, India
3 University of Calcutta, Kolkata, West Bengal 700019, India
Abstract
Landslide susceptibility mapping is crucial for reducing risks in culturally and historically signicant areas like the
Darjeeling Toy Train route, a UNESCO World Heritage site. In this study, the risk of landslides along this road is evalu-
ated using Geographic Information System (GIS) tools and advanced machine learning models, such as Support Vec-
tor Machine (SVM), Gradient Boosting Machine (GBM), Logistic Regression, and Classication and Regression Trees
(CART). It uses a set of 512 landslide and non-landslide sites, with a 70:30 split between training and testing. Within the
research area, thirteen topographical, hydrological, and geological factors linked to landslides are shown as GIS layers to
make maps of landslide susceptibility (LSM). The study area particularly vulnerable to various types of landslides, includ-
ing debris slides, rock falls, and soil slips. ROC–AUC results show that the SVM model did the best (0.813), followed by
GBM (0.807), Logistic Regression (0.797), and CART (0.781). SVM had the highest accuracy rate at 83.2%, followed by
GBM at 81.5% and LR at 80.3%. CART had the lowest overall accuracy rate at 78.6%. Furthermore, confusion matrix
analysis showed that SVM and Logistic Regression were better at nding actual landslide-prone areas, with 84.6% and
82.1% recall rates, respectively. This made them more accurate in predicting high-risk areas. Susceptibility levels were
categorized, revealing high-risk areas like Darjeeling and Rishihat and safer areas like Kurseong and Mohanbari. For
lowering the risk of landslides and protecting this historic route, these results are very useful for land management and
disaster preparation.
Keywords Darjeeling Toy Train · Landslide-prone areas · Support vector machine (SVM) · Gradient boosting machine
(GBM) · Logistic regression · CART · Disaster preparedness
Accepted: 11 December 2024 / Published online: 31 December 2024
© The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2024
Landslide susceptibility assessment for the Darjeeling Toy Train route:
a GIS and machine learning approach
PrasanyaSarkar1· MadhumitaMondal2· AlokSarkar3· Shasanka KumarGayen1
1 3
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
ResearchGate has not been able to resolve any citations for this publication.
Development of an automated method for flood inundation monitoring, flood hazard, and soil erosion susceptibility assessment using machine learning and AHP–MCE techniques
Article
Full-text available
  • Mar 2024
Operational large-scale flood monitoring using publicly available satellite data is possible with the advent of Sentinel-1 microwave data, which enables near-real-time (at 6-day intervals) flood mapping day and night, even in cloudy monsoon seasons. Automated flood inundation area identification in near-real-time involves advanced geospatial data processing platforms, such as Google Earth Engine and robust methodology (Otsu’s algorithm). Objectives The current study employs Sentinel-1 microwave data for flood extent mapping using machine learning (ML) algorithms in Assam State, India. We generated a flood hazard and soil erosion susceptibility map by combining multi-source data on weather conditions and soil and terrain characteristics. Random Forest (RF), Classification and Regression Tool (CART), and Support Vector Machine (SVM) ML algorithms were applied to generate the flood hazard map. Furthermore, we employed the multicriteria evaluation (MCE) analytical hierarchical process (AHP) for soil erosion susceptibility mapping. Summary The highest prediction accuracy was observed for the RF model (overall accuracy [OA] > 82%), followed by the SVM (OA > 82%) and CART (OA > 81%). Over 26% of the study area indicated high flood hazard-prone areas, and approximately 60% showed high and severe potential for soil erosion due to flooding. The automated flood mapping platform is an essential resource for emergency responders and decision-makers, as it helps to guide relief activities by identifying suitable regions and appropriate logistic route planning and improving the accuracy and timeliness of emergency response efforts. Periodic flood inundation maps will help in long-term planning and policymaking, flood management, soil and biodiversity conservation, land degradation, planning sustainable agriculture interventions, crop insurance, and climate resilience studies.
View
Exploring machine learning and statistical approach techniques for landslide susceptibility mapping in Siwalik Himalayan Region using geospatial technology
Article
Full-text available
  • Jan 2024
  • ENVIRON SCI POLLUT R
Landslides are a natural threat that poses a severe risk to human life and the environment. In the Kumaon mountains region in Uttarakhand (India), Nainital is among the most vulnerable areas prone to landslides inflicting harm to livelihood and civilization due to frequent landslides. Developing a landslide susceptibility map (LSM) in this Nainital area will help alleviate the probability of landslide occurrence. GIS and statistical-based approaches like the certainty factor (CF), information value (IV), frequency ratio (FR) and logistic regression (LR) are used for the assessment of LSM. The landslide inventories were prepared using topography, satellite imagery, lithology, slope, aspect, curvature, soil, land use and land cover, geomorphology, drainage density and lineament density to construct the geodatabase of the elements affecting landslides. Furthermore, the receiver operating characteristic (ROC) curve was used to check the accuracy of the predicting model. The results for the area under the curves (AUCs) were 87.8% for logistic regression, 87.6% for certainty factor, 87.4% for information value and 84.8% for frequency ratio, which indicates satisfactory accuracy in landslide susceptibility mapping. The present study perfectly combines GIS and statistical approaches for mapping landslide susceptibility zonation. Regional land use planners and natural disaster management will benefit from the proposed framework for landslide susceptibility maps.
View
Development and assessment of a novel hybrid machine learning-based landslide susceptibility mapping model in the Darjeeling Himalayas
Article
Full-text available
  • Aug 2023
  • STOCH ENV RES RISK A
Natural disasters like landslides risk people's lives and the environment. To mitigate these hazards, scientists employ landslide susceptibility mapping that evaluates zones prone to landslides and identifies contributing factors. This study aimed to introduce ensemble landslide susceptibility models based on a statistical model (analytical hierarchy process) and a machine learning model (support vector machine) for predicting landslides in a part of the Darjeeling district, West Bengal, India. A total of 114 landslide locations were identified and randomly divided into training and validation databases, with proportions of 70% and 30%, respectively. Ten conditioning factors were considered, including rainfall, soil texture, slope, aspect, geomorphology, lithology, curvature, land use and land cover, drainage density, and lineament density. The AHP-SVM model, employing linear, polynomial, radial basis function (RBF), and sigmoid function algorithms, was applied using the training landslide and non-landslide datasets along with the spatial database of conditioning factors. Four landslide susceptibility maps were generated using this model, and their accuracy was assessed using the area under the curve (AUC) of the receiver operating characteristic (ROC) tool with the validation dataset. Among the four ensemble methods tested (AHP-SVM_Sigmoid, AHP-SVM_RBF, AHP-SVM_Polynomial, and AHP-SVM_Linear), the AHP-SVM_Sigmoid model demonstrated the highest degree-of-fit and prediction performance, achieving a prediction capability of 86.2%. Consequently, it is concluded that the AHP-SVM_Sigmoid ensemble model holds promise as a novel technique for spatial landslide prediction in future studies. The findings of this study are valuable for local planning and decision-makers and can be utilized to implement landslide susceptibility methods in other regions.
View
A Multi-Criteria Decision Analysis (MCDA) Approach for Landslide Susceptibility Mapping of a Part of Darjeeling District in North-East Himalaya, India
Article
Full-text available
  • Apr 2023
Landslides are the nation’s hidden disaster, significantly increasing economic loss and social disruption. Unfortunately, limited information is available about the depth and extent of landslides. Therefore, in order to identify landslide-prone zones in advance, a well-planned landslide susceptibility mapping (LSM) approach is needed. The present study evaluates the efficacy of an MCDA-based model (analytical hierarchy process (AHP)) and determines the most accurate approach for detecting landslide-prone zones in one part of Darjeeling, India. LSM is prepared using remote sensing thematic layers such as slope, rainfall earthquake, lineament density, drainage density, geology, geomorphology, aspect, land use and land cover (LULC), and soil. The result obtained is classified into four classes, i.e., very high (11.68%), high (26.18%), moderate (48.87%), and low (13.27%) landslide susceptibility. It is observed that an entire 37.86% of the area is in a high to very high susceptibility zone. The efficiency of the LSM was validated with the help of the receiver operating characteristics (ROC) curve, which demonstrate an accuracy of 96.8%, and the success rate curve showed an accuracy of 81.3%, both of which are very satisfactory results. Thus, the proposed framework will help natural disaster experts to reduce land vulnerability, as well as aid in future development.
View
Comparison of optimized data-driven models for landslide susceptibility mapping
Article
Full-text available
  • Apr 2023
  • Environ Dev Sustain
Locations prone to landslides must be identified and mapped to prevent landslide-related damage and casualties. Machine learning approaches have proven effective for such tasks and have thus been widely applied. However, owing to the rapid development of data-driven approaches, deep learning methods that can exhibit enhanced prediction accuracies have not been fully evaluated. Several researchers have compared different methods without optimizing them, whereas others optimized a single method using different algorithms and compared them. In this study, the performances of different fully optimized methods for landslide susceptibility mapping within the landslide-prone Kermanshah province of Iran were compared. The models, i.e., convolutional neural networks (CNNs), deep neural networks (DNNs), and support vector machine (SVM) frameworks were developed using 14 conditioning factors and a landslide inventory containing 110 historical landslide points. The models were optimized to maximize the area under the receiver operating characteristic curve (AUC), while maintaining their stability. The results showed that the CNN (accuracy = 0.88, root mean square error (RMSE) = 0.37220, and AUC = 0.88) outperformed the DNN (accuracy = 0.79, RMSE = 0.40364, and AUC = 0.82) and SVM (accuracy = 0.80, RMSE = 0.42827, and AUC = 0.80) models using the same testing dataset. Moreover, the CNN model exhibiting the highest robustness among the three models, given its smallest AUC difference between the training and testing datasets. Notably, the dataset used in this study had a low spatial accuracy and limited sample points, and thus, the CNN approach can be considered useful for susceptibility assessment in other landslide-prone regions worldwide, particularly areas with poor data quality and quantity. The most important conditioning factors for all models were rainfall and the distances from roads and drainages.
View
View
Landslide Susceptibility Mapping Based on Multitemporal Remote Sensing Image Change Detection and Multiexponential Band Math
Article
Full-text available
  • Jan 2023
Landslides pose a great threat to the safety of people’s lives and property within disaster areas. In this study, the Zigui to Badong section of the Three Gorges Reservoir is used as the study area, and the land use (LU), land use change (LUC) and band math (band) factors from 2016–2020 along with six selected commonly used factors are used to form a land use factor combination (LUFC), land use change factor combination (LUCFC) and band math factor combination (BMFC). An artificial neural network (ANN), a support vector machine (SVM) and a convolutional neural network (CNN) are chosen as the three models for landslide susceptibility mapping (LSM). The results show that the BMFC is generally better than the LUFC and the LUCFC. For the validation set, the highest simple ranking scores for the three models were obtained for the BMFC (37.2, 32.8 and 39.2), followed by the LUFC (28, 26.6 and 31.8) and the LUCFC (26.8, 28.6 and 20); that is, the band-based predictions are better than those based on the LU and LUC, and the CNN model provides the best prediction ability. According to the four groups of experimental results with ANNs, compared with LU and LUC, band is easier to access, yields higher predictive performance, and provides stronger stability. Thus, band can replace LU and LUC to a certain extent and provide support for automatic and real-time landslide monitoring.
View
Influence of slope position and aspect on the vegetation attributes and treewater relations in forests of the central Himalayas
Article
  • Aug 2023
  • J MT SCI-ENGL
While the need for understanding the effects of topographical factors on forest structure and function is well recognized, comprehensive studies are scarce. This study evaluates the effect of slope aspect and slope position on water relations and forest attributes across six forest types occurring between 400 m and 2600 m altitude in the Central Himalaya (27°-38°N). We found that predawn tree water potential and soil water potential were generally higher on moist north slope aspect (-0.78±0.05 MPa and-3.34±0.18 MPa, respectively) than dry south slope aspect (-0.82±0.18 MPa and-3.77±0.18 MPa, respectively). Across six different forests, these values were higher at hill base (-0.71±0.06 MPa and-2.77±0.19 MPa, tree predawn water potential and soil water potential, respectively) than other topographical positions. The favorable effect of north aspect and hill base was also observed in maintaining soil water and tree water potential during the dry season. Vegetation attributes, such as species richness, unique species and plant density were also generally higher on north slope and hill base than southern aspect and lowest at hill top. Across forest types, the hill base provided shelter to 46 unique species, compared to 16-18 at the other positions, thus emphasizing its importance as refugia for species to survive climate change induced perturbations. The favorable conditions of hill base position not only contribute to increase in alpha diversity, but also to extended species distributional range. Influence of slope position and aspect on the vegetation attributes and tree-water relations in forests of the central Himalayas
View
Prioritization of sub-watersheds of Teesta River according to soil erosion susceptibility using multi-criteria decision-making in Sikkim and West Bengal
Article
  • Jun 2023
Identification of the areas vulnerable to soil erosion through the prioritization of watersheds can help in the planning and execution of suitable conservational measures. In this study, prioritization for soil erosion of 14 sub-watersheds of the Teesta was determined through morphometric parameters using the analytical hierarchical process (AHP) and Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS). The relative weights of various parameters were determined using AHP. For the evaluation of soil erosion hazard, 10 factors are used: bifurcation ration (Rb), circulatory ratio (Rc), basin length (L), stream frequency (Fs), drainage density (Dd), basin perimeter (P), basin width (W), shape factor (Bs), drainage texture (Dt), and elongation ratio (Re). The results demonstrate that sub-watersheds 1 and 4 have been ranked 1 and 2 in terms of highest closeness (cli+{cl}_{i}^{+}) to an ideal solution with 0.774 and 0.434 respectively. These sub-watersheds must be given the highest priority for soil conservation measures, to ensure future sustainable agriculture.
View
Restoration Measures of Landslide Affected Valley Slope—A Case Study of Tindharia Area Along NH 55 in West Bengal
Chapter
  • Feb 2023
The existing road width including the narrow gauge railway line of “NF Railway” were affected partially to fully due to the occurrences of landslides on the valley slopes in total of three numbers of locations within the chainage from km 50 + 155 to 53 + 485 in and around location of “Tindharia” area along NH 55, i.e., Darjeeling–Siliguri road in the state of West Bengal on September 18, 2011. The landslides happened during the rainy season of the said year subsequent to the occurrence of earthquake in the adjoining area of Sikkim. Total height of the landslide affected valley slope in “Tindharia” area was in between 30 m and little more than 100 m. The probable reasons of landslide in the area were anticipated due to frequent occurrences of earthquakes in addition to the heavy rainfall, poor drainage, and also due to the fragile geological conditions of the existing slope. A detailed investigation program covering the topographical, geological cum geotechnical aspects of the affected area and also on the proposed construction materials were undertaken to formulate the suitable rehabilitations measures of the failed valley slopes. The detailed engineering design by analyzing the stability of proposed rebuilding of valley slopes corresponding to widening of the affected road bench was done as per the available standards and guidelines for ensuring the adequate safety under static and seismic conditions. The proposed restoration measures consisted of composite reinforced earth technology by strengthening the existing exposed valley slope surface with adequate steel nailing at first and then rebuilding of the road bench by newly added steep reinforced soil slope using layers of designed polymeric reinforcement in multi-tiered manner covering the entire affected heights. The nailing of the existing slope was proposed to connect suitably with the soil reinforcement of the newly added valley side steep slope. Adequate surface and subsurface drainage measures including the erosion protection schemes were provided for the restored valley slope. The surface cracks generated due to landslide at the existing road bench and also on the hill slope portion were sealed with suitable filling material. This paper presents the details of investigation, design analysis, and actual construction works done for the restoration of one of the landslide affected valley slopes of moderate height of 30–34 m in “Tindharia” area.
View