Scholarlink

<p>Landslides pose a persistent and deadly hazard across India’s diverse terrain, demanding robust, high-resolution national-scale assessments to guide effective risk mitigation and policy decisions. This study addresses a critical research gap in landslide susceptibility and risk assessment across India at a high spatial resolution of 90&#xa0;m × 90&#xa0;m on a national scale. Leveraging a robust landslide inventory comprising 109,504 documented events and incorporating causative factors selected through expert judgment, three models Analytic Hierarchy Process (AHP), Frequency Ratio (FR), and Yule’s Coefficient (Yc) were employed to reduce model bias and identify the most reliable susceptibility and risk mapping approach through comprehensive field validation and map interpretation. The results indicate, very high landslide susceptibility zones cover 4.0% (AHP), 4.2% (Yc), and 4.5% (FR) of India’s land area, while combined high and very high susceptibility classes account for 11.0% (AHP), 10.4% (Yc), and 10.5% (FR). Model evaluation demonstrates robust predictive performance, with accuracies of 89.1% (AHP), 88.9% (Yc), and 90.5% (FR) according to the ROC curve, and 86.7%, 85.6%, and 87.2% respectively via the precision-recall curve. Minor variations in susceptibility extent and accuracy among models underscore the resilience of the input data, however the Yc model provides a a more ground representative output based on factor weighting. Utilizing the Yc model and exposure data, approximately 8,606&#xa0;km² (0.26% of India) are classified as having a high to extremely high landslide risk at a qualitative scale. Overall, the susceptibility analysis indicates that approximately 0.34&#xa0;million km² is categorized as having high and very high susceptibility under Yc model. The analysis identifies Nagaland (55%), Mizoram (53.1%), Arunachal Pradesh (52.1%), Sikkim (45.0%), Uttarakhand (40.0%), Manipur (30.6%), Himachal Pradesh (28.9%), Meghalaya (22.0%), Jammu &amp; Kashmir (20.2%), and Tripura (8.0%) as the ten most landslide-susceptible States and Union Territories , listed in descending order of susceptibility. These findings provide critical insights for disaster risk reduction strategies and policymaking, enhancing landslide prediction, mitigation, and resilience planning across susceptible regions at mapped scale.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

National-scale landslide susceptibility and risk mapping of India using a hybrid data-driven approach

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年12月31日

查看全文
  • Imran Khan,
  • Harish Bahuguna,
  • Ashutosh Kainthola,
  • Rayees Ahmed,
  • Md. Sarfaraz Asgher

摘要

Landslides pose a persistent and deadly hazard across India’s diverse terrain, demanding robust, high-resolution national-scale assessments to guide effective risk mitigation and policy decisions. This study addresses a critical research gap in landslide susceptibility and risk assessment across India at a high spatial resolution of 90 m × 90 m on a national scale. Leveraging a robust landslide inventory comprising 109,504 documented events and incorporating causative factors selected through expert judgment, three models Analytic Hierarchy Process (AHP), Frequency Ratio (FR), and Yule’s Coefficient (Yc) were employed to reduce model bias and identify the most reliable susceptibility and risk mapping approach through comprehensive field validation and map interpretation. The results indicate, very high landslide susceptibility zones cover 4.0% (AHP), 4.2% (Yc), and 4.5% (FR) of India’s land area, while combined high and very high susceptibility classes account for 11.0% (AHP), 10.4% (Yc), and 10.5% (FR). Model evaluation demonstrates robust predictive performance, with accuracies of 89.1% (AHP), 88.9% (Yc), and 90.5% (FR) according to the ROC curve, and 86.7%, 85.6%, and 87.2% respectively via the precision-recall curve. Minor variations in susceptibility extent and accuracy among models underscore the resilience of the input data, however the Yc model provides a a more ground representative output based on factor weighting. Utilizing the Yc model and exposure data, approximately 8,606 km² (0.26% of India) are classified as having a high to extremely high landslide risk at a qualitative scale. Overall, the susceptibility analysis indicates that approximately 0.34 million km² is categorized as having high and very high susceptibility under Yc model. The analysis identifies Nagaland (55%), Mizoram (53.1%), Arunachal Pradesh (52.1%), Sikkim (45.0%), Uttarakhand (40.0%), Manipur (30.6%), Himachal Pradesh (28.9%), Meghalaya (22.0%), Jammu & Kashmir (20.2%), and Tripura (8.0%) as the ten most landslide-susceptible States and Union Territories , listed in descending order of susceptibility. These findings provide critical insights for disaster risk reduction strategies and policymaking, enhancing landslide prediction, mitigation, and resilience planning across susceptible regions at mapped scale.