Scholarlink

<p>Landslides pose significant risks to infrastructure and human life, particularly in mountainous regions like Lahaul and Spiti in Himachal Pradesh, India. This study focuses on assessing the vulnerability of road networks with landslide risks serving as the primary environmental hazard. Using a combination of machine learning algorithms and traditional statistical methods, the study develops road network vulnerability maps to identify segments most at risk of disruption. The models applied include Logistic Regression (LR), Adaboost, Neural Networks (Nnet), SVM Radial, Random Forest (RF), MARS, Information Value (IV), Frequency Ratio (FR), and Weight of Evidence (WoE). The Random Forest (RF) model performed best, achieving an AUC of 0.954, and was used to generate a detailed road vulnerability map. The findings indicate that 60% of National Highway 3 (NH3) and 48.59% of State Highway 26 (SH26) fall within high-risk zones, largely due to slope and proximity to rivers. The results provide critical insights for road planners and disaster management agencies to develop targeted interventions in high-risk areas. The study highlights the importance of integrating landslide susceptibility in road network planning and recommends the future use of real-time data for more accurate predictions.</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

Landslide-induced vulnerability of road networks in Lahaul and Spiti, India: a geospatial study

  • 期刊论文

  • 发表日期：

    2025年5月24日

查看全文
  • Devraj Dhakal,
  • Kanwarpreet Singh,
  • Damandeep Kaur,
  • Sahil Verma,
  • Abdullah H. Alsabhan,
  • Shamshad Alam,
  • Osamah J. Al-sareji,
  • Randeep,
  • Kavita

摘要

Landslides pose significant risks to infrastructure and human life, particularly in mountainous regions like Lahaul and Spiti in Himachal Pradesh, India. This study focuses on assessing the vulnerability of road networks with landslide risks serving as the primary environmental hazard. Using a combination of machine learning algorithms and traditional statistical methods, the study develops road network vulnerability maps to identify segments most at risk of disruption. The models applied include Logistic Regression (LR), Adaboost, Neural Networks (Nnet), SVM Radial, Random Forest (RF), MARS, Information Value (IV), Frequency Ratio (FR), and Weight of Evidence (WoE). The Random Forest (RF) model performed best, achieving an AUC of 0.954, and was used to generate a detailed road vulnerability map. The findings indicate that 60% of National Highway 3 (NH3) and 48.59% of State Highway 26 (SH26) fall within high-risk zones, largely due to slope and proximity to rivers. The results provide critical insights for road planners and disaster management agencies to develop targeted interventions in high-risk areas. The study highlights the importance of integrating landslide susceptibility in road network planning and recommends the future use of real-time data for more accurate predictions.