Scholarlink

<p>Susceptibility, vulnerability, and risk assessment (SVRA) related to landslides of the existing Himalayan township are essential for land-use planning of the area. Therefore, in the present study, a detailed SVRA of the Nainital township in the Lesser Himalaya, India, constituting thirteen municipal wards, has been carried out. Three machine learning approaches, such as Artificial Neural Network (ANN), Logistic Regression (LR), and Support Vector Machine (SVM), were used for the preparation of landslide susceptibility maps. ANN has a slightly higher success and prediction rate accuracy and is used for landslide risk assessment of the area. Intersecting different elements at risk with the landslide susceptibility map, the landslide risk assessment of the area has been evaluated at the local municipality ward level. It exhibits that ~ 44% of the area having ~ 4,757 buildings with habitation of ~ 24,000 persons lies in the areas categorized as having an elevated likelihood of landslides, including those at significant risk, mostly located in fewer municipality wards. These are mainly located in the vicinity of areas like Sher-Ka-Danda, Naina Peak, Balia Nala, Birla Vidhya Mandir School, and Rais Hotel Colony. Around ~ 30% of the area lies in moderate, and ~ 26% in low &amp; very low landslide risk zones. The outcomes from this research can be utilized for sustainable land use planning, at the meso- and micro-scale, and for further development of the region. It is recommended that high and very high-risk zones in the area should be declared ‘green areas’ with no further cutting or intervention of slopes, whereas in the moderate-risk zones, limited construction should be allowed, and most of the developmental activities, if necessary, at all, must be in the low and very-low-risk zones. Besides, an attempt should also be made to minimize the landslide susceptibility by stabilizing the active landslides either structurally or by non-structural means. Further, this risk map should be updated at regular intervals to incorporate any changes in environmental conditions in an area.</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

Application of machine learning (ML) to landslide susceptibility, vulnerability and risk assessment (SVRA): A case study from Nainital township, Kumaun Himalaya, India

  • 期刊论文

  • 发表日期：

    2026年5月23日

查看全文
  • Pratap Ram,
  • Vikram Gupta,
  • Ruchika S. Tandon

摘要

Susceptibility, vulnerability, and risk assessment (SVRA) related to landslides of the existing Himalayan township are essential for land-use planning of the area. Therefore, in the present study, a detailed SVRA of the Nainital township in the Lesser Himalaya, India, constituting thirteen municipal wards, has been carried out. Three machine learning approaches, such as Artificial Neural Network (ANN), Logistic Regression (LR), and Support Vector Machine (SVM), were used for the preparation of landslide susceptibility maps. ANN has a slightly higher success and prediction rate accuracy and is used for landslide risk assessment of the area. Intersecting different elements at risk with the landslide susceptibility map, the landslide risk assessment of the area has been evaluated at the local municipality ward level. It exhibits that ~ 44% of the area having ~ 4,757 buildings with habitation of ~ 24,000 persons lies in the areas categorized as having an elevated likelihood of landslides, including those at significant risk, mostly located in fewer municipality wards. These are mainly located in the vicinity of areas like Sher-Ka-Danda, Naina Peak, Balia Nala, Birla Vidhya Mandir School, and Rais Hotel Colony. Around ~ 30% of the area lies in moderate, and ~ 26% in low & very low landslide risk zones. The outcomes from this research can be utilized for sustainable land use planning, at the meso- and micro-scale, and for further development of the region. It is recommended that high and very high-risk zones in the area should be declared ‘green areas’ with no further cutting or intervention of slopes, whereas in the moderate-risk zones, limited construction should be allowed, and most of the developmental activities, if necessary, at all, must be in the low and very-low-risk zones. Besides, an attempt should also be made to minimize the landslide susceptibility by stabilizing the active landslides either structurally or by non-structural means. Further, this risk map should be updated at regular intervals to incorporate any changes in environmental conditions in an area.