<p>Located within the tectonically active Barail Range of the Eastern Himalayan Foothills and receiving intense seasonal precipitation as part of one of the world’s wettest monsoon belts, the Dima Hasao district of Assam is highly vulnerable to landslides. Therefore, the primary objectives of this study were to identify and evaluate the spatial patterns of landslide-prone areas through the integration of Geographic Information System (GIS) techniques, remote sensing data, and field surveys; to determine the relative importance of natural and human-induced causative factors; and to validate the predictive accuracy of the applied methods. Fourteen key causative factors were analysed using Frequency Ratio (FR), Weight of Evidence (WoE) and Shannon Entropy (SE) models to generate landslide susceptibility maps. Validation based on Receiver Operating Characteristic (ROC) and Area Under the Curve (AUC) demonstrated strong model performance, with the FR model achieving a success rate of 0.734 and prediction rate of 0.706, outperforming WoE and SE. The results indicated that critical influential factors include land use/land cover (notably barren land), proximity to roads, rainfall, lithology, and proximity to faults. Particularly across the central and eastern regions, the district falls within zones of very high landslide susceptibility, where the triggering factors are most prominent. Implementing targeted risk mitigation strategies and sustainable land-use policies in areas identified as highly susceptible to landslides is essential not only for safeguarding local communities and minimizing potential casualties and economic losses but also for promoting resilient development in Dima Hasao.</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

Geospatial modelling of landslide susceptibility using modified frequency ratio, weight of evidence, and shannon entropy: a case study in Dima Hasao district of Northeast India

  • Princi Gogoi,
  • Pixi Gogoi

摘要

Located within the tectonically active Barail Range of the Eastern Himalayan Foothills and receiving intense seasonal precipitation as part of one of the world’s wettest monsoon belts, the Dima Hasao district of Assam is highly vulnerable to landslides. Therefore, the primary objectives of this study were to identify and evaluate the spatial patterns of landslide-prone areas through the integration of Geographic Information System (GIS) techniques, remote sensing data, and field surveys; to determine the relative importance of natural and human-induced causative factors; and to validate the predictive accuracy of the applied methods. Fourteen key causative factors were analysed using Frequency Ratio (FR), Weight of Evidence (WoE) and Shannon Entropy (SE) models to generate landslide susceptibility maps. Validation based on Receiver Operating Characteristic (ROC) and Area Under the Curve (AUC) demonstrated strong model performance, with the FR model achieving a success rate of 0.734 and prediction rate of 0.706, outperforming WoE and SE. The results indicated that critical influential factors include land use/land cover (notably barren land), proximity to roads, rainfall, lithology, and proximity to faults. Particularly across the central and eastern regions, the district falls within zones of very high landslide susceptibility, where the triggering factors are most prominent. Implementing targeted risk mitigation strategies and sustainable land-use policies in areas identified as highly susceptible to landslides is essential not only for safeguarding local communities and minimizing potential casualties and economic losses but also for promoting resilient development in Dima Hasao.