Scholarlink

<p>The present study investigates the link between geological environmental issues and human engineering activities, focusing on how these interactions often trigger landslides that threaten lives and infrastructure. It highlights the significance of landslide susceptibility assessment in urban planning and disaster prevention, using a dataset of 224 historical landslide instances analyzed through remote sensing and field investigation techniques. The study focuses on nineteen factors influencing landslides, employing the weight of evidence (WoE) method to evaluate their spatial correlation and optimize factor selection. Using GRASS GIS and the <i>r.slopeunits</i> method, the study analyzed 180 slope units dataset of different scales to determine the optimal configuration for landslide susceptibility modeling, based on accuracy and a comprehensive index. Multiple sets of landslide susceptibility models, including alternating decision tree (ADT), rotation forest - alternating decision tree (RF-ADT), random subspace - alternating decision tree (RS-ADT), forest by penalizing attribute (FPA), rotation forest - forest by penalizing attribute (RF-FPA), random subspace - forest by penalizing attribute (RS-FPA) and random forest (RAF), based on decision algorithms were established on the basis of the optimal slope units for comparative analysis of landslide susceptibility. The results show that extracting 12,282 slope units using a 200&#xa0;m spatial resolution, a minimum circular variance (cv) of 0.1, and a minimum surface area (amin) of 290,000&#xa0;m² yielded the best internal uniformity and external heterogeneity. Among the tested models, the RF-ADT integrated model showed the best performance (AUROC = 0.791), proving to be a stable and reliable tool for creating landslide susceptibility zoning maps in Shenmu City. These maps are supports government and engineering decisions, providing a valuable resource for safer urban planning and effective disaster mitigation strategies.</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 susceptibility evaluation by spatial data-driven technology under different resolutions of the slope units

  • 期刊论文

  • 发表日期：

    2025年11月1日

查看全文
  • Xia Zhao,
  • Wei Chen,
  • Paraskevas Tsangaratos,
  • Ioanna Ilia

摘要

The present study investigates the link between geological environmental issues and human engineering activities, focusing on how these interactions often trigger landslides that threaten lives and infrastructure. It highlights the significance of landslide susceptibility assessment in urban planning and disaster prevention, using a dataset of 224 historical landslide instances analyzed through remote sensing and field investigation techniques. The study focuses on nineteen factors influencing landslides, employing the weight of evidence (WoE) method to evaluate their spatial correlation and optimize factor selection. Using GRASS GIS and the r.slopeunits method, the study analyzed 180 slope units dataset of different scales to determine the optimal configuration for landslide susceptibility modeling, based on accuracy and a comprehensive index. Multiple sets of landslide susceptibility models, including alternating decision tree (ADT), rotation forest - alternating decision tree (RF-ADT), random subspace - alternating decision tree (RS-ADT), forest by penalizing attribute (FPA), rotation forest - forest by penalizing attribute (RF-FPA), random subspace - forest by penalizing attribute (RS-FPA) and random forest (RAF), based on decision algorithms were established on the basis of the optimal slope units for comparative analysis of landslide susceptibility. The results show that extracting 12,282 slope units using a 200 m spatial resolution, a minimum circular variance (cv) of 0.1, and a minimum surface area (amin) of 290,000 m² yielded the best internal uniformity and external heterogeneity. Among the tested models, the RF-ADT integrated model showed the best performance (AUROC = 0.791), proving to be a stable and reliable tool for creating landslide susceptibility zoning maps in Shenmu City. These maps are supports government and engineering decisions, providing a valuable resource for safer urban planning and effective disaster mitigation strategies.