Scholarlink

<p>The accuracy of landslide susceptibility assessment (LSA) highly depends on the precision of conditioning factors, yet a systematic framework for quantifying errors in both DEM-derived continuous and spatially projected discrete factors is still lacking. To address this gap, this study proposes a novel error inversion framework to objectively determine the optimal spatial precision for both factor types under actual terrain conditions. For continuous factors, Gaussian synthetic surfaces and analytical derivations are combined with simulation experiments to evaluate truncation and propagation errors arising from noise in DEM-derived topography. For discrete factors, rasterization-induced uncertainty is assessed using area, perimeter, and shape-change metrics. By integrating these analyses, the hybrid error inversion model is established to identify the optimal DEM resolution that minimizes overall factor uncertainty. Finally, the Zagunao River Basin in Sichuan Province, Southwest China, is taken as a case to demonstrate the application of this framework in LSA. Results indicate that truncation and DEM topography-derived noise propagation errors for continuous factors exhibit opposing trends with increasing resolution and intersect at an optimal scale. In contrast, discrete factor errors show no systematic resolution dependence but reveal a unique minimum error within a topography resolution sequence. The inversion identifies 13&#xa0;m as the optimal factor resolution for LSA in the study area. The findings provide a transferable methodology for quantifying factor errors, offering practical guidance for optimizing terrain resolution in multi-hazard assessment applications.</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

A novel error inversion method for landslide susceptibility assessment based on topography-derived and spatially projected factors

  • 期刊论文

  • 发表日期：

    2026年5月22日

查看全文
  • Zemin Gao,
  • Xiaoguang Cai,
  • Hao Luo,
  • Mingtao Ding,
  • Zhenjiang Yu,
  • Muhammad Hassan,
  • Xingwang Liu,
  • Lei Duan

摘要

The accuracy of landslide susceptibility assessment (LSA) highly depends on the precision of conditioning factors, yet a systematic framework for quantifying errors in both DEM-derived continuous and spatially projected discrete factors is still lacking. To address this gap, this study proposes a novel error inversion framework to objectively determine the optimal spatial precision for both factor types under actual terrain conditions. For continuous factors, Gaussian synthetic surfaces and analytical derivations are combined with simulation experiments to evaluate truncation and propagation errors arising from noise in DEM-derived topography. For discrete factors, rasterization-induced uncertainty is assessed using area, perimeter, and shape-change metrics. By integrating these analyses, the hybrid error inversion model is established to identify the optimal DEM resolution that minimizes overall factor uncertainty. Finally, the Zagunao River Basin in Sichuan Province, Southwest China, is taken as a case to demonstrate the application of this framework in LSA. Results indicate that truncation and DEM topography-derived noise propagation errors for continuous factors exhibit opposing trends with increasing resolution and intersect at an optimal scale. In contrast, discrete factor errors show no systematic resolution dependence but reveal a unique minimum error within a topography resolution sequence. The inversion identifies 13 m as the optimal factor resolution for LSA in the study area. The findings provide a transferable methodology for quantifying factor errors, offering practical guidance for optimizing terrain resolution in multi-hazard assessment applications.