Scholarlink

<p>Snow disasters, intensified by climate change, pose significant risks to infrastructure, agriculture, and socio-economic systems. However, identifying snow disaster-prone areas remains a challenge due to the complex interactions among climatic, topographic, and exposure factors. This study introduces the Maximum Disaster Spatial Density (MDSD) method, a novel framework that directly integrates observed disaster records into an optimized spatial clustering system. Using datasets covering the entire South Korea from 2009 to 2018, comprising the mean and standard deviation of precipitation, temperature, and the MODIS-based Normalized Difference Snow Index (NDSI), as well as elevation and building density aggregated to a 1&#xa0;km grid, the MDSD iteratively adjusted feature weights to maximize areal disaster density within the top-ranked clusters. The results revealed high snow disaster proneness in Jeju Island and in inland regions such as the southern coast, the northwestern area, and parts of the eastern coast, whereas high-elevation areas showed lower proneness due to reduced exposure despite frequent snowfall. Cluster-level feature analysis highlighted winter precipitation, temperature, and persistent snow cover as primary drivers, while elevation had minimal influence. Building density significantly contributed only to the most vulnerable groups, reflecting its role as an exposure factor. Relative importance analysis showed that precipitation and building density were the most influential factors, followed by temperature, whereas NDSI and elevation had relatively minor effects. Probabilistic validation demonstrated that spatial patterns of snow-prone areas remained consistent across multiple optimization iterations, underscoring the robustness of the framework. Moreover, adjusting the selection criteria for clusters enables flexible application of the results for different disaster management objectives, such as emergency response or long-term mitigation. Overall, the MDSD method substantially improves the reliability, objectivity, and transferability of snow disaster-prone area assessments and provides a scalable framework that can be extended to other disaster types and regions.</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 approach of mapping snow disaster-prone areas based on areal disaster density optimization: a case study of South Korea

  • 期刊论文

  • 发表日期：

    2026年1月24日

查看全文
  • Soohyun Kim,
  • Dongkyun Kim,
  • Jinwook Lee,
  • Sayed M. Bateni,
  • Waqas Ahmad,
  • Jeongha Park

摘要

Snow disasters, intensified by climate change, pose significant risks to infrastructure, agriculture, and socio-economic systems. However, identifying snow disaster-prone areas remains a challenge due to the complex interactions among climatic, topographic, and exposure factors. This study introduces the Maximum Disaster Spatial Density (MDSD) method, a novel framework that directly integrates observed disaster records into an optimized spatial clustering system. Using datasets covering the entire South Korea from 2009 to 2018, comprising the mean and standard deviation of precipitation, temperature, and the MODIS-based Normalized Difference Snow Index (NDSI), as well as elevation and building density aggregated to a 1 km grid, the MDSD iteratively adjusted feature weights to maximize areal disaster density within the top-ranked clusters. The results revealed high snow disaster proneness in Jeju Island and in inland regions such as the southern coast, the northwestern area, and parts of the eastern coast, whereas high-elevation areas showed lower proneness due to reduced exposure despite frequent snowfall. Cluster-level feature analysis highlighted winter precipitation, temperature, and persistent snow cover as primary drivers, while elevation had minimal influence. Building density significantly contributed only to the most vulnerable groups, reflecting its role as an exposure factor. Relative importance analysis showed that precipitation and building density were the most influential factors, followed by temperature, whereas NDSI and elevation had relatively minor effects. Probabilistic validation demonstrated that spatial patterns of snow-prone areas remained consistent across multiple optimization iterations, underscoring the robustness of the framework. Moreover, adjusting the selection criteria for clusters enables flexible application of the results for different disaster management objectives, such as emergency response or long-term mitigation. Overall, the MDSD method substantially improves the reliability, objectivity, and transferability of snow disaster-prone area assessments and provides a scalable framework that can be extended to other disaster types and regions.