Scholarlink

<p>Forest fire (FF) risk zonation is the process of dividing forest areas into different zones based on the likelihood of wildfire risk. This study proposes a method that combines forest fire risk indices (FFRIs) and a biodiversity index (BI) using a logtri-centric fuzzy logic system (LFLS) within a geographic information system (GIS) to identify forest fire risk zones. We examined various factors and fire danger indices, revealing that 44.62% of the incidents showed very high vulnerability, affecting an area of 529 km2. While, 54.11% were classified as high vulnerability category, covering 2818 km2, and 1.26%, as moderate vulnerability, spanning 2852.9 km2. The study involved preprocessing satellite data from Landsat 8, LISS-IV, and ASTER DEM, followed by clustering and feature extraction. Land cover was classified using the SpatianeL Region-based Convolutional Neural Network (SLRCNN), and lightning data were processed. Fire risk was further analysed using Kernel Density Analysis (KDA) and Hotspot Analysis (HA) while the biodiversity index (BI) was utilised to account for the ecological sensitivity. Then, by utilizing the weighted overlay analysis the risk map was generated and the LFLS was used to classify risk zones. The proposed method demonstrates high accuracy and robustness, achieving a low classification error of 0.03, in mapping forest fire risk zones.</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 Significant Framework for Forest Fire Risk Zonation Identification in the Sikkim Himalayas Using LFLS and GIS

  • 期刊论文

  • 发表日期：

    2026年6月3日

查看全文
  • Kapila Sharma,
  • Gopal Thapa

摘要

Forest fire (FF) risk zonation is the process of dividing forest areas into different zones based on the likelihood of wildfire risk. This study proposes a method that combines forest fire risk indices (FFRIs) and a biodiversity index (BI) using a logtri-centric fuzzy logic system (LFLS) within a geographic information system (GIS) to identify forest fire risk zones. We examined various factors and fire danger indices, revealing that 44.62% of the incidents showed very high vulnerability, affecting an area of 529 km2. While, 54.11% were classified as high vulnerability category, covering 2818 km2, and 1.26%, as moderate vulnerability, spanning 2852.9 km2. The study involved preprocessing satellite data from Landsat 8, LISS-IV, and ASTER DEM, followed by clustering and feature extraction. Land cover was classified using the SpatianeL Region-based Convolutional Neural Network (SLRCNN), and lightning data were processed. Fire risk was further analysed using Kernel Density Analysis (KDA) and Hotspot Analysis (HA) while the biodiversity index (BI) was utilised to account for the ecological sensitivity. Then, by utilizing the weighted overlay analysis the risk map was generated and the LFLS was used to classify risk zones. The proposed method demonstrates high accuracy and robustness, achieving a low classification error of 0.03, in mapping forest fire risk zones.