Scholarlink

<p>Wildfires are recognized as a pervasive global threat that contribute to rangeland degradation and disrupt landscape connectivity for vulnerable species such as <i>M. peregrina</i> in southern Iran. The spatial overlap between wildfire probability, fire connectivity, and habitat suitability and connectivity has received limited attention. In this study, wildfire probability was modeled using MODIS active-fire data and environmental variables, including anthropogenic drivers, seasonal temperature dynamics, and topographic features. A weighted ensemble modeling approach based on Area Under the Curve (AUC) was employed to integrate individual models. Variable importance was assessed using Gradient Boosted Trees, identifying Soil Organic Carbon, Digital Elevation Model, and distance from agricultural lands as the most influential factors. Core fire zones were delineated using True Skill Statistic and binary Foreground Area Density analysis. Fire connectivity was evaluated through pairwise circuit theory, and monthly fire frequency trends were tested with the Mann–Kendall test. Spatial overlap among fire connectivity, wildfire probability, and habitat connectivity was quantified using the Universal Image Quality Index. The ensemble model achieved an AUC = 0.90, outperforming individual models. Most wildfires occurred below 1,000&#xa0;m elevation, with January, October, and November showing significant upward trends. Approximately ~ 14,041 km<sup>2</sup> of suitable habitat falls within fire-prone zones. Thirteen fire cores were identified along the periphery of <i>M. peregrina</i> distribution, with the highest current flow observed in the southern and northern sectors. These findings highlight the urgent need for spatially targeted fire management, such as prioritized firebreak placement and targeted monitoring, to safeguard vulnerable habitats in arid landscapes.</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

Spatial overlap between wildfire probability and Moringa peregrina (Forssk.) habitat suitability: a modeling approach in southern Iran

  • 期刊论文

  • 发表日期：

    2025年11月26日

查看全文
  • P. Karami,
  • H. Piri Sahragard

摘要

Wildfires are recognized as a pervasive global threat that contribute to rangeland degradation and disrupt landscape connectivity for vulnerable species such as M. peregrina in southern Iran. The spatial overlap between wildfire probability, fire connectivity, and habitat suitability and connectivity has received limited attention. In this study, wildfire probability was modeled using MODIS active-fire data and environmental variables, including anthropogenic drivers, seasonal temperature dynamics, and topographic features. A weighted ensemble modeling approach based on Area Under the Curve (AUC) was employed to integrate individual models. Variable importance was assessed using Gradient Boosted Trees, identifying Soil Organic Carbon, Digital Elevation Model, and distance from agricultural lands as the most influential factors. Core fire zones were delineated using True Skill Statistic and binary Foreground Area Density analysis. Fire connectivity was evaluated through pairwise circuit theory, and monthly fire frequency trends were tested with the Mann–Kendall test. Spatial overlap among fire connectivity, wildfire probability, and habitat connectivity was quantified using the Universal Image Quality Index. The ensemble model achieved an AUC = 0.90, outperforming individual models. Most wildfires occurred below 1,000 m elevation, with January, October, and November showing significant upward trends. Approximately ~ 14,041 km2 of suitable habitat falls within fire-prone zones. Thirteen fire cores were identified along the periphery of M. peregrina distribution, with the highest current flow observed in the southern and northern sectors. These findings highlight the urgent need for spatially targeted fire management, such as prioritized firebreak placement and targeted monitoring, to safeguard vulnerable habitats in arid landscapes.