Scholarlink

Background <p>Accurate prediction of fire severity is increasingly important in Mediterranean ecosystems, where climate change and rural land abandonment are intensifying fire regimes. This study assesses, for the first time, the potential of pre-fire Prometheus fuel types (FTs) and fuel biophysical variables to predict fire severity and understand the underlying fuel-related drivers with implications for fuel management and risk mitigation strategies. We evaluated four large wildfires (&gt; 257,000&#xa0;ha) that occurred in Spain during the extreme 2022 fire season. Fire severity was estimated using the differenced Normalized Burn Ratio-Enhanced Vegetation Index (dNBR-EVI) from Sentinel-2 Level-2A optical data. We leveraged previously developed estimates of pre-fire Prometheus FTs at 20-m grid size for each study site, derived from physically-based remote sensing techniques. The inversion of PROSAIL and INFORM radiative transfer models (RTMs) were used to retrieve pre-fire fuel biophysical variables (canopy water content -CWC-, canopy chlorophyll content -CCC-, and fractional vegetation cover -FCOVER-) from Sentinel-2 data in grasslands/shrublands and forests, respectively.</p> Results <p>Random forest (RF) regression models showed that fire severity was most accurately predicted when Prometheus FTs and fuel biophysical variables were fitted together (R<sup>2</sup> = 0.71 and RMSE = 9.18% of the observed dNBR-EVI range), outperforming models using either set of predictors alone. The most important predictor was CWC, followed by Prometheus fuel type, FCOVER and CCC, but fire severity variability was largely driven by predictor interactions. High fire severity was only associated with dense and continuous fuel loads when water amount in live fuels was low, as depicted by the strong interaction between CWC and Prometheus FTs. Conversely, fire severity was constrained under high fuel moisture availability, regardless of fuel arrangement. Finally, in areas with moderate to high FCOVER, reduced CCC was associated with markedly higher fire severity, suggesting that dense and physiologically stressed canopies may be more prone to intense fire damage.</p> Conclusions <p>Our findings support that Prometheus FTs combined with fuel biophysical variables, mainly CWC, may strengthen decision-making frameworks for adaptive fire management under current and future global change scenarios.</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

Fuel biophysical variables retrieved from satellite optical data combined with Prometheus fuel types allow for accurate prediction of potential fire severity

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月16日

查看全文
  • José Manuel Fernández-Guisuraga,
  • Carmen Quintano,
  • Alfonso Fernández-Manso,
  • Leonor Calvo

摘要

Background

Accurate prediction of fire severity is increasingly important in Mediterranean ecosystems, where climate change and rural land abandonment are intensifying fire regimes. This study assesses, for the first time, the potential of pre-fire Prometheus fuel types (FTs) and fuel biophysical variables to predict fire severity and understand the underlying fuel-related drivers with implications for fuel management and risk mitigation strategies. We evaluated four large wildfires (> 257,000 ha) that occurred in Spain during the extreme 2022 fire season. Fire severity was estimated using the differenced Normalized Burn Ratio-Enhanced Vegetation Index (dNBR-EVI) from Sentinel-2 Level-2A optical data. We leveraged previously developed estimates of pre-fire Prometheus FTs at 20-m grid size for each study site, derived from physically-based remote sensing techniques. The inversion of PROSAIL and INFORM radiative transfer models (RTMs) were used to retrieve pre-fire fuel biophysical variables (canopy water content -CWC-, canopy chlorophyll content -CCC-, and fractional vegetation cover -FCOVER-) from Sentinel-2 data in grasslands/shrublands and forests, respectively.

Results

Random forest (RF) regression models showed that fire severity was most accurately predicted when Prometheus FTs and fuel biophysical variables were fitted together (R2 = 0.71 and RMSE = 9.18% of the observed dNBR-EVI range), outperforming models using either set of predictors alone. The most important predictor was CWC, followed by Prometheus fuel type, FCOVER and CCC, but fire severity variability was largely driven by predictor interactions. High fire severity was only associated with dense and continuous fuel loads when water amount in live fuels was low, as depicted by the strong interaction between CWC and Prometheus FTs. Conversely, fire severity was constrained under high fuel moisture availability, regardless of fuel arrangement. Finally, in areas with moderate to high FCOVER, reduced CCC was associated with markedly higher fire severity, suggesting that dense and physiologically stressed canopies may be more prone to intense fire damage.

Conclusions

Our findings support that Prometheus FTs combined with fuel biophysical variables, mainly CWC, may strengthen decision-making frameworks for adaptive fire management under current and future global change scenarios.