DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación
Artículo científico (JournalArticle)
Modelling the influence of meteorological drivers on spectral fire severity: a case study in NW Spain
16 pages, 9 figures, 4 tables, Data will be made available on request., Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.agrformet.2026.111405., Fire severity is a key indicator of post-fire ecological impacts. While previous studies have shown that fire-weather conditions can strongly influence fire severity, the direct influence of meteorological variables on severity has not yet been sufficiently explored. This study aims to quantify the influence of meteorological variables on vegetation and soil fire severity measured through spectral burn indices. First, we evaluated the correspondence between field-measured vegetation and soil fire severity (VBS, SBS) and several spectral indices: differenced Normalized Burn Ratio (dNBR), the relativized dNBR (RdNBR), the dNBR enhanced with the Enhanced Vegetation (dNBR-EVI) and the Blue Normalized Difference Vegetation (BNDVI). BNDVI showed the strongest and most consistent agreement with field severity for both VBS and SBS across shrublands and forested stands, while dNBR performed best only for VBS in forested areas. Then, to quantify the influence of meteorological variables on severity, we fitted Generalized Additive Models (GAM) and Random Forests (RF) based on meteorological predictors filtered through correlation and Variance Inflation Factor (VIF) analysis. The results showed that RF models generally provided comparatively lower prediction errors across fire severity descriptors. The strongest RF model predicted performance was obtained for dNBR standard deviation (R² = 0.67), while relatively high predictive performance was also observed for the 75th percentile of BNDVI (R² = 0.62). Variable importance and partial dependence analyses identified atmospheric moisture conditions (d2m), Initial Spread Index (ISI), vapor pressure deficit (VPD), surface soil moisture (Swvl1), and incoming solar radiation (ssrd) as the most influential meteorological predictors. These results advance the integration of atmospheric variables and fuel dryness into severity prediction and demonstrate the potential of spectral indices to improve assessment of post-fire impacts. However, this study focused exclusively on meteorological drivers of fire severity, the findings should be interpreted as evidence of the relevance of meteorological controls rather than as a comprehensive assessment of all drivers of fire severity., This research was funded by the Spanish Research Agency of the Spanish Ministry of Science and Innovation through project ENFIRES-NW, PID2020-116494RR-C42 and through project RESFIRE, PID2023-152690OA-C22, and by the Galician Innovation Agency-funded GAIN Tipo C Project MULCHAR (IN607D 2025/08).
AN would like to thank to the Spanish Ministry of Science and Innovation through the grant JDC2022-049758-I. AR was funded by the ´Ramón y Cajal´ program of the Spanish Ministry of Science and Innovation (RYC2022-036822-I). SP was funded by the ´Ramón y Cajal´ program of the Spanish Ministry of Science and Innovation (RYC2022-035489-I)., Peer reviewed
Proyecto:
AEI / Plan Estatal de Investigación Científica y Técnica y de Innovación 2017-2020 PID2020-116494RR-C42
,
AEI / Plan Estatal de Investigación Científica y Técnica y de Innovación 2021-2023 PID2023-152690OA-C22
,
GAIN IN607D 2025
DOI: http://hdl.handle.net/10261/442706, https://api.elsevier.com/content/abstract/scopus_id/105046658497
HANDLE: http://hdl.handle.net/10261/442706, https://api.elsevier.com/content/abstract/scopus_id/105046658497