Publication: Climate-Driven Variability of Global Burned Area
Authors
Gincheva Norcheva, Andrina
item.page.secondaryauthor
Escuelas::Escuela Internacional de Doctorado
item.page.director
Turco, Marco ; Jerez Rodríguez, Sonia
Publisher
Universidad de Murcia
publication.page.editor
publication.page.department
DOI
item.page.type
info:eu-repo/semantics/doctoralThesis
Description
Abstract
Los incendios forestales son un proceso complejo del sistema terrestre, moldeado por la interacción entre fuentes de ignición, condiciones climáticas y disponibilidad de combustible. Identificar cuánto y dónde la variabilidad climática impulsa los cambios en el área quemada (BA) permite anticipar la evolución de los regímenes de fuego en un clima cambiante y apoyar el desarrollo de estrategias de gestión adaptativa.
Esta tesis doctoral aborda tres retos clave que limitan la comprensión actual de los vínculos entre clima e incendios: (i) la ausencia de registros armonizados y de largo plazo sobre BA; (ii) la limitada comparabilidad entre estimaciones históricas y derivadas de satélite; y (iii) la necesidad de modelos que integren señales climáticas tanto concurrentes como antecedentes en una diversidad de ecorregiones. El estudio se estructura en tres líneas de análisis complementarias.
En primer lugar, se desarrolló el conjunto de datos ONFIRE, integrando registros oficiales mensuales de BA de cinco regiones propensas a incendios (Europa, EE.UU., Canadá, Chile y Australia). Estos datos fueron armonizados en una malla espacial de 1° × 1° y validados frente al producto satelital FireCCI51 y al índice meteorológico de incendios Fire Weather Index (FWI). El conjunto de datos ONFIRE ofrece una cobertura temporal ampliada —de hasta cinco décadas en algunos casos— y respalda estudios a gran escala sobre la dinámica clima–fuego, con datos y código disponibles en acceso abierto y de forma reproducible.
En segundo lugar, se realizó una comparación multifuente de estimaciones de BA en el Mediterráneo occidental. La comparación entre ONFIRE y tres productos satelitales (FireCCI51, MCD64A1 y GFED5) mostró diferencias marcadas en los valores absolutos, especialmente en GFED5, que estimó entre dos y tres veces más BA. No obstante, las fuentes coincidieron en la variabilidad interanual y los patrones espaciales. La concordancia entre ONFIRE y los productos satelitales fue alta durante los periodos anual y de temporada, y las tendencias indicaron un descenso generalizado de la BA en las dos últimas décadas. Estos resultados destacan la utilidad de combinar datos históricos y satelitales para comprender mejor la actividad de incendios en contextos afectados por el cambio climático, y para detectar coincidencias y discrepancias que refuercen una gestión más precisa y fundamentada.
En tercer lugar, se desarrolló un modelo global para cuantificar la influencia de la variabilidad climática sobre la variabilidad interanual de la BA. El modelo combinó predictores concurrentes y antecedentes derivados del FWI y del índice de precipitación-evapotranspiración estandarizado (SPEI), que cuantifica déficits hídricos acumulados, aplicados a escala de ecorregión en todo el mundo. Los resultados muestran que la variabilidad climática explica cerca del 60% de la variación interanual de la BA en el 77% de las áreas combustibles del planeta, duplicando aproximadamente las estimaciones previas de evaluaciones globales. Los resultados destacan la necesidad de considerar la sincronía y acumulación de los efectos climáticos en los análisis clima–fuego, especialmente en enfoques multivariados.
Esta tesis aporta nuevas evidencias sobre los factores que explican la variabilidad global de los incendios y ofrece datos, modelos y código en acceso abierto para impulsar la investigación sobre clima e incendios. Sus resultados son especialmente relevantes ante el creciente riesgo de incendios forestales y pueden apoyar estrategias de gestión y decisión adaptadas a cada territorio.
Además de sus aportaciones científicas, esta tesis se alinea con los Objetivos de Desarrollo Sostenible de las Naciones Unidas —en particular el ODS 13 (Acción por el Clima), el ODS 15 (Vida de Ecosistemas Terrestres) y el ODS 17 (Alianzas para Lograr los Objetivos)— al fomentar la ciencia abierta, el fortalecimiento del monitoreo de incendios y la cooperación internacional frente al aumento del riesgo de incendios forestales.
Wildland fire is a complex Earth system process, shaped by the interaction of ignition sources, weather and climate conditions, and fuel availability. Identifying how much and where climate variability drives changes in burned area (BA) allows anticipating the evolution of fire regimes in a changing climate and guiding adaptive management strategies. Yet, current progress is limited by several structural gaps in available data and modelling approaches. This PhD thesis addresses three interrelated challenges that limit current understanding of climate–fire links: (i) the absence of harmonised long-term BA records; (ii) the limited comparability between historical and satellite-derived estimates; and (iii) the need for models that integrate concurrent and antecedent climate signals across diverse ecoregion contexts. To address these challenges, this work employed three complementary lines of analysis. First, the ONFIRE dataset was developed, integrating official monthly BA records from five fire-prone regions (Europe, USA, Canada, Chile, and Australia). These data were harmonised to a 1° × 1° spatial grid and validated against the FireCCI51 satellite product and the Fire Weather Index (FWI). ONFIRE offers extended temporal coverage—up to five decades in some cases—and supports large-scale studies on fire–climate dynamics with reproducible and openly available data and code. Second, BA estimates were compared across multiple sources in the Western Mediterranean. Using the ONFIRE dataset in combination with three satellite products (FireCCI51, MCD64A1, and GFED5), the analysis revealed major differences in absolute values, especially for GFED5, which estimated two to three times more BA than the other datasets. Nonetheless, all sources showed broadly consistent interannual variability and similar spatial patterns. Correlation analyses confirmed strong agreement between ONFIRE and all three satellite products during the annual and fire seasons. Trend analyses revealed general declines in BA over the past two decades across most sources. These findings highlight the value of integrating historical and satellite-based data to obtain a more comprehensive picture of fire activity in a region increasingly shaped by climate change and anthropogenic transformation. Cross-comparisons between sources improve our ability to detect areas of convergence and uncertainty, supporting more targeted and evidence-based fire management. Third, a global model was developed to quantify the influence of climate variability on interannual BA variability. The model combined concurrent and antecedent predictors derived from FWI and the Standardised Precipitation–Evapotranspiration Index (SPEI), applied across ecoregions worldwide. Results show that climate variability explains nearly 60% of the interannual variation in BA across 77% of the world’s burnable area—a performance that approximately doubles previous estimates from global assessments. The strong contribution of multivariate configurations highlights the importance of considering the timing and accumulation of climatic effects when examining climate–fire links. Overall, this thesis provides new insights into the global drivers of fire variability and contributes open-access datasets, code, and models to advance fire–climate research. These results are especially relevant in the context of rising fire risk and can contribute to the development of management and decision-making strategies. Beyond the scientific findings, this work also contributes to the United Nations Sustainable Development Goals, particularly SDG 13 (Climate Action), SDG 15 (Life on Land), and SDG 17 (Partnerships for the Goals), by promoting open science, robust fire monitoring, and international cooperation in the face of escalating fire risks.
Wildland fire is a complex Earth system process, shaped by the interaction of ignition sources, weather and climate conditions, and fuel availability. Identifying how much and where climate variability drives changes in burned area (BA) allows anticipating the evolution of fire regimes in a changing climate and guiding adaptive management strategies. Yet, current progress is limited by several structural gaps in available data and modelling approaches. This PhD thesis addresses three interrelated challenges that limit current understanding of climate–fire links: (i) the absence of harmonised long-term BA records; (ii) the limited comparability between historical and satellite-derived estimates; and (iii) the need for models that integrate concurrent and antecedent climate signals across diverse ecoregion contexts. To address these challenges, this work employed three complementary lines of analysis. First, the ONFIRE dataset was developed, integrating official monthly BA records from five fire-prone regions (Europe, USA, Canada, Chile, and Australia). These data were harmonised to a 1° × 1° spatial grid and validated against the FireCCI51 satellite product and the Fire Weather Index (FWI). ONFIRE offers extended temporal coverage—up to five decades in some cases—and supports large-scale studies on fire–climate dynamics with reproducible and openly available data and code. Second, BA estimates were compared across multiple sources in the Western Mediterranean. Using the ONFIRE dataset in combination with three satellite products (FireCCI51, MCD64A1, and GFED5), the analysis revealed major differences in absolute values, especially for GFED5, which estimated two to three times more BA than the other datasets. Nonetheless, all sources showed broadly consistent interannual variability and similar spatial patterns. Correlation analyses confirmed strong agreement between ONFIRE and all three satellite products during the annual and fire seasons. Trend analyses revealed general declines in BA over the past two decades across most sources. These findings highlight the value of integrating historical and satellite-based data to obtain a more comprehensive picture of fire activity in a region increasingly shaped by climate change and anthropogenic transformation. Cross-comparisons between sources improve our ability to detect areas of convergence and uncertainty, supporting more targeted and evidence-based fire management. Third, a global model was developed to quantify the influence of climate variability on interannual BA variability. The model combined concurrent and antecedent predictors derived from FWI and the Standardised Precipitation–Evapotranspiration Index (SPEI), applied across ecoregions worldwide. Results show that climate variability explains nearly 60% of the interannual variation in BA across 77% of the world’s burnable area—a performance that approximately doubles previous estimates from global assessments. The strong contribution of multivariate configurations highlights the importance of considering the timing and accumulation of climatic effects when examining climate–fire links. Overall, this thesis provides new insights into the global drivers of fire variability and contributes open-access datasets, code, and models to advance fire–climate research. These results are especially relevant in the context of rising fire risk and can contribute to the development of management and decision-making strategies. Beyond the scientific findings, this work also contributes to the United Nations Sustainable Development Goals, particularly SDG 13 (Climate Action), SDG 15 (Life on Land), and SDG 17 (Partnerships for the Goals), by promoting open science, robust fire monitoring, and international cooperation in the face of escalating fire risks.
publication.page.subject
Citation
item.page.embargo
18-sep-2026
Collections
Ir a Estadísticas
Este ítem está sujeto a una licencia Creative Commons. http://creativecommons.org/licenses/by-nc-nd/4.0/