Joaquín Bedia
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Research Interests
Joaquín Bedia has a total of 202 co-authors. Here are the top co-authors by number of shared publications:
- Jorge Baño-Medina
Center for Western Weather and Water Extremes Scripps Institution of Oceanography University of California San Diego La Jolla CA USA
20 shared publications - Ana Casanueva
Dept. Matemática Aplicada y Ciencias de la Computación (MACC), Universidad de Cantabria, Santander, Spain
16 shared publications Sixto Herrera
16 shared publicationsMaialen Iturbide
15 shared publications- Oscar Mirones
Instituto de Física de Cantabria (IFCA), CSIC-Universidad de Cantabria, Santander, Spain
13 shared publications - Sixto Herrera
Universidad de Cantabria
12 shared publications
Latest Publications
- Open Access
- 12 August 2026
- Published by Elsevier in SSRN Electronic Journal
Fire-Weather Extremes over the Iberian Peninsula: Intensification and Seasonal Reorganization in Deep-Learning Downscaled Projections
Fire weather is projected to intensify across the Mediterranean, yet changes in the duration and timing of hazardous conditions remain poorly quantified, hampered by two limitations of general circulation models (GCMs): their coarse spatial resolution, which blurs regional fire-weather patterns, and their lack of the instantaneous noon variables required by the standard Fire Weather Index (FWI). Here we address both with a deep-learning (DL) downscaling framework: a U-Net maps coarse (1.5$^\circ$) daily mean temperature, relative humidity, and wind speed to high-resolution (0.1$^\circ$) 12 UTC FWI, while retaining the large-scale climate-change signal. Over the historical period (1979--2014), the DL downscaling improves the spatial structure, distribution, and extremes of FWI relative to GCM-based estimates, with the clearest added value for extreme fire-weather conditions. Applied to nine CMIP6 models across multiple SSPs and global warming levels (GWLs), the projections reveal robust, spatially coherent intensification of fire weather, strongest over Mediterranean and continental regions. Beyond this intensification, the results expose a systematic broadening of the fire-weather season: conditions historically confined to June--September increasingly emerge in late spring and persist into early autumn, shifting the ensemble-mean season to May--September at +3$^\circ$C GWL and, in some regions, extending toward May--October under stronger warming. This seasonal expansion shows substantial multi-model agreement and occurs even where absolute fire danger remains moderate, underscoring a change in the timing and duration of fire-weather extremes that GCM-resolution projections alone would miss.
- 12 August 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessFire-Weather Extremes over the Iberian Peninsula: Intensification and Seasonal Reorganization in Deep-Learning Downscaled Projections
Fire weather is projected to intensify across the Mediterranean, yet changes in the duration and timing of hazardous conditions remain poorly quantified, hampered by two limitations of general circulation models (GCMs): their coarse spatial resolution, which blurs regional fire-weather patterns, and their lack of the instantaneous noon variables required by the standard Fire Weather Index (FWI). Here we address both with a deep-learning (DL) downscaling framework: a U-Net maps coarse (1.5$^\circ$) daily mean temperature, relative humidity, and wind speed to high-resolution (0.1$^\circ$) 12 UTC FWI, while retaining the large-scale climate-change signal. Over the historical period (1979--2014), the DL downscaling improves the spatial structure, distribution, and extremes of FWI relative to GCM-based estimates, with the clearest added value for extreme fire-weather conditions. Applied to nine CMIP6 models across multiple SSPs and global warming levels (GWLs), the projections reveal robust, spatially coherent intensification of fire weather, strongest over Mediterranean and continental regions. Beyond this intensification, the results expose a systematic broadening of the fire-weather season: conditions historically confined to June--September increasingly emerge in late spring and persist into early autumn, shifting the ensemble-mean season to May--September at +3$^\circ$C GWL and, in some regions, extending toward May--October under stronger warming. This seasonal expansion shows substantial multi-model agreement and occurs even where absolute fire danger remains moderate, underscoring a change in the timing and duration of fire-weather extremes that GCM-resolution projections alone would miss.
- Open Access
- 22 July 2026
- Published by Copernicus GmbH in Natural Hazards and Earth System Sciences
- Vol. 26 (7) , 3417-3442
Deep Learning Emulation of Multivariate Climate Indices: A Case Study of the Fire Weather Index in the Iberian Peninsula
The Fire Weather Index (FWI) is an essential multivariate climate index for assessing wildfire risk and the associated impacts of climate change, as it provides a quantitative measure of wildfire danger by integrating different critical near-surface fire-weather variables, namely air temperature, relative humidity, wind speed, and precipitation. FWI calculation depends on instantaneous data representing noon local standard times, which are often unavailable in many climate data repositories – particularly in climate projections. In these instances, a “Proxy” of actual FWI is often used, applying the same FWI formulation to daily aggregated values (mean, max, or min), despite known limitations in capturing extremes and temporal dynamics. This study investigates the use of deep learning (DL) models to emulate the reference FWI over the Iberian Peninsula – a predominantly Mediterranean and fire-prone region – using only daily inputs. The emulators are trained and evaluated using ERA5-Land data, which, while not observational ground truth, provides a consistent and high-resolution dataset suitable for controlled inter-comparison. The focus is not on validating FWI against observations, but on assessing the ability of DL models to reproduce the reference FWI more accurately than traditional proxy approaches, using the same input data source. Our results show substantial improvements in spatial accuracy, preservation of temporal sequences, and detection of extreme fire danger events when compared with the corresponding Proxy version. Furthermore, after evaluating different combinations of input variables for DL model training, we find that precipitation can be excluded without substantially affecting accuracy – especially for the upper part of the distribution – an important insight given the challenges climate models face in representing precipitation. These findings highlight the potential of deep learning tools to enhance the usability of FWI in contexts where sub-daily data are unavailable, and set the stage for the emulation of other multivariate climate indices, which are vital for climate impact studies, spatial planning and management, and adaptation decision-making.
- 22 July 2026
- Published by Copernicus GmbH in Natural Hazards and Earth System Sciences
- Vol. 26 (7) , 3417-3442
Open AccessDeep Learning Emulation of Multivariate Climate Indices: A Case Study of the Fire Weather Index in the Iberian Peninsula
The Fire Weather Index (FWI) is an essential multivariate climate index for assessing wildfire risk and the associated impacts of climate change, as it provides a quantitative measure of wildfire danger by integrating different critical near-surface fire-weather variables, namely air temperature, relative humidity, wind speed, and precipitation. FWI calculation depends on instantaneous data representing noon local standard times, which are often unavailable in many climate data repositories – particularly in climate projections. In these instances, a “Proxy” of actual FWI is often used, applying the same FWI formulation to daily aggregated values (mean, max, or min), despite known limitations in capturing extremes and temporal dynamics. This study investigates the use of deep learning (DL) models to emulate the reference FWI over the Iberian Peninsula – a predominantly Mediterranean and fire-prone region – using only daily inputs. The emulators are trained and evaluated using ERA5-Land data, which, while not observational ground truth, provides a consistent and high-resolution dataset suitable for controlled inter-comparison. The focus is not on validating FWI against observations, but on assessing the ability of DL models to reproduce the reference FWI more accurately than traditional proxy approaches, using the same input data source. Our results show substantial improvements in spatial accuracy, preservation of temporal sequences, and detection of extreme fire danger events when compared with the corresponding Proxy version. Furthermore, after evaluating different combinations of input variables for DL model training, we find that precipitation can be excluded without substantially affecting accuracy – especially for the upper part of the distribution – an important insight given the challenges climate models face in representing precipitation. These findings highlight the potential of deep learning tools to enhance the usability of FWI in contexts where sub-daily data are unavailable, and set the stage for the emulation of other multivariate climate indices, which are vital for climate impact studies, spatial planning and management, and adaptation decision-making.
- Top Cited PapersOpen Access
- 13 March 2026
- Published by American Association for the Advancement of Science (AAAS) in Science Advances
- Vol. 12 (11) , eadx9845
The emerging human fingerprint on global extreme fire weather
Extreme fire weather (hot, dry, and windy conditions) has intensified globally, yet formally attributing this trend to anthropogenic climate change remains challenging. Here, we analyze global trends in extreme fire weather days (FWI95d, annual count of days with Fire Weather Index above the 95th percentile) over 1980–2023, using climate model ensembles, observational data, and fingerprint detection techniques. We find that the observed increase in extreme fire weather bears a clear externally forced signal, detectable at 99% confidence above natural variability and attributable to human-induced climate change. This emerging human-induced fingerprint on extreme fire weather highlights a benchmark for climate science and underscores the urgency of integrating these insights into wildfire risk management and adaptation strategies.
- 13 March 2026
- Published by American Association for the Advancement of Science (AAAS) in Science Advances
- Vol. 12 (11) , eadx9845
Top Cited PapersOpen AccessThe emerging human fingerprint on global extreme fire weather
Extreme fire weather (hot, dry, and windy conditions) has intensified globally, yet formally attributing this trend to anthropogenic climate change remains challenging. Here, we analyze global trends in extreme fire weather days (FWI95d, annual count of days with Fire Weather Index above the 95th percentile) over 1980–2023, using climate model ensembles, observational data, and fingerprint detection techniques. We find that the observed increase in extreme fire weather bears a clear externally forced signal, detectable at 99% confidence above natural variability and attributable to human-induced climate change. This emerging human-induced fingerprint on extreme fire weather highlights a benchmark for climate science and underscores the urgency of integrating these insights into wildfire risk management and adaptation strategies.
- Open Access
- 11 February 2026
- Published by Springer Nature in npj Climate and Atmospheric Science
- Vol. 9 (1)
Emerging near-surface extratropical circulation changes due to climate change: a weather typing based global analysis
This study examines how large-scale near-surface circulation is projected to evolve under anthropogenic forcing throughout the 21st century. Using a multi-model ensemble of 41 state-of-the-art Global Climate Models high emissions scenarios, we analyze changes in Weather Type (WT) frequencies, based on Sea-Level Pressure (SLP), as a function of global warming and estimate their Time of Emergence (ToE) from historical variability. Significant and robust trends are identified across key climatic regions, including shifts in anticyclonic types in subtropical high-pressure belts and the Mediterranean, westerly and cyclonic types in the Subantarctic, and unclassified types in the Mediterranean. While most robust signals emerge by the late 21st century, some changes in the Mediterranean are already emerging in the present decade. Overall, results reveal significant circulation changes in the coming decades, affecting dominant near-surface circulation modes globally. A comprehensive dataset of 6-hourly WT projections (2006/15-2100) over mid-latitudes is publicly available as part of this study.
- 11 February 2026
- Published by Springer Nature in npj Climate and Atmospheric Science
- Vol. 9 (1)
Open AccessEmerging near-surface extratropical circulation changes due to climate change: a weather typing based global analysis
This study examines how large-scale near-surface circulation is projected to evolve under anthropogenic forcing throughout the 21st century. Using a multi-model ensemble of 41 state-of-the-art Global Climate Models high emissions scenarios, we analyze changes in Weather Type (WT) frequencies, based on Sea-Level Pressure (SLP), as a function of global warming and estimate their Time of Emergence (ToE) from historical variability. Significant and robust trends are identified across key climatic regions, including shifts in anticyclonic types in subtropical high-pressure belts and the Mediterranean, westerly and cyclonic types in the Subantarctic, and unclassified types in the Mediterranean. While most robust signals emerge by the late 21st century, some changes in the Mediterranean are already emerging in the present decade. Overall, results reveal significant circulation changes in the coming decades, affecting dominant near-surface circulation modes globally. A comprehensive dataset of 6-hourly WT projections (2006/15-2100) over mid-latitudes is publicly available as part of this study.
- Open Access
- 3 October 2025
- Published by Copernicus GmbH
- Vol. 2025, 1-27
Reply on RC1
The Fire Weather Index (FWI) is an essential multivariate climate index for assessing wildfire risk and the associated impacts of climate change, as it provides a quantitative measure of wildfire danger by integrating different critical near-surface fire-weather variables, namely air temperature, relative humidity, wind speed, and precipitation. FWI calculation depends on instantaneous data representing noon local standard times, which are often unavailable in many climate data repositories – particularly in climate projections. In these instances, a "proxy" of actual FWI is often used, applying the same FWI formulation to daily aggregated values (mean, max, or min), despite known limitations in capturing extremes and temporal dynamics. This study investigates the use of deep learning (DL) models to emulate the reference FWI over the Iberian Peninsula – a predominantly Mediterranean and fire-prone region – using only daily inputs. The emulators are trained and evaluated using ERA5-Land data, which, while not observational ground truth, provides a consistent and high-resolution dataset suitable for controlled inter-comparison. The focus is not on validating FWI against observations, but on assessing the ability of DL models to reproduce the reference FWI more accurately than traditional proxy approaches, using the same input data source. Our results show substantial improvements in spatial accuracy, preservation of temporal sequences, and detection of extreme fire danger events when compared with the corresponding proxy version. Furthermore, after evaluating different combinations of input variables for DL model training, we find that precipitation can be excluded without substantially affecting accuracy – especially at the upper end – an important insight given the challenges climate models face in representing precipitation. These findings highlight the potential of deep learning tools to enhance the usability of FWI in contexts where sub-daily data are unavailable, and set the stage for the emulation of other multivariate climate indices, which are vital for climate impact studies, spatial planning and management, and adaptation decision-making.
- 3 October 2025
- Published by Copernicus GmbH
- Vol. 2025, 1-27
Open AccessReply on RC1
The Fire Weather Index (FWI) is an essential multivariate climate index for assessing wildfire risk and the associated impacts of climate change, as it provides a quantitative measure of wildfire danger by integrating different critical near-surface fire-weather variables, namely air temperature, relative humidity, wind speed, and precipitation. FWI calculation depends on instantaneous data representing noon local standard times, which are often unavailable in many climate data repositories – particularly in climate projections. In these instances, a "proxy" of actual FWI is often used, applying the same FWI formulation to daily aggregated values (mean, max, or min), despite known limitations in capturing extremes and temporal dynamics. This study investigates the use of deep learning (DL) models to emulate the reference FWI over the Iberian Peninsula – a predominantly Mediterranean and fire-prone region – using only daily inputs. The emulators are trained and evaluated using ERA5-Land data, which, while not observational ground truth, provides a consistent and high-resolution dataset suitable for controlled inter-comparison. The focus is not on validating FWI against observations, but on assessing the ability of DL models to reproduce the reference FWI more accurately than traditional proxy approaches, using the same input data source. Our results show substantial improvements in spatial accuracy, preservation of temporal sequences, and detection of extreme fire danger events when compared with the corresponding proxy version. Furthermore, after evaluating different combinations of input variables for DL model training, we find that precipitation can be excluded without substantially affecting accuracy – especially at the upper end – an important insight given the challenges climate models face in representing precipitation. These findings highlight the potential of deep learning tools to enhance the usability of FWI in contexts where sub-daily data are unavailable, and set the stage for the emulation of other multivariate climate indices, which are vital for climate impact studies, spatial planning and management, and adaptation decision-making.
- Open Access
- 3 October 2025
- Published by Copernicus GmbH
- Vol. 2025, 1-27
Reply on RC2
The Fire Weather Index (FWI) is an essential multivariate climate index for assessing wildfire risk and the associated impacts of climate change, as it provides a quantitative measure of wildfire danger by integrating different critical near-surface fire-weather variables, namely air temperature, relative humidity, wind speed, and precipitation. FWI calculation depends on instantaneous data representing noon local standard times, which are often unavailable in many climate data repositories – particularly in climate projections. In these instances, a "proxy" of actual FWI is often used, applying the same FWI formulation to daily aggregated values (mean, max, or min), despite known limitations in capturing extremes and temporal dynamics. This study investigates the use of deep learning (DL) models to emulate the reference FWI over the Iberian Peninsula – a predominantly Mediterranean and fire-prone region – using only daily inputs. The emulators are trained and evaluated using ERA5-Land data, which, while not observational ground truth, provides a consistent and high-resolution dataset suitable for controlled inter-comparison. The focus is not on validating FWI against observations, but on assessing the ability of DL models to reproduce the reference FWI more accurately than traditional proxy approaches, using the same input data source. Our results show substantial improvements in spatial accuracy, preservation of temporal sequences, and detection of extreme fire danger events when compared with the corresponding proxy version. Furthermore, after evaluating different combinations of input variables for DL model training, we find that precipitation can be excluded without substantially affecting accuracy – especially at the upper end – an important insight given the challenges climate models face in representing precipitation. These findings highlight the potential of deep learning tools to enhance the usability of FWI in contexts where sub-daily data are unavailable, and set the stage for the emulation of other multivariate climate indices, which are vital for climate impact studies, spatial planning and management, and adaptation decision-making.
- 3 October 2025
- Published by Copernicus GmbH
- Vol. 2025, 1-27
Open AccessReply on RC2
The Fire Weather Index (FWI) is an essential multivariate climate index for assessing wildfire risk and the associated impacts of climate change, as it provides a quantitative measure of wildfire danger by integrating different critical near-surface fire-weather variables, namely air temperature, relative humidity, wind speed, and precipitation. FWI calculation depends on instantaneous data representing noon local standard times, which are often unavailable in many climate data repositories – particularly in climate projections. In these instances, a "proxy" of actual FWI is often used, applying the same FWI formulation to daily aggregated values (mean, max, or min), despite known limitations in capturing extremes and temporal dynamics. This study investigates the use of deep learning (DL) models to emulate the reference FWI over the Iberian Peninsula – a predominantly Mediterranean and fire-prone region – using only daily inputs. The emulators are trained and evaluated using ERA5-Land data, which, while not observational ground truth, provides a consistent and high-resolution dataset suitable for controlled inter-comparison. The focus is not on validating FWI against observations, but on assessing the ability of DL models to reproduce the reference FWI more accurately than traditional proxy approaches, using the same input data source. Our results show substantial improvements in spatial accuracy, preservation of temporal sequences, and detection of extreme fire danger events when compared with the corresponding proxy version. Furthermore, after evaluating different combinations of input variables for DL model training, we find that precipitation can be excluded without substantially affecting accuracy – especially at the upper end – an important insight given the challenges climate models face in representing precipitation. These findings highlight the potential of deep learning tools to enhance the usability of FWI in contexts where sub-daily data are unavailable, and set the stage for the emulation of other multivariate climate indices, which are vital for climate impact studies, spatial planning and management, and adaptation decision-making.
- Open Access
- 28 July 2025
- Published by Copernicus GmbH
- Vol. 2025, 1-27
Comment on egusphere-2025-2349
The Fire Weather Index (FWI) is an essential multivariate climate index for assessing wildfire risk and the associated impacts of climate change, as it provides a quantitative measure of wildfire danger by integrating different critical near-surface fire-weather variables, namely air temperature, relative humidity, wind speed, and precipitation. FWI calculation depends on instantaneous data representing noon local standard times, which are often unavailable in many climate data repositories – particularly in climate projections. In these instances, a "proxy" of actual FWI is often used, applying the same FWI formulation to daily aggregated values (mean, max, or min), despite known limitations in capturing extremes and temporal dynamics. This study investigates the use of deep learning (DL) models to emulate the reference FWI over the Iberian Peninsula – a predominantly Mediterranean and fire-prone region – using only daily inputs. The emulators are trained and evaluated using ERA5-Land data, which, while not observational ground truth, provides a consistent and high-resolution dataset suitable for controlled inter-comparison. The focus is not on validating FWI against observations, but on assessing the ability of DL models to reproduce the reference FWI more accurately than traditional proxy approaches, using the same input data source. Our results show substantial improvements in spatial accuracy, preservation of temporal sequences, and detection of extreme fire danger events when compared with the corresponding proxy version. Furthermore, after evaluating different combinations of input variables for DL model training, we find that precipitation can be excluded without substantially affecting accuracy – especially at the upper end – an important insight given the challenges climate models face in representing precipitation. These findings highlight the potential of deep learning tools to enhance the usability of FWI in contexts where sub-daily data are unavailable, and set the stage for the emulation of other multivariate climate indices, which are vital for climate impact studies, spatial planning and management, and adaptation decision-making.
- 28 July 2025
- Published by Copernicus GmbH
- Vol. 2025, 1-27
Open AccessComment on egusphere-2025-2349
The Fire Weather Index (FWI) is an essential multivariate climate index for assessing wildfire risk and the associated impacts of climate change, as it provides a quantitative measure of wildfire danger by integrating different critical near-surface fire-weather variables, namely air temperature, relative humidity, wind speed, and precipitation. FWI calculation depends on instantaneous data representing noon local standard times, which are often unavailable in many climate data repositories – particularly in climate projections. In these instances, a "proxy" of actual FWI is often used, applying the same FWI formulation to daily aggregated values (mean, max, or min), despite known limitations in capturing extremes and temporal dynamics. This study investigates the use of deep learning (DL) models to emulate the reference FWI over the Iberian Peninsula – a predominantly Mediterranean and fire-prone region – using only daily inputs. The emulators are trained and evaluated using ERA5-Land data, which, while not observational ground truth, provides a consistent and high-resolution dataset suitable for controlled inter-comparison. The focus is not on validating FWI against observations, but on assessing the ability of DL models to reproduce the reference FWI more accurately than traditional proxy approaches, using the same input data source. Our results show substantial improvements in spatial accuracy, preservation of temporal sequences, and detection of extreme fire danger events when compared with the corresponding proxy version. Furthermore, after evaluating different combinations of input variables for DL model training, we find that precipitation can be excluded without substantially affecting accuracy – especially at the upper end – an important insight given the challenges climate models face in representing precipitation. These findings highlight the potential of deep learning tools to enhance the usability of FWI in contexts where sub-daily data are unavailable, and set the stage for the emulation of other multivariate climate indices, which are vital for climate impact studies, spatial planning and management, and adaptation decision-making.
- Open Access
- 21 July 2025
- Published by Copernicus GmbH
- Vol. 2025, 1-27
Comment on egusphere-2025-2349
The Fire Weather Index (FWI) is an essential multivariate climate index for assessing wildfire risk and the associated impacts of climate change, as it provides a quantitative measure of wildfire danger by integrating different critical near-surface fire-weather variables, namely air temperature, relative humidity, wind speed, and precipitation. FWI calculation depends on instantaneous data representing noon local standard times, which are often unavailable in many climate data repositories – particularly in climate projections. In these instances, a "proxy" of actual FWI is often used, applying the same FWI formulation to daily aggregated values (mean, max, or min), despite known limitations in capturing extremes and temporal dynamics. This study investigates the use of deep learning (DL) models to emulate the reference FWI over the Iberian Peninsula – a predominantly Mediterranean and fire-prone region – using only daily inputs. The emulators are trained and evaluated using ERA5-Land data, which, while not observational ground truth, provides a consistent and high-resolution dataset suitable for controlled inter-comparison. The focus is not on validating FWI against observations, but on assessing the ability of DL models to reproduce the reference FWI more accurately than traditional proxy approaches, using the same input data source. Our results show substantial improvements in spatial accuracy, preservation of temporal sequences, and detection of extreme fire danger events when compared with the corresponding proxy version. Furthermore, after evaluating different combinations of input variables for DL model training, we find that precipitation can be excluded without substantially affecting accuracy – especially at the upper end – an important insight given the challenges climate models face in representing precipitation. These findings highlight the potential of deep learning tools to enhance the usability of FWI in contexts where sub-daily data are unavailable, and set the stage for the emulation of other multivariate climate indices, which are vital for climate impact studies, spatial planning and management, and adaptation decision-making.
- 21 July 2025
- Published by Copernicus GmbH
- Vol. 2025, 1-27
Open AccessComment on egusphere-2025-2349
The Fire Weather Index (FWI) is an essential multivariate climate index for assessing wildfire risk and the associated impacts of climate change, as it provides a quantitative measure of wildfire danger by integrating different critical near-surface fire-weather variables, namely air temperature, relative humidity, wind speed, and precipitation. FWI calculation depends on instantaneous data representing noon local standard times, which are often unavailable in many climate data repositories – particularly in climate projections. In these instances, a "proxy" of actual FWI is often used, applying the same FWI formulation to daily aggregated values (mean, max, or min), despite known limitations in capturing extremes and temporal dynamics. This study investigates the use of deep learning (DL) models to emulate the reference FWI over the Iberian Peninsula – a predominantly Mediterranean and fire-prone region – using only daily inputs. The emulators are trained and evaluated using ERA5-Land data, which, while not observational ground truth, provides a consistent and high-resolution dataset suitable for controlled inter-comparison. The focus is not on validating FWI against observations, but on assessing the ability of DL models to reproduce the reference FWI more accurately than traditional proxy approaches, using the same input data source. Our results show substantial improvements in spatial accuracy, preservation of temporal sequences, and detection of extreme fire danger events when compared with the corresponding proxy version. Furthermore, after evaluating different combinations of input variables for DL model training, we find that precipitation can be excluded without substantially affecting accuracy – especially at the upper end – an important insight given the challenges climate models face in representing precipitation. These findings highlight the potential of deep learning tools to enhance the usability of FWI in contexts where sub-daily data are unavailable, and set the stage for the emulation of other multivariate climate indices, which are vital for climate impact studies, spatial planning and management, and adaptation decision-making.
- Open Access
- 10 July 2025
- Published by Copernicus GmbH in EGUsphere
- Vol. 2025, 1-27
Deep Learning Emulation of Multivariate Climate Indices: A Case Study of the Fire Weather Index in the Iberian Peninsula
The Fire Weather Index (FWI) is an essential multivariate climate index for assessing wildfire risk and the associated impacts of climate change, as it provides a quantitative measure of wildfire danger by integrating different critical near-surface fire-weather variables, namely air temperature, relative humidity, wind speed, and precipitation. FWI calculation depends on instantaneous data representing noon local standard times, which are often unavailable in many climate data repositories – particularly in climate projections. In these instances, a "proxy" of actual FWI is often used, applying the same FWI formulation to daily aggregated values (mean, max, or min), despite known limitations in capturing extremes and temporal dynamics. This study investigates the use of deep learning (DL) models to emulate the reference FWI over the Iberian Peninsula – a predominantly Mediterranean and fire-prone region – using only daily inputs. The emulators are trained and evaluated using ERA5-Land data, which, while not observational ground truth, provides a consistent and high-resolution dataset suitable for controlled inter-comparison. The focus is not on validating FWI against observations, but on assessing the ability of DL models to reproduce the reference FWI more accurately than traditional proxy approaches, using the same input data source. Our results show substantial improvements in spatial accuracy, preservation of temporal sequences, and detection of extreme fire danger events when compared with the corresponding proxy version. Furthermore, after evaluating different combinations of input variables for DL model training, we find that precipitation can be excluded without substantially affecting accuracy – especially at the upper end – an important insight given the challenges climate models face in representing precipitation. These findings highlight the potential of deep learning tools to enhance the usability of FWI in contexts where sub-daily data are unavailable, and set the stage for the emulation of other multivariate climate indices, which are vital for climate impact studies, spatial planning and management, and adaptation decision-making.
- 10 July 2025
- Published by Copernicus GmbH in EGUsphere
- Vol. 2025, 1-27
Open AccessDeep Learning Emulation of Multivariate Climate Indices: A Case Study of the Fire Weather Index in the Iberian Peninsula
The Fire Weather Index (FWI) is an essential multivariate climate index for assessing wildfire risk and the associated impacts of climate change, as it provides a quantitative measure of wildfire danger by integrating different critical near-surface fire-weather variables, namely air temperature, relative humidity, wind speed, and precipitation. FWI calculation depends on instantaneous data representing noon local standard times, which are often unavailable in many climate data repositories – particularly in climate projections. In these instances, a "proxy" of actual FWI is often used, applying the same FWI formulation to daily aggregated values (mean, max, or min), despite known limitations in capturing extremes and temporal dynamics. This study investigates the use of deep learning (DL) models to emulate the reference FWI over the Iberian Peninsula – a predominantly Mediterranean and fire-prone region – using only daily inputs. The emulators are trained and evaluated using ERA5-Land data, which, while not observational ground truth, provides a consistent and high-resolution dataset suitable for controlled inter-comparison. The focus is not on validating FWI against observations, but on assessing the ability of DL models to reproduce the reference FWI more accurately than traditional proxy approaches, using the same input data source. Our results show substantial improvements in spatial accuracy, preservation of temporal sequences, and detection of extreme fire danger events when compared with the corresponding proxy version. Furthermore, after evaluating different combinations of input variables for DL model training, we find that precipitation can be excluded without substantially affecting accuracy – especially at the upper end – an important insight given the challenges climate models face in representing precipitation. These findings highlight the potential of deep learning tools to enhance the usability of FWI in contexts where sub-daily data are unavailable, and set the stage for the emulation of other multivariate climate indices, which are vital for climate impact studies, spatial planning and management, and adaptation decision-making.
- Open Access
- 20 May 2025
- Published by Forum XXI in European Public & Social Innovation Review
- Vol. 10, 1-22
Apoyo audiovisual en un modelo de aula invertida en las prácticas de Matemáticas en Grados de Ingeniería
Introducción: El estudio se enfoca en la incorporación de recursos audiovisuales y la estrategia del aula invertida en prácticas de ordenador. Estas técnicas son efectivas para enriquecer el proceso de enseñanza-aprendizaje, proporcionando a los estudiantes un enfoque más interactivo Metodología: Se presenta la metodología implementada durante el curso 2023-24 por el grupo de innovación docente "InMaths" en la Universidad de Cantabria. La metodología incluyó la producción de videos, su integración en clases prácticas de ordenador y el análisis del impacto de los mismos en el aprendizaje del uso de software en matemáticas. Resultados: Los resultados de 300 encuestas realizadas en todas las materias del estudio indican que la implementación del aula invertida con soporte audiovisual mejora la motivación y la participación del alumnado en su aprendizaje. Los alumnos valoran positivamente los videos y creen que esta estrategia les prepara mejor para las actividades de prácticas. Discusión: Los resultados son coherentes con investigaciones anteriores que indican que la estrategia del aula invertida puede mejorar la preparación y participación del alumnado aunque en la experiencia de evaluación realizada con grupos de control y experimental solo se ha observado una mejora en la calificación en algunas asignaturas y grupos de estudiantes. Conclusiones: La metodología utilizada está muy bien percibida por el alumnado y puede mejorar sus habilidades para resolver problemas académicos a través del uso de herramientas de software.
- 20 May 2025
- Published by Forum XXI in European Public & Social Innovation Review
- Vol. 10, 1-22
Open AccessApoyo audiovisual en un modelo de aula invertida en las prácticas de Matemáticas en Grados de Ingeniería
Introducción: El estudio se enfoca en la incorporación de recursos audiovisuales y la estrategia del aula invertida en prácticas de ordenador. Estas técnicas son efectivas para enriquecer el proceso de enseñanza-aprendizaje, proporcionando a los estudiantes un enfoque más interactivo Metodología: Se presenta la metodología implementada durante el curso 2023-24 por el grupo de innovación docente "InMaths" en la Universidad de Cantabria. La metodología incluyó la producción de videos, su integración en clases prácticas de ordenador y el análisis del impacto de los mismos en el aprendizaje del uso de software en matemáticas. Resultados: Los resultados de 300 encuestas realizadas en todas las materias del estudio indican que la implementación del aula invertida con soporte audiovisual mejora la motivación y la participación del alumnado en su aprendizaje. Los alumnos valoran positivamente los videos y creen que esta estrategia les prepara mejor para las actividades de prácticas. Discusión: Los resultados son coherentes con investigaciones anteriores que indican que la estrategia del aula invertida puede mejorar la preparación y participación del alumnado aunque en la experiencia de evaluación realizada con grupos de control y experimental solo se ha observado una mejora en la calificación en algunas asignaturas y grupos de estudiantes. Conclusiones: La metodología utilizada está muy bien percibida por el alumnado y puede mejorar sus habilidades para resolver problemas académicos a través del uso de herramientas de software.