Wildfire behaviour is becoming increasingly volatile, expressed through more frequent and pronounced episodes of anomalously intense fire activity. Conventional approaches to characterising wildfire intensity, particularly those based on satellite-derived fire radiative power, remain strongly tied to fire occurrence, making it difficult to distinguish between periods of widespread burning and genuinely anomalous or extreme fire behaviour.
Here, we demonstrate that departures from expected relationships between fire detections and energy release reveal increasingly frequent and pronounced episodes of anomalous fire behaviour. These anomalies are characterised not only by elevated total energy release, but by shifts in the upper tail of the fire energy distribution, indicating a growing prevalence of disproportionately intense fire activity. Importantly, this behaviour does not occur uniformly, but instead reflects distinct modes of fire activity, ranging from widespread high-energy burning to more localised, extreme events.
Across multiple regions and scales, these patterns point to a broader increase in wildfire volatility, defined by growing variability and instability in fire energy behaviour over time. This suggests that fire regimes are not only intensifying but becoming less predictable and more prone to rapid departures from expected conditions. Within this framework, Mediterranean ecosystems emerge as exhibiting episodic volatility, characterised by punctuated extreme fire activity, underscoring their sensitivity to climate-driven risk.
This emerging volatility has important implications for wildfire management under a changing climate. As fire behaviour becomes increasingly unstable, existing assumptions around fire growth, intensity, and spread may become less reliable, complicating forecasting, planning, and response. Recognising and accounting for increasing volatility will be critical for adapting wildfire management strategies to a future in which extreme and anomalous fire behaviour plays a more dominant role.
How to cite: Derler, Z., Little, K., Flannigan, M., Belcher, C., and Kettridge, N.: Fire regimes are becoming increasingly volatile: implications for managing extreme fire events in the Mediterranean, 19th Plinius Conference on Mediterranean Risks, Murcia, Spain, 6–9 Oct 2026, Plinius19-12, https://doi.org/10.5194/egusphere-plinius19-12, 2026.
Background: Wildfires in Southern Europe arise from interacting climatic, ecological, and socio-ecological mechanisms.
Drought enhances fuel desiccation, land cover change reshapes vegetation flammability, and the expansion
of the wildland–urban interface (WUI) increases ignition pressure and exposure. Understanding how these drivers
overlap spatially and temporally is essential to identify where reinforcing or decoupled processes shape wildfire
dynamics. This study develops a spatially explicit framework integrating drought variability (Standardized Precipitation–
Evapotranspiration Index, SPEI), Landscape Flammability Classes (LFC), and Wildland–Urban Interface (WUI)
expansion to detect statistically supported hotspots and coldspots of wildfire occurrence across biogeographical
regions.
Methods: We analysed a 20-year dataset (2001–2020) at 12-km resolution, combining number of fire and fire size
with long-term trends in three macro-drivers (SPEI, LFC, WUI). Temporal changes in wildfire parameters were estimated
using generalized linear models with False Discovery Rate (FDR) correction. Trends in SPEI and WUI were
assessed through Mann–Kendall and Kendall tau tests supported by Theil–Sen slope estimation, while LFC change
was derived from land-cover transitions between 2000 and 2018. Finally, a spatial co-occurrence analysis classified
each grid cell as a hotspot, coldspot, or mismatch area based on the degree of alignment between macro-driver
trajectories and wildfire trends.
Results: Significant and candidate hotspots were concentrated in the Anatolian, Continental, and Mediterranean
bioregions, where over 30% of burned areas showed concurrent increases in drought intensity, landscape flammability,
and WUI expansion. The Anatolian bioregion exhibited the strongest increases in WUI (+ 73.4%) and LFC (+ 67.2%),
while the Continental region was dominated by widespread drying (90.5% of its area). Coldspots were mainly located
in the Atlantic region, reflecting coherent declines in both wildfire parameters and macro-drivers. Mismatch zones,
encompassing more than half of the burned area, revealed high spatial variability where fire dynamics diverged
from macro-driver trends, indicating the influence of local-scale or non-mechanistic processes.
Conclusions: The joint analysis of climate variability, fuel continuity and human pressures shows that wildfire patterns
in Southern Europe depend on how the long-term trends of these macro-drivers align or decouple. Hotspots reveal
regions where multiple drivers evolve in the same direction as wildfire activity, whereas mismatch areas highlight
the role of local processes that interrupt large-scale relationships. The framework provides a quantitative basis
How to cite: Cappelluti, O., Ascoli, D., Oliveira, S., Lafortezza, R., Chiriacò, M. V., Sanesi, G., and Elia, M.: Intertwined drivers of wildfire occurrence in Southern Europe: climate, landscape flammability and WUI expansion, 19th Plinius Conference on Mediterranean Risks, Murcia, Spain, 6–9 Oct 2026, Plinius19-20, https://doi.org/10.5194/egusphere-plinius19-20, 2026.
Accurate wildfire prediction is becoming increasingly critical as climate change drives warmer and drier conditions worldwide. The complex, non-linear interactions among meteorological factors, fuel characteristics, and landscape structure make wildfire risk a primary candidate for advanced machine learning (ML) approaches that integrate Earth Observation (EO) and climate data. In contrast to traditional operational risk systems commonly based only on weather conditions, these ML-EO systems can be trained on a much richer context, allowing the generation of not only more accurate wildfire occurrence risk maps but also the potential corresponding impacts at a much higher spatial resolution.
To demonstrate this paradigm shift, we present a multi-scale set of modeling developments that address the wildfire lifecycle from onset to impact. We begin at the landscape scale with a 100m calibrated daily probability of occurrence model that combines the Fire Weather Index (FWI), high-resolution land cover data, and historical fire event records. For all subsequent models, we deliberately neglect the highly unpredictable ignition component by focusing exclusively on burned areas to predict potential behavior in the event of a fire. Within this framework, we first explore the estimation of active fire intensity, evaluating how environmental drivers can enable the prediction of potential Fire Radiative Power (FRP). Moving further to evaluate the final physical consequences on the landscape, we introduce a 30m potential wildfire severity model for predicting vegetation damage in ecosystems; just from initial conditions, the model successfully identifies critical thresholds and skillfully predicts which specific areas within a fire's perimeter would be most severely burned. Finally, to capture the key role that spatial context plays in fire behavior, we explore Convolutional Neural Networks (CNNs) aiming to learn fire connectivity patterns directly from historical events and enabling the modelling of valuable variables for fire managers, such as the size of a potential wildfire event. Together, these developments mark a significant step toward an operational, high-resolution and comprehensive wildfire risk pipeline strengthening both early-warning capabilities and long-term resilience planning.
How to cite: Peña-Izquierdo, J., Perpinyà-Vallès, M., Cendagorta-Galarza, D., Florindo, C., Huertas, C., Teruel, D., Folguera, G., Llort, J., and Romero, L.: From Occurrence to Impact: A Multi-Scale Machine Learning Pipeline for High-Resolution Wildfire Modelling, 19th Plinius Conference on Mediterranean Risks, Murcia, Spain, 6–9 Oct 2026, Plinius19-106, https://doi.org/10.5194/egusphere-plinius19-106, 2026.
Climate change is increasing the frequency, duration, and intensity of droughts and heat waves across Mediterranean regions, contributing to more severe wildfire seasons. Portugal has experienced several extreme wildfire events in recent decades, highlighting the need for improved understanding of fuel moisture dynamics and their relationship with fire risk. This study presents a field-based monitoring approach using a network of in-situ sensors installed in Portugal to continuously measure Fine Fuel Moisture Content (FMC). The sensors record FMC (%) at 10-minute intervals, providing high-resolution information on fuel moisture variations under different meteorological conditions. The objective is to quantify the relationship between FMC dynamics and wildfire danger under increasingly frequent drought and heatwave conditions. Continuous monitoring allows the identification of critical moisture thresholds associated with increased fuel flammability and potential fire ignition. Preliminary observations indicate rapid decreases in FMC during prolonged drought periods and heatwave events, with fuel moisture reaching critical levels associated with increased fuel flammability and ignition potential.. The collected data are being analysed to characterize FMC variability, identify critical moisture thresholds, and assess the potential integration of FMC measurements into operational wildfire danger assessment systems.. The study further examines how projected climate change conditions may influence fuel moisture availability and contribute to increased wildfire danger and extreme fire behaviour in Mediterranean environments. The preliminary results highlight the value of real-time FMC monitoring for wildfire prevention, early warning systems, and decision support in fire management.. This approach contributes to a better understanding of climate-driven changes in fuel moisture dynamics and supports the development of adaptation strategies aimed at increasing the resilience of Mediterranean forest ecosystems to wildfire risk.
How to cite: Raposo, J., Rodrigues, A., Raposo, H., Reis, L., and Morais, A.: Monitoring Fine Fuel Moisture Content Using In-Situ Sensors for Wildfire Risk Assessment under Climate Change Conditions in Portugal, 19th Plinius Conference on Mediterranean Risks, Murcia, Spain, 6–9 Oct 2026, Plinius19-82, https://doi.org/10.5194/egusphere-plinius19-82, 2026.
Fire lookout towers remain a critical component of early detection systems in Mediterranean countries. However, the effectiveness of tower networks is constrained by topographic visibility limitations, creating significant surveillance gaps in forested landscapes. This study presents a national-scale GIS-based visibility analysis of Türkiye's 548 forest fire lookout towers and identifies priority surveillance deficit zones by integrating historical fire records, forest stand characteristics, and road network accessibility.
Viewshed analysis was conducted using a 30-metre resolution NASA SRTM digital elevation model under curved-earth parameters, with an observer height of 6 metres and a smoke detection height of 100 metres, consistent with operational standards defined in the General Directorate of Forestry's Communiqué No. 285. The resulting binary visibility surface was overlaid with three datasets: the 12 regional forest directorate boundaries, a national forest stand map comprising 3,998,699 polygons with species composition and structural attributes, and a georeferenced dataset of 30,883 wildfire incidents recorded between 2013 and 2023.
Results indicate that the existing tower network provides visibility coverage over 47.6% of Türkiye's forested territory, leaving 52.4% as blind spots. Regional analysis reveals substantial disparities, ranging from near-complete coverage in the Aegean and Marmara regions (Kütahya 99.2%, Çanakkale 98.5%) to near-zero coverage in northeastern regions (Artvin 0.0%, Erzurum 0.1%). Species-level analysis shows that fire-prone pine species such as Pinus brutia (86.6%) and Pinus nigra (87.8%) are relatively well covered, whereas Picea orientalis (0.2%) and Pinus sylvestris (43.5%) stands fall predominantly within blind spots. Analysis of historical fire records reveals that 75.1% of all recorded fires occurred within visible zones, while 24.9% (n=7,618) originated in areas invisible to any tower.
Optimized Hot Spot Analysis (Getis-Ord Gi*) applied to 5,356 wildfire incidents occurring in forested blind spots identified 251 statistically significant priority surveillance deficit zones. These zones are concentrated predominantly in Şanlıurfa (69 hotspots, 502 fires), Elazığ (46 hotspots, 381 fires), Adana, Trabzon, and Hatay. Road network accessibility analysis using OpenStreetMap data confirmed that 99.2% of identified hotspot zones are accessible via the existing road network, based on a connectivity threshold of 4,000 metres derived from the road-access pattern of the existing tower network.
These findings reveal structural surveillance gaps in Türkiye's fire detection network, particularly in southeastern and eastern Anatolia and the Black Sea region. The study provides a data-driven framework for prioritising surveillance deficit zones and recommends evaluating complementary detection technologies such as unmanned aerial vehicles and automated camera systems for areas where conventional tower infrastructure is insufficient.
How to cite: Küçükosmanoğlu, M. A. and İnan, M.: National-Scale GIS-Based Visibility Analysis of Fire Lookout Towers and Identification of Surveillance Deficit Zones in Türkiye, 19th Plinius Conference on Mediterranean Risks, Murcia, Spain, 6–9 Oct 2026, Plinius19-56, https://doi.org/10.5194/egusphere-plinius19-56, 2026.
Catastrophic wildfires are increasing, encroaching further into populated areas. It is extremely important to be prepared when wildfire strikes. By preparing the home and property for wildfire and knowing what to do if evacuation is necessary, the safety and the survivability can be dramatically increased.
Get Ready
Hardening the home: a home can be hardened by retrofitting it with ignition-resistant or noncombustible materials to protect against the threat of flying embers, direct flame contact and radiant heat exposure.
Defensible space: defensible space is the buffer created between a building on the property and the plants, brush, trees, or other combustible items in the near vicinity. This buffer helps to keep wildfire away from the home by reducing the fire’s intensity and slowing or halting the spread of wildfire. Creating this space also provides protection for the firefighters defending the home.
Fire smart landscaping: taking it into consideration can be crucial to reduce the spread of wildfire around the home.
Get set
Create a wildfire action plan: such plan must be prepared and familiar to all members of the household well in advance of a wildfire.
Keep necessary items ready if immediate evacuation required: pets, important papers, prescriptions, PC, plastic cards, etc.
Prepare an emergency supply kit: put together an Emergency Supply Kit/Go Bag before a wildfire occurs and keep it easily accessible to take when evacuation occurs.
Be prepared for power outages: power outages may occur before and during the threat of a wildfire. It’s important to be prepared and know what actions to take when leaving the home during a power outage.
Have a family communication plan: have one in place in case family members get separated.
Insurance preparedness: maintain insurance, know what your policy covers, make home inventory before evacuating.
Go!
Prepare pre-evacuation steps: when evacuation is anticipated and if time allows check one more time outside the house, inside the house, animals.
Implement evacuation steps: review the evacuation checklist. Take the Emergency Supply Kit. Cover up to protect against heat and flying embers. Locate and take pets.
Know when to evacuate: leave when evacuation is recommended by fire officials to avoid being caught in fire, smoke or road congestion. In an intense wildfire, emergency personnel may not have time to knock on every door. If danger is imminent, the best course of action is to evacuate. If advised to leave, don’t hesitate!
Anticipate about animal evacuation: after family and home safety, don’t forget about pets and livestock. Their chances of surviving a wildfire can be increased with some advanced planning.
Know what to do if trapped: while in vehicle, while on foot, while in the home.
Returning home after a wildfire: always check with officials before attempting returning home after a wildfire. Be aware of dangers that exist after a wildfire.
How to cite: Katay, Z.: Wildfire preparedness, awareness and action plan - ready, set, go !, 19th Plinius Conference on Mediterranean Risks, Murcia, Spain, 6–9 Oct 2026, Plinius19-2, https://doi.org/10.5194/egusphere-plinius19-2, 2026.
Wildfires occurring on slopes with canyons represent some of the most challenging and hazardous scenarios for firefighting operations. The interaction between topography, wind and fire dynamics can lead to sudden changes in fire behaviour, including rapid acceleration, increases in fire intensity and eruptive fire behaviour, significantly reducing the time available for operational response and escape.
This study analyses a set of wildfire case studies that occurred in Portugal and the United States, including the Armamar (1985), Yarnell Hill (2013), Caramulo (2013) and Serra da Estrela (2022) wildfires. These events were selected because they exhibited significant fire behaviour changes associated with canyon topography and, in several cases, resulted in firefighter fatalities. The analysis was based on official reports, fire progression records, field observations and photographic evidence, complemented by knowledge gained from experimental studies on fire propagation in slopes with canyons.
The results reveal consistent fire behaviour patterns despite differences in fuel characteristics, meteorological conditions and geographical location. In all cases, the approach of the fire to a canyon was followed by a marked increase in the rate of spread and fire intensity, often accompanied by abrupt changes in propagation direction and eruptive fire behaviour. These transitions contributed to dangerous operational conditions and, in the cases of Armamar, Yarnell Hill and Caramulo, were associated with fatal firefighter entrapments. Observations from the Serra da Estrela wildfire further demonstrate that canyon-induced fire acceleration can occur even without significant changes in weather conditions.
The analysed case studies highlight the critical role of canyon topography in wildfire behaviour and firefighter safety. The findings emphasise the importance of recognising high-risk topographic configurations, improving situational awareness and incorporating fire behaviour knowledge into operational planning and decision-making. These lessons can contribute to safer wildfire suppression strategies in complex terrain.
How to cite: Rodrigues, A., Antunes, F., and Raposo, J.: Fire behaviour and safety challenges in wildfires occurring on slopes with canyons: case studies, 19th Plinius Conference on Mediterranean Risks, Murcia, Spain, 6–9 Oct 2026, Plinius19-107, https://doi.org/10.5194/egusphere-plinius19-107, 2026.
Under climate change and shifting fire regimes, a wildfire can trigger forest responses ranging from full recovery to shifts toward non-forest states. Understanding the trajectories of forests after fire and the effectiveness of post-fire management and restoration strategies is essential for designing sustainable forest management. However, our capacity to learn from local field studies is limited as they lack the capacity to extrapolate results to broader scales, and scientific syntheses often suffer from non-standardised data generation. Such limitations can be tacked by conducting coordinated, distributed studies, in which many sampling sites are collaboratively implemented following a common protocol across large geographic regions. This approach allows responding research questions broadly as well as robustly assessing the drivers of heterogeneity.
The FireTran project aims to establish an international distributed research network across Mediterranean countries to monitor post-fire forest trajectories and experimentally test the effect of small-scale restoration treatments in modulating these trajectories. The project aims to collaboratively monitor dozens of sites across the region –under small local effort– to quantify regeneration dynamics and assess the drivers of heterogeneity at plot, site, and regional scales. Additionally, voluntary experimental treatments (tree planting, seeding, soil transfer) can be implemented at a subset of participating sites. Remote sensing will help locate field plots and, through integration with field data, it will allow to assess large-scale drivers through spatial upscaling.
The FireTran project will soon initiate a local pilot phase in Granada. The study protocol will build on the lessons learned therein and learn from consultation with scientists and stakeholders. It will be published to invite scientists from Mediterranean countries to participate in the distributed study by establishing local sites. The protocol will establish the scientific aims, network structure and governance, field methods, requirements for co-authorship, and the data management approach. The presentation will provide a general overview of the project, invite the audience to provide input for the protocol, and screen their prospective willingness to participate by establishing local field sites.
How to cite: Leverkus, A. B. and Beierkuhnlein, C.: A new Mediterranean network to assess post-fire forest trajectories: Invitation for input and participation, 19th Plinius Conference on Mediterranean Risks, Murcia, Spain, 6–9 Oct 2026, Plinius19-117, https://doi.org/10.5194/egusphere-plinius19-117, 2026.
The summer of 2025 marked an exceptional wildfire season in Europe, with total burned area reaching approximately 1 million hectares by the end of August, the highest value on record. More than half of this burned area was concentrated in Northwest Iberia, where multiple large fires developed over only a few weeks during an intense heatwave across southwestern Europe.
Building on Sánchez-Hernández et al. (2025), this contribution analyses the 2025 Northwest Iberian fires as an example of emerging Mediterranean wildfire risk. Using EFFIS burned area data and Fire Weather Index information, we show that August 2025 displayed the most extreme monthly fire-weather conditions in the region during 1985–2025. Burned area and fire weather were strongly associated, but their relationship was non-linear, indicating that extreme fire weather is necessary but not sufficient to produce extreme burned area.
The fires also showed marked vegetation selectivity, with shrublands contributing disproportionately to burned area, suggesting an important role of fine-fuel continuity and landscape-scale fuel accumulation. Overall, the 2025 fires illustrate how extreme meteorological hazard, continuous fuels and territorial vulnerability can interact to generate near-synchronous large fires that exceed suppression capacity. These results underline the need for integrated risk-reduction strategies combining climate mitigation, land-use planning, fuel management and community resilience.
Reference
Sánchez-Hernández, G., Turco, M., Repeto-Deudero, I., Royé, D., Baudena, M., Montávez, J. P., ... & Pausas, J. G. (2025). Record-breaking 2025 European wildfires concentrated in Northwest Iberia. Global Change Biology, 31(12), e70649.
How to cite: Turco, M., Sánchez-Hernández, G., Repeto-Deudero, I., Royé, D., Baudena, M., Montávez, J. P., Pietroiusti, R., Provenzale, A., Santin, C., Torres-Vázquez, M. Á., and Pausas, J. G.: Record-breaking 2025 wildfires in Northwest Iberia: extreme fire weather, fuel continuity and emerging Mediterranean fire risk , 19th Plinius Conference on Mediterranean Risks, Murcia, Spain, 6–9 Oct 2026, Plinius19-47, https://doi.org/10.5194/egusphere-plinius19-47, 2026.
As the MODIS era approaches its operational end, maintaining the long-term consistency of global burned area (BA) records is essential for fire-climate research, emissions assessment and climate applications. Here, we present MRBA60, a harmonised global monthly BA dataset designed to ensure continuity between MODIS-based FireCCI51 and Sentinel-3-based FireCCIS311.
MRBA60 should be used as a combined time series: MRBA60H provides the harmonised historical component for 2003-2018, while FireCCIS311 provides the continuation from 2019 onwards. The combination of both components is referred to as MRBA60, providing a continuous global BA record at 0.25°. For the historical period, biome- and month-specific Random Forest models were trained during the 2019-2024 overlap between FireCCI51 and FireCCIS311, using auxiliary predictors related to active fires, fire radiative power, climate, vegetation and spatial gradients. Model outputs were then adjusted through quantile mapping to better reproduce the FireCCIS311 BA distribution, while preserving historical extreme events not represented in the calibration period.
During 2003-2024, MRBA60 estimates a mean global BA of approximately 5.5 Mkm² yr⁻¹, compared with 4.4 Mkm² yr⁻¹ from FireCCI51. This represents a cumulative increase of ~24.5 Mkm² relative to FireCCI51, equivalent to ~1.1 Mkm² additional BA per year and a relative increase of 25%. Differences are seasonally heterogeneous: the largest increases occur in Sep-Oct-Nov and Mar-Apr-May, while Jun-Jul-Aug shows the smallest differences. Despite this increase in absolute BA magnitude, MRBA60 preserves the long-term global decline in burned area, with a cumulative reduction of -25.1% between 2003 and 2024, compared with -33.3% estimated by FireCCI51. Overall, MRBA60 provides a traceable and physically consistent time series for analysing BA trends across the MODIS-Sentinel transition, and is designed to support studies of fire regimes, climate-model evaluation, fire-climate attribution, emissions estimation and ecological impact assessment. MRBA60H is available through CEDA: https://catalogue.ceda.ac.uk/uuid/db75c5f51ee240ae8743355dcebbb9b9/. FireCCIS311 is available through CEDA: https://catalogue.ceda.ac.uk/uuid/da8e669a74334c82a56e0b470bc4ef04/.
Acknowledgements: This work was developed within the framework of the European Space Agency Fire Climate Change Initiative (FireCCI) project, contract Nº 4000126706/19/I-NB. We acknowledge CEDA for data archiving and distribution, and the providers of the satellite, climate and validation datasets used in the development and evaluation of MRBA60.
How to cite: Torres-Vázquez, M. Á., Segura-García, C., Turco, M., Khairoun, A., Pettinari, M. L., Solano-Romero, E., García, M., Oliva, P., W. Jones, M., and Chuvieco, E.: MRBA60: A harmonised global burned area record bridging the MODIS and Sentinel-3 eras (2003-present), 19th Plinius Conference on Mediterranean Risks, Murcia, Spain, 6–9 Oct 2026, Plinius19-103, https://doi.org/10.5194/egusphere-plinius19-103, 2026.
