Reliable geolocation of thermal hotspots, such as smoldering embers that can reignite after vegetation fire suppression, deep-seated peat fires, or underground coal seam fires, is critical to prevent fire resurgence, limit prolonged greenhouse gas emissions, and mitigate environmental and health impacts. This study develops and tests an algorithm to estimate the GPS positions of thermal hotspots detected in infrared images acquired by an unmanned aerial vehicle (UAV), designed to operate over flat and mountainous terrain. Its originality lies in a reformulated Bresenham traversal of the digital elevation model (DEM), combined with a lightweight, ray-tracing-inspired strategy that efficiently detects the intersection of the optical ray with the terrain by approximating the ray altitude at the cell level. UAV flight experiments in complex terrain were conducted, with thermal image acquisitions performed at 60 m and 120 m above ground level and simulated hotspots generated using controlled heat sources. The tests were carried out with two thermal cameras: a Zenmuse H20T mounted on a Matrice 300 UAV flown both with and without Real-Time Kinematic (RTK) positioning, and a Matrice 30T UAV without RTK. The implementation supports both real-time and post-processed operation modes. The results demonstrated robust and reliable geolocation performance, with mean positional errors consistently below 4.2 m for all the terrain configurations tested. A successful real-time operation in the test confirmed the suitability of the algorithm for time-critical intervention scenarios. Since July 2024, the post-processed version of the method has been in operational use by the Corsica fire services.

With global warming, the wildfire season tends to get longer, causing fatalities and devastating damage to human property. Although many countries have implemented fire risk prevention measures, particularly in Wildland Urban Interfaces (WUI), this finding shows that there are weaknesses in the prevention measures. This is mainly due to a lack of knowledge about WUI fire exposure conditions. This paper presents field-scale experiments to characterise the burning of rockrose-reconstructed hedges (6 x 1 x 1 m) close to a building in order to provide experimental data on heat release rate (HRR), flame front geometry, and heat fluxes to the building. The mean horizontal flame extent was 4.4 (+/- 0.7) m with values up to 5.5 m. These values are generally higher than the minimum distance to be maintained between vegetation and buildings in most countries. The fire intensity ranged from 283 to 3479 kW/m, resulting in maximum values at 3 m from the hedge of up to 45.4 kW/m2 for the total heat flux and 38.1 kW/m2 for the radiant heat flux. However, the flame duration is short, averaging 112.8 (+/- 27.1) s. This type of exposure is far from those used to test the fire resistance worldwide. Therefore, the data obtained in this study are crucial for improving fire risk prevention methods worldwide, whether for fuel management in defensible zones or for testing building materials to make buildings more resistant to wildfires.

With climate change, complex fire scenarios at Wildland Urban Interfaces (WUI) have become a growing concern for authorities worldwide. Such scenarios involve unfavorable wind conditions, which greatly increase the fire spread near communities and makes the intervention of firefighters even more complex. To design effective fire safety regulations to protect communities, it is important to have a clear understanding of the role of wind in the burning of vegetation to help identify the vectors of fire spread within the defensible zone. In this context, this paper attempts to relate the properties of the flames generated by the burning of a reconstructed rockrose hedge to the atmospheric wind turbulence using four field-scale experiments. The experimental setup includes 3 visible cameras to record the flame geometry, extracted by post-processing; 3 pairs of radiant and total heat fluxes to record the heat fluxes downstream of the hedge, and an anemometer to measure the wind speed and direction. The results show that the wind at the experimental site is composed of wall-bounded turbulent structures coupled with strong and short intermittent events such as gusts. Regarding the fire, the four replicates show similarities in phase durations and dynamics, but revealed dependencies on Reynolds number and turbulent kinetic energy when averaging over the fully developed flame phase. In addition, a slight flapping motion of the ensemble flame was observed and associated with wave numbers of the order of the turbulent cascade in the kinetic energy power spectra. Finally, the analysis of wind and fire intermittencies revealed that strong and short wind events generate large fluctuations of flame geometry and received heat fluxes that increase with the characteristic time of the event considered.

AbstractWildfire pressure involves today to implement silvicultural practices that provide a good compromise between reducing fire risk and maintaining ecological functioning. Thinning reduces tree density and low branches, but results in the deposition of a considerable biomass of woody debris on the ground (up to 4800 g m2 in this study). They can be eliminated by prescribed burning, but this raises questions about the fire intensity that can be generated and the impact on soil fauna. We undertook the monitoring of a thinning and prescribed burning operation, separated and combined, in November 2020, in a Pinus laricio stand prone to fire risk, located in Bavella, Corsica. Fuel load was determined, and temperature measurements in the soil were performed using K‐type thermocouples. Soil arthropod populations were monitored using pitfall traps, in particular Collembola, Acari, Aranae, and Coleoptera. The combination of thinning and burning resulted in a fire intensity of 75.8 versus 8.4 kW m−1 for burning alone. Maximum temperature rise measured at −2 cm below the surface was less than 5°C for both treatments. The combination of thinning and burning did not result in higher fire intensity at ground level than burning alone, and the soil showed high insulation capacity. Most of the woody debris that burned was small‐diameter, and large‐diameter debris remained unconsumed. This burning, performed during a period of low biological activity, had no effect on soil arthropods, and the presence of large debris may have provided refuge areas. Collembola group was the faster to recover, and were followed by cohorts of predators in summer, especially Acari. Our results suggest that a combination of burning and thinning in autumn may be beneficial for fire prevention. However, the decomposition of woody debris in relation to fire risk, and the occurrence of pests after these treatments need to be monitored.

The aim of this work is to improve the understanding of the fire behaviour of building materials used at the wildland urban interface using laboratory scale experiments. To this end, wood cladding was exposed to a propagating flame front across a decking slab. This slab element was ignited by a radiant panel. Three slab configurations were studied: thermoplastic slabs and wooden slabs with slats oriented parallel or perpendicular to the cladding. The vertical distance between the slabs and the cladding varied from 0 to 20 cm and two types of batten arrangement were used to separate the cladding from the supporting cellular concrete. The experiments first focused on the measurement of the heat fluxes received by a non-combustible wall in place of the cladding, depending on the type of slab used. Radiation was the dominant heat flux measured on the wall and it increased up to 35 kW m-2 as the flame front approached to the cladding. The spreading of the fire from the slab to the cladding was then investigated. Ignition of the cladding occurred in all configurations tested. The slab configuration (material or orientation) has little effect on the combustion of the cladding once it has been ignited. In addition, the distance between the deck slab and the cladding did not have a significant impact on the fire behaviour of the cladding, even at the 20 cm distance recommended by the standards.

This study is part of an ongoing project on modeling vegetation fires and their impacts on dwellings at wildland-urban interfaces. As a first step, fire tests were conducted at laboratory scale on both wooden wool litter and rockrose non-natural hedge. To this end, a 0.5 m x 0.5 m x 1 m hedge reconstituted from rockrose branches was burned under a Large Scale Heat Release (LSHR) apparatus. A litter of wooden wool was used before and under the hedge to mimic an ignition from an underneath burning herbaceous layer. These fire tests provided experimental data for mass loss and heat release rate, which were used as a basis for comparison with the simulations performed by using the Wildland-urban interface Fire Dynamics Simulator (WFDS). To numerically model the rockrose hedge, the distribution of the different particle classes (leaves, twigs of various diameters) was determined as well as their physicochemical properties. The comparison of the experimental data and predicted quantities showed a good agreement, whether for wooden wool burnings alone or mixed wooden wool and rockrose hedge.

The aim of this work is to improve the understanding of the fire behaviour of building materials used in the wildland urban interface at a large scale. To this end, wood cladding was exposed to the heat flux from the burning of a cistus hedge ignited by a spreading fire front in a fuel bed of excelsior litter. Six experiments were performed with different configurations: hedge/cladding distance, cistus bulk density, hedge length. During the experiments, the assessment of the cladding ignition was observed by camera. Wind direction and velocity were measured continuously since the tests were conducted outdoors and wind is a significant driver of flame spread. Heat fluxes and temperatures were also measured in order to compare the intensity of the thermal aggression on the cladding coming from the burning hedge. When the distance between the hedge and the cladding was 3 m, the wood did not ignite even when the wind was in the direction of propagation towards the cladding. In these cases, the maximum heat flux was 18 kW m−2. This value did not allow the ignition of the wood cladding. When the distance was reduced, the ignition of the cladding depended on the wind direction. When the cladding was ignited, the heat fluxes from the burning hedge could exceed 100 kW m−2.

This paper reports two experimental fires conducted at field-scale in Corsica, across a particular mountain shrubland. The orientation of the experimental plots was chosen in such a way that the wind was aligned along the main slope direction in order to obtain a high intensity fire. The first objective was to study the high intensity fire behavior by evaluating the propagation conditions related to its speed and intensity, as well as the geometry of the fire front and its impact on different targets. Therefore, an experimental protocol was designed to determine the properties of the fire spread using UAV cameras and its impact using heat flux gauges. Another objective was to study these experiments numerically using a fully physical fire model, namely FireStar3D. Numerical results concerning the fire dynamics, particularly the ROS, were also compared to other predictions of the FireStar2D model. The comparison with experimental measurements showed the robustness of the 3D approach with a maximum difference of 5.2% for the head fire ROS. The fire intensities obtained revealed that these experiments are representative of high intensity fires, which are very difficult to control in the case of real wildfires. Other parameters investigated numerically (flame geometry and heat fluxes) were also in fairly good agreement with the experimental measurements and confirm the capacity of FireStar3D to predict surface fires of high intensity.

A safe separation distance (SSD) needs to be considered during firefighting activities (fire suppression or people evacuation) against wildfires. The SSD is of critical interest for both humans and assets located in the wildland–urban interfaces (WUI). In most cases, the safety zone models and guidelines assume a flat terrain and only radiant heating. Nevertheless, injuries or damage do not result exclusively from radiant heating. Indeed, convection must be also considered as a significant contribution of heat transfer, particularly in the presence of the combined effects of sloping terrain and a high wind velocity. In this work, a critical case study is considered for the village of Sari-Solenzara in Corsica (France). This site location was selected by the operational staff since high-intensity fire spread is likely to occur in the WUI during wind-blown conditions. This study was carried out for 4 m high shrubland, a sloping terrain of 12° and a wind speed of 16.6 m/s. The numerical simulations were performed using a fully physical fire model, namely, FireStar2D, to investigate a case of fire spreading, which is thought to be representative of most high wildfire risk situations in Corsica. This study is based on the evaluation of the total (radiative and convective) heat flux received by two types of targets (human bodies and buildings) located ahead of the fire front. The results obtained revealed that the radiation was the dominant heat transfer mode in the evaluation of the SSD. In addition, the predictions were consistent with the criterion established by the operational experts, which assumes that in Corsica, a minimum SSD of 50 m is required to keep an equipped firefighter without injury in a fuelbreak named ZAL. This numerical work also provides correlations relating the total heat flux to the SSD.

Among the vectors of fire propagation towards buildings in the WUI, ornamental hedges have been identified as one of the main elements [...]

This paper reported a high intensity experimental fire conducted during a field-scale experiment on a steep sloped terrain (28) as part of a winter prescribed burns campaign managed by the local firefighter service in the north-western region of Corsica. The rate of spread (ROS) of fire, measured using UAV cameras (thermal and visible), was evaluated at 0.45 m/s. The experiment was numerically reproduced using a completely physical 2D model, namely FireStar2D, and the comparison with the experimental measurements mainly concerned the fire ROS and the heat fluxes received by three distant targets placed at the end of the plot. The results analysis shows that the considered fire has a wind-driven regime of propagation with a fire intensity higher than 7 MW/m. The numerical results are in fairly good agreement with the experimental measurements, within 11% difference for the ROS and 5% for the heat fluxes, validating consequently the relevance of the numerical approach to tackle such high-intensity wildfires. Despite the unfavorable wind and humidity conditions for fire propagation (U = 1.67 m/s and RH = 82%), this experiment confirms that such fire can exhibit a dangerous behavior due to the steep slope of the terrain.

With global warming and rising occurrence of drought conditions, the risk of fire spreading at Wildland–Urban Interfaces (WUI) increases every year [...]

Like many Mediterranean regions, Corsica is frequently exposed to forest fire risk [...]

Field-scale experiments have been conducted on steep sloped terrains in Speluncatu and Letia, north-western and southern regions of Corsica. This work lies within the GOLIAT project framework and it was provided by the Fire and Rescue Service of North Corsica and the Corsican DFCI (Défense de la Forêt Contre l’Incendie) Group. This work reported high intensity fires propagating through shrub vegetation areas (Genista Salzmannii) lying between 60 cm and 85 cm. These sites were selected because of the density of the vegetation, the high slope angle values with a wind direction aligned with the main slope, which can generate a fire close to wildfire behaviour. A detailed experimental protocol is used in order to determine the propagation conditions and the fire behaviour using UAV cameras and heat flux gauges. In order to investigate the different phenomena encountered in these types of fires, numerical simulations were conducted using a complete physical fire model, based on multiphase formulation, namely FireStar2D. Numerical predictions were used to examine the fire front dynamics related to the fire’s rate of spread and fireline intensity. Despite the unfavourable wind and humidity conditions, experimental results analysis showed that the fireline intensity was higher than 7 MW/m, which means that these fires fall into the category of the very high fire severity. Numerical results predicting the fire’s rate of spread, fireline intensity and fire impact were in good agreement with the experimental data.

The objective of this preliminary study is to highlight the potential impacts of prescribed fire on soil organism communities in Pinus laricio forests, a species endemic to Corsica. For this purpose, a control plot, and a plot burned on 10th November 2020, were delimited in Bavella (South Corsica). The intensity of the burning was characterized using K-Type thermocouples at different depths in the soil. Soil organisms were collected with pitfall trap and Berlèse method, then identified to the order. Results show a decrease of all orders on the burned plot, with important potential of survival of the organisms in the soil, since heat transfer is weak (+10°C max at 3cm depth) and the burned surfaces heterogeneous. A rapid recolonization is indeed observed in the spring. The contribution of organic matter, minerals, and free ecological niches could encourage this phenomenon. In order to better understand the heterogeneity of the plots and to highlight the specificities of Pinus laricio forests, a study at the sub-plot level is in progress, as well as an identification of organisms up to the family level.

SummaryThis study evaluates two kinetic mechanisms to predict, on matter‐level and material‐level scales, the thermal degradation of thin plates of oak and eucalyptus as representing vegetative fuel involved in wildfires. The lumped mechanism considers four successive steps, whereas the simplified one includes two steps. The kinetic parameters of these mechanisms were identified with thermogravimetric analysis in air for heating rates between 2 and 30 °C min−1. On the matter‐level scale, kinetic mechanisms were tested for heating rates inside and outside the parameters optimization range. The simulations of the mass loss were close to the experimental data. The lumped mechanism gave better results than the simplified one. On the material‐level scale, the mass loss and temperature recorded during thermal degradation of thin wooden plates, with a heating cone, were compared to simulations performed with GPYRO. When only the gasification stage occurred, both mechanisms predicted results close to the experimental data. When char oxidation occurred, the beginning of the gasification stage was well predicted, whereas some differences appeared during the transition between the gasification and char oxidation stages. The performance of both mechanisms indicates that a two‐step mechanism is capable of modeling the thermal degradation of thin wooden plates on a material scale.

The GOLIAT project is a consortium of academics and firefighting operators and land-use planning professionals of Corsica. One goal of GOLIAT project is to provide four operational decision support tools. To reach this goal, a survey of past fires occurred in Corsica since the twentieth century beginning is made. This inventory contributes to build up a database with a web display interface easy to use as fire patterns history. A fire behavior and impact simulator prototype for vegetation fires, a geolocation tool for hot spots using UAV images, and a guide of good practices of prescribed fires in the undergrowth are building. At the same time, experimental fires are carried out to improve knowledge about high intensity fire and the experimental results were compared to the predictions provided by a complete physical 3D model, namely FireStar3D.

Wildfires stand as one of the most relevant natural disasters worldwide, particularly more so due to the effect of climate change and its impact on various societal and environmental levels. In this regard, a significant amount of research has been done in order to address this issue, deploying a wide variety of technologies and following a multi-disciplinary approach. Notably, computer vision has played a fundamental role in this regard. It can be used to extract and combine information from several imaging modalities in regard to fire detection, characterization and wildfire spread forecasting. In recent years, there has been work pertaining to Deep Learning (DL)-based fire segmentation, showing very promising results. However, it is currently unclear whether the architecture of a model, its loss function, or the image type employed (visible, infrared, or fused) has the most impact on the fire segmentation results. In the present work, we evaluate different combinations of state-of-the-art (SOTA) DL architectures, loss functions, and types of images to identify the parameters most relevant to improve the segmentation results. We benchmark them to identify the top-performing ones and compare them to traditional fire segmentation techniques. Finally, we evaluate if the addition of attention modules on the best performing architecture can further improve the segmentation results. To the best of our knowledge, this is the first work that evaluates the impact of the architecture, loss function, and image type in the performance of DL-based wildfire segmentation models.