Wildfires represent one of the most devastating natural disasters, bearing relevant environmental and socioeconomic impacts. The Mediterranean region is characterized by large and recurring summer wildfires that often jeopardize people’s safety. Currently, wildfire management largely (if not entirely) relies on wildfire suppression, despite growing evidence of its inefficiency to control the larger and more intense wildfires [1]. Moreover, climate change is expected to significantly affect the Mediterranean region and further exacerbate such hazard, even if global warming does not exceed 1.5°C (target of the Paris Agreement) [2]. Hence, fire prevention measures based on landscape fuel reduction strategies are crucial to decrease the magnitude of the impacts of future wildfires.

Here, we used FlamMap, a widely applied fire spread simulation system, to estimate fire spread and behaviour properties in the Monchique region, a highly fire-prone area, located in Southern Portugal. Five weather scenarios were defined based on hierarchical clustering analysis of temperature, relative humidity, wind speed and direction data derived from the spreading days of large wildfires (larger than 100 ha) between 2001 and 2019. Complex networks were generated from fireline intensity and rate of spread estimates (proxies for the difficulty of suppression and safety) with the main goal of decreasing landscape fire hazard. More precisely, we aimed to: i) evaluate how different weather scenarios/conditions affect landscape connectivity; ii) identify the location of fuel treatments; and iii) assess the impact of the proposed fuel breaks on the fire properties. These challenges were addressed under the perspective of connectivity indexes and metrics from the field of network science.

The results show that, as expected, weather conditions affect both the amount of area with more intense wildfires and wildfire connectivity, with more severe weather conditions presenting the greatest hazards. Additionally, the identified optimal locations of fuel treatments were compared against the locations previously proposed for fuel breaks and the potential impact on fire properties of both was evaluated. Further analysis of the effectiveness of different management options (fraction of landscape treatment and extent of each intervention) will be assessed under the previously identified weather scenarios, considering the extent of high-intensity classes of fires and multiple landscape connectivity indexes. Based on our results, we discuss the best strategies to reduce wildfire hazard for different criteria and under different weather scenarios. Moreover, both methods can be used to assess fire transmission between land uses and then to identify the key values exposed. We demonstrate that combining network graphs and fire spread simulations have a large potential to support more informed decision-making and significantly wildfire impact mitigation.

References

[1] Moreira, F., Ascoli, D., Safford, H. et al. (2020) Wildfire management in Mediterranean-type regions: paradigm change needed. Environmental Research Letters, 15, 011001. https://doi.org/10.1088/1748-9326/ab541e

[2] Turco, M., Rosa-Cánovas, J.J., Bedia, J. et al. (2018) Exacerbated fires in Mediterranean Europe due to anthropogenic warming projected with non-stationary climate-fire models. Nature Communications 9, 3821. https://doi.org/10.1038/s41467-018-06358-z

How to cite: Aparício, B. A., Sá, A. C. L., Santos, F. C., Bruni, C., and Pereira, J. M. C.: Combining wildfire behaviour simulations and complex network theory to support decision-making: A case-study in a Mediterranean region, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-980, https://doi.org/10.5194/egusphere-egu21-980, 2021.

Anthropogenic factors and climate change induced severe forest fires that are reoccurring globally. Because of the large spatial scale, frequent occurrence, and danger involved with the forest fires, remote sensing-based approaches are best suited to study this phenomenon. However, there are few studies addressing the temporal effects of the various factors associated with the  forest fire. In this work, by using Analytical Hierarchy Process (AHP), a multi-criteria decision support system and geostatistical methods namely Getis-Ord Gi* statstic and Mann Kendall trend test, we have developed a framework to understand the temporal dynamics of forest fire hazard and associated factors by demarcating the hotspots of forest fire using freely available datasets . The proposed framework has been applied on the Similipal Biosphere Reserve (SBR), Odisha, India. With an area of 5569 km², the SBR is the sixth largest biosphere reserve in India, comprising of a national park, bird sancturary, tiger reserve, and elephant corridor. Due to its biodiversity and importance in terms of rich and endemic species of flora and fauna, SBR was brought into the umbrella of world network of biosphere reserve under the Man and Biosphere (MAB) programme of UNESCO in the year 2008. Although being a biosphere of international importance, the SBR annually experiences nearly 12 km² of fire damage.Through this work, the most significant clusters of forest fire hotspots have been demarcated. We have used factors related to topographical, climatic, and physical characteristics of forest to generate forest fire hazard index at annual scale for 28 years (1988 – 2018) using AHP method. The geostatistical methods were applied on the generated annual fire hazard index data to demarcate the zones of emerging hotspots of forest fire. The results indicate that temporally, the trend of forest fire hazard in buffer zone of the area (Similipal Sanctuary) is decreasing, whereas in core area (Similipal National Park), it is increasing and corelates with the temporal trend of vegetation density in the whole area. However, vegetation density and land surface temperature in the core area does not changes significantly with time. The emerging hotspot analysis shows that most of the region (32% of the total area) is exhibiting an oscillating behaviour with respect to the fire hazard over the studied time-period of 28 years, with more than 50% of the years showing increasing trends of fire hazard. A total of 186 km² of the region is persistently a hotspot of fire hazard in studied time-period. Overall, 11% of the study area is either under persistent fire hazard or showing increasing trend of fire hazard. However, in the central part of the SNP, the fire hazard is decreasing with time. The same region also observes intense rain, and this could be a factor for the observed decrement in the fire hazard. The results can be used for mitigating the fire hazard of the SBR, alsothe proposed framework can be applied globally to any region with dense vegetation for fire hazard spatiotemporal assessments.

How to cite: Laha, A., Singh, S., Mishra, U., and Singh, M.: Estimating spatiotemporal dynamics of forest fire hazard using Analytical Hierarchy Process and geostatistical methods in Similipal Biosphere Reserve, India, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-890, https://doi.org/10.5194/egusphere-egu21-890, 2021.

The importance of implementing preventive fuel reduction strategies to build wildfire resilient landscapes has been increasingly present in the Portuguese politicians’ agenda. Science-based information is crucial to guide decision-makers, to better allocate resources, to decrease the projected increasing impacts of large wildfires following climate change, and to ensure the sustainability of environmental resources. Currently, fuel management is implemented without prior evaluation of wildfire exposure or optimization of strategic location of landscape treatments units, impairing a greater reduction in wildfire hazard and losses.

Prescribed burning can be used to create spatial fuel discontinuities in the landscape thus, to mitigate wildfire impacts. This study proposes to evaluate wildfire exposure in a large and diverse fire-prone area (~193 000ha) containing the Cabreira Mountain, located in Northwestern Portugal.  The main goal is to locate vegetation patches where fuel management can decrease landscape connectivity, fire spread (ROS) and fireline intensity (FLI), simultaneously reducing wildfire burn probability (BP). To address this, we run simulations using the FlamMap MTT fire spread model and quantify landscape connectivity using indexes from the graph theory, under different weather scenarios. Input data on fuels and topography were assembled in a binary landscape file at 100m spatial resolution.

Fire regime analysis was done for burned areas larger than 100 ha, from 2001 to 2019. Using the national fire ignition database and satellite data, the dates of active fire progression and fire durations are calculated. Daily weather variables (temperature, relative humidity, wind speed and direction) corresponding to those dates are compiled. To calibrate the fire model, we compare the observed and the estimated distributions of fire sizes, and the observed fire frequency with the estimated BP. A hierarchical clustering analysis identified three historical weather scenarios. Besides these a 95^th percentile extreme weather scenario is also defined.

Results show a strong relationship between wind speed and landscape connectivity. The contribution of old, burned Pine stands and shrubland areas, mainly located at the east part of the Cabreira Mountain, is high for the overall landscape connectivity. For the extreme weather scenario, assessment of the impact of different fuel treatment extents (Treatment Optimization Model), from 5 to 30%, on the landscape connectivity and on the decrease of the FLI values showed that with a 20% of fuel treatment area (~39 000ha): 1) landscape connectivity decreases 85%; 2) the proportion of the two most extreme FLI classes decreases to ~10% within the study area.

Based on the results, we discuss the best strategies to reduce wildfire hazard for multi criteria based on the studied fire regime and under different weather scenarios, providing information to support a fire management plan. This study explores the potential of fire spread models and graph theory to assess wildfire landscape connectivity and to identify the landcover patches that mostly contribute to that, to determine optimal landscape treatment proportion and to evaluate the impact of treatment locations on the decrease of wildfire properties, ultimately leading to a more comprehensive and effective wildfire management strategy.

How to cite: Sá, A. C. L., Aparicio, B. A., Bruni, C., Benali, A., Silva, F., Salis, M., and Pereira, J. M. C.: Towards supporting prescribed fire management decisions in the Cabreira Mountain, Portugal, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-13648, https://doi.org/10.5194/egusphere-egu21-13648, 2021.

The rate of spread (ROS) of wildfires is an important parameter for understanding fire-atmospheric interactions and developing fire-spread models, but it is also vital for firefighting operations to ensure the safety of firefighters (Plucinski 2017, Stow 2019). Spatial ROS observations are usually carried out by using visible and thermal satellite imagery of wildfires estimating the ROS on a time scale of hours to days for large fires (>100 ha) or repeated passing with an airborne thermal infrared imager for higher spatial and temporal resolution (Viedma et al. 2015, Stow 2014). For fire experiments in highly controlled conditions like laboratory fires or during light fuel prescribed burns, ROS estimation usually involves lag-correlation of temperature point measurements (Finney 2010, Johnston 2018). However, these methodologies are not applicable to fast-spreading grass or bush fires because of their temporal and spatial limitations. Instantaneous spatial ROS of these fires is needed to understand rapid changes in connection with the three major drivers of the fire: fuel, topography and atmospheric forcings.

We are presenting a new approach towards a spatial ROS product which includes newly developed image tracking methods based on thermal and visible imagery collected from unmanned aerial vehicles to estimate instantaneous, spatial ROS of fast spreading grass or bush fires. These techniques were developed using imagery from prescribed wheat-stubble burns carried out in Darfield, New Zealand in March 2018 (Finney 2018). Results show that both the visible and thermal tracking techniques produce similar mean ROS; however they differ in limitations and advantages. The visible-spectrum tracking method clearly identifies the flaming zone and provides accurate ROS measurements especially at the fire front. The thermal tracking technique is superior when resolving dynamics and ROS within the flaming zone because it resolves smaller scale structures within the imagery.

References:

Finney, M. et al. 2010: An Examination of Fire Spread Thresholds in Discontinuous Fuel Beds.” International Journal of Wildland Fire, 163–170.

Finney, M. et al. 2018: New Zealand prescribed fire experiments to test convective heat transfer in wildland fires. In Advances in Forest Fire Research, Imprensa da Universidade de Coimbra: Coimbra, 2018.

Johnston, J. M., et al. 2018:  Flame-Front Rate of Spread Estimates for Moderate Scale Experimental Fires are Strongly Influenced by Measurement Approach. Fire 1: 16–17

Plucinski M., et al. 2017: Improving the reliability and utility of operational bushfire behaviour predictions in Australian vegetation. Environmental Modelling & Software 91, 1-12.

Stow, D., et al. 2014: Measuring Fire Spread Rates from Repeat PassAirborne Thermal Infrared Imagery. Remote Sensing Letters 5: 803–881.

Stow, D., et al. 2019: Assessing uncertainty and demonstrating potentialfor estimating fire rate of spread at landscape scales based on time sequential airbornethermal infrared imaging, International Journal of Remote Sensing, 40:13, 4876-4897

Viedma, O., et al. 2015:  Fire Severity in a Large Fire in a Pinus Pinaster Forest Is Highly Predictable from Burning Conditions, Stand Structure, and Topography. Ecosystems18: 237–250.

How to cite: Schumacher, B., Melnik, K., Katurji, M., Clifford, V., Zhang, J., Mcnair, H., and Pearce, G.: Instantaneous spatio-temporal rate of spread of fast spreading wildfires - a new approach from visible and thermal image processing, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-6502, https://doi.org/10.5194/egusphere-egu21-6502, 2021.

Disastrous wildfires have occurred in many parts of the world during the last two years (2019 and 2020), most notably in South America, Australia, the United States, and regions north of the polar circle. Such extreme wildfire events pose a pervasive threat to human lives and property and have thus been widely recognized in the global media. This study focusses on large-scale developments in fire activity. It investigates the occurrence of burnt areas regarding several relevant parameters, namely fire extent, fire severity and fire seasonality. The entirety of those parameters allows an extensive insight regarding large-scale, long-term fire activity trends.

The burnt area derivation process, which is fully automated, is described in the literature (see reference below). The analysis is based on an extensive set of satellite data, specifically 9,612 granules of the MODIS MOD09/MYD09 product in conjunction with 3,503 tiles of the OLCI (Ocean and Land Colour Instrument) instrument onboard Sentinel-3.

The study design consists of two parts:

Firstly, the long-term temporal variability in fire activity, covering the time span from 2000 until 2020, is analyzed for the study region of New South Wales, Australia.

Secondly, the large-scale spatial variability is investigated by comparing the New South Wales extreme events in 2019/2020 with events of comparable magnitude in California, US and the Siberian taiga.

The study shows that New South Wales features an upward trend regarding the extent of yearly affected area, as well as a shift towards a prolongated end of the fire season towards the Autumn months. It also shows the exceptionality of the Australian wildfire activity in comparison with other geographical regions.

Reference:

Nolde, Michael; Plank, Simon; Riedlinger, Torsten. "An Adaptive and Extensible System for Satellite-Based, Large Scale Burnt Area Monitoring in Near-Real Time." Remote Sensing 12.13 (2020): 2162.

How to cite: Nolde, M., Mueller, N., Strunz, G., Fichtner, F., Plank, S., and Riedlinger, T.: Wildfire extreme events: Large-scale developments in fire activity of New South Wales, Australia, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-8805, https://doi.org/10.5194/egusphere-egu21-8805, 2021.

Plant biomes and climatic zones are characterized by a specific type of fire regime which can be determined from the history of fires in the area and it is a synergy mainly of the climatic conditions and the functional characteristics of the types of vegetation. They correspond also to specific phenology types, a feature that can be useful for various applications related to vegetation monitoring, especially when remote sensing methods are used. Both the assessment of fire regime from the reconstruction of fire history and the monitoring of post-fire evolution of the burned areas can be studied with satellite remote sensing based on satellite time series images. The free availability of (i) Landsat satellite imagery by US Geological Survey (USGS, (ii) Sentinel-2 satellite imagery by ESA and (iii) MODIS satellite imagery by NASA / USGS allow low-cost data acquisition and processing (eg 1984-present) which otherwise would require very high costs. The purpose of this work is to determine the fire regime as well as the patterns of post-fire evolution of burned areas in selected vegetation/climate zones for the entire planet by studying the phenology of the landscape with time series of satellite images. More specifically, the three research questions we are negotiating are: (i) the reconstruction of the history of fires in the period 1984-2017 and the determination of fire regimes with Landsat and Sentinel-2 satellite data , (ii) the assessment of pre-fire phenological pattern of vegetation and (iii) the monitoring and comparative evaluation of post-fire evolution patterns of the burned areas.

Acknowledgements

This research has been co-financed by the Operational Program "Human Resources Development, Education and Lifelong Learning" and is co-financed by the European Union (European Social Fund) and Greek national funds.

How to cite: Koutsias, N., Karamitsou, A., Nioti, F., and Coutelieris, F.: Fire regimes and post-fire evolution of burned areas in selected plant biomes of the planet studying the phenology of the landscape with time series of satellite images, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-15173, https://doi.org/10.5194/egusphere-egu21-15173, 2021.

Large wildfires are amongst the most destructive natural disasters in southern Europe, posing a serious threat to both human lives and the environment.

Although wildfire simulations and fire risk maps are already very a useful tool to assist fire managers in their decisions, the complexity of fire spread and ignition mechanisms can greatly hinder their accuracy. An important step in improving the reliability of wildfire prediction systems is to implement additional drivers of fire spread and fire risk in simulation models.

Despite their recognized importance as factors influencing fuel flammability and fire spread, soil moisture and live fuel moisture content are rarely implemented in the simulation of large wildfires due to the lack of sufficient and accurate data. Fortunately, new satellite products are giving the opportunity to assess these parameters on large areas with high temporal and spatial resolution.

The purpose of this study is twofold. First, we aimed to evaluate the capabilities of satellite data to estimate soil moisture and live fuel moisture content in different landcovers.  Secondly, we focused on the potential of these estimates for assessing fire risk and fire spread patterns of large wildfires in Portugal. Ultimately, the goal of this study is to implement these estimated variables in fire spread simulations and fire risk maps.

We compared datasets retrieved from Sentinel 1, SMAP (Soil Moisture Active Passive radiometer) and MODIS (Moderate Resolution Imaging Spectrometer) missions. Several estimators of LFMC based on spectral indices were tested and their patterns were compared with field data. Based on these estimators, we assessed the impact of LFMC and soil moisture on the extent and occurrence of large wildfires. Finally, we built a database of detailed historical wildfire progressions, which we used to evaluate the influence of soil moisture and LFMC on the velocity and direction of the fire spread.

How to cite: Briquemont, F. and Benali, A.: Soil Moisture and Live Fuel Moisture Content as key remote sensing variables to unlock improved wildfire predictions, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-15159, https://doi.org/10.5194/egusphere-egu21-15159, 2021.

The study of complex interactions between fire and atmospheric dynamics of the earth system is drawing increasing attention in recent years, especially when fire seasons are extended due to global warming, where the historical daily burnt area data played a pivotal role in analyzing wildfire regimes change. Existing products could not fully meet the temporal requirements: daily burnt area data in global fire emissions database (GFED4) starts from mid-2000 using MODIS while ESA Fire Climate Change Initiative (FireCCILT10) Dataset from 1982 to 2017 is provided on a monthly grid.

Advanced Very High Resolution Radiometer (AVHRR) series of sensors are widely used to develop pre‐MODIS daily historical records. However, compared to MODIS, the AVHRR sensor has a lower radiometric and geometric quality and is missing Short Wave Infrared (SWIR) band. To address the data quality problem, this research study presents a time-series mapping method for daily burned area using AVHRR composite. Daily fire-sensitive indices are calculated to develop a time-series data composite which is masked by the burnable surface of GLASS_GLC land cover product. Then, Continuous Change Detection and Classification (CCDC) time-series model, which originally implemented on Landsat data monitoring land cover change, is revised to detect an abrupt change in the time-series data composite and remove noise, ensuring temporal consistency. The image of a time-series breakpoint is further classified using a spatial contextual method to distinguish biomass burning from other forest degradation change like a landslide and is used to generate burned area probability map.

The methodology is verified in California, US, where fuel aridity increased during 1984–2015 driven by anthropogenic climate change. The samples are collected based on the National Monitoring Trends in Burn Severity（MTBS）Burned Areas Boundaries Dataset from 1984 – 2018 and California Department of Forestry and Fire Protection's Fire and Resource Assessment Program (FRAP) fire perimeters from since 1950. Primary results show that the proposed method can effectively detect burned area on daily basis with CCDC algorithm reducing the complexity of change detection.

How to cite: Lou, S., Liao, Y., Liu, Y., and Bai, Y.: A daily burned area mapping method using AVHRR time-series data, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-3976, https://doi.org/10.5194/egusphere-egu21-3976, 2021.

Fire severity, defined as the degree of environmental change caused by a fire, is a critical fire regime attribute of interest to fire emissions modelling and post-fire rehabilitation planning. Remotely sensed fire severity is traditionally assessed by the differenced normalized burned ratio (dNBR). This spectral index captures fire-induced reflectance changes in the near infrared (NIR) and short-wave infrared (SWIR) spectral regions. This study evaluates a spectral index based on a band combination including the NIR and mid infrared (MIR) spectral regions, the differenced normalized difference vegetation index (dNDVI_MID), to assess fire severity. This evaluation capitalized upon the unique opportunity stemming from the pre- and post-fire airborne acquisitions over the Rim (2013) and King (2014) fires in California with the MODIS/ASTER (MASTER) instrument. The field data consists of 85 Geometrically structured Composite Burn Index (GeoCBI) plots. In addition, six different index combinations, respectively three with a NIR-SWIR combination and three with a NIR-MIR combination, were evaluated based on the optimality of fire-induced spectral displacements. The optimality statistic ranges between zero and one, with values of one representing pixel displacements that are unaffected by noise. Results show that the dNBR demonstrated a stronger relationship with GeoCBI field data when field measurements over the two fire scars were combined than the dNDVI_MID approaches. The results yielded an R² of 0.68 based on a saturated growth model for the best performing dNBR index, whereas the performance of the dNDVI_MID indices was clearly lower with an R² = 0.61 for the best performing dNDVI_MID index. The dNBR also outperformed the dNDVI_MID in terms of spectral optimality across both fires. The best performing dNBR index yielded the optimality statistics of 0.56 over the Rim and 0.60 over the King fire. The best performing dNDVI_MID, index recorded optimality values of 0.49 over the Rim and 0.46 over the King fire. We also found that the dNBR approach led to considerable differences in the form of the relationship with the GeoCBI between the two fires, whereas the dNDVI_MID approach yielded comparable relationships with the GeoCBI over the two fires. This suggests that the dNDVI_MID approach, despite its slightly lower performance than the dNBR, may be a more robust method for estimating and comparing fire severity over large regions. This premise needs additional verification when more air- or spaceborne imagery with NIR and MIR bands will become available with a spatial resolution that allows ground truthing of fire severity. 

How to cite: van Gerrevink, M. and Veraverbeke, S.: Evaluating the near and mid infrared bi-spectral space for assessing fire severity and comparison with the differenced normalized burn ratio, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-2238, https://doi.org/10.5194/egusphere-egu21-2238, 2021.

In 2015, the particularly strong dry season in Indonesia, caused by an exceptional strong El Niño, led to severe peatland fires. Due to the high carbon content of peatland, these fires are characterised by high volatile organic compound (VOC) biomass burning emissions. The resulting primary and secondary pollutants are efficiently transported to the upper troposphere/lower stratosphere (UTLS) by the developing Asian monsoon anticyclone (ASMA) and the general upward transport in the intertropical convergence zone (ITCZ). In this study, we assess the importance of these VOC emissions for the composition of the lower troposphere and the UTLS by performing multiple chemistry simulations using the global atmospheric model ECHAM/MESSy (EMAC). In a first step, we find that EMAC properly captures the exceptional strength of the Indonesian fires based on the comparison of modelled columns of the biomass burning marker hydrogen cyanide (HCN) to spaceborne measurements from the Infrared Atmospheric Sounding Interferometer (IASI). In the lower troposphere, the increase in VOC levels is higher in Indonesia compared to other biomass burning regions. This directly impacts the oxidation capacity and leads to a high reduction in hydroxyl radicals (OH) and nitrogen oxides (NO_x). In general, an increase in ozone (O₃) is predicted close to the peatland fires. However, particular high concentrations of phenols lead to an O₃ depletion in eastern Indonesia. By employing the detailed in-cloud OVOC oxidation scheme Jülich Aqueous-phase Mechanism of Organic Chemistry (JAMOC), we find that the predicted changes are dampened and that by ignoring these processes, global models tend to overestimate the impact of such extreme pollution events. The upward transport in the ASMA and the ITCZ leads to elevated VOC concentrations in the UTLS region. This also results in a depletion of lower stratospheric O₃. We find that this is caused by a high destruction of O₃ by phenoxy radicals and by the increased formation of NO_x reservoir species, which dampen the chemical production of O₃.

How to cite: Rosanka, S., Franco, B., Clarisse, L., Coheur, P.-F., Wahner, A., and Taraborrelli, D.: Organic pollutants from Indonesian peatland fires: regional influences and its impact on lower the stratospheric composition, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-10203, https://doi.org/10.5194/egusphere-egu21-10203, 2021.

Although wildfires in Ireland are not extensive, information on their impacts in terms of atmospheric emissions and pollutants, and habitat losses is essential.  Current ground-based wildfire data are limited by their incompleteness, inconsistency in reporting, and a lack of timeliness. Additional data on fire alerts are drawn from international satellite derived databases such as NASA’s Fire Information for Resource Management System (FIRMS) and the European Forest Fire Information System (EFFIS) to produce a more consistent national summary. However, these databases exploit thermal anomalies derived from low spatial resolution satellite imagery, which can result in a large number of omissions of small, short-lived fires, especially when extensive cloud-cover persists, as is common in Ireland. To overcome these limitations, a new approach is proposed whereby data from the Copernicus Atmosphere Monitoring Service (CAMS) are used to identify atmospheric pollutant anomalies that may be associated with a wildfire, with Sentinel-2 pre- and post-fire imagery providing a more detailed account of the area burned and the vegetation cover affected. An inventory of fire events in Ireland reported by local and social media and the FIRMS and EFFIS databases from 2015-2020 was compiled. The average hourly concentration of selected pollutants (CO, O3, PM2.5, PM10, SO2, NOx) was derived from the CAMS European air quality analysis product at the location of each fire shortly before, during, and after the event. The average concentrations for the same period from the years excluding the year of the fire being studied were compared to the pollutant concentrations observed during the event. Preliminary results suggest that the concentration of PM2.5, PM10, SO2, and NOx show the clearest deviations from the baseline during the occurrence of a fire. Clear-sky Sentinel-2 images preceding and after selected fires were identified, and a number of different indices (NBR, dNBR, RdNBR, dMIRBI) calculated and combined to delineate burn event areas. Post-processing was undertaken to remove errors due to water, shadow and cloud cover, and eliminate features less than 0.4ha in size. Preliminary results show that burn scars can be clearly distinguished and their areas calculated, including fire events omitted from the 2015-2020 inventory. However, false alarms arise from natural land cover change, especially agricultural activity, and attempts to exclude these are being explored using the national mapping agency’s object-oriented digital mapping data model, PRIME2. Further analysis of the Sentinel-2 imagery to map the habitats burned is in progress, with a particular focus on identifying the location of gorse (Ulex europaeus), which is highly flammable in dry summer conditions due to the presence of deadwood. Atmospheric chemistry colleagues are undertaking a field campaign during 2021 to monitor the air quality during a burn event, along with laboratory measurements in a burn chamber, from which emissions factors for gorse can be calculated. Subsequently, it is hoped that detailed estimates of emissions from upland wildfires can be derived leading to improved national GHG inventories, and an assessment of these events made in terms of atmospheric impacts on population centres and environmental impacts on habitats and biodiversity.

How to cite: Cawkwell, F., Chalencon, E., Postma, T., Dwyer, N., Martin, B., and Serbin, G.: Copernicus data for wildfire mapping and monitoring in Ireland, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-8083, https://doi.org/10.5194/egusphere-egu21-8083, 2021.

Fire is an integral and predictable component of ecological functioning and dynamics in fire-prone biomes. However, the relationships and potential feedback between fire and its drivers are complex, as they depend on the temporal and spatial scales adopted when analyzing the fire regime. A remote sensing approach allows the characterization of fire regimes with larger spatial coverage and temporal homogeneity, especially where fire records are rare, as in the Brazilian savannas (Cerrado). The Cerrado is a mosaic of soil types and topographic settings, with varying regional climate patterns, resulting in a variety of fire resistant/sensitivity vegetation types, and recent disturbances, mostly due to increasing economic and agricultural development, along with changes in climate, are disrupting its natural fire patterns. Most studies characterizing fire activity in Cerrado are either performed at the biome-level or focus on very specific locations with results then extrapolated over the whole biome, which may mask important regional patterns. Here, we aim to characterize the regional fire patterns into the Cerrado’s 19 ecoregions, previously defined based on biophysical parameters which do not include fire. 

We use burned area (BA), fire radiative power and individual fire scar data based on MODIS products (respectively, MCD64A1, MCD14ML and Global Fire Atlas) to evaluate inter and intra annual cycles, spatial anomalies and trends of BA, fire intensity and fire size (small fires: <1000ha, medium: 1000-5000ha and large fires: >5000ha) in each ecoregion from 2001 to 2019. 

Our results show a marked north-south BA gradient, with higher annual BA contributions from the northern ecoregions. These ecoregions are mainly located in the latest agricultural frontier, MATOPIBA, where there are more vegetation remnants that are under high anthropogenic pressure due to recent economic development. Conversely, ecoregions showing low BA are highly fragmented and  have been historically deforested for longer periods. Most fires are of low intensity and higher intensity fires occur towards the end of the dry season period (June to October). Moreover, there are considerable differences in extremely intense events, especially in the eastern ecoregions. We also found that temporal and spatial patterns are highly variable, depending on fire scars size. Infrequent medium and large scars account for most of BA compared to common very small and small scars. Overall, fire seasonality varies substantially depending on fire size class: larger scars occur over a 2-month period within the dry season, whereas the remaining classes are increasingly scattered along the year. BA is increasing and fire intensity decreasing over MATOPIBA’s ecoregions, while in southern ecoregions, is the opposite, with a decreasing over BA and an increase of fire intensity. Smaller scars are overall decreasing, whereas medium and larger scars show positive trends over central and northern ecoregions. 

This study highlights the importance of understanding the diversity of fire dynamics in Cerrado to better inform and prepare refined-scale fire management strategies in light of current regional ecosystem disturbances and future climate change. 

The study was funded by CNPQ (grant 441971/2018-0) and P. S. Silva is supported by FCT (grant SFRH/BD/146646/2019).

How to cite: Silva, P. S., Nogueira, J., Rodrigues, J. A., Santos, F. L. M., Daldegan, G. A., Pereira, A. A., DaCamara, C. C., Pereira, J. M. C., Peres, L. F., Schmidt, I. B., and Libonati, R.: Unveiling the diversity of regional burning patterns at the Brazilian savanna, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-10008, https://doi.org/10.5194/egusphere-egu21-10008, 2021.

The central South American forest is one of the area most affected by wildfires in the world. Because of climate changes and land use management, these events are becoming more frequent and extended in the last years. For example, in 2019 Bolivia faced an extremely extensive wildfire event that had a serious ecological impact in the department of Santa Cruz. This region, called Chiquitania and characterized by a mosaic where wet tropical forests, dry tropical forests and savannas alternate, accounts for more than two-thirds of the total wildfires in the country. Despite Bolivia is between the top-ten countries with the highest expected risk in terms of annual burned forest area, the literature on wildfires here is quite limited, also because of the scarcity of available data and resources. To fill this gap, as part of the present study, we implemented an accurate dataset of burned areas, based on MODIS wildfire product, occurred in the entire Santa Cruz region in the period 2010-2019. Predisposing factors, such as topography, land use and ecoregions, were also collected in the form of digital spatial data. This information allowed assessing the susceptibility to wildfires on the entire region, with a special focus on the municipality of San Ignacio de Velasco. The analysis was performed using Random Forest (RF), an ensemble-learning algorithm based on decision trees, capable of learning from and make predictions on data by modeling the hidden relationships between a set of input and output variables. The goodness of fit was estimated by the area under the ROC (receiver operating characteristic) curve (AUC), selecting the validation dataset by using a 5-folds cross validation procedure. In addition, the last three years of observed burned areas were kept out during the medialization stage and used to test if the implemented model gives good predictions on new data. As result, we obtained a probabilistic output from RF indicating the probability for an area to burn in the future, which allowed elaborating the susceptibility maps. For San Ignacio de Velasco it resulted an AUC of 0.8, while for the entire Santa Cruz the AUC was of 0.73. Likewise, the predictive capabilities of the model gave quite good results, better at municipality that at regional level. The detailed investigation of the relative importance of each categorical class belonging to the variables ecoregions and land use reveals that “Flooded savanna” and “Shrub or herbaceous cover, flooded, fresh/saline/brakish water” are respectively the classes most related with wildfires. This important outcome confirms recent findings, that seasonally wet and dry climate, coupled with hydrologic controls on the vegetation, create in this ecoregion favorable conditions to the ignition and spreading of large wildfires during the driest period, when the biomass is abundant. The occurrence of large fires, initiated by slash-and-burn practice getting out of control, is predicted to increase in the near future and the development of new tools for fire risk assessment and reduction is thus needed. 

How to cite: Tonini, M., Bustillo Sanchez, M., Mapelli, A., and Fiorucci, P.: Susceptibility wildfire assessment in Bolivia (Santa Cruz): an approach based on Random Forest ensemble learning algorithm , EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-5653, https://doi.org/10.5194/egusphere-egu21-5653, 2021.

Wildfire disturbances severely modifies the ecosystem structure and natural regeneration processes. Predicting mid- to long-term post-fire vegetation recovery patterns, is pivotal to improve post-fire management and restoration of burned areas forest ecosystems management. Currently, many research efforts have been conducted, in order to monitor and predict wildfires, using Machine Learning and Remote sensing techniques. Instead, the method proposed in this study combines Satellite images and Data Mining algorithm to process data collected by time series and regional forest dataset to predict post-fire vegetation recovery patterns. For this reason, we analysed Normalized Burn Ratio (NBR) patterns from Landsat Time series (LTS), to assess post-fire vegetation recovery for several wildfires that occurred in three different forest Corine Land Cover classes (311, 312, 313) in the Basilicata region during the period 2005-2012. Random Forest model, was used to classify the observed recovery patterns and investigate the influence of burn severity, topographic variables, climate and spectral vegetation indices on post-fire recovery. Image acquisition and Random Forest classifier was undertaken in Google Earth Engine (GEE). Results of bootstrapping classification, across forest type, show high percentage for high recovered (HR) classes and moderate recovered (MR) classes and moderate-low percentage for low (LR) and unrecovered (UR) classes. Specifically, in the holm- and cork-oak and oak forests show medium to high recovery rates, while Mediterranean pine and conifer-oak forests show the slowest recovery rates. Different post-fire recovery patterns are related to fire severity, vegetation type and post-fire environmental conditions. Our methodology shows that post-fire recovery classification, using RF classifier provides a robust method for both local and broad scale monitoring for mid- to long-term recovery response.

Keywords: Wildfires, Post-fire recovery, Landsat Time Series (LTS), Google Earth Engine, Wildfires, Machine Learning, Random Forest.

How to cite: Maria Floriana, S., Marco, B., Angelo, R., and Angelo, N.: Predicting post-fire vegetation recovery patterns in three different forest types, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-12800, https://doi.org/10.5194/egusphere-egu21-12800, 2021.

In recent years, forest fires have become more frequent in central Europe. As the frequency and magnitude of future extreme weather events such as droughts are projected to increase, also the trend of increasing fire frequency in temperate forests is expected to continue. However, knowledge about fire behavior and spread dynamics in these forests is scarce. One of the key drivers of fire behavior is the availability of flammable vegetation, i.e. fuels. In the project ErWiN, we aim to describe the amount and distribution of fuels in different forest types in Southwestern Germany. Detailed field inventories of fuels in all vertical strata of the stands allow a first classification into different fuel types, which can be used in fire behavior simulations to obtain estimates of fire spread and intensity. In a further step, deep learning algorithms will be trained on recognizing these fuel types on GNSS located photos of forest stand situations to provide an efficient solution for mapping fuels in the field. By coupling field data with detailed remotely sensed information on forest structure obtained from airborne laserscanning, continuous fuel maps will be derived. Such fuel maps in turn allow landscape-scale analysis of fire behavior and can be useful in forest management decisions as well as in developing firefighting strategies. We thus hope to make a contribution to a better understanding of fuel-driven fire risk in central European forests and to facilitate the operational use of fire behavior models. In this contribution we present the concept developed in the ErWiN project and present first results obtained from the field survey of fuel types in Southwestern Germany.

How to cite: Labenski, P., Ewald, M., and Fassnacht, F. E.: Classifying and mapping fuels in central European forests, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-8886, https://doi.org/10.5194/egusphere-egu21-8886, 2021.

Forest fires affect Mediterranean ecosystem, often affecting protected areas. Because these normally harbour vegetation in a better conservation state and more continuous in space, it is important to determine how they burn compared to other areas. In this study we modelled fire ignition likelihood in west-central Spain as a function of biophysical and anthropogenic variables, with a special focus on natural areas that have been recently protected by the EU Natura 2000 Network. During the 2001-2015 period more than 9000 ignitions (≥1ha) were recorded in the Spanish National Forest Fire Statistics (EGIF). We characterized each ignition point with a series of biophysical (topography, radiation and land use-land cover [LULC] types) and anthropogenic (distance to highways and roads, population density, farm density, protected areas, and forest interfaces [WUI, WAI, WGI]) variables. We built and compared statistical models of fire likelihood using the MaxEnt software for three different fire sizes: ≥ 1ha (n=9089), ≥ 10ha (n=1927) and ≥ 100ha (n=292) using a 50% random test percentage in each model. Models for the likelihood of having small and medium fires (≥ 1ha and ≥ 10ha) showed the lowest performance (AUC = 0.65, AUC = 0.73). Biophysical variables barely showed importance in explaining fire activity (except for radiation). Conversely, anthropic variables like distance to roads and settlements, population density, and farm density were important predictors. Models for fires ≥ 100ha showed the best performance (AUC = 0.84). Large fire likelihood was mainly explained by biophysical variables like radiation, elevation and some LULC types (e.g., grasslands, agrarian, shrublands, and oak forests), compared to those of anthropic origin. Protected areas showed the greatest contribution to explain the ignitions of large fires. Our models highlight the different relations of biophysical and anthropogenic variables with the likelihood of fire ignitions according to their final size. 

How to cite: Arellano-del-Verbo, G., R. Urbieta, I., and Moreno, J. M.: Human and biophysical influence on fire ignition likelihood in protected areas as a function of fire size in west-central Spain, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-10893, https://doi.org/10.5194/egusphere-egu21-10893, 2021.

Weather and climate play an important role in shaping global fire regimes and geographical distributions of burnable areas. At the global scale, fire danger is likely to increase in the near future due to warmer temperatures and changes in precipitation patterns, as projected by the Fifth Assessment Report (AR5) of the Intergovernmental Panel on Climate Change (IPCC). There is a need to develop the most reliable projections of future climate-driven fire danger to enable decision makers and forest managers to take both targeted proactive actions and to respond to future fire events.

Climate change projections generated by Earth System Models (ESMs) provide the most important basis for understanding past, present and future changes in the climate system and its impacts. ESMs are, however, subject to systematic errors and biases, which are not fully taken into account when developing risk scenarios for wild fire activity. Projections of climate-driven fire danger have often been limited to the use of single models or the mean of multi-model ensembles, and compared to a single set of observational data (e.g. one index derived from one reanalysis).

Here, a comprehensive global evaluation of the representation of a series of fire weather indicators in the latest generation of ESMs is presented. Seven fire weather indices from the Canadian Forest Fire Weather Index System were generated using daily fields realisations simulated by 25 ESMs from the 6^th Coupled Model Intercomparison Project (CMIP6). With reference to observational and reanalysis datasets, we quantify the capacity of each model to realistically simulate the variability, magnitude and spatial extent of fire danger. The highest-performing models are identified and, subsequently, the limitations of combining models based on independency and equal performance when generating fire danger projections are discussed. To conclude, recommendations are given for the development of user- and policy-driven model evaluation at spatial scales relevant for decision-making and forest management.

How to cite: Gallo Granizo, C., Eden, J., Dieppois, B., and Blackett, M.: Assessing the capacity of Earth System Models to simulate spatiotemporal variability in fire weather indicators, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-12208, https://doi.org/10.5194/egusphere-egu21-12208, 2021.

The Canadian Forest Fire Danger Rating System (CFFDRS) is used to assess and predict the fire behavior in various forest ecosystems all over the world. The Fire Weather Index (FWI) module of the CFFDRS models the relationship between meteorology and forest fires. It was observed in our earlier study that the values of the FWI and its related parameters were considerably different from the other countries that use the model for their operational fire weather simulation. In this study we evaluate the model performance over Indian climate for a period of 10 years 1996-2005 under various weather scenarios. The daily meteorological data from ECMWF’s ERA5 reanalysis has been used as inputs to the fire model and the active fire data from MODIS Terra and Aqua satellites over the study period has been used to evaluate the capability of model to simulate fire danger. As India has many different climatic zones, we evaluated the behavior fire model parameters over 5 forest zones namely Himalayan, Deciduous, Western Ghats, Thorn forests and North Eastern forests based on the Roy et al. 2016 Land Use Land Cover data and Koppen climatic zones.  The analysis was narrowed down over only the forest areas of the zones so as to remove any chances of including the non-forest fires detected by the satellite. Results show that the FWI shows a strong correlation with forest fires if the model is correctly spun up and appropriately calibrated. A spin up time of minimum 60 days was found to be appropriate for stabilization of FWI components like Duff Moisture Code (DMC) and Drought Code (DC). Sensitivity studies showed that temperature and relative humidity are the key controlling factors of forest fires over India and that the parameters depict high interannual seasonality due to relatively lower values during the Indian monsoon season.

This study is one of the first attempts to use fire models to simulate fire behavior over India. It can serve as a launchpad for further work on fire hazard prediction and effects of climate change on fire hazard in India.

How to cite: Barik, A. and Baidya Roy, S.: Sensitivity of the CFFDRS Fire Weather Index parameters for Indian weather conditions, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-14231, https://doi.org/10.5194/egusphere-egu21-14231, 2021.

Fire strongly depends on the weather, especially in Mediterranean climate regions with rainy winters but dry and hot summers, as in Portugal. Fire weather indices are commonly used to assess the current and/or cumulative effect of weather conditions on fuel moisture and fire behaviour. The Daily Severity Rating (DSR) is a numeric rating of the difficulty of controlling fires, based on the Canadian Fire Weather Index (FWI), developed to accurately assess the expected efforts required for fire suppression. Recently, the 90th percentile of DSR (90pDSR) was identified as a good indicator of extreme fire weather and well related to the burnt area in some regions of the Iberian Peninsula. The purposes of this work were: 1) to verify if this threshold is adequate for all continental Portugal; 2) to identify and characterize local variations of this threshold, at a higher spatial resolution; and, 3) to analyse other variables that can explain this spatial heterogeneity.

We used fire data from the Portuguese Institute for the Conservation of Nature and Forests and weather data from ERA5, for the 2001 – 2019 study period. We also used the Land Use and Occupation Charter for 2018 (COS2018), provided by the Directorate-General for Territory, to assess land use and cover in Portugal. The meteorological variables to compute the DSR are air temperature, relative humidity, wind speed and daily accumulated precipitation, at 12 UTC. DSR percentiles (DSRp) were computed for summer period (between 15^th May and 31^st October) and combined with large (>100 ha) burnt areas (BA), with the purpose to identify which DSRp value is responsible of a large amount of BA (80 or 90%). Cluster analysis was performed using the relation between DSRp and BA, in each municipality of Continental Portugal.

Results reveal that the 90pDSR is an adequate threshold for the entire territory. However, at the municipalities’ level, some important differences appear between DSRp thresholds that explain 90 and 80% of the total BA. Cluster analysis shows that these differences justified the existence of several statistically significant clusters. Generally, municipalities where large fires take place in high or very high DSRp are located in north and central coastal areas, Serra da Estrela, Serra de Montejunto and Algarve. In contrast, clusters where large fires where registered with low DSRp appear in northern and central hinterland. COS2018 data was assessed to analyse if and how the vegetation cover type influences the clusters’ distribution and affects the relationship between DSRp and total BA. Preliminary results expose a possible vegetation influence, especially between forests and shrublands.

How to cite: Calheiros, T., Benali, A., Silva, J. N., Pereira, M., and Nunes, J. P.: Fire weather thresholds and burnt area in Portugal, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-14916, https://doi.org/10.5194/egusphere-egu21-14916, 2021.

Weather and climate extreme events contribute to the increase of wildfire risk. A recent study carried out in Mainland Portugal for the period 1981 – 2017, using Standardized Precipitation Index (SPI) and Standardized Precipitation Evapotranspiration Index (SPEI) to assess drought conditions, revealed that drought affects 70% of the months and a very strong relationship between the occurrence of drought and the spatio-temporal distribution of extreme wildfires (> 5,000 ha). These results raised additional scientific questions that need to be answered, such as: Is the relationship between droughts and fires equally strong for wildfires of smaller size? The study was carried out at the country level, but what are the regions where the relationship is more and less strong? Therefore the objective of this study is to assess the influence of drought on fire incidence, considering all wildfires or classes of wildfire sizes and in each of the 278 counties of Continental Portugal characterized by different features (landscape, weather/climate, drought and fire incidence). This study benefits from the existence of long and reliable meteorological and wildfire datasets. The methodology comprises cluster analysis, contingency tables, accuracy metrics, statistical measures of association to test the independence and help find interactions between these two natural hazards. Main results include: (i) the characterization of spatio-temporal distribution of drought number, duration, severity, intensity, extension; (ii) wildfire space-time distribution within drought periods and affected area; and, (iii) the assessment of the relationship between droughts and wildfires at county scale. The authors believe that the findings of this study are very useful for the definition of adaptation and mitigation strategies for the impacts of droughts in wildfire occurrence and to assess the climatic wildfire hazard/risk.

Acknowledgements

This work was supported and conducted in the framework of the FEMME project (PCIF/MPG/0019/2017) funded by FCT - Portuguese Foundation for Science and Technology. The study was also supported by: i) National Funds by FCT - Portuguese Foundation for Science and Technology, under the project UIDB/04033/2020; and, ii) National Funds by FCT - Portuguese Foundation for Science and Technology, under the project UID/AMB/50017/2019. Data were provided by the European Forest Fire Information System – EFFIS (http://effis.jrc.ec.europa.eu) of the European Commission Joint Research Centre.

How to cite: Pereira, M. and Parente, J.: Drought and wildfires in Portugal at the county scale, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-15135, https://doi.org/10.5194/egusphere-egu21-15135, 2021.

The climate in the Boreal area is warming at a pace that is exceeding the global average. Both temperature and precipitation is projected to increase due to climate change. The gross primary production in the forested area is also projected to increase, as well as the soil respiration. The burned area is sensitive to the meteorological forcing and the risk of ignition depends on the amount and properties of the litter. Overall climate change has a potential to increase the fire risk in the Boreal forests.

The effects of projected climate change on forest fires in Fennoscandia, and in parts of Russia adjacent to Finland, were simulated with the JSBACH-SPITFIRE. JSBACH is the land model in the Earth system models of the Max-Planck Institute for Meteorology. SPITFIRE is a mechanistic fire model, driven by meteorology, vegetation cover, fuel load and fuel properties. The model simulates fire risk, number of fires and burned area fraction. SPITFIRE uses ignition rates and distinguishes between ignition events caused by lightning and humans. Ignition events result in fire only when enough fuel is present, and the fuel is sufficiently dry. The JSBACH-SPITFIRE model was driven by downscaled and bias corrected meteorological data from the EURO-CORDEX initiative, for the period from 1951 to 2100. The model domain was the land area within 55-71°N and 5-38°E. A subset of the EUR-44 domain was regridded to 0.5° resolution for our model domain. The global driving models used for producing the EURO-CORDEX data used here were CanESM2, CNRM-CM5, MIROC5. We selected driver models that represent mid-range regarding the projected change in temperature and precipitation for Finland under RCP4.5 and RCP8.5. We used daily bias corrected data of precipitation and temperature from 1951 to 2100 for both RCP4.5 and RCP8.5 climate change projections. In addition, daily data of relative humidity, wind speed, longwave and shortwave radiation were used for the historical (1951-2005) and scenario period (2006-2100).

Preliminary results indicate that the increase in temperature, which affects the drying rate of the fuel, is the major factor for driving the changes in forest fires in the simulations.

How to cite: Backman, L., Aalto, T., Aalto, J., Markkanen, T., Thölix, L., and Lasslop, G.: Effect of climate change on wildfires in Fennoscandia, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-16030, https://doi.org/10.5194/egusphere-egu21-16030, 2021.