Wildfires have shaped ecosystems for millions of years, with plant functional traits playing a key role in fire behaviour and severity. Morphological and physiological traits, particularly at the leaf and shoot levels, influence flammability by determining fuel composition and structure within both canopy and litter layers. These traits are hypothesized not only to affect critical fire dynamics, such as the likelihood of surface fires transitioning into crown fires, with significant consequences for fire intensity and ecosystem impacts, but also influence the evolution of fire-related traits.
This study investigates how leaf- and shoot-level morphology influences flammability in canopy and litter contexts across six dominant conifer families: Araucariaceae, Cupressaceae, Pinaceae, Podocarpaceae, Taxaceae, and Taxodiaceae. Flammability properties were assessed using fire calorimetry to measure ignitability, flame spread, and variability in the rate of energy release from combustion. Results indicate that while shed plant parts (e.g., leaves and shoots) shape fire behaviour by influencing bulk density, aeration, and flame spread rate—ultimately affecting burn sustainability and total energy release—shoot-level traits in isolation, including leaf shape and the arrangement of leaves within shoots, do not consistently predict flammability in canopy material.
Our findings highlight the dynamic interplay between plant morphology, fire regimes, and evolutionary pressures. Traits such as leaf size, shape, and arrangement contribute to fuel structure, driving patterns of fire behaviour that influence long-term plant fitness and survival. This underscores the importance of reconciling fire behaviour, plant functional traits, and the evolutionary history of fire adaptations across phylogenies.
With global change drivers intensifying fire regimes, understanding the relationship between plant flammability, fire regimes, and the acquisition of fire-related traits is increasingly critical. Non-fire-adapted species may face heightened extinction risks, threatening ecosystem stability. Quantifying the intrinsic flammability of plant traits is therefore essential for informing fire management, guiding conservation strategies, and ensuring the long-term sustainability of vegetative communities in a changing climate.
How to cite: Koll, R. and Belcher, C.: Morphological drivers of flammability in canopy and litter contexts across conifer families, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16840, https://doi.org/10.5194/egusphere-egu25-16840, 2025.
Forest fires are the primary disturbance agent in global boreal forests, and they play a significant role in shaping their composition and structure. Boreal forests are also considered a carbon sink but rising temperatures in high-latitude regions are likely increasing wildfire activity, raising concerns that they may become net carbon emitters. Climate change has also increased the frequency and intensity of fire weather in high-latitude boreal forests and is expected to increase the frequency of lightning, a major source of ignition, which could potentially lead to a substantial increase in burned areas. Lightning-ignited wildfires (LIW) pose unique challenges due to their ability to (i) smoulder for long periods of time undetected, (ii) form fire clusters, and (iii) resist suppression efforts. Understanding drivers of ignition is critical for ignition prediction and for optimizing resource allocation for fire managers. Understanding the dynamics of LIWs is, however, challenging due to lack of spatially explicit data that would allow for pan-Boreal analyses of ignition drivers.
Current LIW research is thus heavily concentrated in regions with detailed fire data (like North America). In a past study, we filled this data gap by introducing the Temporal Minimum Distance (TMin) method, a new approach to match lightning strikes to wildfires without ignition location data (Engle et al. 2024). The TMin method outperformed current methodologies like the Daily Minimum Distance and the Maximum Index A by identifying 74.71% of fires in boreal forests. Using this method, a comprehensive dataset - BoLtFire - was developed, encompassing 6,228 fires larger than 200 ha spanning across the entire boreal forest from 2012 to 2022. When benchmarked to agency reference datasets, BoLtFire performed reasonably well, with an overall commission error of 30.06% and omission error of 53.63%, but global extent.
To model lighting ignition efficiency, the BoLtFire dataset was enhanced to include location data for over 6,000 lightning strikes that did not result in a fire. This expanded dataset also now integrates “ignition drivers,” identified through modelling over 80 different lightning characteristic, climatic, topographic, and fuel-related variables to identify the most influential factors in the ignition process. This enriched dataset provides valuable insights into why certain lightning events trigger wildfires, while others do not. It thus enables more accurate ignition prediction and improved wildfire management strategies. This expanded dataset provides new opportunities to model ignition and spread dynamics for wildfires in boreal forests, deepening our understanding of lightning-driven fire activity. By addressing key knowledge gaps and advancing methodological approaches, this research contributes to a more comprehensive framework for mitigating the growing risks of wildfires in boreal regions and their potential impacts on one of the most important land carbon sinks.
References:
Engle, B., Bratoev, I., Crowley, M. A., Zhu, Y., and Senf, C.: Distribution and Characteristics of Lightning-Ignited Wildfires in Boreal Forests – the BoLtFire database, Earth Syst. Sci. Data Discuss. [preprint], https://doi.org/10.5194/essd-2024-465, in review, 2024.
How to cite: Engle, B., Bratoev, I., Crowley, M. A., Zhu, Y., and Senf, C.: Identifying Ignition Drivers of Lightning-Ignited Wildfires in Boreal Forests, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-3090, https://doi.org/10.5194/egusphere-egu25-3090, 2025.
Lightning-induced ignitions play a major role shaping the frequency, patterns and characteristics of wildfires in several regions across the globe, including extreme wildfire events (e.g., Góis wildfire in 2017 in Portugal) and fire seasons, such as 2019-20 in Australia, 2020 in California, and 2023 in Canada. The attention to lightning-ignited wildfires has been growing in recent years. Studies on LIWs frequently associate lightning and wildfire data to discern or approximate the place and moment of fire ignition. This typically requires to select the lightning strike responsible for the ignition.
Currently, several methods are applied to select the most likely lightning strike causing the ignition. However, this selection is complicated by, at least, two aspects. First, the spatial uncertainty of fire and lightning data (e.g., the location errors of detected lightning events). Second, the holdover phenomenon. Holdover time, commonly defined as the time between lightning-induced fire ignition and fire detection, can range from a few minutes to several days, and more rarely to some weeks or even months. Long holdover times are associated to the presence of a smoldering phase that hinders the detection of these lightning fires.
Here, we present a novel method that uses location accuracy information from lightning location networks, as well as expected distributions of holdover time, to assess the probabilities of lightning igniting wildfires. Our method computes a probability metric, which is the product of two independent probabilities: a spatial and a temporal probability. The spatial component assesses the probability of a cloud-to-ground lightning event striking within a given area surrounding the fire discovery point, while the temporal component evaluates the probability of a lightning-ignited wildfire undergoing a certain holdover time. The lightning event with the maximum probability metric value is then selected as the most likely ignition source. We applied this method in three study areas: Switzerland, Catalonia (Spain), and California and Nevada (USA). The results were compared with lightning selections identified by the index of proximity, one of the currently most common methods to select the most likely ignition source of lightning-induced wildfires.
The initial results indicate that the probability metric yields a different selection of lightning events, in comparison with the index of proximity, for a great proportion of wildfires, with considerable differences across the study areas. We suggest that the probability metric provides a solid alternative to current methods. The probability metric offers some advantages: (1) it simplifies some methodological decisions despite the need for additional computations; (2) it is flexible and can be adapted to different types of lightning and fire data (e.g., fire perimeters); (3) it has a more robust theoretical basis than current methods; and (4) the lightning selection can be enhanced over time due to continuous improvements in lightning and fire databases.
How to cite: Moris, J. V., Hunt, H. G. P., Álvarez-Álvarez, P., Conedera, M., Gordillo-Vázquez, F. J., Lapierre, J., Pérez-Invernón, F. J., Pineda, N., Pezzatti, G. B., Veraverbeke, S., and Ascoli, D.: A new probabilistic method to identify fire-igniting lightning events, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-3335, https://doi.org/10.5194/egusphere-egu25-3335, 2025.
Synchronous extreme fire weather significantly heightens wildfire ignition and spread risk, potentially overwhelming firefighting efforts. Despite evidence of increasing fire weather extremes in a warming climate, the spatial-temporal synchronicity of these conditions remains understudied outside North America. This research investigates historical and projected changes in the synchronicity of extreme fire weather in Europe, employing the Fire Weather Index (FWI) from 1981–2022 and climate scenarios representing temperature increases (1°C to 6°C) and precipitation changes (-40% to +60%).
Our findings reveal Central Europe as a significant hotspot, with synchronicity increases up to 389%, and the Mediterranean region experiencing a 66% rise. Synchronicity trends are driven by rising temperatures and shifting atmospheric circulation patterns, particularly in summer and autumn. Future projections suggest compounded fire risks across broader regions, requiring enhanced transnational coordination. This study emphasizes the growing need for proactive fire management strategies tailored to increasing synchronicity, including shared resource mechanisms like RescEU, and highlights the value of integrating synchronicity assessments into regional climate adaptation planning. This abstract is based on findings from a study accepted for publication in Environmental Research Letters.
Acknowledgements
A.G. thanks to the Ministerio de Ciencia, Innovación y Universidades of Spain for Ph.D. contract FPU19/06536. A.G., M.A.T-V., and M.T. acknowledge the support of the ONFIRE project, grant PID2021-123193OB-I00, funded by MCIN/AEI/10.13039/501100011033 and by “ERDF A way of making Europe”. M.T. acknowledges funding by the Spanish Ministry of Science, Innovation, and Universities through the Ramón y Cajal Grant Reference RYC2019-027115-I. This work was supported by the project ‘Climate and Wildfire Interface Study for Europe (CHASE)’ under the 6th Seed Funding Call by the European University for Well-Being (EUniWell).
How to cite: Gincheva, A., Torres-Vázquez, M. Á., Di Giuseppe, F., Moreno Torreira, A., Jerez, S., and Turco, M.: Rising Synchronicity of Extreme Fire Weather Across Europe in a Warming Climate , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-4352, https://doi.org/10.5194/egusphere-egu25-4352, 2025.
Rural fires are recurrent in Southern Europe due to climate conditions, land use change, or a combination of both. Wet and mild winters and dry and warm summers favour the growth of vegetation and its subsequent low moisture content, increasing fuel availability. In Portugal, between 15 and 20 September 2024, severe wildfires burned more than 145,000 hectares and caused the death of more than 9 people. In Greece a major fire, stated as the largest recorded in the EU, started near the city of Alexandroupolis on August 21, with around 80.000 hectares burnt, mainly affecting the Dadia Forest and causing the death of almost two tens of migrants. Despite the crucial role played by dry fuel conditions fostering the propagation of wildfires, favourable meteorological conditions and fuel accumulation are related to the recorded fire activity and burned area. The influence of spring meteorological conditions on fire season burned area through their effect on fuel accumulation and dryness is assessed. The link between hot temperature and water availability in spring and the increased risk of summer flammability and fire spread through their influence on vegetation gross productivity is evaluated using satellite-derived data. The important role of fuel accumulation during the early growing season in fire-prone regions is highlighted in the case of Portugal in 2024 and Greece in 2023 and reinforces the crucial importance of fuel management for the definition of effective fire prevention measures in the context of warmer and drier conditions forecasted for southern European Countries.
Acknowledgements: This work is supported by the Portuguese Fundação para a Ciência e Tecnologia, FCT, I.P./MCTES through national funds (PIDDAC): UID/50019/2025 and LA/P/0068/2020 https://doi.org/10.54499/LA/P/0068/2020 and also on behalf of DHEFEUS -2022.09185.PTDC and the project FAIR- 2022.01660.PTDC).
How to cite: Gouveia, C., Finuras, M., Russo, A., and Ermitão, T.: The role of dry and heat extremes on vegetation dynamics in the recent fire seasons in Southern Europe, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-19493, https://doi.org/10.5194/egusphere-egu25-19493, 2025.
Fires have great ecological, social, and economic impact. However, fire prediction and management remain challenges due to a limited understanding of their roles in the Earth system. Fires over southern Mexico and Central America (SMCA) are a good example of this, greatly impacting local air quality and regional climate. Here we report that the spring peak (April–May) of fire activities in this region has a distinct quasi-biennial signal based on multiple satellite datasets measuring different fire characteristics. The variability is initially driven by quasi-biennial variations in precipitation. Composite analysis indicates that strong fire years correspond to suppressed ascending motion and weakened precipitation over the SMCA. The anomalous precipitation over the SMCA is further found to be mostly related to the East Pacific–North Pacific (EP-NP) pattern 2 months prior to the fire season. The positive phase of the EP-NP leads to enhanced precipitation over the eastern US but suppressed precipitation over the SMCA, similar to the spatial pattern of precipitation differences between strong and weak fire years. Meanwhile, the quasi-biennial signals in precipitation and fires appear to be amplified by their interactions through a positive feedback loop at short timescales. Model simulations show that in strong fire years, more aerosol particles are released and transported downstream over the Gulf of Mexico and the eastern US, where suspended light-absorbing aerosols warm the atmosphere and cause the ascending motion of the air aloft. Subsequently, a compensating downward motion is formed over the region of the fire source and ultimately suppresses precipitation and intensifies fires. Statistical analysis shows the different durations of the two-way interaction, where the fire suppression effect of precipitation lasts for more than 20 d, while fire leads to a decrease in precipitation at shorter timescales (3–5 d). This study demonstrates the importance of fire–climate interactions in shaping the fire activities on an interannual scale and highlights how precipitation–fire interactions at short timescales contribute to the interannual variability in both fire and precipitation.
How to cite: Liu, Y., Qian, Y., and Wang, M.: Fire–precipitation interactions amplify the quasi-biennial variability in fires over southern Mexico and Central America , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-5454, https://doi.org/10.5194/egusphere-egu25-5454, 2025.
Synchronous fires, that is fires co-occurring at different geographical locations within a few days of each other, challenge the distribution of firefighting resources among regions and can have more severe impacts on human health, infrastructure and environmental systems than individual fire events. However, so far very little is known about the occurrence, spatial patterns and the atmospheric drivers of synchronous fires in Europe.
In this work, we use fire observations from a global fire event dataset FRYv2.0 to (1) detect fire synchronicity between ten European regions during 2001–2020 and (2) link the occurrence of synchronous fires to seven dominant European-Atlantic weather regimes. To detect fire synchronicity, we apply complex network theory and an event synchronicity statistical framework to identify significant links between the ten regions. To analyze the relationship between synchronous fire events and dominant weather regimes, we use a conditional probability-based measure calculating the dependency of synchronous fires –between each region pair– on seven common European weather regimes. We perform 2000 block permutations to test the statistical significance of these dependencies. Lastly, we use the CERRA reanalysis data to analyze the seasonal anomalies of relevant atmospheric variables under each weather regime, including temperature, wind speed, precipitation and relative humidity.
We find multiple significant connections between regions across Europe showing fire synchronicity in spring, summer and fall. We show that (1) northern and western regions in Europe experience fire synchronicity in spring under the influence of blocking regimes (i.e., European and Scandinavian Blocking) which promote warm and dry conditions; (2) eastern regions show fire synchronicity in spring and fall during the Zonal Regime under warm and dry conditions; and (3) fire synchronicity in southern regions are significantly modulated by Scandinavian Troughs due to positive wind speed anomalies and dry conditions in spring and fall as well as by Atlantic Ridges due to positive wind speed anomalies in summer.
Our work reveals significant fire synchronicity across Europe with significant links to atmospheric circulation patterns. As the seven weather regimes have predictability on weekly to monthly time scales, our work might help to develop early warning systems for elevated risks of synchronous fires under climate change and improve fire emergency preparedness across different European regions.
How to cite: Li, X., Wood, R., and Brunner, M.: Linking fire synchronicity in Europe to persistent weather regimes, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-3285, https://doi.org/10.5194/egusphere-egu25-3285, 2025.
Wildfires pose a significant threat to ecosystems, human life, and infrastructure, particularly in South America, where diverse climatic and environmental factors contribute to their occurrence. Climate change has exacerbated extreme weather conditions such as intense heat and drought, leading to a global increase in the frequency and intensity of wildfires. Countries like Brazil have experienced significant rises in wildfire damage, highlighting the urgent need for predictive models that accurately assess future wildfire risks to mitigate their impact effectively. This thesis addresses this need by developing a wildfire risk prediction system leveraging deep learning methods and remote sensing data.
Using Earth Observation (EO) APIs, the system avoids downloading and storing vast amounts of satellite imagery, enabling efficient data acquisition and preprocessing. The study focuses on key variables that influence wildfire activity, including dynamic variables such as Land Surface Temperature (LST), Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), radiation, Leaf Area Index (LAI), evapotranspiration (ET), wind speed, and temperature, as well as static variables like land cover, Digital Elevation Model (DEM), and population density. The system is designed to predict wildfire risk for the next day and up to eight days, offering a robust tool for proactive wildfire management.
Given the stochastic and nonlinear nature of wildfire phenomena, this research employs advanced deep learning techniques, including Random Forests (RF), Long Short-Term Memory networks (LSTM), and Convolutional LSTM (ConvLSTM) models, to predict wildfire risk in near real-time. Active fire data from MODIS products, along with their burn dates, serve as the basis for training datasets. Non-fire points are generated by mapping the land cover distribution of fire points, ensuring balanced datasets for model training. Variables are extracted and classified into dynamic and static categories to capture both temporal variability and fixed geographical characteristics.
The objectives of this research are threefold: (1) to investigate existing remote sensing-based wildfire management methodologies and identify enhancements through the integration of data cubes and deep learning; (2) to develop a scalable platform for efficient data acquisition, preprocessing, and risk prediction using deep learning algorithms; and (3) to evaluate the system’s accuracy, efficiency, and scalability with real-world datasets and disaster scenarios.
Preliminary results highlight the effectiveness of integrating remote sensing data with deep learning models for wildfire risk prediction. Dynamic variables such as EVI, LST, and NDVI, along with human influence factors like Global Human Modification Index (gHM), emerged as key predictors, demonstrating the interplay of environmental and anthropogenic drivers in wildfire occurrences. Seasonal analysis from 2021 to 2024 revealed a strong correlation between fire activity, elevated temperatures, and declining vegetation indices from November to April. The Random Forest model achieved 83% accuracy, while the LSTM model showed promise with 75% accuracy, emphasizing the potential of both static and temporal data. These findings lay a robust foundation for enhancing wildfire risk management through advanced machine-learning approaches.
How to cite: Sanchez Vargas, J. A., Heisig, J., Painho, M., and Gharun, M.: Development of a Wildfire Risk Prediction System based on Deep Learning Methods and Remote Sensing, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-7107, https://doi.org/10.5194/egusphere-egu25-7107, 2025.
Recent advancements in machine learning (ML) have significantly broadened its applications, including the potential to transition from forecasting fire weather to predicting actual fire activity. In this study, we demonstrate the feasibility of this transition using an operational forecasting system. By integrating data on human and natural ignitions along with observed fire activity, data-driven models effectively address the persistent overprediction of fire danger in fuel-limited biomes. This results in fewer false alarms and more informative outputs compared to traditional methods.
A key factor driving this improvement is the availability of global datasets for fuel dynamics and fire detection, which were not accessible during the development of earlier physics-based models. We find that the enhanced predictive skill of ML models stems largely from the comprehensive characterization of fire processes provided by these datasets, rather than from the complexity of the ML methods themselves.
As enthusiasm gather around data-driven approaches, our findings highlight the critical importance of high-quality training data in improving forecast accuracy. While the rapid advancement of ML techniques generates excitement, there is a risk of undervaluing the essential role of data acquisition and, where necessary, its creation through physical modeling. Our results underscore that investing in robust datasets is indispensable and should not be overlooked in the pursuit of very complex algorithm.
How to cite: Di Giuseppe, F., Mc Norton, J., Wetterhall, F., and Lombardi, A.: The real drivers of the ML revolution in fire forecasting, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-8637, https://doi.org/10.5194/egusphere-egu25-8637, 2025.
Machine learning (ML) models are widely used to predict wildfire occurrence and susceptibility (Brys et al., 2025). However, while these models excel at prediction, they often fail to provide insights into their inner workings or uncover the causal pathways driving wildfires. This study addresses this limitation by extending ML models beyond prediction to explore the drivers and causal pathways underlying wildfire occurrence. Our primary aim is to identify meaningful, interpretable patterns from wildfire data.
We developed a novel multi-stage clustering methodology inspired by Cooper et al. (2021) and Cohen et al. (2024). This approach integrates feature attribution (SHAP values), dimensionality reduction (UMAP), hierarchical clustering (HDBSCAN), and causal discovery methods: PC and FCI (Spirtes et al., 2001), and DirectLiNGAM (Shimizu et al., 2011). The causal methods were enhanced with prior background knowledge to derive meaningful insights. We used datasets from Italy (Cilli et al., 2022) and the Netherlands.
A central feature of our methodology is the use of SHAP values to define subgroups and derive causal pathways. SHAP values reduce noise in the feature space while preserving critical information for clustering. By reducing multidimensional SHAP values to two dimensions with UMAP, we improved clustering performance and interpretability. The resulting clusters were described using concise, non-overlapping decision rules based on the original variables, eliminating the need for manual filtering commonly required in clustering raw feature space. The identified clusters revealed specific relationships between wildfire drivers and occurrence. For each cluster, we applied advanced causal discovery techniques to derive probable causal pathways, aligning the findings with the knowledge of stakeholders and domain experts. These actionable and interpretable explanations offer practical utility.
Findings from the case studies demonstrate that supervised clustering effectively characterizes wildfire occurrence by linking it to influencing factors. Furthermore, the approach provides valuable insights into cluster-specific causal pathways. The methodology translates complex relationships into simple causal logic, offering stakeholders and domain experts the necessary context to understand the model's behavior.
Brys, C., La Red Martínez, D.L. & Marinelli, M. Machine learning methods for wildfire risk assessment. Earth Science Informatics 18, 148 (2025). https://doi.org/10.1007/s12145-024-01690-z
Cilli, R., Elia, M., D’Este, M., Giannico, V., Amoroso, N., Sanesi, G., Lombardi, A., Pantaleo, E., Monaco, A., Tangaro, S., Bellotti, R. & Lafortezza, R. (2022). Explainable artificial intelligence (XAI) detects wildfire occurrence in the Mediterranean countries of Southern Europe. Scientific Reports 12, 16349. https://doi.org/10.1038/s41598-022-20347-9
Cohen, J., Huan, X. & Ni, J. (2024). Shapley-based explainable AI for clustering applications in fault diagnosis and prognosis. Journal of Intelligent Manufacturing, 35, 4071-4086. https://doi.org/10.1007/s10845-024-02468-2
Cooper, A., Doyle, O. & Bourke, A. (2021). Supervised clustering for subgroup discovery: An application to covid-19 symptomatology. Communications in Computer and Information Science, 1525, 408–422. https://doi.org/10.1007/978-3-030-93733-1_29
Shimizu, S., Inazumi, T., Sogawa, Y., Hyvärinen, A., Kawahara, Y., Washio, T., Hoyer, P. O., & Bollen, K. (2011). DirectLiNGAM: A direct method for learning a linear non-Gaussian structural equation model. The Journal of Machine Learning Research, 12, 1225–1248. https://doi.org/10.48550/arXiv.1101.2489
Spirtes, P., Glymour, C. & Scheines, R. (2001). Causation, Prediction, and Search. Second Edition. MIT Press. https://doi.org/10.7551/mitpress/1754.001.0001
How to cite: Korving, H. and Van Marle, M.: Decoding Wildfires - Extracting Interpretations and Causal Pathways of Catalysts for Wildfire Occurrence from Machine Learning Models, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-12889, https://doi.org/10.5194/egusphere-egu25-12889, 2025.
In this study, we report the preliminary findings from a series of sans-human wildfire simulations using a process based dynamic global to regional vegetation model (DGVM), LPJ-GUESS v 4.1, coupled with the SIMple FIRE Model (SIMFIRE) and the wildfire combustion model (BLAZE), where we investigate the performance of the DGVM to reenact a specific wildfire instance in a Mediterranean catchment. For this, we compared the simulated burned area (BAs) to that in the actual event (BAo) in Manavgat, Antalya, Türkiye. The DGVM spatially captured the fire instance, albeit with a much smaller BA as a result. In July 2021, the largest single wildfire incidence for this region for the last two centuries occurred. The wildfire scorched an area of 60.000 ha.s where the dominant vegetation types were fire adapted dry conifer forests (mainly Pinus brutia) and Mediterranean shrubs. Previous years’ precipitation patterns had encouraged fuel build up, and the extreme heat of the summer of 2021, coupled with the seasonal drought and strong winds provided suitable environmental conditions for the wildfire’s spread. The ignitions in this specific case were intentional, majority were targeted arsons, and a plausible reason behind the ultimate extent of the BA. Here, we show the simulation results from our sans-human model runs using ERA5-Land reanalysis dataset, and compare BAs to BAo for this catchment for 2021. Our ultimate aim in these series of experiments where the ignition source is non-human is initially to decipher the dynamics, and later to develop a methodology to assess the human influence in BA in Mediterranean type ecosystems in the Eastern Mediterranean Basin, under a changing climate.
How to cite: Ekberzade, B., Avcıoğlu, A., and Görüm, T.: Determining the human signal in burned area under a changing climate, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-2313, https://doi.org/10.5194/egusphere-egu25-2313, 2025.
The temporal coverage from ˜2000 to present of global burned area satellite observations limits many aspects of fire research. As a result, global fire models are often being used to investigate past and future fire behaviour. Unfortunately, the limited temporal coverage of the observations also hinders the development and evaluation of these fire models. The current generation of global fire models are capable of simulating some characteristics of regional fire behaviour, such as mean state and seasonality, well. However, the performance of these models differs greatly from region to region, and aspects such as extreme fire behaviour are not well represented yet.
Here, we propose a new, data-driven fire model that predicts burned area from the same input parameters that are passed to global fire models. We trained LSTMs to model burned area from GFED5. We split our data according to the IPCC regions and perform a region-based cross-validation, that is, we train different LSTMs on different region-splits of the data. We then compose the predictions of these different models so that for each region the predictions are made by LSTMs that have never seen any data during training and validation from that region before. Our model outperforms all fire models on a global scale and in most IPCC regions. With our model, we can improve our understanding of past fire behavior and simulate future fire trends.
How to cite: Lampe, S., Gudmundsson, L., Kraft, B., Le Saux, B., Hantson, S., Kelley, D., Humphrey, V., Chuvieco, E., and Thiery, W.: Modelling global burned area with deep learning, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-15236, https://doi.org/10.5194/egusphere-egu25-15236, 2025.
Peatland fires pose significant environmental and societal challenges. We recently advanced the Canadian Fire Weather Index (FWI) system for northern peatlands by integrating peatland-specific hydrological data derived from assimilating Soil Moisture and Ocean Salinity (SMOS) L-band brightness temperature observations into the NASA Catchment Land Surface model with its peatland modules, ‘PEATCLSM’. This novel FWIpeat (Mortelmans et al. 2024) was evaluated using satellite-based fire presence data over boreal peatlands from 2010 through 2018, demonstrating improved estimation of peatland fire presence.
Here, we extend the use of this renewed FWIpeat system by integrating it into a machine learning framework to gain deeper insights into when, where, and why peatlands burn. We utilize an XGBoost algorithm trained on peatland burned area data from 2012-2023, incorporating a suite of predictors, including (i) peatland distribution characteristics, (ii) peatland groundwater table, (iii) lightning occurrence, (iv) meteorological data, (v) vegetation properties, and (vi) socio-economic factors. This approach enables proactive fire risk management strategies and contributes to a comprehensive assessment of peatland fire vulnerability and resilience. Preliminary results indicate the importance of peatland groundwater table and lightning occurrence in estimating peat burned area.
Mortelmans, J., Felsberg, A., De Lannoy, G. J. M., Veraverbeke, S., Field, R. D., Andela, N., and Bechtold, M.: Improving the fire weather index system for peatlands using peat-specific hydrological input data, Nat. Hazards Earth Syst. Sci., 24, 445–464, https://doi.org/10.5194/nhess-24-445-2024, 2024.
How to cite: Mortelmans, J., De Lannoy, G., Dunmire, D., Veraverbeke, S., Waddington, J., Scholten, R., and Bechtold, M.: Modeling peat burned area and understanding its drivers with machine learning, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-15841, https://doi.org/10.5194/egusphere-egu25-15841, 2025.
Within the framework of the project PolarRES (POLAR Regions in the Earth System) we assess the impact of climate change on the ecosystems of the terrestrial northern high latitudes by making use of a range of high resolution regional climate simulations. These regional simulations were themselves driven by global climate simulations selected following the storyline approach described in Levine et al. 2024 from the set of CMIP6 SSP3-7.0 simulations, namely NorESM2-MM and CNRM-ESM2-1. These define two extremes in the climatic envelope of the CMIP6 simulations. While NorESM2-MM shows a high warming of the Barents-Kara seas but a low Arctic tropospheric warming CNRM-ESM2-1 shows the opposite. The storyline approach is a comprehensive way of defining pathways for physical outcomes of climate change that are observable in the region of interest and can directly be linked to certain consequences.
The 2nd generation Dynamic Global Vegetation Model (DGVM) LPJ-GUESS with its wildfire model SIMFIRE-BLAZE was applied using the high-resolution meteorological forcing from the regional climate models (RCMs) to investigate the potential impacts on both vegetation and the development of wildfires as well as the role of uncertainty implied by the variability of the forcing data.
It can clearly be stated that wildfire activity will increase significantly under the given scenarios driven mainly by shifts in vegetation distribution, i.e. northward migration of both treeline as well as shrubs and grasses. These effects differ regionally, depending on both, the storyline and the RCMs.
We present the findings from an envelope of potential future climate forcings depicting the impact of climate depending on the regionally observable effects of Arctic tropospheric warming and the Barents-Kara Seas warming, making use of the storyline approach as a comprehensive indicator for regional future change.
The results of this assessment will directly influence the research conducted in the project GreenFeedBack (GREENhouse gas fluxes and earth-system FEEDBACKs), which focusses on enhancing the knowledge on GHG dynamics in the boreal high latitude terrestrial and marine ecosystems.
Levine, X. Jet al. : Storylines of summer Arctic climate change constrained by Barents–Kara seas and Arctic tropospheric warming for climate risk assessment, Earth Syst. Dynam., 15, 1161–1177, https://doi.org/10.5194/esd-15-1161-2024, 2024
How to cite: Nieradzik, L., Lee, H., Levine, X., Miller, P., Mooney, P., Mottram, R., and Wårlind, D.: Assessing the impact of climate change on boreal high latitude wildfire using a storyline approach, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17962, https://doi.org/10.5194/egusphere-egu25-17962, 2025.
Whilst global burned area continues to decline, recent climate warming has led to an increase in the occurrence and intensity of extreme fires. Humanity must adapt to this new reality. Two proposed management options are a) prescribed livestock grazing, and b) prescribed fire use. Both methods promise cost-efficient means to reduce fire intensity, fire-induced vegetation mortality, and carbon emissions by reducing and fragmenting fuel loads. However, at present, there has been no systematic global assessment of the efficacy of these interventions. Reasons for this include a lack of data to understand their present-day distribution and impact as well as a lack of formal model structures to represent their uptake under future scenarios.
Here, we present two applications of the newly developed global, agent-based Wildfire Human Agency Model (WHAM!)1 to assess the potential effect of prescribed grazing and prescribed fire as adaptations to future fire regimes. Firstly, to explore the effect of prescribed livestock grazing on global fire regimes, we share a representation of livestock grazing intensity in WHAM! and its integration with the generalised linear models of Haas et al., (2). Secondly, we present work on a tight coupling of WHAM! with the JULES-INFERNO dynamic global vegetation model, focusing on parameterisation of how managed human fire use reduces fire-induced vegetation mortality.
Overall, early results suggest both management options already play a significant role in reducing global fire intensity and highlight the importance of considering dynamic human responses to a changing climate in global projections of future fire regimes.
1Perkins, O… et al. (2024). GMD.
2Haas, O. et al., (2022). Env. Res let.
How to cite: Perkins, O., Haas, O., Kasoar, M., Kelley, D., Teixeira, J. C. M., Voulgarakis, A., and Millington, J. D. A.: Adapting to fire in a warming climate: towards global assessment of prescribed grazing and prescribed fire, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-18640, https://doi.org/10.5194/egusphere-egu25-18640, 2025.
Agricultural straw burning is a significant source of greenhouse gas emissions, adversely affecting regional human health and air quality. Understanding the spatiotemporal patterns of agricultural fires is crucial for developing effective emissions reduction strategies in cropland to mitigate climate change. Although it is reported that cropland fires have been decreasing over the past two decades, the trends of global cropland fires on seasonal and diurnal scales remain poorly quantified, limiting a complete understanding of their spatiotemporal dynamics. This study analyzes global cropland fire activity from 2003 to 2020 at annual, seasonal, and diurnal scales, using multiple satellite-based burned area datasets, active fire products, and cropland classification datasets. The results show that from 2003 to 2020, global cropland burned area, active fire detections, and fire intensity all exhibited significant decreasing trends (p < 0.05), with relative changes of -43.5%, -30.3%, and -3.5%, respectively. The most significant decreases in cropland burned area and active fire detections occurred in Africa, while the largest decline in fire intensity was observed in Asia. Moreover, cropland fire activity displayed notable seasonal and diurnal variations. On the seasonal scale, the largest declines in cropland burned area, active fire detections, and fire intensity were observed in December, August, and November, respectively. Notably, fire intensity showed a significant increasing trend (p < 0.05) in April and September. On the diurnal scale, the decrease in cropland active fire detections was primarily driven by daytime activity; however, the rate of decline in fire intensity at night was about 1.5 times that during the day. These findings offer valuable insights into the comprehensive spatiotemporal patterns of global cropland fires, providing a foundation for more effective cropland management and carbon mitigation strategies.
How to cite: Wang, J., Li, J., Zhao, J., Zhong, X., Wang, M., He, J., and Yue, C.: Spatiotemporal changes in global cropland fire activity from 2003 to 2020, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-1776, https://doi.org/10.5194/egusphere-egu25-1776, 2025.
The recovery time of ecosystems following wildfire significantly influences carbon sequestration rates, land-atmosphere exchanges, and hydrological processes. Post-fire recovery has been studied at local scales but there is a lack of comprehensive global-scale analyses. We used solar-induced chlorophyll fluorescence (SIF) to quantify the recovery of photosynthetic activity after more than 10,000 fires from diverse ecosystems. We used the relaxed lasso technique to identify key determinants of the length of time required for post-fire recovery, and used these to build a linear regression model. Our results show that vegetation characteristics, fire properties, and post-fire climatic conditions all influence recovery time. Gross primary production (GPP) is the most important determinant of recovery time: ecosystems with higher GPP recover faster. Fires with greater intensity and duration, which cause more extensive vegetation damage, are associated with longer recovery times. Post-fire climate also affects recovery time: anomalously high temperatures and temperature seasonality, and increased number of dry days, cause slower recovery, while above-average precipitation accelerates recovery. Recovery times vary between different biomes, potentially reflecting variations in plant fire adaptations: ecosystems with a higher abundance of resprouting plants recover more rapidly. These findings provide a global perspective on how vegetation responds to fire disturbances, offering insights into carbon and water cycle dynamics under changing climatic conditions.
How to cite: Shen, Y., Prentice, C., and Harrison, S.: Global Drivers of Post-Fire Ecosystem Recovery: Insights from Solar-Induced Chlorophyll Fluorescence, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-3308, https://doi.org/10.5194/egusphere-egu25-3308, 2025.
Wildfires are an increasing concern for climate change, air quality and recognized for their substantial impacts on atmospheric composition. In addition to significant emissions of carbon dioxide (CO2) and particular matter (PM), biomass burning events are characterized by substantial non-CO2 emissions, which encompass a wide range of species. These emissions significantly influence atmospheric chemistry at a regional to global scale. Particularly in regions with ample fuel sources and hot, dry, or windy meteorological conditions, surface fires can lead to high-intensity crown fires and frequent downwind spotting. In certain circumstances, the intense formation of crown fires triggers the development of pyrocumulonimbus (PyroCb) atop smoke columns, which ascend to the upper troposphere and lower stratosphere (UTLS) and thus promote the dispersion of the fire emissions within wide regions. On August 2, 2019, the Williams Flats Fire ignited due to lightning from early morning thunderstorms in eastern Washington, USA. The main fire activity occurred between August 2 and August 9. On August 8, the high intensity crown fires led to the formation of a PyroCb. This event was observed and probed by the joint NOAA and NASA FIREX-AQ field campaign, providing a unique observation dataset. In this study, we utilize the Weather Research and Forecasting Model (WRF) to assess the impact of the Williams Flats fires on the atmospheric composition. In particular, we couple the representation of detailed multi-phase chemistry (WRF-CHEM) with WRF’s fire spread model (WRF-FIRE), employing WRF’s Large Eddy Simulation capabilities to resolve turbulence at resolutions of 100 m. In this presentation, results from WRF-FIRE-CHEM simulations with and without aqueous-phase chemistry will be shown to isolate its effects on the long-range transport of trace gases and aerosols.
How to cite: Rosanka, S., Juliano, T., Carlton, A. M., and Barth, M.: Large eddy simulations of the Williams Flat Fire: Aqueous chemistry in pyrocumulous clouds, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-12104, https://doi.org/10.5194/egusphere-egu25-12104, 2025.
In recent years, large wildfires have spread in Arctic regions as a consequence of ongoing climate change. Arctic organic soils are comparatively shallow but may be ancient, thus thousands of years old carbon may be released in smoldering and deeply burning fires. In Greenland, a land known for its icy expanse, fires are extremely rare. However, in summer 2019, the second-largest wildfire ever recorded on the island occurred at the Kangerluarsuk Tulleq fjord in southwestern Greenland. This study aims to produce pioneering in-field data on this tundra fire, focusing on three key aspects: 1) combustion, 2) burn depth, and 3) the age of the carbon released. Understanding whether the released carbon is modern or old is crucial due to different implications for the global carbon cycle and climate. To estimate carbon losses from the Kangerluarsuk Tulleq tundra fire, we established 14 sampling plots in burned areas and at unburned control sites. The selection of sampling plots was guided by a differenced Normalized Burn Ratio (dNBR) map generated using Sentinel-2 data and field reconnaissance. Within each plot, we assessed fire severity to estimate the above-ground carbon loss. For below-ground carbon loss estimation and burn depth analysis, organic soil samples were collected at burned plots and compared with unburned ones. To explore the vegetation succession and burned vegetation type, organic soil profiles (n=10) were extracted down to the mineral ground using a soil box corer and were studied by light-microscopy. Subsamples (n=20) from burned soil horizons were selected for radiocarbon dating to determine the age of carbon released in the fire. Our preliminary results suggest that soil carbon loss was higher than previously reported at an Alaskan tundra fire site with a mean value of 6.718 ± 0.9 kg of C m-2. The mean burn depth was 9.0 ± 1.8 cm, and soil thaw depths during the 2024 summer were approximately 24 cm deeper in the 2019 burned area compared to unburned tundra. Expected radiocarbon results will indicate the maximum age of the carbon released by the fire. Vegetation succession measurements show that post-fire surfaces were predominantly colonized by pioneering non-Sphagnum bryophytes, Cyperaceae, and Ericaceae. The acquired results are first of a kind from a Greenland tundra fire and produce essential data for global climate modeling to assess the climate impacts of increasing Arctic wildfires.
How to cite: Granqvist, S., Diaz, L., Veraverbeke, S., Pilkama, E., Väliranta, M., and Ruppel, M.: Carbon emissions of an unprecedented Greenland wildfire, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-15318, https://doi.org/10.5194/egusphere-egu25-15318, 2025.
The Copernicus Atmosphere Monitoring Service CAMS is using ECMWF's Integrated Forecasting System IFS-COMPO with fire emissions from its Global Fire Assimilations System GFAS to monitor and forecast the effect of smoke from vegetation fires, resp. biomass burning, on atmospheric composition. The simulated atmospheric composition fields are routinely validated against observations including from satellites, aircraft and ground stations.
The emissions calculation by the operational GFAS version 1.2 have recently been updated for use in the upcoming HTAP3 multi-model, multi-pollutant study of fire impacts (Whaley et al. 2024), creating the dataset GFAS4HTAP. It is based on the dry matter burnt estimates of GFASv1.2, and uses an updated spurious signal mask, ESA CCI land cover data for 2018, a global peat map (Xu et al. 2018) and emission factors from NEIVA (Shahid et al. 2024) to calculate emission fluxes for various smoke constituents for 2003-2024. An additional GFAS-based dataset has been created by calibration against GFED5beta.
Global comparisons of dry matter, resp. biomass, combustion rates of the three GFAS-based inventories with GFED4s, GFED5beta, and the two variants of FINN2.5 reveal that these inventories can be roughly classified into one group of "traditional" inventories with lower fire activity, resp. emissions, and another of "more recent" inventories with higher fire activity. The pyrogenic carbon monoxide emission estimates from an inversion of satellite observations of atmospheric composition (Zheng et al. 2019) lie between these two groups in terms of global annual values. However, at a global level, they are more consistent with the "more recent" inventories during the late boreal summer peak of the global fire activity and with the "traditional" inventories during periods of lower fire activity.
In order to gain more insight from independent validation, we here present simulations with IFS-COMPO for 2019 based on the three GFAS-based inventories and compare these with atmospheric observations of carbon monoxide, nitrogen dioxide and aerosol optical depth. We find that the best agreement of simulation and observations is achieved by different inventories for different regions, seasons and smoke constituents. However, the emissions of the GFAS4HTAP dataset appears to lead to the overall most balanced atmospheric composition simulation. This supports the group of "traditional" inventories mentioned above.
How to cite: Kaiser, J., Huijnen, V., Remy, S., Ytre-Eide, M. A., De Jong, M. C., Zheng, B., and Wiedinmyer, C.: Evaluation of fire emissions for HTAP3 with CAMS GFAS and IFS-COMPO, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-8708, https://doi.org/10.5194/egusphere-egu25-8708, 2025.
Vegetation fires are well known as an important source of aerosol particles originating from the combustion of carbonaceous material. Much less known is that these fires can also efficiently inject soil-dust particles into the atmosphere, raised by the strong fire-induced winds. These soil-dust particles and the likely co-emitted biogenic particles are potent cloud condensation nuclei (CCN) and ice nucleating particles (INPs), and can substantially alter the cloud microphysics and thus impact the Earth’s radiation budget. Fires are an integral component of the Earth system that affect different landscapes around the globe. As they are supposed to get more frequent and more severe along with the ongoing global warming, a better knowledge of these specific fire emissions is crucial to understand their impacts on weather and climate.
Therefore, this work investigates the potential of wildfires to emit soil-dust particles on a global scale as a part of the newly established Leibniz ScienceCampus “BioSmoke” (‘smoke and bioaerosols in a changing climate’). As this particular dust emission pathway is not considered by the state-of-the-art dust emission models, a parameterization describing fire-induced dust emission fluxes has been developed and implemented into the global aerosol-climate model ICON-HAM. Fire-dust emissions are modelled as a function of the fire radiative power (FRP), the ambient wind conditions, and further soil-surface properties, including the soil type and a vegetation-dependent surface roughness correction.
Multi-year ICON-HAM simulations have revealed that fire-related dust emissions can account for up to one fifth of the total global dust emissions with strong regional and seasonal variations, both as the result of a varying fire activity and the local soil-surface conditions that can foster or impede also fire-dust emission significantly. In regions where the classic wind-driven dust emissions from arid, unvegetated soil surfaces are rather low but wildfires occur frequently, e.g., in large parts of the Southern hemisphere, fire-related dust emissions can add substantially to the atmospheric aerosol load and affect the local radiation budget there.
How to cite: Wagner, R., Tegen, I., and Schepanski, K.: Vegetation fires as a source of soil-dust particles – a global model perspective, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-8307, https://doi.org/10.5194/egusphere-egu25-8307, 2025.
The global wildland fire management community faces pressing climate change and operational challenges and requires improved capabilities in existing modelling tools or the development of novel decision support tools to limit the negative impact of wildfires and to increase use of prescribed burning where appropriate. This presentation will discuss limitations in the existing approaches to incorporating fuel structure effects in different model types (empirical, semi-empirical, detailed physics-based). In particular, novel experimental data will be presented addressing previously identified limitations [1] in the description of surface fuel beds in one of the most widely-used semi-empirical models; the Rothermel model, which underpins many current operational models.
The Rothermel model [2] involves a conservation of energy approach, incorporating separate terms to describe energy release rate in the combustion zone (reaction intensity) and energy transferred to the unburnt fuel (propagating flux), and incorporates a number of empirical closure terms. The reaction intensity is empirically based, with the underpinning experimental measurements described in Frandsen and Rothermel [3]. By measuring the mass loss rate in a section of a fuel bed, Frandsen and Rothermel were able to characterize the intensity distribution within the combustion zone. However, the interacting effects of simultaneously varying fuel loading and packing ratio were not systematically considered, complicating efforts to understand the interacting effects of fuel loading and bulk density.
This study presents a series of laboratory-based flame spread experiments (no wind) involving excelsior fuel beds of varying structural conditions (Fuel Height: 0.02 to 0.12 m, Bulk Density: 3.3 to 20 kg/m3, Fuel Loading: 0.2 to 0.4 kg). The reaction intensity was calculated via a similar procedure to that described by Frandsen & Rothermel [2] as ‘Method 2’, in which the longitudinal length of the mass measurement region is greater than or equal to the combustion zone depth.
Clear trends in the peak mass loss rate and profile with bulk density were observed with a significant reduction at lower fuel loadings (0.2 kg/m2), and the reaction time was observed to increase at higher bulk densities along with a lengthening in the reaction intensity distribution region (further behind the combustion wave front). These results, along with existing observations of the trailing, in-depth combustion region in porous fuel beds, can be used to further investigate the observed tendency for underprediction of spread rates when the Rothermel model is applied to compressed fuel bed scenarios and has practical implications for other fire behaviour modelling applications. For example, improved characterisation of the overall combustion wave may enable improved modelling of smoke generation, surface-to-crown fire transition, and fuel consumption (e.g. to evaluate prescribed fire effectiveness).
[1] Z. Campbell-Lochrie, M. Gallagher, N. Skowronski, R.M. Hadden, The effect of fuel bed structure on Rothermel model performance, Int. J. of Wildland Fire. 33 (2023).
[2] R.C. Rothermel, A Mathematical Model for Predicting Fire Spread in Wildland Fuels, Research Paper INT-115, USDA Forest Service.,1972.
[3] W.H. Frandsen, R.C. Rothermel, Measuring the energy-release rate of a spreading fire, Combust Flame 19 (1972) 17–24.
How to cite: Campbell-Lochrie, Z.: Revisting Intensity of Combustion Waves to Address Outstanding Issues in Wildfire Modelling, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-1390, https://doi.org/10.5194/egusphere-egu25-1390, 2025.
China is a significant region for crop cultivation. For a long time, there has been a common practice of burning crop residues during the post-harvest period (from May to October). The smoke emitted from straw burning contains both types of ozone precursors, including nitrogen oxides (NOx=NO+NO2) and volatile organic compounds (VOCs), and can be transported over long distances. During the transport process, secondary formation or consumption of ozone precursors occurs within the smoke plumes. After the smoke plume mixes with the atmosphere in the downwind urban area, it will lead to changes in the local ozone formation sensitivity. However, due to the nonlinear relationship between ozone and its precursors, the changes in ozone levels in downwind cities are not as straightforward as expected.
Here, we explore the temporal evolution of urban ozone and its precursors on smoke-affected days using multi-source satellite-derived fire event tracking datasets, which are screened by a semi-quantitative absorbing aerosol index (AAI), tropospheric NO2 and HCHO columns measurements from OMI, fire points from Himawari-8, and ground-level O3 monitoring dataset. We aimed to understand the associations between urban ground-level O3 concentrations and crop residue burning events in China. Our analysis revealed that no consistent changes were shown in urban O3 on smoke-affected days. In addition, there was an increase in NO2, while HCHO and O3 decreased in cities after mixing with smoke that had taken a long transport time. Our findings suggest that the O3 formation sensitivity within aged smoke tends to be controlled by VOC-limited regime. We hypothesize that the large amount of NOx carried by aged smoke consumes urban VOCs and O3, while producing NO2 locally. When fresh smoke, which is mainly controlled by the NOx-limited regime, enters urban environments rich in NOx, it leads to an increase in O3 concentration. Our analysis may contribute to an improved understanding of the influence of straw burning on urban ozone levels in China.
How to cite: Wang, W., van der A, R., Ding, J., Cheng, T., and Wang, C.: Exploring the effect of straw burning on urban ozone levels based on multi-source satellites in northern China, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-1314, https://doi.org/10.5194/egusphere-egu25-1314, 2025.
Forest and vegetation fires are one of the major sources of air pollution and have triggered air quality issues in many regions of Pacific Asia. Here we isolate the fire-specific PM2.5 from monitoring concentrations using an observation-driven approach in the region. The total PM2.5 in Pacific Asia exhibited a rapid declining trend from 2014 to 2021, while fire-specific PM2.5 decreased in early years but begun to reverse, leading to an increasing proportions of fire-specific PM2.5 in recent years. The inconsistency between the decreasing number of fire points and the rising levels of fire-specific PM2.5 may be attributed to a shift in dominant sources of fire emissions in Pacific Asia, moving from anthropogenic agriculture fires to wildfires. Fire-related PM2.5 poses a significant public health threat in Pacific Asia, contributing to approximately 334,300 premature deaths each year. Our assessment highlights the disproportionate impact of fire-specific PM2.5 on poverty populations, indicating a pressing need for more attentions and researches in these regions. Based on the positive correlation between vapor pressure deficit and fire-specific PM2.5, this study suggests that without further regulation and policy intervention, the contributions of fire-specific PM2.5 to air pollution in Pacific Asia are likely to continue increasing under the influence of future climate change.
How to cite: Lu, H., Xie, M., Wang, N., and Liu, B.: The contribution of fires to PM2.5 and population exposure in Pacific Asia, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-2467, https://doi.org/10.5194/egusphere-egu25-2467, 2025.
In 2024, Bolivia experienced the worst year of fires since 2002, when Aqua MODIS began collecting data. According to estimates, more than 15 million hectares were burned this year. A sunphotometer sitting in the Bolivian lowlands recorded AOD values higher than two for several continuous days indicating the degradation of the air quality in the region. A unique set of instruments located in the Bolivian Andes recorded the transport of smoke produced by this biomass burning. Very high values of atmospheric tracers like carbon monoxide, equivalent black carbon, and others have been measured as high as 5240 m asl at the Chacaltaya GAW station (CHC, 16.35ºS, 68.13ºW, 5240 m asl) and other sites around it both in the Altiplano and adjacent high altitude valleys. Although transport to these sites was observed previously, usually the events lasted one or two days. However, in 2024 longer periods of consecutive days with smoke arriving from the lowlands were observed for a second year in a row. Similar high values were observed in CHC in October of 2023, a year with less than half of fires in the country. The conditions that led to the transport of smoke to the mountains in the Andean Cordillera will be discussed, as well as the possible effects of the associated deforestation in terms of water availability for the central Andes.
How to cite: Andrade, M., Ticona, L., Velarde, F., Guzman, D., Blacutt, L., Forno, R., Gutierrez, R., Moreno, I., Avila, F., Uzu, G., Goloub, P., Ramonet, M., Laurent, O., Wiedensohler, A., Weinhold, K., Krejci, R., Aliaga, D., Whiteman, D., and Laj, P.: Intense transport of smoke to the Central Andes: Insights from a unique set of instruments located in the Bolivian Andean Cordillera, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-14573, https://doi.org/10.5194/egusphere-egu25-14573, 2025.
South eastern Mediterranean region of Türkiye is well known with intense industrialization, shipping activities, agriculture and livestock production in addition to urban emission sources, thus struggle with significant air pollution problems. In addition to criteria pollutants, combination of these sources also results in high ammonia (NH3) levels in the region.
NH3 is released into the atmosphere mainly from agriculture, including nitrogen-based fertilizer applications and livestock farming as well as from several industries and from biomass burning. Atmospheric NH3 plays a significant role in the formation of secondary inorganic particulate matter (PM), which negatively impacts on human health and ecosystems and indirectly influences climate change by altering radiative forcing. Climate change has increased the frequency and intensity of wildfires globally, which became another significant source of NH3 over the Eastern Mediterranean, because the region is among the most sensitive regions. Besides wildfires, agricultural residue burning, although prohibited, also contributes to overall NH3 levels.
Biomass burning contributes to atmospheric pollutants, as the combustion process emits nitrogen and carbon compounds from organic matter. In this study, multi-satellite derived retrievals were utilized, including IASI Level-2 NH3 and CO, TROPOMI Level-2 NO2, CO, and HCHO along with VIIRS S-NPP Fire Radiative Power product to investigate biomass burning related NH3 levels. Products were processed at a 1x1km2 gridded resolution to analyse spatio-temporal variations from 2019 to 2023, especially focusing on intense fire time intervals. While NH3 levels were generally high during the summer over the region, the 2021 summer stood out with exceptionally high levels, coinciding with intense wildfires recorded that year. Similarly, CO levels revealed elevated levels during the same period, further strengthening the common impact of these extreme events. Further, fires detected over some areas by the VIIRS product were associated with residue-burning practices, as the area predominantly consists of agricultural lands.
The aim of the study is to investigate the impacts of fire-related NH3 levels and quantify NH3 enhancements during these fire events in the region. In this context, NH3 to other pollutant ratios will be examined and temporal variation between different biomes will be classified. Air quality and climate change impact studies over the Mediterranean are critically important, with the absence of ground-based NH3 measurements, satellite retrievals have to be utilized more to investigate the sensitivity of the region to extreme biomass burning events with the growing impacts of climate change.
Keywords: ammonia, carbon monoxide, nitrogen dioxide, biomass burning, wildfires
Acknowledgements: IASI is a joint mission of EUMETSAT and the Centre National d'Etudes Spatiales (CNES, France). The authors acknowledge the AERIS data infrastructure for providing access to the IASI data in this study and ULB-LATMOS for the development of the retrieval algorithms. This study was supported by the Scientific and Technological Research Council of Türkiye under the grant number 123Y364.
How to cite: Saracoglu, S., Alban, A. M., Tokgoz, S., and Kaynak, B.: Investigation of the intense wildfire events and NH3 levels over the Eastern Mediterranean, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17646, https://doi.org/10.5194/egusphere-egu25-17646, 2025.
For thirty years, the European Research Infrastructure IAGOS (In-Service Aircraft for a Global Observing System) has been equipping commercial aircraft with instruments to measure atmospheric composition on long-haul flights around the world. Ten aircraft are currently equipped with IAGOS instruments to measure ozone, and the precursors carbon monoxide and nitrous oxides from the surface to the upper-troposphere during landing and take-off at worldwide airports, and at cruise altitude where we observe the long-range transport of polluted airmasses. We analyse the transport of biomass burning pollutants from the intense Canadian wildfire seasons of 2023 and 2024 which impacted air-quality in North America and in Europe, and the extreme wildfires over the Amazon in 2024 that impacted air quality in South American cities. The significance of these events is interpreted within the context of the 30-year climatology. The events will be compared with forecasts and analyses from the Copernicus Atmosphere Monitoring Service's global and regional models (projects CAMS2_82 and CAMS2_83) and we further highlight the role of IAGOS in developing air-quality networks in susceptible urban areas (project RI-URBANS) and the impacts of heatwaves and wildfires on air-quality in a changing climate (project IRISCC).
How to cite: Bennouna, Y., Clark, H., Wolff, P., Thouret, V., Blot, R., Nédélec, P., and Boulanger, D.: Impact of wildfires on air quality as seen by IAGOS in-situ measurements, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-18252, https://doi.org/10.5194/egusphere-egu25-18252, 2025.
Wildfires exert a important influence on the chemical composition of the atmosphere, thereby impacting air quality, ecosystem, and climate forcing. The substantial emission of pollutants from such fires, coupled with their long-range transport, has the potential to counteract the progress achieved in reducing anthropogenic emissions. Numerous studies show that the increase in the frequency and intensity of fires offsets the general trend towards improved air quality observed in regions influenced by wildfires. These studies also caution of an increasing risk of the population being exposed to extreme levels of aerosols and ozone. In addition to their regional impacts, the plumes from the most intense fires can be transported on a continental or even hemispheric scale, thereby imposing health constraints on regions that are not generally affected by widespread, frequent or intense fires.
The northern hemisphere is home to a group of biomes that are particularly sensitive to hydro-meteorological conditions, and therefore to the effects of climate change on burned areas. The majority of the most intense fires of the last two decades have occurred in North America and in the boreal regions of Asia.
This study assesses the impact of fires on the variability of total CO, total PAN and AOD in the Northern Hemisphere using 16 years (2008-2023) of observations from the IASI/MetOp and MODIS/Terra and Aqua satellite instruments. More specifically, the variability in the number of detected plumes of extreme values of CO, PAN and aerosol from fires is studied.
The trajectories of these plumes are estimated using only satellite observations and are used to assess the contribution of the different regions of the Northern Hemisphere to the variability of atmospheric composition. The potential impact of the long-range transport of the identified plumes on air quality is estimated using observations of the altitude of the plumes obtained from both active CALIOP observations and passive IASI observations.
The chemical composition of the identified plumes is characterised using IASI observations of ammonia (NH3), formic acid (HCOOH), methanol (CH3OH) and ozone (O3).
How to cite: Ehret, A., Turquety, S., Lecomte, G., Franco, B., George, M., Clarisse, L., Van Damme, M., Clerbaux, C., and Coheur, P.: Satellite observation of long-range transport of wildfires plumes in the northern hemisphere in 2008-2023, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-18688, https://doi.org/10.5194/egusphere-egu25-18688, 2025.
Globally, approximately half of landscape fire emissions originate from savannas and grasslands. Furthermore, our observations in South Africa indicated major secondary aerosol formation in near-fire plume ageing. However, the measurements in South Africa are affected by anthropogenic emissions from the Highveld region, except for a clean sector towards the semi-arid Karoo region. Aiming for a savanna environment with minimal anthropogenic influence we set up a new measurement site in the Okavango delta area in northern Botswana in August 2024.
For the active savanna fire season in 2024, we operated online measurements of aerosol chemical composition with an aerosol chemical speciation monitor (ACSM), an online gas chromatograph coupled to an MS detector (GC-MS) for volatile organic compounds and a single particle soot photometer (SP2) for refractive BC. Measurements of aerosol particle size distribution with a differential mobility particle sizer (DMPS), aerosol absorption with a multi angle absorption photometer (MAAP), as well as CO and CO2 concentrations will continue for the next couple of years at least.
For fresh plumes, initial analysis shows a strong decrease in submicron aerosol emission factor (EFPM1) with increasing modified combustion efficiency, i.e. with increasing flaming fraction. The EFPM1 values are in good agreement with previous observations in southern African savanna and with recent laboratory experiments that we carried out in collaboration with University of Eastern Finland. Analysis of ageing effects on the fire plumes in a clean savanna environment is ongoing.
How to cite: Vakkari, V., Mmereki, B. T., Koolebogile, D., van Niekerk, C. P. E., Le, V., Letsatle, M., Jaars, K., and van Zyl, P. G.: A new measurement site in northern Botswana to observe savanna fire plumes, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-19104, https://doi.org/10.5194/egusphere-egu25-19104, 2025.
Large-scale, frequent forest fires have a detrimental effect on the environment, the quality of the air, and human health. In the present study, from 2001 to 2020, March (1,857.5 counts/month) and April (922.8 counts/month) saw around 70% of the region's annual forest fires. Unusually high numbers of forest fires have been reported in some years, including 2009, 2012, and 2017. A thorough investigation is conducted into the contribution of numerous climate extremes and persistently rising temperatures to the rise in forest fire activity over central India. Forest fire activity doubled and tripled during the non-fire (July–January) and forest fire (February–June) seasons, respectively, over the warmer period from 2006 to 2020. A severe heat wave, an unusual drought, and an exceptionally powerful El Nino occurred in central India between 2015 JASONDJ and 2018 FMAMJ. These events are thought to have contributed to an upsurge in forest fires. The quinquennial spatiotemporal changes in forest fire characteristics, including average fire intensity and fire count density, were also evaluated. Significantly high soil temperature, low soil moisture content, poor evapotranspiration, and low normalized difference vegetation index are statistically associated with high near-surface air temperature and low precipitation during FMAMJ. This makes the climate much drier, which encourages a lot of forest fires in the Central Indian region.
How to cite: Sonwani, S., Saxena, P., and Jain, M.: Forest Fire Variability Over the Central India Region from 2001–2020, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-1122, https://doi.org/10.5194/egusphere-egu25-1122, 2025.
Wildfire emissions are a significant environmental concern, especially as climate change is expected to increase the frequency and intensity of extreme wildfires. Numerical weather and chemical transport models often struggle to reliably capture the injection height of wildfire plumes, a key parameter for transport that determines the impact on air quality and climate.
This study uses the ICON-ART numerical model to analyze fire-atmosphere feedbacks and their impact on the aerosol plume. The Australian New Year’s wildfire event of 2019/2020, a period of extreme wildfires and pyro-convection, is chosen as the case study. The simulations are performed with a grid spacing of 6.6 km. At this resolution, convection cannot be resolved, so a plume rise model is employed to parameterize the injection height. However, the resolution is sufficiently fine to account for the impact of the fire on meteorological variables.
Our simulations reveal that fire-induced moisture release leads to increased cloud formation under near-saturation conditions, but the overall impact on plume development is small. In contrast, fire-induced heat release significantly increases the mass-weighted height from the start, driven by sensible heat release, increased injection height, and enhanced convective cloud formation.
Comparison with observations shows that accounting for the heat release by the fire enables the simulation of the observed plume heights. These implementations have the strongest effect on the first simulation day, when the fires are most intense, and are negligible on the last simulation day. For fires with lower intensity, the plume rise model performs well without additional implementations.
How to cite: Muth, L., Vogel, B., Vogel, H., and Hoshyaripour, G.: Modeling Fire-Atmosphere Feedbacks: Insights from the 2019/2020 Australian Wildfires, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-1707, https://doi.org/10.5194/egusphere-egu25-1707, 2025.
Southeast Asia experiences widespread wildfires and biomass burning events during the dry season (January to April), leading to poor air quality, haze, and smog. NASA conducted the Airborne and Satellite Investigation of Asian Air Quality (ASIA-AQ) flight campaign in February and March 2024 to study the contribution of smoke to urban air quality through a multi-faceted observational approach (aircraft, satellite, and ground). The campaign deployed the NASA DC-8 aircraft, equipped with instruments from the Langley Aerosol Research Group (LARGE) and other teams, to measure real-time aerosol microphysical and optical properties, trace gases, and meteorological parameters. During the campaign in the Philippines, South Korea, Thailand, and Taiwan, it was noted that the northern region of Thailand was predominantly impacted by agricultural residue burning and wildfires. Here, we present the variations of vertical and horizontal profiles of aerosol properties and biomass burning tracers, alongside meteorological data to assess the impacts of local conditions and potential pollution pathways. Key findings will include observed variability in aerosols properties, the role of absorbing and scattering aerosols, boundary layer dynamics, and regional pollution transport across the ASIA-AQ domain.
How to cite: Roy, S., Gallo, F., Wiggins, E. B., Ziemba, L. D., Jordan, C., Winstead, E. L., Shook, M. A., DiGangi, J. P., Diskin, G. S., Choi, Y., Miech, J. A., Wojnowski, W., Piel, F., Swift, S. J., Wisthaler, A., and Moore, R. H.: Wildfires and biomass burning in northern Thailand: Observations from ASIA-AQ Campaign, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-2016, https://doi.org/10.5194/egusphere-egu25-2016, 2025.
Incidents in nuclear facilities can lead to the emission of a radioactive plume dis-
persing into the atmosphere. In such events, the highest radionuclide concentration
is usually located near the source at distances ranging from a few meters to several
hundred meters. It is, therefore, crucial to be able to accurately predict these levels
of near-source concentrations.
One challenge arises from the thermal characteristics of the source, which regulate
the initial dispersion of the plume. In the case of a non-thermal gas release, the
dispersion of the plume is driven by atmospheric conditions, related to wind and
atmospheric instability, and is influenced by local surface characteristics such as
roughness and the presence of obstacles. In contrast, when the gas is emitted from
a hot source such as a fire, the released gas first rises in the atmosphere up to a
so-called ‘injection height’ due to buoyant forces. The injection height is reached at
a certain distance from the source and doesn’t only depends on the properties of the
hot source but also on the atmospheric conditions (e.g. downdraft effects). The gas
then disperses like in a non-thermal gas release.
While CFD modelling can offer an accurate description of the plume dispersion, its
processing speed is not suitable for use in emergency situations. In contrast, existing
analytical models can provide rapid results, but their injection height parametriza-
tions may lack comprehensive coverage. So far, analytical models have rarely been
validated against field measurements, and few field experiments have been conducted
to improve their parameterization.
The goal of this presentation is twofold, first to present a field experiment on the
atmospheric plume dispersal of a gas released from a hot source, and second to
evaluate an analytical model of plume dispersal against the experiment, with a
particular focus on the Atmospheric Transfer Coefficient of the released gas.
The field experiment was conducted in May 2024 on a flat terrain near Vire (Nor-
mandy, France), under unstable and neutral atmospheric conditions.
The source comprised a burner (PYROS) that generated a propane fire with an average heat
release rate of between 450 kW and 750 kW . Helium was injected into the plume
to serve as a tracer gas. During 15-minute observation periods, helium concentra-
tions in the air were measured at ground level at distances from the source ranging
from 40 m to 400 m, as well as at various altitudes, using air sampling points at-
tached to a rope lifted vertically by a drone. Additionally, atmospheric turbulence
characteristics were also measured using ultrasonic anemometers.
The analytical model employs Heskestad’s formulas to determine the fire character-
istics and Briggs’ dispersion parameters to characterise the Gaussian dispersion of
the plume when buoyant forces become negligible.
How to cite: Mendez, A., Manon, G., Dupont, S., and Laguionie, P.: Near-field atmospheric dispersion of a gas emitted from a hot source : a comparison between analytical modelling and in situ measurements, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-2253, https://doi.org/10.5194/egusphere-egu25-2253, 2025.
Biomass-burning organic aerosol(s) (BBOA) are rich in brown carbon (BrC), which significantly absorbs solar irradiation and potentially accelerates global warming. Despite its importance, the multiphase photochemistry of BBOA after light absorption remains poorly understood due to challenges in determining the oxidant concentrations and the reaction kinetics within aerosol particles. In this study, we explored the photochemical reactivity of BBOA particles in multiphase S(IV) oxidation to sulfate. We found that sulfate formation in BBOA particles is predominantly driven by photosensitization involving the triplet excited states (3BBOA*) instead of iron, nitrate, and S(IV) photochemistry. Rates in BBOA particles are three orders of magnitude higher than those observed in the bulk solution, primarily due to the fast interfacial reactions. Our results highlight that the chemistry of 3BBOA* in particles can greatly contribute to the formation of sulfate, as an example of the secondary pollutants. Photosensitization of BBOA will likely become increasingly crucial due to the intensified global wildfires.
How to cite: Liang, Z., Zhou, L., Chang, Y., Qin, Y., and Chan, C. K.: Biomass Burning Organic Aerosols as a Pool of Atmospheric Reactive Triplets to Drive Multiphase Sulfate Formation, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-3113, https://doi.org/10.5194/egusphere-egu25-3113, 2025.
Open biomass burning affects many aspects of the Earth system, including atmospheric chemistry and composition. Due to its impact on human health, we focus on the contribution of biomass burning emissions to fine particulate matter (PM2.5) concentrations on a global, annual mean basis, particularly the lesser-studied secondary inorganic component. We use the EMEP MSC-W WRF atmospheric chemistry transport model to show that biomass burning leads to increased ammonium nitrate (NH4NO3) concentrations in densely populated regions not necessarily associated with large-scale fire activity. This is prominent in the eastern USA, northwestern Europe, the Indo-Gangetic Plane and eastern China, where NH4NO3 contributes between 29 and 51% to annual mean biomass burning-derived PM2.5. Pyrogenic CO and NOx (NO and NO2) emissions alter the global-scale oxidising capacity of the atmosphere, affecting how local-scale anthropogenic NOx and NH3 emissions lead to formation of NH4NO3. These teleconnections can locally increase, by up to a factor of two, the contribution of biomass burning emissions to PM2.5 concentrations, which measurements alone cannot detect. This will become relatively more important as anthropogenic sources of PM2.5 are reduced, and with potentially intensified biomass burning occurrences under climate change.
How to cite: Tan, D. Y. T., Heal, M. R., Vieno, M., Stevenson, D. S., Reis, S., and Nemitz, E.: Changes in atmospheric oxidising capacity cause teleconnections between biomass burning and NH4NO3 formation, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6847, https://doi.org/10.5194/egusphere-egu25-6847, 2025.
Wildfire smoke is increasingly recognized as a significant source of air pollution that leads to public health issues. Over the past few decades, air pollution in Canada has been reduced due to effective regulations. However, fine particulate emissions (i.e., particles with an aerodynamic diameter of less than 2.5 μm (PM2.5)) from wildfires have shown upward trends as climate change exacerbates the frequency and likelihood of wildfires. According to the Canadian Interagency Forest Fire Centre (CIFFC) in 2021, there were 18% more fire starts and nearly a 61% increase in the total area burned compared to the past 10-year average in Canada. The emissions inventories used for modeling the impact of fires on air quality and climate exhibit several discrepancies in emissions estimates, primarily due to the different types of satellite products used for identifying fires and measuring burned area, as well as differences in emission factors describing the vegetative fuels burned. This variability of fire emission inventories leads to uncertainties in predicting air quality. Using the GEOS-Chem chemical transport model, we studied how differences in emissions estimates among three commonly used global biomass burning inventories—the Global Fire Emissions Database 4 (GFED4), the Global Fire Assimilation System (GFAS), and the Quick-Fire Emissions Database 2 (QFED2)—and a newly developed regional biomass burning emission inventory, the Canadian Forest Fire Emissions Prediction System (CFFEPS), affect modeled concentrations of PM2.5 during the 2021 wildfire season in Canada. To examine the sensitivity of simulated PM2.5 to different biomass burning emission datasets, we compared them with ground based PM2.5 data from 70 NAPS (National Air Pollution Surveillance) stations across Canada, from east to west. The simulated PM2.5 concentrations showed significant variation in model performance based on the geographic location of the monitoring stations, particularly between the western and eastern regions of Canada. These findings indicate the importance of considering the strengths and weaknesses of each fire inventory, as some inventories may more accurately represent fire emissions in certain regions than others.
How to cite: Ashraf, S., Hayes, P., Stevens, R., and Chen, J.: Evaluating the Effect of Variability in Biomass Burning Emissions Inventories on Modeled Smoke Concentrations: Insights from the 2021 Canadian Wildfire Season, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-7351, https://doi.org/10.5194/egusphere-egu25-7351, 2025.
As the climate warmings, the frequency and intensity of wildfires have escalated in recent decades. While the adverse effects of wildfires on air quality are well-documented, their influence on atmospheric ozone in China remains unclear. Here, we apply deep learning and a trajectory-fire interception method (TFIM) to estimate wildfire contributions to ozone concentrations in Chinese cities from 2015 to 2023. Our findings indicate that wildfires influenced 15.1 ± 9.3% of all days during this period, with a wildfire-induced ozone concentration averaging 6.8 μg m-³. Over the nine-year study period, these concentrations exhibited a modest upward trend, increasing by 0.091 μg m⁻³ annually. Regions such as Southwest China, the Qinghai-Tibet Plateau, and Northwest China experienced the highest levels of wildfire-induced ozone. We further utilize SHapley Additive exPlanations algorithms to investigate driving factor behind wildfire-induced ozone. The burnt area, aging hour, and injection height of smoke have a large effect on wildfire-induced ozone concentrations. Finally, we evaluated the health impacts of wildfire-induced ozone, highlighting its significant implications for public health in affected regions.
How to cite: Liu, S.: Explainable deep learning reveal the contribution of wildfire to ozone in China, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-8751, https://doi.org/10.5194/egusphere-egu25-8751, 2025.
Fires as a disturbance regime are an important component of ecosystems, and are involved in many feedback loops within these systems such as climate-carbon feedbacks. The changing climate can influence fire regimes in multiple ways, both directly and indirectly. For example, changing weather patterns can directly alter the occurrence and timing of fire weather days. Weather patterns also influence vegetation growth and ecosystem composition, leading to changes in fuel availability and flammability. Meanwhile, humans also partially shape fire regimes via accidental and managed ignitions as well as various suppression measures.
In this study, we use 35 years of remote sensing data to establish global pyromes; regions of similar fire regimes, via their fire characteristics. This length of data period allows for the allocation of pyromes across multiple time segments, and for changes in their prevalence and spatial distribution to be observed. We have found that the majority of pyrome transitions occurring are shifts towards smaller or less frequent fires, and these transitions are widespread across the globe. However, some regions such as the Northern high latitudes, the Western United States, and Northern Australia are shown to experience larger or more frequent fires in the final observation segment of the study.
Following on from this, we use statistical methods to investigate relationships between pyromes and a wide variety of non-fire properties, including climate, vegetation, and human influence. This allows for inference of the most relevant drivers of pyrome change, both climatic and non-climatic. Initial results suggest for example, that population density is a more important predictor for pyromes with small and medium sized fires. However, there are significant challenges to disentangling the effects of such complex drivers within a relatively short observational period. Nevertheless, it is possible to build a picture of plausible fire regime evolution in regions with shifting environmental components.
How to cite: Butler, E., Sippel, S., and Bastos, A.: Global fire regimes, their non-fire characteristics, and changes in time., EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-9010, https://doi.org/10.5194/egusphere-egu25-9010, 2025.
Wildfires have a significant influence on the Earth system through perturbing the carbon cycle and also emitting large quantities of short lived climate forcers (SLCFs) such as aerosol precursors (black and organic carbon) and gases that can lead to ozone formation (carbon monoxide, nitrogen oxides). SLCFs are important as they affect the Earth’s radiative balance, influencing climate, and also can have important impacts on air quality in the near-surface atmosphere. Climate change and human interference also have important effects on the size, magnitude and duration of wildfires, which are important to understand further, particularly in the context of a changing climate. Such influences are potentially important in the northern high latitudes, where wildfires have been increasing in magnitude and frequency over the last few decades. Here, we present an evaluation of the representation of high latitude wildfires in a configuration of UKESM with an interactive fire module (INFERNO) coupled to chemistry, aerosol and radiation schemes. We also show results from sensitivity studies analysing the influence of model process drivers on high latitude wildfires and their impacts on atmospheric composition over the recent past, including from changes in climate, socio-economic factors and underlying vegetation properties.
The baseline configuration of UKESM coupled with INFERNO shows an underestimation of burnt area from high latitude wildfires over the period 2000 to 2015 compared to that reported by GFED4s. The sensitivity scenarios show that this underestimation is found to be strongly driven by the human suppression factor included within INFERNO. The underestimation in burnt area is also reflected in the emission of SLCFs from high latitude wildfires e.g. CO, with implications for both climate and air quality. The INFERNO fire scheme does not currently include the representation of peat fires, which are important sources in the high latitude. When we include a representation of SLCF emissions from high latitude peat fires, the magnitude and temporal variability of such emissions are much improved in the model and compare better with those in GFED4s. Including this additional source also increases the contribution of wildfires to particulate air pollution and the degradation in surface air quality simulated by the model over the northern high latitudes. The interactive fire model coupled within UKESM is shown to underestimate high latitude wildfires due to missing sources and the representation of human interactions in this region. This has important consequences for regional air quality and climate in an area of the world experiencing rapid changes to its climate.
How to cite: Turnock, S., Teixeira, J., Burton, C., Blackford, K., Arnold, S., and O'Connor, F.: The Sensitivity of High Latitude Wildfires and their impacts on Atmospheric Composition to underlying driving processes in the UK’s Earth System Model (UKESM), EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-9049, https://doi.org/10.5194/egusphere-egu25-9049, 2025.
The backscattering linear depolarization ratio (LDR) is a key parameter to identify particle types. Previous studies on smoke LDR have shown significant differences in their measurements, with the magnitudes varying widely under different study scenarios. Single-particle models involving internally mixed black carbon (BC) are applied to assess the LDR of smoke aerosols. However, handicaps have been found to apply such models to describe the bulk optical properties of aerosols, because of their overlook of the contribution of externally mixed organic carbon (OC) to the LDR. Smoke aerosols typically consist of a low proportion of BC particle population and a high proportion of externally mixed OC particle population. If the spherical assumption is applied to the calculation of smoke LDRs, the LDRs turned to be extremely low even approach zero. This leads to difficulties in explaining the observed variability and higher levels of smoke LDR. We conducted a prescribed burning experiment in Xichang, Sichuan Province, China, and did onsite measurement on the LDR of smoke at a wavelength (λ) of 532 nm using atmospheric laser lidar. Field smoke particles were collected using a single-particle sampler and the morphology of particles was then characterized by the transmission electron microscope (TEM). The results indicated that the LDR of local smoke varied between 0 and 20.1%, with rapid fluctuations. The TEM images confirmed the coexistence of both internally mixed BC and externally mixed OC in the smoke aerosols, with OC displaying an ellipsoidal morphology even on copper grids. Using the discrete dipole approximation, we subsequently evaluated the LDR of individual BC and OC. Based on light scattering theory, we further quantified the bulk LDRs of the aerosol aerosols. The results shown that the smoke LDR ranged from 0.0% to 28.2% in λ = 532 nm while accounting for the effect of externally mixed OC. The LDR is slightly influenced by BC and is significantly affected by the externally mixed OC. Furthermore, the LDR is primarily governed by the morphology and particle size distribution of the externally mixed OC. It is concluded that the high levels and rapid variations in the LDRs of smoke can be largely attributed by the non-sphericity and particle size distribution of externally mixed OC. This study advances the methodologies for LDR measurements and evaluations of smoke aerosols from biomass burning.
How to cite: Qin, Z., Zhang, Q., Wang, H., and Zhang, Y.: The role of non-sphericity of externally mixed organic carbon in altering the backscattering linear depolarization ratio of smoke aerosols from biomass burning, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-9452, https://doi.org/10.5194/egusphere-egu25-9452, 2025.
Fire is a natural phenomenon that modulates form, function, diversity and distribution of plant species affecting ecosystem dynamics. Global warming and land use change are altering fire regimens potentially threating ecosystem functioning and species persistence. However, pyrogeographical studies aiming to understand differences across fire regimens are usually not considering the role played of plant functional traits. Here, based on a recent pyroregionalization in Italy and using species distribution data from the Italian National Forest Inventory and trait values from public databases we assessed if: 1) species distribution across different pyroregions is affected by fire regime, 2) species in different pyroregions exhibit distinct fire-related trait values, and, if so, 3) trait differences suggest better abilities to cope with fire in species distributed in more fire-prone regions (e.g. thicker bark). Our results tend to positively answer our questions suggesting the necessity of including fire-related traits when studying pyroregions. Noticeably, our study showed that the most fire-prone pyroregions collapse into one region from a functional perspective, with species characterized by highly similar trait values and indicative of fire adaptations.
How to cite: Costa-Saura, J. M., Midolo, G., Ricotta, C., Baudena, M., Calfapietra, C., Elia, M., Fiorucci, P., Mereu, S., Sirca, C., Spano, D., Vivaldo, G., and Ottaviani, G.: Fire proneness of Mediterranean pyroregions is positively linked to tree functional traits indicative of fire-modulated responses, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-10591, https://doi.org/10.5194/egusphere-egu25-10591, 2025.
Rising temperatures and changing climate conditions have increased wildfire risk across the world, including in regions such as The Netherlands that have not historically faced these threats. With this trend expected to continue, understanding risk perceptions among individuals with little to no wildfire experience becomes crucial for mitigating the impacts and designing effective risk communication strategies.
Recent advancements in Artificial Intelligence (AI) wildfire mapping tools have proven highly effective in identifying areas susceptible to wildfires, particularly in detecting low-probability incidents by uncovering subtle patterns often missed by traditional methods. For example, machine learning (ML) wildfire risk maps developed by MEJOR Technologies have accurately predicted wildfire locations in The Netherlands in the past. Despite the potential, the use of these tools as communication instruments to improve wildfire risk perception among the public remains largely unexplored.
Through an online randomised experiment conducted among a sample of residents in the Veluwe area of The Netherlands, we empirically assess how AI-generated labelling (AI label, ML label, or no label) and information presentation formats (map, text, or combined) affect individuals’ perceived wildfire risk. Additionally, we investigate whether perceived trustworthiness in technologies and emotion mediate these effects, providing deeper insights into the cognitive and affective processes that shape how individuals in this area perceive wildfire risk. By leveraging our results, policy makers and AI mapping developers can design effective communication interventions and improve public preparedness in the face of wildfires. While our findings are specific to wildfires in the Veluwe area, they may also hold relevance for understanding the perception of other low-probability hazards among individuals with little to no prior exposure.
How to cite: Mijailovic, M., Ranucci, A., Geib, C., Nardelli, B., Koppen, E., Tamura, F., and Kandathil Parambil, P.: Using AI-enabled wildfire risk maps to communicate risk: the role of labelling, information presentation, perceived trustworthiness and emotion in shaping perceived risk in Veluwe, Netherlands, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-13378, https://doi.org/10.5194/egusphere-egu25-13378, 2025.
Fire plays a critical role in the Earth system, both as a driver and responder to climate change. Variations in vegetation cover and ignition patterns, influenced by climate, affect fire behavior, while fire emissions impact climate by altering radiative fluxes and cloud properties. Despite these interactions, most global climate models fail to fully represent the dynamic interplay between vegetation, fire, and climate. In this study, we leverage the prognostic fire module from GFDL’s Land Model (LM4.1), which includes dynamic vegetation processes, to interactively calculate biomass burning emissions and injection heights. Emissions are then coupled with the atmospheric chemistry and aerosol component (AM4.1) in GFDL’s Earth System Model version 4.1 (ESM4.1). The model calculates fire radiative power (FRP) from fire spread rates and fuel content, using it alongside atmospheric parameters like boundary layer height and Brunt-Väisälä frequency in the Sofiev injection height scheme. Fire emissions are calculated using carbon release rates from biomass estimated by the land model and emission factors from Akagi et al. (2011) and Andreae and Merlet (2001), and these emissions are integrated directly into the atmospheric model for interactive coupling.
We conducted a coupled simulation in AMIP mode and compared the modeled emissions with the observation-based Global Fire Emissions Database (GFED4.1s). Preliminary results show a promising agreement for global fire emissions of trace gases and aerosols during the 1997–2014 period, with seasonal variability falling within the error margins of observed emissions. We then compared results from interactive fire emissions experiment with a fixed fire emission experiment to analyze the direct radiative effects of fire-emitted aerosols. By treating fire emissions as an interactive component of the Earth system, rather than as a prescribed external forcing, this approach enables a more comprehensive representation of fire-climate feedback and enhances the assessment of radiative effects from fire aerosols.
How to cite: Pouyaei, A., Ginoux, P., Shevliakova, E., and Malyshev, S.: Interactive Fire Emissions Coupled with Climate and Chemistry in GFDL’s Earth System Model version 4.1, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-13861, https://doi.org/10.5194/egusphere-egu25-13861, 2025.
Wildfires emit a complex mixture of trace gases and aerosols that significantly impact air quality, climate, and atmospheric chemistry. Key trace gases include carbon dioxide (CO₂), carbon monoxide (CO), nitric oxide (NO), methane (CH₄), and volatile organic compounds (VOCs). Wildfire-generated aerosols predominantly consist of organic carbon (OC), black carbon (BC), and secondary organic aerosols (SOA). Over recent decades, the frequency and intensity of wildfires, particularly in the western United States, have risen due to warmer temperatures and prolonged periods of drought. This trend has led to increased fire activity and smoke emissions, causing wildfires to be a growing contributor to regional and global aerosol forcing, in turn affecting the Earth's radiation budget and climate system. However, substantial uncertainties remain in estimating the composition and quantity of wildfire emissions.
Large variability in biomass burning aerosol estimates across different fire emission inventories poses challenges for accurate air quality and climate impact assessments. To address these challenges, we leverage observational data from the FIREX-AQ and WE-CAN campaigns to investigate how wildfire characteristics such as individual fire size, fire radiative power, and fuel composition influence the chemical composition of wildfire emissions, particularly VOCs. We then develop and apply an artificial neural network in tandem with dimensionality reduction methods to estimate smoke chemistry utilizing fire characteristics. Our machine learning model's results are compared with existing observations and current fire emission inventories to improve our understanding of wildfire emissions and their impacts.
How to cite: Zhang, Y., Barth, M., Emmons, L., Kelp, M., Juliano, T., Tang, W., Hornbrook, R., and Apel, E.: Understanding Wildfire Emissions: From Composition to Variability, and their Link to Fire Characteristics , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-14193, https://doi.org/10.5194/egusphere-egu25-14193, 2025.
The recent surge in forest fires has significantly impacted atmospheric chemistry, carbon cycles, and climate. Wildfires release CO2 along with various reactive species such as CO, volatile organics, and nitrogen oxides. While the effects of CO2 emissions on the carbon cycle and climate, as well as the impact of reactive species emissions on air quality and health, have been extensively studied, this research demonstrates that reactive species emitted from wildfires create a positive climate feedback through the “fire-chemistry-methane” mechanism. In this process, chemical reactions of reactive carbon species suppress the concentration of hydroxyl radicals, extending the lifetime of heat-trapping methane. The significance of this feedback is suggested by observations of multiple proxy gases for global atmospheric oxidation (i.e., methyl chloroform, methane, and CO) during recent extreme forest fire events. By coupling a fire-ecosystem model and an atmospheric chemistry model, we quantify the effect of this feedback in the future. We find that additional warming caused by this mechanism rivals that of wetland methane feedback and fire CO2 feedback by the 2050s under an intermediate climate scenario. Our analysis highlights the critical role of atmospheric chemistry in regulating fire-climate interactions and the methane budget.
How to cite: Chen, W., Zhang, Y., Zou, Y., and Zhang, Z.: Climate feedback of forest fires amplified by atmospheric chemistry, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16179, https://doi.org/10.5194/egusphere-egu25-16179, 2025.
The interactions of different components of the Earth system, such as between the biosphere and the atmosphere, are still poorly understood. A major issue is understanding the consequences of increasing wildfire activity in a changing climate. Smoke particles and gases emitted from such fires affect air quality and the Earth’s radiation balance, and can potentially affect the formation of clouds and precipitation. Understanding links between biodiversity and type of vegetation, smoke emission and the atmospheric distribution and processing of these particles and gases is key for assessing potential impacts and future changes. Addressing the depth of processes in the interconnected atmosphere-climate-vegetation system requires a combination of expertise from various scientific disciplines. The new Leibniz ScienceCampus “Smoke and bioaerosols: Release, processes, and impacts in a changing climate” (BioSmoke) located in Leipzig, Germany will combine expertise in atmospheric and biodiversity research as well as atmospheric processes at several research institutions including the Leibniz Institute for Tropospheric Research and Leipzig University to study effects of the release of aerosol particles from vegetation. To this end, combustion experiments in the laboratory, field measurements of aerosol properties, and remote sensing and modelling of particle emission, transport, and atmospheric effects are envisioned. We will present an overview of the planned projects within the ScienceCampus.
How to cite: Tegen, I., Wagner, R., and Tesche, M.: Introducing the Leibniz Science Campus “Smoke and bioaerosols: Release, processes, and impacts in a changing climate”, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17284, https://doi.org/10.5194/egusphere-egu25-17284, 2025.
Peatlands, despite covering only 3% of the terrestrial surface, are one of the world's most important carbon storage environments, accumulating around 25% of the total soil carbon. However, climate change is increasing the vulnerability of these carbon-rich ecosystems to fire, with potentially severe implications for the global climate. Warmer and drier conditions, driven by climate change, are expected to intensify and increase the frequency of peat fires, potentially transforming peatlands from carbon sinks into net sources of greenhouse gas emissions. Such a shift could trigger a positive feedback loop, accelerating climate change through the release of vast amounts of sequestered carbon into the atmosphere.
While incorporating peatland fire feedbacks into Earth System Models (ESMs) is essential for accurate climate projections, the majority of the existing models lack a representation of peat fires, limiting their ability to predict future climate dynamics effectively. Understanding the smouldering behaviour of organic soils, their ignition probability, and how these processes can be represented at a global scale is essential. The current state-of-the-art approach to compute peat combustibility, established by Frandsen (1997) and applied in recent peat fire modelling efforts (e.g., INFERNO-peat), relies on a parameterization derived from a single peat type, hampering its global applicability. Frandsen (1997), by conducting experiments on natural peat samples developed an empirical model for smouldering ignition probability based on three key properties of peat: moisture content, inorganic content, and bulk density.
Our study proposes an improved method for calculating peat combustibility by optimizing the coefficients in Frandsen’s model and investigating the ignition limits of diverse peat samples. The optimization process utilized experimental data from seven distinct peat types. First, we established through inverse modelling a link between inorganic content, bulk density and critical moisture content, the moisture threshold above which smouldering cannot be self-sustained. Then we determined the probability distribution of self-sustained smouldering, as a function of moisture content, around the critical moisture content, also employing inverse modelling. The combination of both optimizations yielded consistent coefficients, providing a more robust framework for modelling peat ignition probability.
By improving the representation of peat ignition probability using experimental data from both previous studies and our own experiments, this work aims to upgrade the simulation of peat fires in fire models and ESMs, enhancing our understanding of the impacts of such fires on future atmospheric composition, radiative forcing, and climate.
How to cite: Tarasi, D., Kasoar, M., Mulyasih, H., Castagna, A., Rein, G., and Voulgarakis, A.: An improved approach for simulating peat ignition probability using experimental data, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-18848, https://doi.org/10.5194/egusphere-egu25-18848, 2025.
Wildfires increasingly threaten European ecosystems and communities, highlighting the necessity for effective predictive metrics to enhance fire risk management strategies. This study aims to compare the effectiveness of Vapor Pressure Deficit (VPD) and the Fire Weather Index (FWI) in forecasting wildfire occurrence and the extent of burned areas across various European forest types. Utilizing the European Forest Fire Information System (EFFIS) for comprehensive fire event data and the ERA5 reanalysis dataset from the European Centre for Medium-Range Weather Forecasts (ECMWF) for meteorological variables, daily VPD and FWI values will be derived for multiple fire seasons spanning from 2000 to 2024.
The research will explore how VPD and FWI each predict wildfire occurrence and burned area, with a focus on different forest types are categorized according to the CORINE Land Cover classification into broadleaf, conifer, and mixed forests while encompassing a range of climatic regions across Europe. VPD calculation methods are generally more straightforward and require fewer input parameters. In contrast FWI system is more complex, requiring a broader range of input data to compute its numerous indices.
By comparing these two metrics across diverse forest types and biomes, the study seeks to determine the most effective indicators for wildfire prediction in Europe. The findings are intended to inform policymakers and fire management agencies, aiding in the development of targeted early warning systems and adaptive fire management strategies. This comparative assessment will contribute to a deeper understanding of the climatic drivers of wildfires and support efforts to mitigate their impacts under changing environmental conditions.
How to cite: Shatto, C. and Samimi, C.: Comparative Assessment of Vapor Pressure Deficit and Fire Weather Index in Predicting Wildfire Occurrence and Burned Area Across European Forest Types Using EFFIS and ERA5 Data, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-19262, https://doi.org/10.5194/egusphere-egu25-19262, 2025.
After a wildfire event, ashes and pollutants from burns are transported to public supply reservoirs and other water systems, altering the physical and chemical properties of the water. Turbidity is a water parameter that can be applied in environmental monitoring studies to assess water quality in public supply reservoirs, especially in fire-prone regions such as the Brazilian Cerrado. So, this work aims to answer the following question: What is the impact of the increase in burned area on water turbidity in public supply reservoirs? This study aims to investigate the relationship between environmental variables obtained through remote sensing, such as the burned area product (MODIS-MCD64A1) and turbidity data derived from the red band (620-670 nm) of MODIS Terra Surface Reflectance (Daily Global, 250m resolution), using a global algorithm and statistical analyses to derive insights over the period from 2001 to 2023 in public supply reservoirs of Cerrado.There is variability in both positive and negative turbidity anomalies from 2001 to 2023. However, in some years, positive turbidity anomalies were observed along burned areas. The insights provide the initial understanding of the relationship between burned areas and water quality, and also provide valuable support for water supply managers and the public.
How to cite: da Silva Nemirovsky, A. K., Augusto Sander de Carvalho, L., and Libonati, R.: Assessing increased turbidity in reservoirs due to wildfires, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-20164, https://doi.org/10.5194/egusphere-egu25-20164, 2025.
