Articles | Volume 26, issue 6
https://doi.org/10.5194/nhess-26-2871-2026
© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
https://doi.org/10.5194/nhess-26-2871-2026
© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Research article
| 
17 Jun 2026
Research article |
| 17 Jun 2026
| 17 Jun 2026
Predicting spatio-temporal wildfire propagation with dynamic firebreaks
Jiahe Zheng
Department of Mathematics, Imperial College London, London SW7 2AZ, UK
Zhengsen Xu
Schulich School of Engineering, Department of Geomatics Engineering, University of Calgary, 2500 University Dr NW, Calgary, T2N1N4, Alberta, Canada
Rossella Arcucci
Data Science Institute, Department of Computing, Imperial College London, London SW7 2BX, UK
Department of Earth Science & Engineering, Imperial College London, London SW7 2BX, UK
Sandy P. Harrison
Leverhulme Centre for Wildfires, Environment, and Society, London SW7 2AZ, UK
Geography & Environmental Science, University of Reading, Reading RG6 6EU, UK
Lincoln Linlin Xu
Schulich School of Engineering, Department of Geomatics Engineering, University of Calgary, 2500 University Dr NW, Calgary, T2N1N4, Alberta, Canada
Sibo Cheng
CORRESPONDING AUTHOR
CEREA, ENPC, EDF R&D, Institut Polytechnique de Paris, France
Related authors
No articles found.
Mengmeng Liu, Iain Colin Prentice, and Sandy P. Harrison
Clim. Past, 22, 205–226, https://doi.org/10.5194/cp-22-205-2026, https://doi.org/10.5194/cp-22-205-2026, 2026
Short summary
Short summary
Dansgaard-Oeschger events were large and rapid climate transitions that occurred multiple times during the last glacial period. We reconstructed the global pattern during these events using fossil pollen records, and found a north-south antiphasing of temperature change, and reduced seasonality in the northern extratropics.
Wenbo Yu, Anirbit Ghosh, Tobias Sebastian Finn, Rossella Arcucci, Marc Bocquet, and Sibo Cheng
Geosci. Model Dev., 19, 1027–1054, https://doi.org/10.5194/gmd-19-1027-2026, https://doi.org/10.5194/gmd-19-1027-2026, 2026
Short summary
Short summary
We introduce the first denoising diffusion model for wildfire spread prediction, a new kind of generative AI model that learns to simulate fires not just as one fixed outcome, but as a range of possible scenarios. This allows us to capture the inherent uncertainty of wildfire dynamics. Our model produces ensembles of forecasts that reflect physically meaningful distributions of where fire might go next.
Marc Bocquet, Tobias Sebastian Finn, Sibo Cheng, and Alban Farchi
EGUsphere, https://doi.org/10.5194/egusphere-2026-245, https://doi.org/10.5194/egusphere-2026-245, 2026
Short summary
Short summary
Deep learning has been used to discover new data assimilation methods to track chaotic dynamical systems. Strikingly, these data assimilation networks can match the accuracy of ensemble-based methods using only a single state forecast. This paper first investigates the reasons for this efficacy, and then shows that their performance in more nonlinear regimes matches that of the most accurate scalable data assimilation methods, with greater efficiency and robustness.
Jierong Zhao, Boya Zhou, Sandy P. Harrison, and Colin Prentice
Earth Syst. Dynam., 16, 1655–1669, https://doi.org/10.5194/esd-16-1655-2025, https://doi.org/10.5194/esd-16-1655-2025, 2025
Short summary
Short summary
We used eco-evolutionary optimality modelling to examine how climate and CO2 impacted vegetation at the Last Glacial Maximum (LGM; 21 000 years ago) and the mid-Holocene (MH; 6000 years ago). Low CO2 at the LGM was as important as climate in reducing tree cover and productivity and in increasing C4 plant abundance. Climate had positive effects on MH vegetation, but the low CO2 was a constraint on plant growth. These results show it is important to consider changing CO2 to model ecosystem changes.
Luke Sweeney, Sandy P. Harrison, and Marc Vander Linden
Biogeosciences, 22, 4903–4922, https://doi.org/10.5194/bg-22-4903-2025, https://doi.org/10.5194/bg-22-4903-2025, 2025
Short summary
Short summary
Changes in tree cover across Europe during the Holocene are reconstructed from fossil pollen data using a model developed with modern observations of tree cover and modern pollen assemblages. There is a rapid increase in tree cover after the last glacial period, with maximum cover during the mid-Holocene and a decline thereafter; the timing of the maximum and the speed of the increase and subsequent decrease vary regionally, likely reflecting differences in climate trajectories and human influence.
Arianna Olivelli, Rossella Arcucci, Mark Rehkämper, and Tina van de Flierdt
Earth Syst. Sci. Data, 17, 3679–3699, https://doi.org/10.5194/essd-17-3679-2025, https://doi.org/10.5194/essd-17-3679-2025, 2025
Short summary
Short summary
In this study, we use machine learning models to produce the first global maps of Pb concentrations and isotope compositions in the ocean. In line with observations, we find that (i) the surface Indian Ocean has the highest levels of pollution, (ii) pollution from previous decades is sinking in the North Atlantic and Pacific oceans, and (iii) waters carrying Pb pollution are spreading from the Southern Ocean throughout the Southern Hemisphere at intermediate depths.
Kieran M. R. Hunt and Sandy P. Harrison
Clim. Past, 21, 1–26, https://doi.org/10.5194/cp-21-1-2025, https://doi.org/10.5194/cp-21-1-2025, 2025
Short summary
Short summary
In this study, we train machine learning models on tree rings, speleothems, and instrumental rainfall to estimate seasonal monsoon rainfall over India over the last 500 years. Our models highlight multidecadal droughts in the mid-17th and 19th centuries, and we link these to historical famines. Using techniques from explainable AI (artificial intelligence), we show that our models use known relationships between local hydroclimate and the monsoon circulation.
Fang Li, Xiang Song, Sandy P. Harrison, Jennifer R. Marlon, Zhongda Lin, L. Ruby Leung, Jörg Schwinger, Virginie Marécal, Shiyu Wang, Daniel S. Ward, Xiao Dong, Hanna Lee, Lars Nieradzik, Sam S. Rabin, and Roland Séférian
Geosci. Model Dev., 17, 8751–8771, https://doi.org/10.5194/gmd-17-8751-2024, https://doi.org/10.5194/gmd-17-8751-2024, 2024
Short summary
Short summary
This study provides the first comprehensive assessment of historical fire simulations from 19 Earth system models in phase 6 of the Coupled Model Intercomparison Project (CMIP6). Most models reproduce global totals, spatial patterns, seasonality, and regional historical changes well but fail to simulate the recent decline in global burned area and underestimate the fire response to climate variability. CMIP6 simulations address three critical issues of phase-5 models.
Nikita Kaushal, Franziska A. Lechleitner, Micah Wilhelm, Khalil Azennoud, Janica C. Bühler, Kerstin Braun, Yassine Ait Brahim, Andy Baker, Yuval Burstyn, Laia Comas-Bru, Jens Fohlmeister, Yonaton Goldsmith, Sandy P. Harrison, István G. Hatvani, Kira Rehfeld, Magdalena Ritzau, Vanessa Skiba, Heather M. Stoll, József G. Szűcs, Péter Tanos, Pauline C. Treble, Vitor Azevedo, Jonathan L. Baker, Andrea Borsato, Sakonvan Chawchai, Andrea Columbu, Laura Endres, Jun Hu, Zoltán Kern, Alena Kimbrough, Koray Koç, Monika Markowska, Belen Martrat, Syed Masood Ahmad, Carole Nehme, Valdir Felipe Novello, Carlos Pérez-Mejías, Jiaoyang Ruan, Natasha Sekhon, Nitesh Sinha, Carol V. Tadros, Benjamin H. Tiger, Sophie Warken, Annabel Wolf, Haiwei Zhang, and SISAL Working Group members
Earth Syst. Sci. Data, 16, 1933–1963, https://doi.org/10.5194/essd-16-1933-2024, https://doi.org/10.5194/essd-16-1933-2024, 2024
Short summary
Short summary
Speleothems are a popular, multi-proxy climate archive that provide regional to global insights into past hydroclimate trends with precise chronologies. We present an update to the SISAL (Speleothem Isotopes
Synthesis and AnaLysis) database, SISALv3, which, for the first time, contains speleothem trace element records, in addition to an update to the stable isotope records available in previous versions of the database, cumulatively providing data from 365 globally distributed sites.
Synthesis and AnaLysis) database, SISALv3, which, for the first time, contains speleothem trace element records, in addition to an update to the stable isotope records available in previous versions of the database, cumulatively providing data from 365 globally distributed sites.
Huiying Xu, Han Wang, Iain Colin Prentice, and Sandy P. Harrison
Biogeosciences, 20, 4511–4525, https://doi.org/10.5194/bg-20-4511-2023, https://doi.org/10.5194/bg-20-4511-2023, 2023
Short summary
Short summary
Leaf carbon (C) and nitrogen (N) are crucial elements in leaf construction and physiological processes. This study reconciled the roles of phylogeny, species identity, and climate in stoichiometric traits at individual and community levels. The variations in community-level leaf N and C : N ratio were captured by optimality-based models using climate data. Our results provide an approach to improve the representation of leaf stoichiometry in vegetation models to better couple N with C cycling.
Esmeralda Cruz-Silva, Sandy P. Harrison, I. Colin Prentice, Elena Marinova, Patrick J. Bartlein, Hans Renssen, and Yurui Zhang
Clim. Past, 19, 2093–2108, https://doi.org/10.5194/cp-19-2093-2023, https://doi.org/10.5194/cp-19-2093-2023, 2023
Short summary
Short summary
We examined 71 pollen records (12.3 ka to present) in the eastern Mediterranean, reconstructing climate changes. Over 9000 years, winters gradually warmed due to orbital factors. Summer temperatures peaked at 4.5–5 ka, likely declining because of ice sheets. Moisture increased post-11 kyr, remaining high from 10–6 kyr before a slow decrease. Climate models face challenges in replicating moisture transport.
Olivia Haas, Iain Colin Prentice, and Sandy P. Harrison
Biogeosciences, 20, 3981–3995, https://doi.org/10.5194/bg-20-3981-2023, https://doi.org/10.5194/bg-20-3981-2023, 2023
Short summary
Short summary
We quantify the impact of CO2 and climate on global patterns of burnt area, fire size, and intensity under Last Glacial Maximum (LGM) conditions using three climate scenarios. Climate change alone did not produce the observed LGM reduction in burnt area, but low CO2 did through reducing vegetation productivity. Fire intensity was sensitive to CO2 but strongly affected by changes in atmospheric dryness. Low CO2 caused smaller fires; climate had the opposite effect except in the driest scenario.
Giulia Mengoli, Sandy P. Harrison, and I. Colin Prentice
EGUsphere, https://doi.org/10.5194/egusphere-2023-1261, https://doi.org/10.5194/egusphere-2023-1261, 2023
Preprint archived
Short summary
Short summary
Soil water availability affects plant carbon uptake by reducing leaf area and/or by closing stomata, which reduces its efficiency. We present a new formulation of how climatic dryness reduces both maximum carbon uptake and the soil-moisture threshold below which it declines further. This formulation illustrates how plants adapt their water conservation strategy to thrive in dry climates, and is step towards a better representation of soil-moisture effects in climate models.
Caili Zhong, Sibo Cheng, Matthew Kasoar, and Rossella Arcucci
Nat. Hazards Earth Syst. Sci., 23, 1755–1768, https://doi.org/10.5194/nhess-23-1755-2023, https://doi.org/10.5194/nhess-23-1755-2023, 2023
Short summary
Short summary
This paper introduces a digital twin fire model using machine learning techniques to improve the efficiency of global wildfire predictions. The proposed model also manages to efficiently adjust the prediction results thanks to data assimilation techniques. The proposed digital twin runs 500 times faster than the current state-of-the-art physics-based model.
Mengmeng Liu, Yicheng Shen, Penelope González-Sampériz, Graciela Gil-Romera, Cajo J. F. ter Braak, Iain Colin Prentice, and Sandy P. Harrison
Clim. Past, 19, 803–834, https://doi.org/10.5194/cp-19-803-2023, https://doi.org/10.5194/cp-19-803-2023, 2023
Short summary
Short summary
We reconstructed the Holocene climates in the Iberian Peninsula using a large pollen data set and found that the west–east moisture gradient was much flatter than today. We also found that the winter was much colder, which can be expected from the low winter insolation during the Holocene. However, summer temperature did not follow the trend of summer insolation, instead, it was strongly correlated with moisture.
Yicheng Shen, Luke Sweeney, Mengmeng Liu, Jose Antonio Lopez Saez, Sebastián Pérez-Díaz, Reyes Luelmo-Lautenschlaeger, Graciela Gil-Romera, Dana Hoefer, Gonzalo Jiménez-Moreno, Heike Schneider, I. Colin Prentice, and Sandy P. Harrison
Clim. Past, 18, 1189–1201, https://doi.org/10.5194/cp-18-1189-2022, https://doi.org/10.5194/cp-18-1189-2022, 2022
Short summary
Short summary
We present a method to reconstruct burnt area using a relationship between pollen and charcoal abundances and the calibration of charcoal abundance using modern observations of burnt area. We use this method to reconstruct changes in burnt area over the past 12 000 years from sites in Iberia. We show that regional changes in burnt area reflect known changes in climate, with a high burnt area during warming intervals and low burnt area when the climate was cooler and/or wetter than today.
Sandy P. Harrison, Roberto Villegas-Diaz, Esmeralda Cruz-Silva, Daniel Gallagher, David Kesner, Paul Lincoln, Yicheng Shen, Luke Sweeney, Daniele Colombaroli, Adam Ali, Chéïma Barhoumi, Yves Bergeron, Tatiana Blyakharchuk, Přemysl Bobek, Richard Bradshaw, Jennifer L. Clear, Sambor Czerwiński, Anne-Laure Daniau, John Dodson, Kevin J. Edwards, Mary E. Edwards, Angelica Feurdean, David Foster, Konrad Gajewski, Mariusz Gałka, Michelle Garneau, Thomas Giesecke, Graciela Gil Romera, Martin P. Girardin, Dana Hoefer, Kangyou Huang, Jun Inoue, Eva Jamrichová, Nauris Jasiunas, Wenying Jiang, Gonzalo Jiménez-Moreno, Monika Karpińska-Kołaczek, Piotr Kołaczek, Niina Kuosmanen, Mariusz Lamentowicz, Martin Lavoie, Fang Li, Jianyong Li, Olga Lisitsyna, José Antonio López-Sáez, Reyes Luelmo-Lautenschlaeger, Gabriel Magnan, Eniko Katalin Magyari, Alekss Maksims, Katarzyna Marcisz, Elena Marinova, Jenn Marlon, Scott Mensing, Joanna Miroslaw-Grabowska, Wyatt Oswald, Sebastián Pérez-Díaz, Ramón Pérez-Obiol, Sanna Piilo, Anneli Poska, Xiaoguang Qin, Cécile C. Remy, Pierre J. H. Richard, Sakari Salonen, Naoko Sasaki, Hieke Schneider, William Shotyk, Migle Stancikaite, Dace Šteinberga, Normunds Stivrins, Hikaru Takahara, Zhihai Tan, Liva Trasune, Charles E. Umbanhowar, Minna Väliranta, Jüri Vassiljev, Xiayun Xiao, Qinghai Xu, Xin Xu, Edyta Zawisza, Yan Zhao, Zheng Zhou, and Jordan Paillard
Earth Syst. Sci. Data, 14, 1109–1124, https://doi.org/10.5194/essd-14-1109-2022, https://doi.org/10.5194/essd-14-1109-2022, 2022
Short summary
Short summary
We provide a new global data set of charcoal preserved in sediments that can be used to examine how fire regimes have changed during past millennia and to investigate what caused these changes. The individual records have been standardised, and new age models have been constructed to allow better comparison across sites. The data set contains 1681 records from 1477 sites worldwide.
Alexander Kuhn-Régnier, Apostolos Voulgarakis, Peer Nowack, Matthias Forkel, I. Colin Prentice, and Sandy P. Harrison
Biogeosciences, 18, 3861–3879, https://doi.org/10.5194/bg-18-3861-2021, https://doi.org/10.5194/bg-18-3861-2021, 2021
Short summary
Short summary
Along with current climate, vegetation, and human influences, long-term accumulation of biomass affects fires. Here, we find that including the influence of antecedent vegetation and moisture improves our ability to predict global burnt area. Additionally, the length of the preceding period which needs to be considered for accurate predictions varies across regions.
Cited articles
Alexandridis, A., Vakalis, D., Siettos, C. I., and Bafas, G. V.: A cellular automata model for forest fire spread prediction: the case of the wildfire that swept through Spetses Island in 1990, Appl. Math. Comput., 204, 191–201, https://doi.org/10.1016/j.amc.2008.06.046, 2008. a, b, c, d
Alexandridis, A., Russo, L., Vakalis, D., Bafas, G., and Siettos, C.: Wildland fire spread modelling using cellular automata: evolution in large-scale spatially heterogeneous environments under fire suppression tactics, Int. J. Wildland Fire, 20, 633–647, https://doi.org/10.1071/WF09119, 2011. a, b
Altamimi, A., Lagoa, C., Borges, J. G., McDill, M. E., Andriotis, C., and Papakonstantinou, K.: Large-scale wildfire mitigation through deep reinforcement learning, Frontiers in Forests and Global Change, 5, 734330, https://doi.org/10.3389/ffgc.2022.734330, 2022. a, b
Cheng, S., Jin, Y., Harrison, S. P., Quilodrán-Casas, C., Prentice, I. C., Guo, Y.-K., and Arcucci, R.: Parameter flexible wildfire prediction using machine learning techniques: Forward and inverse modelling, Remote Sens.-Basel, 14, 3228, https://doi.org/10.3390/rs14133228, 2022a. a
Cheng, S., Prentice, I. C., Huang, Y., Jin, Y., Guo, Y.-K., and Arcucci, R.: Data-driven surrogate model with latent data assimilation: application to wildfire forecasting, J. Comput. Phys., 464, 111302, https://doi.org/10.1016/j.jcp.2022.111302, 2022b. a, b, c
Cheng, S., Guo, Y., and Arcucci, R.: A generative model for surrogates of spatial-temporal wildfire nowcasting, IEEE Transactions on Emerging Topics in Computational Intelligence, 7, 1420–1430, https://doi.org/10.1109/TETCI.2023.3298535, 2023. a
Cunningham, C. X., Williamson, G. J., and Bowman, D. M.: Increasing frequency and intensity of the most extreme wildfires on Earth, Nature Ecology & Evolution, 8, 1420–1425, https://doi.org/10.1038/s41559-024-02452-2, 2024. a
Drury, S. A., Rauscher, H. M., Banwell, E. M., Huang, S., and Lavezzo, T. L.: The interagency fuels treatment decision support system: functionality for fuels treatment planning, Fire Ecol., 12, 103–123, https://doi.org/10.4996/fireecology.1201103, 2016. a, b, c
Gimenez, A., Pastor, E., Zarate, L., Planas, E., and Arnaldos, J.: Long-term forest fire retardants: a review of quality, effectiveness, application and environmental considerations, Int. J. Wildland Fire, 13, 1–15, https://doi.org/10.1071/WF03001, 2004. a
Goldberg, E.: Industry Standard: Understanding the Science of Chemical Retardants, https://www.iawfonline.org/article/ industry-standard-understanding-the-science-of-chemical-retardants/ (last access: 31 May 2026), 2022. a
Hersbach, H., Bell, B., Berrisford, P., Hirahara, S., Horányi, A., Muñoz-Sabater, J., Nicolas, J., Peubey, C., Radu, R., Schepers, D., Simmons, A., Soci, C., Abdalla, S., Abellan, X., Balsamo, G., Bechtold, P., Biavati, G., Bidlot, J., Bonavita, M., De Chiara, G., Dahlgren, P., Dee, D., Diamantakis, M., Dragani, R., Flemming, J., Forbes, R., Fuentes, M., Geer, A., Haimberger, L., Healy, S., Hogan, R. J., Hólm, E., Janisková, M., Keeley, S., Laloyaux, P., Lopez, P., Lupu, C., Radnoti, G., de Rosnay, P., Rozum, I., Vamborg, F., Villaume, S., and Thépaut, J.-N.: The ERA5 global reanalysis, Q. J. Roy. Meteor. Soc., 146, 1999–2049, https://doi.org/10.1002/qj.3803, 2020. a
Hilton, J. E., Miller, C., Sullivan, A. L., and Rucinski, C.: Effects of spatial and temporal variation in environmental conditions on simulation of wildfire spread, Environ. Modell.Softw., 67, 118–127, https://doi.org/10.1016/j.envsoft.2015.01.015, 2015. a
Huot, F., Hu, R. L., Goyal, N., Sankar, T., Ihme, M., and Chen, Y.-F.: Next day wildfire spread: a machine learning dataset to predict wildfire spreading from remote-sensing data, IEEE T. Geosci. Remote, 60, 1–13, https://doi.org/10.1109/TGRS.2022.3192974, 2022. a, b
Jain, P., Coogan, S. C., Subramanian, S. G., Crowley, M., Taylor, S., and Flannigan, M. D.: A review of machine learning applications in wildfire science and management, Environ. Rev., 28, 478–505, https://doi.org/10.1139/er-2020-0019, 2020. a
Kondylatos, S., Prapas, I., Ronco, M., Papoutsis, I., Camps-Valls, G., Piles, M., Fernández-Torres, M.-Á., and Carvalhais, N.: Wildfire danger prediction and understanding with deep learning, Geophys. Res. Lett., 49, e2022GL099368, https://doi.org/10.1029/2022GL099368, 2022. a
Kondylatos, S., Prapas, I., Camps-Valls, G., and Papoutsis, I.: Mesogeos: a multi-purpose dataset for data-driven wildfire modeling in the Mediterranean, Zenodo, https://doi.org/10.5281/zenodo.7473331, 2023. a
Liang, H., Zhang, M., and Wang, H.: A neural network model for wildfire scale prediction using meteorological factors, IEEE Access, 7, 176746–176755, https://doi.org/10.1109/ACCESS.2019.2957837, 2019. a
Liu, C., Fu, R., Xiao, D., Stefanescu, R., Sharma, P., Zhu, C., Sun, S., and Wang, C.: Enkf data-driven reduced order assimilation system, Eng. Anal. Bound. Elem., 139, 46–55, https://doi.org/10.1016/j.enganabound.2022.02.016, 2022. a
Masrur, A., Yu, M., and Taylor, A.: Capturing and interpreting wildfire spread dynamics: attention-based spatiotemporal models using ConvLSTM networks, Ecol. Inform., 82, 102760, https://doi.org/10.1016/j.ecoinf.2024.102760, 2024. a
Meng, Q., Lu, H., Huai, Y., Xu, H., and Yang, S.: Forest fire spread simulation and fire extinguishing visualization research, Forests, 14, 1371, https://doi.org/10.3390/f14071371, 2023. a
Miao, X., Li, J., Mu, Y., He, C., Ma, Y., Chen, J., Wei, W., and Gao, D.: Time series forest fire prediction based on improved transformer, Forests, 14, 1596, https://doi.org/10.3390/f14081596, 2023. a
Murray, L., Castillo, T., Carrasco, J., Weintraub, A., Weber, R., de Diego, I. M., González, J. R., and García-Gonzalo, J.: Advancing forest fire prevention: deep reinforcement learning for effective firebreak placement, arXiv [preprint], https://doi.org/10.48550/arXiv.2404.08523, 2024. a
Mutthulakshmi, K., Wee, M. R. E., Wong, Y. C. K., Lai, J. W., Koh, J. M., Acharya, U. R., and Cheong, K. H.: Simulating forest fire spread and fire-fighting using cellular automata, Chinese J. Phys., 65, 642–650, https://doi.org/10.1016/j.cjph.2020.04.001, 2020. a, b, c
Natekar, S., Patil, S., Nair, A., and Roychowdhury, S.: Forest fire prediction using LSTM, in: 2021 2nd International Conference for Emerging Technology (INCET), IEEE, https://doi.org/10.1109/INCET51464.2021.9456113, 2021. a, b
Neidermeier, A., Zagaria, C., Pampanoni, V., West, T., and Verburg, P.: Mapping opportunities for the use of land management strategies to address fire risk in Europe, J. Environ. Manage., 346, 118941, https://doi.org/10.1016/j.jenvman.2023.118941, 2023. a
Ntaimo, L., Zeigler, B. P., Vasconcelos, M. J., and Khargharia, B.: Forest Fire Spread and Suppression in DEVS, SIMULATION, 80, 479–500, https://doi.org/10.1177/0037549704050918, 2004. a
Oliveras, I., Prat-Guitart, N., Spadoni, G. L., Hsu, A., Fernandes, P., Puig-Gironès, R., Ascoli, D., Bilbao, B., Bacciu, V., Brotons, L., Carmenta, R., de Miguel, S., Gonçalves, L., Humphrey, G., Ibarnegaray, V., Jones, M., Machado, M., Millán, A., Falleiro, R., and Armenteras, D.: Integrated fire management as an adaptation and mitigation strategy to altered fire regimes, Communications Earth & Environment, 6, 202, https://doi.org/10.1038/s43247-025-02165-9, 2025. a
Pan, S., Cheng, A., Sun, Y., Kang, K., Pais, C., Zhou, Y., and Shen, Z.-J. M.: Using drone swarm to stop wildfire: A predict-then-optimize approach, arXiv [preprint], https://doi.org/10.48550/arXiv.2411.16144, 2024. a
Plucinski, M., Gould, J., McCarthy, G., and Hollis, J.: The Effectiveness and Efficiency of Aerial Firefighting in Australia, Part 1, Tech. Rep. A0701, Bushfire Cooperative Research Centre, ISBN 0-643-06534-2, 2007. a
Shi, X., Chen, Z., Wang, H., Yeung, D.-Y., Wong, W.-K., and Woo, W.-C.: Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcasting, in: Advances in Neural Information Processing Systems 28 (NeurIPS 2015), 802–810, https://proceedings.neurips.cc/paper/2015/hash/07563a3fe3bbe7e3ba84431ad9d055af-Abstract.html (last access: 6 June 2026), 2015. a
Singla, S., Mukhopadhyay, A., Wilbur, M., Diao, T., Gajjewar, V., Eldawy, A., Kochenderfer, M., Shachter, R., and Dubey, A.: WildfireDB: An Open-Source Dataset Connecting Wildfire Spread with Relevant Determinants, Zenodo, https://doi.org/10.5281/zenodo.5636429, 2021. a
Spadoni, G. L., Moris, J. V., Vacchiano, G., Elia, M., Garbarino, M., Sibona, E., Tomao, A., Barbati, A., Sallustio, L., Salvati, L., Ferrara, C., Francini, S., Bonis, E., Dalla Vecchia, I., Strollo, A., Di Leginio, M., Munafò, M., Chirici, G., Romano, R., Corona, P., Marchetti, M., Brunori, A., Motta, R., and Ascoli, D.: Active governance of agro-pastoral, forest and protected areas mitigates wildfire impacts in Italy, Sci. Total Environ., 890, 164281, https://doi.org/10.1016/j.scitotenv.2023.164281, 2023. a
UNEP: UNEP in 2022 [Annual Report], Tech. rep., United Nations Environment Programme, https://www.unep.org/resources/annual-report-2022 (last access: 6 June 2026), 2023. a
Weise, D. R. and Biging, G. S.: A qualitative comparison of fire spread models incorporating wind and slope effects, Forest Sci., 43, 170–180, https://doi.org/10.1093/forestscience/43.2.170, 1997. a
Wildfire Today: Ferguson Fire image, https://wildfiretoday.com/wp-content/uploads/2018/08/FergusonFire_130amPDT_8-3-2018.jpg (last access: 29 December 2025), 2018. a
Wildfire Today: Bear Fire image, https://wildfiretoday.com/tag/bear-fire/ (last access: 29 December 2025), 2020. a
Xu, Z., Li, J., Cheng, S., Rui, X., Zhao, Y., He, H., Guan, H., Sharma, A., Erxleben, M., Chang, R., and Xu, L. L.: Deep learning for wildfire risk prediction: integrating remote sensing and environmental data, ISPRS J. Photogramm., 227, 632–677, https://doi.org/10.1016/j.isprsjprs.2025.06.002, 2025. a
Yu, W., Ghosh, A., Finn, T. S., Arcucci, R., Bocquet, M., and Cheng, S.: A probabilistic approach to wildfire spread prediction using a denoising diffusion surrogate model, Geosci. Model Dev., 19, 1027–1054, https://doi.org/10.5194/gmd-19-1027-2026, 2026. a
Zheng, J.: Predicting spatio-temporal wildfire propagation with firebreak placement: code and data, Zenodo [code/data set], https://doi.org/10.5281/zenodo.16419810, 2025. a
Short summary
We introduce the first AI model that predicts wildfire spread with the placement of both permanent and temporary firebreaks. Our spatiotemporal model learns from simulation data to capture how fire interacts with changing suppression efforts over time. Our model runs fast enough for near real-time use and performs well across different wildfire events. This approach could lead to better tools for helping decision-makers understand where and when firebreaks are most effective.
We introduce the first AI model that predicts wildfire spread with the placement of both...
Altmetrics
Final-revised paper
Preprint