Register      Login
International Journal of Wildland Fire International Journal of Wildland Fire Society
Journal of the International Association of Wildland Fire
RESEARCH ARTICLE

Simulation-based high-resolution fire danger mapping using deep learning

Frédéric Allaire https://orcid.org/0000-0003-3564-1564 A B , Jean-Baptiste Filippi https://orcid.org/0000-0002-6244-0648 C * , Vivien Mallet A B and Florence Vaysse D
+ Author Affiliations
- Author Affiliations

A Institut national de recherche en informatique et en automatique (INRIA), 2 rue Simone Iff, 75012 Paris, France.

B Sorbonne Université, Laboratoire Jacques-Louis Lions, 75005 Paris, France.

C Centre national de la recherche scientifique (CNRS), Sciences pour l’Environnement – Unité Mixte de Recherche 6134, Università di Corsica, Campus Grossetti, 20250 Corte, France.

D Météo-France, Centre régional Sud-Est, 2 bd Château Double, 13090 Aix-en-Provence, France.

* Correspondence to: filippi@univ-corse.fr

International Journal of Wildland Fire 31(4) 379-394 https://doi.org/10.1071/WF21143
Submitted: 27 March 2021  Accepted: 18 February 2022   Published: 6 April 2022

© 2022 The Author(s) (or their employer(s)). Published by CSIRO Publishing on behalf of IAWF.

Abstract

Wildfire occurrence and behaviour are difficult to predict locally for the next day. In the present work, we propose relying on fire spread simulations to provide a fire danger index representative of the potential for fire spread that includes not only weather but also surrounding vegetation and orography. This is achieved using an artificial neural network emulator called DeepFire, trained based on simulated fire sizes. To determine how relevant this index can be in the assessment of next-day fire danger, the application of DeepFire to fire danger mapping using actual weather forecasts is studied. DeepFire forecasts for 13 fairly large fires that occurred in Corsica are analysed and compared with corresponding forecasts using another fire danger index used in operational conditions, highlighting the differences in terms of precision and the expected results of such predictions. The weather forecasts from which the weather inputs of DeepFire are determined have high spatial resolution and high frequency, which also applies to the fire danger predictions. Additionally, input uncertainty is propagated through DeepFire, resulting in ensembles of emulated fire size. Several approaches are proposed to analyse the results and provide fire danger maps and ratings using this new simulation-based prediction system.

Keywords: Corsica, deep learning, fire danger, fire weather, high resolution, potential fire size, probability distributions, wildfire risk, wildfire simulation.


References

Allaire F, Filippi JB, Mallet V (2020) Generation and evaluation of an ensemble of wildland fire simulations. International Journal of Wildland Fire 29, 160–173.
Generation and evaluation of an ensemble of wildland fire simulations.Crossref | GoogleScholarGoogle Scholar |

Allaire F, Mallet V, Filippi JB (2021a) Emulation of wildland fire spread simulation using deep learning. Neural Networks 141, 184–198.
Emulation of wildland fire spread simulation using deep learning.Crossref | GoogleScholarGoogle Scholar | 33906084PubMed |

Allaire F, Mallet V, Filippi JB (2021b) Novel method for a posteriori uncertainty quantification in wildland fire spread simulation. Applied Mathematical Modelling 90, 527–546.
Novel method for a posteriori uncertainty quantification in wildland fire spread simulation.Crossref | GoogleScholarGoogle Scholar |

Amodei M, Sanchez I, Stein J (2015) Verification of the French operational high-resolution model AROME with the regional Brier probability score. Meteorological Applications 22, 731–745.
Verification of the French operational high-resolution model AROME with the regional Brier probability score.Crossref | GoogleScholarGoogle Scholar |

Benali A, Ervilha AR, Sá ACL, Fernandes PM, Pinto RMS, Trigo RM, Pereira JMC (2016) Deciphering the impact of uncertainty on the accuracy of large wildfire spread simulations. Science of The Total Environment 569–570, 73–85.
Deciphering the impact of uncertainty on the accuracy of large wildfire spread simulations.Crossref | GoogleScholarGoogle Scholar |

Benali A, Sá ACL, Ervilha AR, Trigo RM, Fernandes PM, Pereira JMC (2017) Fire spread predictions: Sweeping uncertainty under the rug. Science of The Total Environment 592, 187–196.
Fire spread predictions: Sweeping uncertainty under the rug.Crossref | GoogleScholarGoogle Scholar |

Bradshaw LS, Deeming JE, Burgan RE, Cohen JD (1984) ‘The 1978 National Fire-Danger Rating System: technical documentation. General Technical Report INT-169. ’ (USDA Forest Service, Intermountain Forest and Range Experiment Station: Ogden, UT)

Burgan RE, Andrews PL, Bradshaw LS, Chase CH, Hartford RA, Latham DJ (1997) Current status of the wildland fire assessment system (WFAS). Fire Management Notes 57, 14–17.

Di Giuseppe F, Vitolo C, Krzeminski B, Barnard C, Maciel P, San-Miguel J (2020) Fire Weather Index: the skill provided by the European Centre for Medium-Range Weather Forecasts ensemble prediction system. Natural Hazards and Earth System Sciences 20, 2365–2378.
Fire Weather Index: the skill provided by the European Centre for Medium-Range Weather Forecasts ensemble prediction system.Crossref | GoogleScholarGoogle Scholar |

Filippi JB, Morandini F, Balbi JH, Hill DR (2010) Discrete Event Fronttracking Simulation of a Physical Fire-spread Model. Simulation 86, 629–646.
Discrete Event Fronttracking Simulation of a Physical Fire-spread Model.Crossref | GoogleScholarGoogle Scholar |

Finney MA, Grenfell IC, McHugh CW, Seli RC, Trethewey D, Stratton RD, Brittain S (2011a) A Method for Ensemble Wildland Fire Simulation. Environmental Modeling & Assessment 16, 153–167.
A Method for Ensemble Wildland Fire Simulation.Crossref | GoogleScholarGoogle Scholar |

Finney MA, McHugh CW, Grenfell IC, Riley KL, Short KC (2011b) A simulation of probabilistic wildfire risk components for the continental United States. Stochastic Environmental Research and Risk Assessment 25, 973–1000.
A simulation of probabilistic wildfire risk components for the continental United States.Crossref | GoogleScholarGoogle Scholar |

Hodges JL, Lattimer BY (2019) Wildland Fire Spread Modeling Using Convolutional Neural Networks. Fire Technology 55, 2115–2142.
Wildland Fire Spread Modeling Using Convolutional Neural Networks.Crossref | GoogleScholarGoogle Scholar |

Jain P, Coogan SCP, Subramanian SG, Crowley M, Taylor S, Flannigan MD (2020) A review of machine learning applications in wildfire science and management. Environmental Reviews 28, 478–505.
A review of machine learning applications in wildfire science and management.Crossref | GoogleScholarGoogle Scholar |

Lac C, Chaboureau JP, Masson V, Pinty JP, Tulet P, Escobar J, Leriche M, Barthe C, Aouizerats B, Augros C, Aumond P, Auguste F, Bechtold P, Berthet S, Bielli S, Bosseur F, Caumont O, Cohard JM, Colin J, Couvreux F, Cuxart J, Delautier G, Dauhut T, Ducrocq V, Filippi JB, Gazen D, Geoffroy O, Gheusi F, Honnert R, Lafore JP, Brossier CL, Libois Q, Lunet T, Mari C, Maric T, Mascart P, Mogé M, Molinié G, Nuissier O, Pantillon F, Peyrillé P, Pergaud J, Perraud E, Pianezze J, Redelsperger JL, Ricard D, Richard E, Riette S, Rodier Q, Schoetter R, Seyfried L, Stein J, Suhre K, Taufour M, Thouron O, Turner S, Verrelle A, Vié B, Visentin F, Vionnet V, Wautelet P (2018) Overview of the Meso-NH model version 5.4 and its applications. Geoscientific Model Development 11, 1929–1969.
Overview of the Meso-NH model version 5.4 and its applications.Crossref | GoogleScholarGoogle Scholar |

Lawson B, Armitage O (2008) ‘Weather Guide for the Canadian Forest Fire Danger Rating System.’ (Natural Resources Canada, Canadian Forest Service, Northern Forestry Centre: Edmonton, Alberta)

Miller C, Hilton J, Sullivan A, Prakash M (2015) SPARK – A Bushfire Spread Prediction Tool. In ‘Environmental Software Systems. Infrastructures, Services and Applications’. (Eds R Denzer, RM Argent, G Schimak, J Hřebíček) pp. 262–271. (Springer International Publishing: Cham)
| Crossref |

Parisien M, Kafka V, Hirsch K, Todd J, Lavoie S, Maczek P (2005) Mapping Wildfire Susceptibility with the BURN-P3 Simulation Model. Natural Resources Canada, Information Report NOR-X-405, Canadian Forest Service, Northern Forestry Centre, Edmonton, Alberta.

Parisien MA, Dawe DA, Miller C, Stockdale CA, Armitage OB (2019) Applications of simulation-based burn probability modelling: a review. International Journal of Wildland Fire 28, 913–926.
Applications of simulation-based burn probability modelling: a review.Crossref | GoogleScholarGoogle Scholar |

Pinto RMS, Benali A, Sá ACL, Fernandes PM, Soares PMM, Cardoso RM, Trigo RM, Pereira JMC (2016) Probabilistic fire spread forecast as a management tool in an operational setting. SpringerPlus 5, 1205
Probabilistic fire spread forecast as a management tool in an operational setting.Crossref | GoogleScholarGoogle Scholar |

Radke D, Hessler A, Ellsworth D (2019) FireCast: Leveraging Deep Learning to Predict Wildfire Spread. In ‘Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence, IJCAI-19’. pp. 4575–4581. (Artificial Intelligence Organization). https://doi.org/
| Crossref |

Rothermel RC (1972) ‘A mathematical model for predicting fire spread in wildland fuels. Research Paper INT-115.’ (USDA, Intermountain Forest and Range Experiment Station: Ogden, UT). p. 40

Van Wagner CE (1987) Development and structure of the Canadian Forest Fire Weather Index System. Canadian Forestry Service, Ottawa, Ontario, Forestry Technical Report 35.

Zhou T, Ding L, Ji J, Yu L, Wang Z (2020) Combined estimation of fire perimeters and fuel adjustment factors in FARSITE for forecasting wildland fire propagation. Fire Safety Journal 116, 103167
Combined estimation of fire perimeters and fuel adjustment factors in FARSITE for forecasting wildland fire propagation.Crossref | GoogleScholarGoogle Scholar |