Fire hazard is often mapped as a static conditional probability of fire characteristics’ occurrence. We developed a dynamic product for operational risk management to forecast the probability of occurrence of fire radiative power in the locally possible near-maximum fire intensity range. We applied standard machine learning techniques to remotely sensed data. We used a block maxima approach to sample the most extreme fire radiative power (FRP) MODIS retrievals in free-burning fuels for each fire season between 2001 and 2020 and associated weather, fuel, and topography features in northwestern south America. We used the random forest algorithm for both classification and regression, implementing the backward stepwise repression procedure. We solved the classification problem predicting the probability of occurrence of near-maximum wildfire intensity with 75% recall out-of-sample in ten annual test sets running time series cross validation, and 77% recall and 85% ROC-AUC out-of-sample in a twenty-fold cross-validation to gauge a realistic expectation of model performance in production. We solved the regression problem predicting FRP with 86% r² in-sample, but out-of-sample performance was unsatisfactory. Our model predicts well fatal and near-fatal incidents reported in Peru and Colombia out-of-sample in mountainous areas and unimodal fire regimes, the signal decays in bimodal fire regimes.

1. Scott J, Thompson M, Calkin D. A wildfire risk assessment framework for land and resource management [Internet]. Gen. Tech. Rep. RMRS-GTR-315. U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station; 2013. Available from: https://www.fs.usda.gov/rm/pubs/rmrs_gtr315.pdf.

2. Finney MA, McAllister SS, Grumstrup TP, et al. Wildland fire behaviour: Dynamics, principles and processes. CSIRO Publishing; 2021.

3. Finney M, Grenfell IC, McHugh CW, et al. A method for ensemble wildland fire simulation. Environmental Modeling and Assessment 2011; 16(2): 153–167. doi: 10.1007/s10666-010-9241-3.

4. Stephens SL, Burrows N, Buyantuyev A, et al. Temperate and boreal forest mega-fires: Characteristics and challenges. Frontiers in Ecology and the Environment 2014; 12(2): 115–122. doi: 10.1890/120332.

5. Scott AC, Holloway R, Bowman D, et al. Fire on Earth: An introduction. Wiley-Blackwell; 2014.

6. Preisler HK, Riley KL, Stonesifer CS, et al. Near-term probabilistic forecast of significant wildfire events for the Western United States. International Journal of Wildland Fire 2016; 25(11): 1169–1680. doi: 10.1071/WF16038

7. Werth PA, Potter BE, Alexander ME, et al. Synthesis of knowledge of extreme fire behavior: Volume 2 for fire behavior specialists, researchers, and meteorologists [Internet]. 2016. Available from: https://www.firescience.gov/projects/09-2-01-11/project/09-2-01-11_Synthesis_Extreme_FB_vol2.pdf.

8. Jolly WM, Freeborn PH. Towards improving wildland firefighter situational awareness through daily fire behaviour risk assessments in the US Northern Rockies and Northern Great Basin. International Journal of Wildland Fire 2017; 26(7): 574–586. doi: 10.1071/WF16153.

9. Jolly WM, Freeborn PH, Page WG, Butler BW. Severe fire danger index: A forecastable metric to inform firefighter and community wildfire risk management. Fire 2019; 2(3): 47. doi: 10.3390/fire2030047.

10. Podschwit H, Cullen A. Patterns and trends in simultaneous wildfire activity in the United States from 1984 to 2015. International Journal of Wildland Fire 2020; 29(12): 1057–1071. doi: 10.1071/WF19150.

11. Laurent P, Mouillot F, Vanesa M, et al. Varying relationships between fire radiative power and fire size at a global scale. Biogeosciences 2019; 16(2): 275–288. doi: 10.5194/bg-16-275-2019.

12. Tedim F, Leone V, Amraoui M, et al. Defining extreme wildfire events: Difficulties, challenges, and impacts. Fire 2018; 1(1): 9. doi: 10.3390/fire1010009.

13. Tedim F, Leone V, Mcgee T. Extreme wildfire events and disasters: Root causes and new management strategies. Elsevier; 2020.

14. Wilson CC. Fatal and near-fatal forest fires the common denominators. The International Fire Chief 1977; 43(9): 9–10.

15. Page WG, Freeborn PH, Butler BW, Jolly WM. A classification of US wildland firefighter entrapments based on coincident fuels, weather, and topography. Fire 2019; 2(4): 52. doi: 10.3390/fire2040052.

16. Page WG, Freeborn PH, Butler BW, Jolly WM. A review of US wildland firefighter entrapments: Trends, important environmental factors and research needs. International Journal of Wildland Fire 2019; 28(8): 551–569. doi: 10.1071/WF19022.

17. Brownlee J. Imbalanced classification with Python: Better metrics, balance skewed classes, cost-sensitive learning [Internet]. Available from: https://books.google.com.co/books?id=jaXJDwAAQBAJ.

18. Ling CX, Sheng VS. Cost-sensitive learning. In: Sammut C, Webb GI (editors). Encyclopedia of machine learning. Springer; 2010. p. 231–235.

19. Cox DR. The regression analysis of binary sequences. Journal of the Royal Statistical Society: Series B (Methodological) 1958; 20(2): 215–242. doi: 10.1111/j.2517-6161.1958.tb00292.x.

20. King G, Nielsen R. Why propensity scores should not be used for matching. Political Analysis 2019; 27(4): 435–454. doi: 10.1017/pan.2019.11.

21. Iacus SM, King G, Porro G. Multivariate matching methods that are monotonic imbalance bounding. Journal of the American Statistical Association 2011; 106(493): 345–361. doi: 10.1198/jasa.2011.tm09599.

22. Gertler PJ, Martinez S, Premand P, et al. Impact evaluation in practice. 2nd ed. Inter-American Development Bank and World Bank; 2016.

23. Good practice guidance. SDG indicator 15.3.1 [Internet]. Available from: https://www.unccd.int/sites/default/files/relevant-links/2021-03/Indicator_15.3.1_GPG_v2_29Mar_Advanced-version.pdf.

24. MapBiomas general “handbook”. Algorithm theoretical basis document (ATBD) [Internet]. 2022. Available from: https://mapbiomas-br-site.s3.amazonaws.com/Metodologia/ATBD_Collection_6_v1_January_2022.pdf.

25. Algorithm theoretical base document (ATBD). RAISG-MapBiomas Amazon—Collection 4 (Spanish) [Internet]. 2022. Available from: https://s3.amazonaws.com/amazonia.mapbiomas.org/atbd/atbd%20general/ATBD_General_MapBiomas_Amazonia_4.0.pdf.

26. MapBiomas chaco general “handbook”. Algorithm theoretical basis document (ATBD) [Internet]. 2020. Available from: https://mapbiomas-br-site.s3.amazonaws.com/MapBiomas%20CHACO/Colecao_2/ATBD/ATBD_-_MapBiomas_Chaco_Collection2.pdf.

27. Buchhorn M, Smets B, Bertels L, et al. Copernicus global land service: Land cover 100 m: Collection 3: Epoch 2015: Globe. Zenodo 2020; 1–14. doi: 10.5281/zenodo.3939038.

28. Sulla-Menashe D, Friedl MA. User guide to collection 6 MODIS land cover (MCD12Q1 and MCD12C1) product [Internet]. 2018. Available from: https://lpdaac.usgs.gov/documents/101/MCD12_User_Guide_V6.pdf.

29. Giglio L, Descloitres J, Justice CO, Kaufman YJ. An enhanced contextual fire detection algorithm for MODIS. Remote Sensing of Environment 2003; 87(2–3): 273–282. doi: 10.1016/S0034-4257(03)00184-6.

30. Lizundia-Loiola J, Otón G, Ramo R, Chuvieco E. A spatio-temporal active-fire clustering approach for global burned area mapping at 250 m from MODIS data. Remote Sensing of Environment 2020; 236: 111493. doi: 10.1016/j.rse.2019.111493.

31. Giglio L, Boschetti L, Roy D, et al. Collection 6 MODIS burned area product user’s guide [Internet]. 2020. Available from: https://lpdaac.usgs.gov/documents/875/MCD64_User_Guide_V6.pdf.

32. Giglio L, Justice C, Boschetti L, Roy D. MCD64A1 MODIS/Terra+Aqua Burned Area Monthly L3 Global 500 m SIN Grid [Internet]. USGS. Available from: https://lpdaac.usgs.gov/products/mcd64a1v006/.

33. Giglio L, Schroeder W, Hall J, Justice C. MODIS Collection 6 and Collection 6.1 active fire product user’s guide [Internet]. 2021. Available from: https://lpdaac.usgs.gov/documents/1005/MOD14_User_Guide_V61.pdf.

34. Schmit TJ, Griffith P, Gunshor MM, et al. A closer look at the ABI on the goes-r series. Bulletin of the American Meteorological Society 2017; 98(4): 681–698. doi: 10.1175/BAMS-D-15-00230.1.

35. Schroeder W, Csiszar I, Giglio L, et al. Early characterization of the active fire detection products derived from the next generation NPOESS/VIIRS and GOES-R/ABI instruments. In: Proceedings of the International Geoscience and Remote Sensing Symposium (IGARSS); 2010 Jul 25–30; Honolulu, USA. p. 2683–2686.

36. Picotte JJ, Dockter D, Long J, et al. LANDFIRE remap prototype mapping effort: Developing a new framework for mapping vegetation classification, change, and structure. Fire 2019; 2(2): 35. doi: 10.3390/fire2020035.

37. Reeves MC, Ryan KC, Rollins MG, Thomas TG. Spatial fuel data products of the LANDFIRE project. International Journal of Wildland Fire 2009; 18(3): 250–267. doi: 10.1071/WF08086.

38. Hansen MC, Potapov PV, Moore R, et al. High-resolution global maps of 21st-century forest cover change. Science 2013; 342(6160): 850–854. doi: 10.1126/science.1244693.

39. Alexander ME, Cruz MG. Fireline intensity. In: Manzello S (editor). Encyclopedia of wildfires and wildland-urban interface (WUI) fires. Springer; 2018. p. 1210.

40. Cheng L, Yajun L, Chang Z, et al. The method of evaluating sub-pixel size and temperature of fire spot in AVHRR data. Journal of Applied Meteorological Science 2004; 15(3): 273–280.

41. Peterson D, Wang J, Ichoku C, Hyer EJ. Sub-pixel fractional area of wildfires from MODIS observations: Retrieval, validation, and potential applications. In: Proceedings of the 34th International Symposium on Remote Sensing of Environment—The GEOSS Era: Towards Operational Environmental Monitoring; 2011 Apr 10–15; Sydney, Australia.

42. Gumbel EJ. Statistics of extremes [Internet]. Available from: https://mathscinet.ams.org/mathscinet/relay-station?mr=0096342.

43. Muñoz-Sabater J, Dutra E, Agustí-Panareda A, et al. ERA5-Land: A state-of-the-art global reanalysis dataset for land applications. Earth System Science Data 2021; 13(9): 4349–4383. doi: 10.5194/essd-13-4349-2021.

44. Lawson BD Armitage OB. Weather guide for the Canadian forest fire danger rating system [Internet]. Available from: https://cfs.nrcan.gc.ca/pubwarehouse/pdfs/29152.pdf.

45. Jolly WM, Cochrane MA, Freeborn PH, et al. Climate-induced variations in global wildfire danger from 1979 to 2013. Nature Communications 2015; 6: 7357. doi: 10.1038/ncomms8537.

46. Keetch JJ, Byram GM. A drought index for forest fire control [Internet]. U.S.D.A. Forest Service Research Paper SE - 38; 1968. Available from: https://www.srs.fs.usda.gov/pubs/rp/rp_se038.pdf.

47. Heinsch FA, Andrews PL, Tirmenstein D. How to generate and interpret fire characteristics charts for the U.S. fire danger rating system [Internet]. Gen. Tech. Rep. RMRSGTR-363. Fort Collins, CO: U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station; 2017. Available from: https://www.fs.usda.gov/rm/pubs_series/rmrs/gtr/rmrs_gtr363.pdf.

48. Jolly WM, Butler BW, Forthofer J. Assessing topography and wind alignment for firefighter safety [Internet]. Available from: file:///C:/Users/Administrator/Downloads/156361.pdf.

49. Huang S, Tang L, Hupy JP, et al. A commentary review on the use of normalized difference vegetation index (NDVI) in the era of popular remote sensing. Journal of Forestry Research 2020; 32(5): 1–6. doi: 10.1007/s11676-020-01155-1.

50. Keeley JE. Fire intensity, fire severity and burn severity: A brief review and suggested usage. International Journal of Wildland Fire 2009; 18(1): 116–126. doi: 10.1071/WF07049.

51. Gao BC. NDWI—A normalized difference water index for remote sensing of vegetation liquid water from space. Remote Sensing of Environment 1996; 58(3): 257–266. doi: 10.1016/S0034-4257(96)00067-3.

52. Armenteras D, Dávalos LM, Barreto JS, et al. Fire-induced loss of the world’s most biodiverse forests in Latin America. Science Advances 2021; 7(33): 2–10. doi: 10.1126/sciadv.abd3357.

53. Barreto JS, Armenteras D. Open data and machine learning to model the occurrence of fire in the ecoregion of “Llanos Colombo–Venezolanos”. Remote Sensing 2020; 12(23): 3291. doi: 10.3390/rs12233921.

54. Jain P, Coogan SCP, Subramanian S, et al. A review of machine learning applications in wildfire science and management. Environmental Reviews 2020; 28(3): 73. doi: 10.1139/er-2020-0019.

55. Suradhaniwar S, Kar S, Durbha S, Jagarlapudi A. Time series forecasting of univariate agrometeorological data: A comparative performance evaluation via one-step and multi-step ahead forecasting strategies. Sensors 2021; 21(7): 2430. doi: 10.3390/s21072430.

56. Pedregosa F, Varoquaux G, Gramfort A, et al. Scikit-learn: Machine learning in python. Journal of Machine Learning Research 2011; 12(85): 2825–2830. doi: 10.1289/EHP4713.

57. Boehmke B, Greenwell BM. Hands-on machine learning with R. 1st ed. Chapman and Hall/CRC; 2019.

58. Ledolter J, Kardon RH. Focus on data: Statistical design of experiments and sample size selection using power analysis. Investigative Ophthalmology and Visual Science 2020; 61(8): 11. doi: 10.1167/IOVS.61.8.11.

59. Alzate DF, Carrillo GAA, Barbosa EOR, et al. REGNIE interpolation for rain and temperature in the Andean, Caribbean, and Pacific regions of Colombia. Colombia Forestal 2018; 21(1): 102–118. doi: 10.14483/2256201X.11601.

60. Gorelick N, Hancher M, Dixon M, et al. Google earth engine: Planetary-scale geospatial analysis for everyone. Remote Sensing of Environment 2016; 202; 18–27. doi: 10.1016/j.rse.2017.06.031.

61. Earth Engine Apps. Available from: https://anmarkos.users.earthengine.app/view/forecasting-wildfire-hazard-across-nw-south-america-app1.

62. Earth Engine Apps. Available from: https://anmarkos.users.earthengine.app/view/forecasting-wildfire-hazard-across-nw-south-america-app2.