A neural network model to study factors impacting the selection of hazardous fuel treatment types in Colorado’s national forests

Shayne Magstadt; Yu Wei

doi:10.1071/WF24024

RESEARCH ARTICLE (Open Access)

Previous Next Contents Vol 34(1)

A neural network model to study factors impacting the selection of hazardous fuel treatment types in Colorado’s national forests

Shayne Magstadt ^A ^* and Yu Wei ^A

+ Author Affiliations

- Author Affiliations

^A Department of Forest and Rangeland Stewardship, Warner College of Natural Resources, Colorado State University, 1472 Campus Delivery, Fort Collins, CO 80523, USA.

^* Correspondence to: Shayne.Magstadt@colostate.edu

International Journal of Wildland Fire 34, WF24024 https://doi.org/10.1071/WF24024

Submitted: 2 February 2024 Accepted: 2 December 2024 Published: 8 January 2025

© 2025 The Author(s) (or their employer(s)). Published by CSIRO Publishing on behalf of IAWF. This is an open access article distributed under the Creative Commons Attribution 4.0 International License (CC BY).

Abstract

Background

Selecting suitable fuel treatment locations and types is important for reducing wildfire risks in Colorado’s national forests. Understanding factors influencing these decisions is needed for effective management.

Aims

This study analyses land management data from 2015 to 2024 using a single multiclass neural network model to understand the drivers influencing fuel treatment decisions in 11 Colorado national forests.

Methods

We utilise Forest Activity Tracking System data, incorporating variables such as wildfire risk, landscape features and human influences. The model employs a feed-forward backpropagation technique to train a neural network model on the spatial dataset.

Key results

The model identifies significant factors associated with past fuel treatment decisions, including burn probability, wildfire hazard potential, conditional flame length and proximity to structures. Analysis reveals the importance of these variables in shaping treatment selection strategies, the model AUC (area under curve) of 0.91 indicating strong predictive performance across the six treatment categories.

Conclusions

Neural networks provide a robust method for analysing past fuel treatment choices. Accurately identifying key factors, this approach provides suggestions to improve future fuel treatment decisions.

Implications

This approach can enhance wildfire mitigation planning across Colorado’s national forests. The findings support more informed wildfire mitigation strategies, with potential applications extending to broader forest management practices.

Keywords: artificial neural networks, Colorado national forests, data-driven decision making, geospatial analysis, hazardous fuel treatment, landscape analysis, machine learning, mechanical fuel treatment, prescribed fire, spatial data analysis.

References

Abadi M, Agarwal A, Barham P, Brevdo E, Chen Z, Citro C, Corrado GS, Davis A, Dean J, Devin M, Ghemawat S, Goodfellow I, Harp A, Irving G, Isard M, Jia Y, Jozefowicz R, Kaiser L, Kudlur M, Levenberg J, Mané D, Monga R, Moore S, Murray D, Olah C, Schuster M, Shlens J, Steiner B, Sutskever I, Talwar K, Tucker P, Vanhoucke V, Vasudevan V, Viégas F, Vinyals O, Warden P, Wattenberg M, Wicke M, Yu Y, Zheng X (2015) TensorFlow: large-scale machine learning on heterogeneous systems. (Google Brain) Available at https://www.tensorflow.org/

Adab H (2017) Landfire hazard assessment in the Caspian Hyrcanian forest ecoregion with the long-term MODIS active fire data. Natural Hazards 87, 1807-1825.
| Crossref | Google Scholar |

Addington RN, Tavernia BG, Caggiano MD, Thompson MP, Lawhon JD, Sanderson JS (2020) Identifying opportunities for the use of broadcast prescribed fire on Colorado’s Front Range. Forest Ecology and Management 458, 117655.
| Crossref | Google Scholar |

Agee JK (2002) The fallacy of passive management managing for firesafe forest reserves. Conservation in Practice 3(1), 18-26.
| Crossref | Google Scholar |

Agee JK, Skinner CN (2005) Basic principles of forest fuel reduction treatments. Forest Ecology and Management 211(1–2), 83-96.
| Crossref | Google Scholar |

Ager AA, Evers CR, Day MA, Alcasena FJ, Houtman R (2021) Planning for future fire: scenario analysis of an accelerated fuel reduction plan for the western United States. Landscape and Urban Planning 215, 104212.
| Crossref | Google Scholar |

Alcasena FJ, Ager AA, Salis M, Day MA, Vega-Garcia C (2018) Optimizing prescribed fire allocation for managing fire risk in central Catalonia. Science of The Total Environment 621, 872-885.
| Crossref | Google Scholar | PubMed |

Alin A (2010) Multicollinearity. Wiley Interdisciplinary Reviews: Computational Statistics 2(3), 370-374.
| Crossref | Google Scholar |

Alonso-Betanzos A, Fontenla-Romero O, Guijarro-Berdiñas B, Hernández-Pereira E, Andrade MIP, Jiménez E, Soto JLL, Carballas T (2003) An intelligent system for forest fire risk prediction and firefighting management in Galicia. Expert Systems with Applications 25(4), 545-554.
| Crossref | Google Scholar |

Altmann A, Toloşi L, Sander O, Lengauer T (2010) Permutation importance: a corrected feature importance measure. Bioinformatics 26(10), 1340-1347.
| Crossref | Google Scholar | PubMed |

Barros AM, Ager A, Day M, Palaiologou P (2019) Improving long-term fuel treatment effectiveness in the National Forest System through quantitative prioritization. Forest Ecology and Management 433, 514-527.
| Crossref | Google Scholar |

Bergstra J, Bengio Y (2012) Random search for hyper-parameter optimization. Journal of Machine Learning Research 13(2), 281-305.
| Google Scholar |

Berry AH, Donovan G, Hesseln H (2006) Prescribed burning costs and the WUI: economic effects in the Pacific Northwest. Western Journal of Applied Forestry 21(2), 72-78.
| Crossref | Google Scholar |

Bishop CM (1995) ‘Neural networks for pattern recognition.’ (Oxford University Press)

Bishop C (2006) ‘Pattern recognition and machine learning. Vol. 2.’ pp. 35–42. (Springer Google Scholar)

Bisquert M, Caselles E, Sánchez JM, Caselles V (2012) Application of artificial neural networks and logistic regression to the prediction of forest fire danger in Galicia using MODIS data. International Journal of Wildland Fire 21(8), 1025-1029.
| Crossref | Google Scholar |

Brennan TJ, Keeley JE (2015) Effect of mastication and other mechanical treatments on fuel structure in chaparral. International Journal of Wildland Fire 24(7), 949-963.
| Crossref | Google Scholar |

Cao X, Yousefzadeh R (2023) Extrapolation and AI transparency: why machine learning models should reveal when they make decisions beyond their training. Big Data & Society 10(1), 20539517231169731.
| Crossref | Google Scholar |

Carvalho DV, Pereira EM, Cardoso JS (2019) Machine learning interpretability: a survey on methods and metrics. Electronics 8(8), 832.
| Crossref | Google Scholar |

Cheng T, Wang J (2008) Integrated spatio-temporal data mining for forest fire prediction. Transactions in GIS 12(5), 591-611.
| Google Scholar |

Chollet F (2021) ‘Deep learning with Python.’ (Simon and Schuster)

De Vasconcelos MP, Silva S, Tome M, Alvim M, Pereira JC (2001) Spatial prediction of fire ignition probabilities: comparing logistic regression and neural networks. Photogrammetric Engineering and Remote Sensing 67(1), 73-81.
| Google Scholar |

De Veaux RD, Ungar LH (1994) Multicollinearity: a tale of two nonparametric regressions. In ‘Selecting models from data: artificial intelligence and statistics IV’. (Eds P Cheeseman, RW Oldford) pp. 393–402. (Springer)

Dillon GK, Menakis J, Fay F (2014) Wildland fire potential: a tool for assessing wildfire risk and fuels management needs. In ‘Proceedings of the large wildland fires conference’. pp. 60–76. (Department of Agriculture, Forest Service, Rocky Mountain Research Station)

Dillon GK, Gilbertson-Day JW (2020) Wildfire Hazard Potential for the United States (270-m), version 2020. 3rd Edition. (Forest Service Research Data Archive: Fort Collins, CO) 10.2737/RDS-2015-0047-3

Dillon GK, Scott JH, Jaffe MR, Olszewski JH, Vogler KC, Finney MA, Short KC, Riley KL, Grenfell IC, Jolly MW, Brittain S (2023) ‘Spatial datasets of probabilistic wildfire risk components for the United States (270m).’ (Forest Service Research Data Archive) 10.2737/RDS-2016-0034-3

Egmont-Petersen M, de Ridder D, Handels H (2002) Image processing with neural networks—a review. Pattern Recognition 35(10), 2279-2301.
| Crossref | Google Scholar |

Fernandes PM, Botelho HS (2003) A review of prescribed burning effectiveness in fire hazard reduction. International Journal of Wildland Fire 12(2), 117-128.
| Crossref | Google Scholar |

Feurer M, Hutter F (2019) Hyperparameter optimization. In ‘Automated machine learning: methods, systems, challenges’. (Eds F Hutter, L Kotthoff, J Vanschoren) pp. 3–33. (Springer)

Finney MA (2001) Design of regular landscape fuel treatment patterns for modifying fire growth and behavior. Forest Science 47(2), 219-228.
| Crossref | Google Scholar |

Finney MA (2006) An overview of FlamMap fire modeling capabilities. In ‘Fuels management – how to measure success: Conference Proceedings RMRS-P-41’, 28-30 March 2006; Portland, OR. (Eds PL Andrews, BW Butler, comps) pp. 213–220. (US Department of Agriculture, Forest Service, Rocky Mountain Research Station: Fort Collins, CO)

Goodfellow I, Bengio Y, Courville A (2016) ‘Deep learning.’ (MIT Press)

Guo C, Pleiss G, Sun Y, Weinberger KQ (2017) On calibration of modern neural networks. In ‘International conference on machine learning. Proceedings of Machine Learning Research 70:1321-1330. Available at https://proceedings.mlr.press/v70/guo17a.html

Haas JR, Calkin DE, Thompson MP (2015) Wildfire risk transmission in the Colorado Front Range, USA. Risk Analysis 35(2), 226-240.
| Crossref | Google Scholar | PubMed |

Harmon ME (1984) Survival of trees after low-intensity surface fires in Great Smoky Mountains National Park. Ecology 65(3), 796-802.
| Crossref | Google Scholar |

Hecht-Nielsen R (1992) Theory of the backpropagation neural network. In ‘Neural networks for perception’. (Ed. H Wechsler) pp. 65–93. (Elsevier)

Hood SM, Varner JM, Jain TB, Kane JM (2022) A framework for quantifying forest wildfire hazard and fuel treatment effectiveness from stands to landscapes. Fire Ecology 18(1), 33.
| Crossref | Google Scholar |

Jafari Goldarag Y, Mohammadzadeh A, Ardakani A (2016) Fire risk assessment using neural network and logistic regression. Journal of the Indian Society of Remote Sensing 44(6), 885-894.
| Crossref | Google Scholar |

Jain P, Coogan SC, Subramanian SG, Crowley M, Taylor S, Flannigan MD (2020) A review of machine learning applications in wildfire science and management. Environmental Reviews 28(4), 478-505.
| Crossref | Google Scholar |

Jiménez-Valverde A (2012) Insights into the area under the receiver operating characteristic curve (AUC) as a discrimination measure in species distribution modelling. Global Ecology and Biogeography 21(4), 498-507.
| Crossref | Google Scholar |

Johnson MC, Peterson DL, Raymond CL (2007) Managing forest structure and fire hazard—A tool for planners. Journal of Forestry 105(2), 77-83.
| Crossref | Google Scholar |

Kalantar B, Ueda N, Idrees MO, Janizadeh S, Ahmadi K, Shabani F (2020) Forest fire susceptibility prediction based on machine learning models with resampling algorithms on remote sensing data. Remote Sensing 12(22), 3682.
| Crossref | Google Scholar |

Kalies EL, Kent LLY (2016) Tamm Review: are fuel treatments effective at achieving ecological and social objectives? A systematic review. Forest Ecology and Management 375, 84-95.
| Crossref | Google Scholar |

Kingma DP, Ba J (2014) Adam: a method for stochastic optimization. arXiv preprint arXiv:1412.6980 [Preprint].
| Google Scholar |

Kuhn M, Johnson K (2013) ‘Applied predictive modeling.’ (Springer)

Liang H, Zhang M, Wang H (2019) A neural network model for wildfire scale prediction using meteorological factors. IEEE Access 7, 176746-176755.
| Crossref | Google Scholar |

Liu Z, Wimberly MC, Lamsal A, Sohl TL, Hawbaker TJ (2015) Climate change and wildfire risk in an expanding wildland–urban interface: a case study from the Colorado Front Range Corridor. Landscape Ecology 30(10), 1943-1957.
| Crossref | Google Scholar |

Mansfield ER, Helms BP (1982) Detecting multicollinearity. The American Statistician 36(3a), 158-160.
| Crossref | Google Scholar |

Matsypura D, Prokopyev OA, Zahar A (2018) Wildfire fuel management: network-based models and optimization of prescribed burning. European Journal of Operational Research 264(2), 774-796.
| Crossref | Google Scholar |

Menard SW (2010) ‘Logistic regression: from introductory to advanced concepts and applications.’ (Sage)

Minas J, Hearne J, Martell D (2015) An integrated optimization model for fuel management and fire suppression preparedness planning. Annals of Operations Research 232, 201-215.
| Google Scholar |

National Interagency Fire Center (NIFC) (2020) ‘Spatial fireline records, 2020.’ (National Interagency Fire Center: Boise, ID, USA) Available at http://www.nifc.gov

Nino-Adan I, Portillo E, Landa-Torres I, Manjarres D (2021) Normalization influence on ANN-based models performance: a new proposal for Features’ contribution analysis. IEEE Access 9, 125462-125477.
| Crossref | Google Scholar |

O’Malley T, Bursztein E, Long J, Chollet F, Jin H, Invernizzi L, et al. (2019) Keras Tuner. Available at https://github.com/keras-team/keras-tuner

Patro S, Sahu KK (2015) Normalization: a preprocessing stage. arXiv preprint arXiv:1503.06462 [Preprint].
| Google Scholar |

Pedregosa F, Varoquaux G, Gramfort A, Michel V, Thirion B, Grisel O, Blondel M, Prettenhofer P, Weiss R, Dubourg V, et al. (2011) Scikit-learn: machine learning in Python. Journal of Machine Learning Research 12, 2825-2830.
| Google Scholar |

Pham BT, Jaafari A, Avand M, Al-Ansari N, Dinh Du T, Yen HPH, Phong TV, Nguyen DH, Le HV, Mafi-Gholami D, et al. (2020) Performance evaluation of machine learning methods for forest fire modeling and prediction. Symmetry 12(6), 1022.
| Crossref | Google Scholar |

Rachmawati R, Ozlen M, Reinke KJ, Hearne JW (2016) An optimisation approach for fuel treatment planning to break the connectivity of high-risk regions. Forest Ecology and Management 368, 94-104.
| Crossref | Google Scholar |

Reinhardt ED (2003) ‘The fire and fuels extension to the forest vegetation simulator.’ (US Department of Agriculture, Forest Service, Rocky Mountain)

Reinhardt ED, Keane RE, Calkin DE, Cohen JD (2008) Objectives and considerations for wildland fuel treatment in forested ecosystems of the interior western United States. Forest Ecology and Management 256(12), 1997-2006.
| Crossref | Google Scholar |

Rodrigues M, De la Riva J (2014) An insight into machine-learning algorithms to model human-caused wildfire occurrence. Environmental Modelling & Software 57, 192-201.
| Google Scholar |

Rollins MG (2009) LANDFIRE: a nationally consistent vegetation, wildland fire, and fuel assessment. International Journal of Wildland Fire 18(3), 235-249.
| Crossref | Google Scholar |

Rosenblatt F (1958) The perceptron: a probabilistic model for information storage and organization in the brain. Psychological Review 65(6), 386-408.
| Crossref | Google Scholar | PubMed |

Sakr GE, Elhajj IH, Mitri G (2011) Efficient forest fire occurrence prediction for developing countries using two weather parameters. Engineering Applications of Artificial Intelligence 24(5), 888-894.
| Crossref | Google Scholar |

Scott JH, Gilbertson-Day JW, Moran C, Dillon GK, Short KC, Vogler KC (2020) Wildfire risk to communities: spatial datasets of landscape-wide wildfire risk components for the United States. (Forest Service Research Data Archive: Fort Collins, CO) Updated 25 November 2020. 10.2737/RDS-2020-0016

Seabold S, Perktold J (2010) Statsmodels: econometric and statistical modeling with Python. In ‘9th Python in Science Conference’. Austin, 28 June–3 July, 2010, pp. 57-61. 10.25080/Majora-92bf1922-011

Shalev-Shwartz S, Ben-David S (2014) ‘Understanding machine learning: from theory to algorithms.’ (Cambridge University Press)

Theobald DM, Romme WH (2007) Expansion of the US wildland–urban interface. Landscape and Urban Planning 83(4), 340-354.
| Crossref | Google Scholar |

Thomas D, Butry D, Gilbert S, Webb D, Fung J, et al. (2017) The costs and losses of wildfires. NIST Special Publication 1215(11), 1-72.
| Google Scholar |

Thompson MP, Calkin DE (2011) Uncertainty and risk in wildland fire management: a review. Journal of Environmental Management 92(8), 1895-1909.
| Crossref | Google Scholar | PubMed |

US Congress (2003) Healthy Forest Restoration Act of 2003. (US Government Publishing Office)

US Congress (2021) Infrastructure Investment and Jobs Act. Available at https://www.govinfo.gov/app/details/PLAW-117publ58

US Department of Agriculture (2022) ‘Wildfire crisis strategy: confronting the wildfire crisis.’ (US Department of Agriculture, Forest Service: Washington, DC) Available at https://www.fs.usda.gov/sites/default/files/fs_media/fs_document/Confronting-the-Wildfire-Crisis.pdf [accessed 1 April 2024]

US Department of Agriculture (2023) ‘Black Diamond Landscape Resiliency and Risk Reduction Project.’ (US Forest Service) Available at https://www.fs.usda.gov/project/?project=62591

USDA–NRCS (2014) Geospatial data gateway. (US Department of Agriculture, Natural Resources Conservation Service)

USDA–USDI (2000) A report to the President in response to the wildfires of 2000. Available at www.fireplan.gov\president.cfm

Vasilakos C, Kalabokidis K, Hatzopoulos J, Kallos G, Matsinos Y (2007) Integrating new methods and tools in fire danger rating. International Journal of Wildland Fire 16(3), 306-316.
| Crossref | Google Scholar |

Vittinghoff E, Glidden DV, Shiboski SC, McCulloch CE (2012) 'Regression methods in biostatistics: linear, logistic, survival, and repeated measures models.’ (Springer Science & Business Media)

Wei Y, Rideout D, Kirsch A (2008) An optimization model for locating fuel treatments across a landscape to reduce expected fire losses. Canadian Journal of Forest Research 38(4), 868-877.
| Crossref | Google Scholar |

Weir JR, Kreuter UP, Wonkka CL, Twidwell D, Stroman DA, Russell M, Taylor CA (2019) Liability and prescribed fire: perception and reality. Rangeland Ecology & Management 72(3), 533-538.
| Google Scholar |

Woods K, Bowyer KW (1997) Generating ROC curves for artificial neural networks. IEEE Transactions on Medical Imaging 16(3), 329-337.
| Crossref | Google Scholar | PubMed |

Zhang Z, Sabuncu M (2018) Generalized cross entropy loss for training deep neural networks with noisy labels. In ‘Advances in neural information processing systems’. (Eds S Bengio, H Wallach, H Larochelle, K Grauman, N Cesa-Bianchi, R Garnett) p. 31. (Curran Associates, Inc.)