Wildfires are becoming more frequent and intense, threatening ecosystems, communities, and economies. In North America, fire activity increased dramatically in recent years, with over 61% of U.S. wildfires occurring since 2000. In response to increasing wildfire severity, new predictive tools are essential for mitigating damage and improving emergency response times. Traditional models such as the Analytical Hierarchy Process can be biased due to their reliance on pairwise matrices, and Random Forest has proven to be a viable alternative. This study presented a predictive model using Random Forest regression to estimate the burn area of forest fires based on climatic variables. The model was trained on historical fire data across the United States, integrating temperature, wind speed, relative humidity, and burn acreage and fire start coordinates from nearly 9000 fires. Data preprocessing included normalization, Gaussian noise, and non-linear scaling. The model achieved an R2 value of 0.8421, indicating a strong correlation between climate variables and fire burn area. While the model effectively predicted smaller fires, it exhibited underestimation for larger fires due to averaging effects in decision trees. Limitations included the exclusion of anthropogenic factors and data imbalances. Future research should integrate anthropogenic variables, explore deep learning, and account for extreme events. This model can provide valuable insights for policymakers and emergency response teams, aiding in wildfire risk management.

References

• Cermak, A. (2024, October 23). Six trends to know about fire season in the western U.S. - NASA Science. NASA Science.
• MacCarthy, J. (n.d.-b). The latest data confirms: for est fires are getting worse. World Resources Institute. https://www.wri.org/insights/global-trends-forest-fires
• Giuseppe, F. D., McNorton, J., Lombardi, A., & Fredrik Wetterhall. (2025). Global data-driven prediction of fire activity. Nature Communications, 16(1). https://doi.org/10.1038/s41467-025-58097-7
• Wfca_Teila. (2022, November 9). How Does Humidity Affect a Fire? WFCA. https://wfca.com/wildfire-articles/how-does-humidity-affect-wildfire/
• Moore, A. (2025). How Strong Winds Complicate Wildfire Containment Efforts. College of Natural Resources News. https://cnr.ncsu.edu/news/2025/01/strong-winds-wildfires/
• Milanović, S., Marković, N., Pamučar, D., Gigović, L., Kostić, P., & Milanović, S. D. (2020). Forest Fire Probability Mapping in Eastern Serbia: Logistic Regression versus Random Forest Method. Forests, 12(1), 5. https://doi.org/10.3390/f12010005
• Matsuki, K., Kuperman, V., & Van Dyke, J. A. (2016). The Random Forests statistical technique: An examination of its value for the study of reading. Scientific Studies of Reading, 20(1), 20–33. https://doi.org/10.1080/10 888438.2015.1107073
• Mehmood, K., Anees, S. A., Luo, M., Akram, M., Zubair, M., Khan, K. A., & Khan, W. R. (2024). Assessing Chilgoza Pine (Pinusgerardiana) forest fire severity: Remote sensing analysis, correlations, and predictive modeling for enhanced management strategies. Trees Forests and People, 16, 100521. https://doi. org/10.1016/j.tfp.2024.100521
• Naderpour, M., Rizeei, H. M., & Ramezani, F. (2021). Forest fire Risk prediction: A spatial deep neural Net work-Based framework. Remote Sensing, 13(13), 2513. https://doi.org/10.3390/rs13132513
• Zhao, P., Zhang, F., Lin, H., & Xu, S. (2021). GIS-Based Forest fire Risk Model: a case study in Laoshan National Forest Park, Nanjing. Remote Sensing, 13(18), 3704. https://doi.org/10.3390/rs13183704
• Data.gov. (2025, January 31). Department of Agriculture - National USFS Fire Occurrence Point (Feature Layer). https://catalog.data.gov/dataset/national-usfs-fire-occurrence-point-feature-layer-d3233
• Historical Weather API | Open-Meteo.com. (n.d.). https://open-meteo.com/en/docs/historical-weather-api
• Wood, D. A. (n.d.). Prediction and data mining of burned areas of forest fires: Optimized data matching and mining algorithm provides valuable insight. Artificial Intelligence in Agriculture, 5, 24–42. https://doi.org/10.1016/j.aiia.2021.01.004
• Malik, A., Rao, M. R., Puppala, N., Koouri, P., Thota, V. a. K., Liu, Q., Chiao, S., & Gao, J. (2021). Data-Driven wildfire risk prediction in Northern California. Atmosphere, 12(1), 109 https://doi.org/10.3390/atmos12010109
• Sivakumar, M., Parthasarathy, S., & Padmapriya, T. (2024). Trade-off between training and testing ratio in machine learning for medical image processing. PeerJ Computer Science, 10, e2245. https://doi.org/10.7717/peerj-cs.2245
• Lee, S. (2025). Mastering OOB Score in ML. Num beranalytics.com. https://www.numberanalytics.com/blog/mastering-oobscore-in-ml
• Gao, C., Shi, C., Li, J., Li, J., Zhang, X., Huang, X., Shi, F., Yang, J., Bai, Y., & Liu, X. (2025). Lightning-ignited wildfire prediction in the boreal forest of northeast China. Global and Planetary Change, 253, 104948. https://doi.org/10.1016/j.gloplacha.2025.104948
• Branco, P., Torgo, L., & Ribeiro, R. P. (2017, October 11). SMOGN: a Pre-processing Approach for Imbalanced Regression. Proceedings.mlr.press; PMLR.