Gao et al. (2026) Distinct seasonality and trends of natural and anthropogenic wildfires in boreal forests

Identification

• Journal: Agricultural and Forest Meteorology
• Year: 2026
• Date: 2026-08-26
• Authors: Cong Gao, Jinbao Li, Keyan Fang, Xu Zhang, Qi Zhang, Chunming Shi
• DOI: 10.1016/j.agrformet.2026.111436

Research Groups

• Department of Geography, The University of Hong Kong, Hong Kong SAR, China
• Institute for Climate and Carbon Neutrality, The University of Hong Kong, Hong Kong SAR, China
• Key Laboratory of Humid Subtropical Eco-geographical Process (Ministry of Education), College of Geographical Sciences, Fujian Normal University, Fuzhou, China
• Key Laboratory of Environmental Change and Natural Disaster of Ministry of Education, Faculty of Geographical Science, Beijing Normal University, Beijing, China

Short Summary

This study employs a novel machine-learning attribution framework to distinguish natural and anthropogenic wildfires in boreal forests and characterize their distinct seasonal patterns and temporal variations. The results demonstrate that human activities drive a significant increase in anthropogenic fires in spring, whereas natural fires prevail in summer.

Objective

• To develop a robust framework for quantifying the contributions of natural and anthropogenic factors to wildfires in boreal forests.
• To attribute fire causes on an individual fire basis using satellite observations.

Study Configuration

• Spatial Scale: Boreal forests of North America and Eurasia, with a spatial resolution of 0.25° × 0.25°.
• Temporal Scale: 2002-2022, with daily temporal resolution.

Methodology and Data

• Models used: eXtreme Gradient Boosting (XGBoost) algorithm.
• Data sources: Satellite-derived fire dataset from the Global Fire Atlas (GFA), in-situ fire records from North America, and environmental data from the European Centre for Medium-Range Weather Forecasts (ECMWF) fifth generation reanalysis (ERA5).

Main Results

• The machine-learning framework achieves a mean accuracy of 85.4% in attributing natural and anthropogenic fires in boreal forests of North America.
• Anthropogenic fires peak during spring and autumn, with minimal occurrence in summer.
• Natural fires prevail in summer, accounting for 85.3% of total natural fires.
• The number of anthropogenic fires exhibits a decreasing trend from 2002 to 2022, while natural fires show a non-significant increase over the same period.

Contributions

• This study provides a scalable and data-driven approach for quantifying human impacts on boreal fire regimes.
• The framework enables robust attribution of fire causes on an individual fire basis using satellite observations.
• The results highlight the importance of considering both natural and anthropogenic factors in understanding boreal wildfire dynamics.

Funding

• Not specified.

Citation

@article{Gao2026Distinct,
  author = {Gao, Cong and Li, Jinbao and Fang, Keyan and Zhang, Xu and Zhang, Qi and Shi, Chunming},
  title = {Distinct seasonality and trends of natural and anthropogenic wildfires in boreal forests},
  journal = {Agricultural and Forest Meteorology},
  year = {2026},
  doi = {10.1016/j.agrformet.2026.111436},
  url = {https://doi.org/10.1016/j.agrformet.2026.111436}
}