A review of new science, in plain language

Geophysical Research Letters  ·  28 August 2026  ·  From issue 28

Event‐Level Attribution of Wildfire Ignition Cause: A Machine‐Learning Approach

A computer model guesses wildfire causes, human or lightning, with 94.2% accuracy.

Kairui Qiu, Yan Yu et al.

A computer program can guess if a wildfire started from lightning or from people. It learned this by studying over one million past fires from around the world. It looked at weather, land shape, and how much people had changed the area nearby.

The number 94.2% accuracy

How often the model correctly guessed if a fire was started by lightning or people, in past records.

Tested on past records only, not live fires as they happen.

Integrity screen: passed (3 checks) Checked 29 August 2026. Retraction record: none. DOI resolves at doi.org. Metadata record found (Geophysical Research Letters). Read the source

This screen checks process signals. It does not tell you the findings are true.

Get it free, every two days

Five papers free every issue, about 15 emails a month. No spam. Leave in one click.

Also in issue 28 15 papers
More from Geophysical Research Letters