Early warning

A roving crew installs cameras on California’s highest, craggiest ridgelines to catch wildfires early. Since 2023 artificial intelligence has scanned the network’s footage, and it now spots 51 per cent of the state’s wildfires first.

Federal agencies in the United States already use drones to map wildfires and detect hotspots. The Forest Service is now testing ground robots, and researchers are training machine-learning models on drone imagery to predict how fast a fire will spread.

Firefighter and researcher Marc Castellnou sends weather balloons into towering wildfire plumes to measure temperature, wind and humidity, opening a new way to anticipate when a fire turns extreme – and why some keep burning fiercely at night.

An EU project is building a multi-risk early warning system that forecasts who and what a hazard will affect, not just where it will strike.

This article reports public reactions to a UK emergency phone alert on wildfire risk, highlighting concerns over alert fatigue, unclear guidance and the need for effective crisis communication.

Southwest Research Institute (SwRI) developed a wildfire early detection system that combines real-time modelling, remote sensing and multiple datasets to help emergency officials decide when to alert at-risk communities and where to deploy resources.

Researchers at University of Montana announced a free high-tech tool used to better understand forests and rangelands across the country. It can identify areas at high risk of wildfire and high-severity fire where catastrophic events can occur.