Large forest fires regularly affect large parts of the world and cost both human lives and enormous economic losses.
But now researchers have trained an AI to analyze which measures work best against various forest fires in Indonesia.
The AI was able to predict where forest fires would occur in 80-95 percent of cases and suggest measures that would reduce the risk by 50-76 percent.
The model only works for peatlands in Indonesia at the moment, but the researchers believe that the method works for other areas as well and that it's just a matter of giving the AI new data to train on.
The model uses 31 different variables to analyze where fires can occur. It can then create a map that shows how likely it is for a fire to occur in different areas.
The measures proposed by the AI may concern recreating wetlands or changing cultivation methods on certain lands. That can have a major impact on the local population's livelihood. So the researchers see the AI primarily as a tool for decision-makers.
"We have tried to quantify how different strategies can work. It's more about informing decision-makers than coming up with direct solutions," says Matti Kummu, professor at Aalto University in Finland and the one who led the research, in a press release.
Marco Borsari
Linn Winge
