High-resolution satellite imagery has been used to map every tree in Africa, demonstrating a technique that could help improve monitoring of deforestation around the world.
Florian Reiner at the University of Copenhagen, Denmark, and his colleagues used images from satellites operated by the US company Planet and machine learning models to map canopy cover across the entire African continent.
Modern satellites typically capture tree canopy at 30 meter resolution – good for measuring the size of forests, but less good for mapping individual trees and small thickets.
The satellite data used by Reiner and his colleagues had a resolution of 3 meters, allowing the study to map all trees, including those that are not part of a forest.
The results suggest that 30% of all trees in Africa are not found in a forest, but rather are scattered across farmland, savanna and urban areas.
Many countries in Africa lack dense forests, but still have plenty of trees, says Reiner. “These trees are extremely important to local ecosystems, to people, to the economy.”
Similar research has also been conducted map canopy cover across Europewhich reveals that in some countries up to 24 percent of tree cover is outside of forests.
By tracking each tree or thicket, researchers can begin to monitor how well those trees are coping with climate change, Reiner says, or if they’re vulnerable to deforestation. It could also improve tracking of reforestation efforts, which are gaining popularity as a way to remove carbon dioxide from the atmosphere.
“At the local level, being able to continuously monitor when and where trees are disappearing or reappearing can generate more actionable insights,” says Jean Francois at the Alan Turing Institute in London.
The study is a proof of concept rather than a map ready for immediate commercial use, Reiner explains. “It’s research work. it shows what could be done,” he says.
But he is already working with colleagues to expand the monitoring approach to cover the entire global canopy: “We hope this will be seen as a way forward in monitoring tree resources.