Why are we losing tropical forests? Satellites can tell!

Every year, we lose around 5 million hectares of forest globally, an area larger than the entire Switzerland. The vast majority happens in the tropics where some of the most valuable forest ecosystems on the planet are located. Tropical forest loss is associated with significant carbon emissions, loss of biodiversity and severely impacts local communities.
To tackle tropical forest loss effectively, it is necessary to understand the driving forces behind it. Policies targeted at specific drivers - such as agriculture expansion or mining - have been successful in reducing deforestation in the Brazilian Amazon. But since drivers vary depending on the region, these measures cannot be readily applied elsewhere with the same effect.
In a new study published in Environmental Research Letters, researchers from EPFL’s Environmental Computational Science and Earth Observation Laboratory, developed a machine learning model that uses time series of satellite images to determine the drivers of forest loss across the tropics. The model is designed to overcome specific challenges associated with this task, such as high cloud coverage.
The map of drivers produced by the model shows fundamental differences in forest loss patterns in South America, Sub-Saharan Africa and Southeast Asia, and also allows to observe variations at a smaller scale. It highlights the need for locally targeted measures and can also serve as a source of recent, up-to-date data for the research community working with tropical forests.