More than 1.8 billion individual trees can be found in the West African Sahara, Sahel and sub-humid zone, according to a report in Nature this week. A combination of high-resolution satellite imaging and deep learning has revealed a relatively high density of tree coverage for this arid area.
Accurate mapping of non-forest trees could offer insights into how they shape the environment and support life around them. However, counting individual trees and shrubs can be challenging as commonly available satellite technologies have resolutions of around 10–30 metres, which is not good enough to pick out individual trees. Previous attempts to map dryland areas have suggested that they are largely free of trees and shrubs.
Martin Brandt, Rasmus Fensholt, Kjeld Rasmussen, Ole Mertz, Christin Abel, Laura Rasmussen, Sizhuo Li and colleagues present a detailed count of trees and shrubs (with canopies of more than 3m2) in the West African Sahara, Sahel and sub-humid zone, over an area of 1.3 million km2. They use more than 11,000 high-resolution satellite images (0.5-m spatial resolution) and deep learning techniques to detect more than 1.8 billion individual trees. They show that tree size and density are associated with rainfall and land use, with canopy cover increasing from 0.1% in hyper-arid areas to 13.3% in sub-humid regions.
Isolated trees and shrubs in arid regions have a number of important roles, such as contributing to biodiversity, offering shelter and food resources, and helping to maintain the environment (for example, by preventing erosion and acting as carbon sinks). The approach described by Brandt and colleagues suggests a way to monitor trees outside forests globally, and to explore their role in mitigating degradation, climate change and poverty.