Sub-meter commercial satellite data for semi-arid tree carbon restoration

We report recent progress mapping large numbers of tree crowns at the 50 cm scale and converting these into tree carbon using Maxar imagery in the Sahel Zone of Africa south of the Sahara. While 50% of Sahelian trees have diameters < 4 m, the only appropriate satellite data to accurately map small diameter trees are provided by the DigitalGlobe satellites. Coupling extensive field-data collections, machine learning, and high-performance computing, we determined the wood, foliage, and root carbon for every tree in the 0 mm to 1000 mm/year rainfall zone in our study. This methodology has potential for successful carbon restoration in semi-arid areas.

Figure 1. Monitoring at the level of single trees from Khombele, Senegal. a, A 50 cm-scale image from 2002; and b, a 50 cm satellite image from 2021 showing an agroforestry area at the same location. Tree cover has increased between 2002 and 2021 and the carbon density of both areas was calculated and increased from 6 to 10 Mg ha-1. A large number of trees grow on farmlands keeping the soils fertile and reducing the need for fallow periods. The greyscale of the background images indicates the carbon density per hectare while the color scale shows the carbon content of individual trees. This is a good example of the tree restoration monitoring potential of our approach.