Carbon Stock Changes of African Forests in the 21st Century

Abstract:

Climate change mitigation through the reduction of emissions from deforestation and degradation (REDD+) has presented a great potential for developing countries globally. In Africa, in particular, REDD+ has the benefits of reducing the extensive rate of woody vegetation disturbance, driven mainly by subsistence agriculture, timber harvesting, charcoal making and fuel wood collection. However, estimates of carbon stocks and emissions that can be used for implementing climate mitigation at local to jurisdictional scales are missing. Here we use a combination of ground, airborne, and satellite observations of vegetation structure to map woody vegetation live biomass carbon stocks annually from 2000-2019. We integrate samples of lidar forest structure from extensive airborne data and samples from GLAS, ICESAT-2, and GEDI data in a deep convolutional neural network (CNN) machine leaning model to map above ground carbon stocks at 1-km spatial resolution. We examine the carbon stock changes across the landscape along with observations of disturbance to provide annual carbon emissions spatially and along the gradient of vegetation from savanna and shrublands to vast areas of woodlands (e.g. miombo) and to the extensive remaining rainforests of the Congo Basin.