The Effect of Dominant Land-Cover Transitions in Shaping Trajectories of Global Forest Change

Land-cover transitions from forest to non-forest land-covers have led to substantial loss and fragmentation of global forests. While average rates of forest change are well studied, less attention has been paid to spatio-temporal trajectories of forest loss and recovery. These trajectories likely differ by land-cover transition types, and may be constrained by the initial amount and spatial arrangement of the non-forest land-cover. In this study, we distinguish between abrupt and gradual forest change trajectories, evaluating how they vary across land-cover transitions and how initial non-forest land-cover conditions mediate the amount and rate of forest change. To this end, we use annual global land-cover maps (1992–2020) from the European Space Agency’s Climate Change Initiative. We find that agricultural expansion (cropland) and land degradation (bare land) drive forest loss, while (re)forestation of wetland, shrubland and grassland land-covers are key contributors to forest recovery. Forest loss rates were lower when non-forest land-cover was initially fragmented, except for settlements and bare land, where fragmentation accelerated loss. Conversely, forest recovery was highest in wetlands and shrublands, through fragmentation generally reduced the amount and rate of forest gain. Our results suggest that identification of the main mode of land-cover transition, and the initial spatial arrangement of the non-forest land-cover, provide diagnostic tools to constrain future spatio-temporal trajectories of forest change. This insight can improve our ability to link remotely-sensed land-cover changes to the underlying drivers of global forest change.