Revealing the proximate drivers behind global tree cover loss using multisourced remote sensing products during 2000–2020

Abstract

Understanding the proximate drivers of tree cover loss is crucial for guiding forest management policies, while significant challenges remain in accurately and comprehensively identifying these drivers on a global scale. In this study, we developed a decision tree framework using multisourced remote sensing products to attribute the global tree cover loss to five human drivers and three natural drivers. Accuracy assessment results based on visually interpreted samples indicate that the driver classification of this study achieved a high overall accuracy of approximately 81.93 % in the pantropics, demonstrating the reliability of the proposed methodology. Further analysis revealed that human drivers were responsible for nearly 79.99 % of global tree cover loss between 2000 and 2020, with natural drivers accounting for the remaining 20.01 %. Among the eight drivers, agricultural encroachment and forestry activity were the most significant, contributing approximately 44.38 % and 11.91 % of total loss, respectively. Moreover, the rates of tree cover loss due to these eight drivers have increased significantly, with nearly a doubling observed in most cases; and these trends have shown no signs of stopping. Therefore, it is imperative to implement more robust forest management policies targeting specific drivers to curb tree cover loss by 2030.