Rapid satellite monitoring reveals complexity of tropical tree cover loss drivers at fine scale

Satellite-based tropical forest monitoring is widely adopted to inform where and when tree cover loss occurs. However, timely and fine-scale information on the direct drivers of tree cover is currently lacking, which limits detailed understanding of forest changes and possibilities to avert illegal or unsustainable activity. Here, we introduce driver-specific monthly monitoring of tropical tree cover loss at 10 m spatial scale, distinguishing eight human-induced and natural driver classes across the Amazon, Congo Basin and Southeast Asian forest for 2022-2024. We attributed direct drivers to pantropical forest disturbance alerts in a simulated near real-time scenario, with deep learning models based on radar and optical satellite imagery. The results reveal strong local-scale complexities, with many different drivers commonly co-occurring within small areas and time frames. The fine spatial scale provides detailed insight into hotspots of small-scale drivers such as road development, selective logging, mining, shifting cultivation, and natural disturbances, while the high temporal resolution reveals their sequences in time, their interactions, and seasonal variations. Timely monitoring of these fine-scale patterns and interactions supports a better understanding of the causes and impacts of forest changes, and strengthens the actionability of forest disturbance alerts for conservation efforts and policies.