Mapping individual tree dynamics in the hyper-arid northern Saudi Arabia

Hyper-arid drylands host sparse yet ecologically important tree populations that play essential roles in microclimate regulation, biodiversity support, and human livelihoods. However, the distribution and dynamics of these trees remain poorly quantified due to their small crown sizes and scattered spatial patterns, which pose challenges to traditional satellite-based monitoring approaches. Our study addresses this knowledge gap by integrating very high-resolution (0.5 m) satellite imagery with deep learning methods to map more than 11 million individual trees across 29,437 km2 of northern Saudi Arabia and to track their growth and mortality between 2017 and 2024.