Forest ecosystems globally are undergoing unprecedented changes driven by both natural events and anthropogenic activities. Events such as wildfires, logging, insect infestations, and extreme climate events disrupt forest structure, biodiversity, and global carbon balances. Understanding when, where, and how these disturbances occur—as well as monitoring the subsequent long-term ecosystem recovery—is critical for sustainable forest management, ecological restoration, and evaluating ecosystem resilience under ongoing global environmental change.

Our research group focuses on leveraging dense time-series Earth observation data and modern spatial analytics to map, quantify, and model forest disturbance and recovery dynamics. Our work emphasizes tracking monthly to interannual canopy changes, classifying distinct disturbance agents (such as fire, selective logging, and clearcutting), and evaluating post-disturbance vegetation trajectories. By combining high-resolution satellite remote sensing with deep learning frameworks, we aim to bridge the gap between regional observation and pixel-level ecological diagnosis, providing actionable scientific insights to anticipate future forest landscape transformations.