This canopy dieback map product was developed under the CIFOA Monitoring Program to support forest health assessment following extreme climate events. Using Sentinel-2 satellite imagery and a Random Forest machine learning model, canopy condition was classified into three standardised classes: live, partially dead, and dead. The model was trained using high-resolution manual classifications and applied across the NSW North Coast to map the spatial extent of canopy dieback observed after the 2023 early-season drought. The dataset contains 57 Tifs and one QGIS Layer Settings file.