Forest Degradation Derived by a Newly Developed Sentinel-1 Change Detection Approach

Forest degradation often surpasses the area affected by deforestation and is therefore important under REDD+ monitoring, reporting and verification (MRV). Many studies dealing with forest degradation monitoring require the collection of sample data for algorithm training. The newly developed synthetic aperture radar (SAR)-based Forest Canopy Disturbance Monitoring (FCDM-radar) approach, implemented in Google Earth Engine and processing Copernicus Sentinel-1 data, allows the operational monitoring of evergreen and deciduous forest types without the need of collecting training data for calibration purposes. With a spatial resolution of 10 m it is able to detect small-scale forest canopy disturbance events such as single tree removals of 0.03 ha. Within a study area in central Cambodia (Prey Lang Wildlife Sanctuary), covering evergreen to semi-deciduous forests which are strongly affected by illegal selective logging activities, small-scale forest disturbances were successfully detected with an overall accuracy of 96.3% (user accuracy of 94.0%; producer accuracy of 79.1%).