Sentinel-1 SAR Coherence as a Conservative Alert Layer for Tree Cover Loss in Estonia

Abstract

Monitoring tree cover loss at national scale requires methods that are weather-independent and operationally scalable. Here, we evaluate Sentinel-1 SAR interferometric coherence as a conservative alert layer for tree cover loss in Estonia, using multitemporal Airborne Laser Scanning (ALS) data and visually interpreted samples for validation. Coherence time series (VV polarisation, 12-day repeat) were analysed with the Continuous Change Detection and Classification (CCDC) algorithm and calibrated against wall-to-wall ALS-derived tree cover loss maps. At pixel level, coherence-based detection achieved User’s Accuracy of about 77–78% and Producer’s Accuracy of about 41% for tree cover loss, indicating that when loss is flagged it is often correct, yet a substantial fraction of true loss pixels remains undetected. Patch-level analysis is more encouraging: under a conservative definition (≥20% overlap between detected and ALS patches), coherence detects about 33% of patches ≤1 ha and 46% of patches >1 ha, confirming greater reliability for larger disturbances. Temporal validation with ~1200 visually interpreted samples confirms that correctly detected patches are typically identified within several months of the reference loss date, despite uncertainties in the timing of optical reference data. Overall, Sentinel-1 coherence alone is not suitable as a standalone national tree cover loss map, but it can support conservative, cloud-independent alerting for larger disturbance patches and complement optical and backscatter-based monitoring systems.