Northern Mongolian Forest Dynamics Monitoring Using Copernicus Satellite

Mongolian’s boreal forests form the southern edge of the Siberian taiga and face increasing pressure from wildfires and logging. This study evaluates forest cover dynamics in northern Mongolia using single -date Sentinel 2 imagery acquired between 2020 and 2025. Forest and non-forest areas, together with dominant forest classes, were mapped using Random Forest classifier to evaluate spatial and temporal patterns of change. Total forest area remained relatively stable during the study period, ranging from 4,987 km² to 5,017 km², which represents 95.7–96.3% of the 5,211.6 km² study area. Among the mapped forest classes, Manchurian birch (Betula platyphylla) occupied the largest area (2,060–2,247 km²), followed by Siberian larch (Larix sibirica) (1,302– 1,455 km²), Scots pine (Pinus sylvestris) (567–764 km²), Siberian pine (Pinus sibirica) (505–644 km²), and Siberian spruce (Picea obovata) (205–294 km²). However, due to spectral overlap among species, these class-specific estimates should be interpreted as indicative rather than definitive. Classification performance was high, with overall accuracy ranging from 0.959 to 0.971 and Kappa values between 0.949 and 0.964. The results indicate that Sentinel-2 imagery is a practical and cost-effective tool for annual forest-cover monitoring in northern Mongolia, although species-level estimates should be interpreted cautiously due to mixed forest stands.