Wall-to-wall mapping of Fagus orientalis Lipsky distribution across Türkiye’s forests using satellite remote sensing

Beech forests are a crucial component of Türkiye’s ecosystems, yet their timely and efficient monitoring remains challenging and time-consuming. Recent advances in remote sensing have made tree species classification increasingly feasible, but one-class classification of individual species remains difficult due to limited and uneven reference data. In this study, we implement a practical method for mapping and monitoring beech forests in Türkiye using Sentinel-2 imagery. We further evaluate the performance of different seasonal image datasets to determine the most suitable input for national-scale beech forest classification and to produce a wall-to-wall beech map for the country. The results show that summer imagery provides the best classification performance, reaching an overall accuracy of 81% and F1-score of 0.85. The developed model was then applied nationwide and integrated into an interactive Google Earth Engine (GEE) application. By applying the model to imagery from 2016 to 2025, the tool generates annual beech forest maps and enables both spatial and temporal assessment of beech distribution. The results also show that model performance varies across forest types, highlighting the importance of ecological and structural context for classification success. The resulting map product can support local field measurements and provides a useful basis for both current assessments and historical forest monitoring studies.