There are a limited number of studies addressing the forest status, its extent, location, type and composition over a larger area at the regional or national levels. The dense time series and a wide swath of Sentinel-2 data are a good basis for forest mapping and tree species identification over a large area. This study presents the results of the classification of the forest/non-forest cover, forest type (broadleaf and coniferous) and the identification of eight tree species (beech, oak, alder, birch, spruce, pine, fir, and larch) using the multi-temporal Sentinel-2 data in combination with topographic information. The study was conducted over the large mountain area located in southern Poland. The Random Forest classifier was used to first derive a forest/non-forest map. Second, the forest was classified into broadleaf and coniferous. Finally, the tree species classification was carried out following two approaches: (i) Non-stratified, where all species were classified together within the forest mask and (ii) stratified, where the broadleaf and coniferous tree species were classified separately within the forest type masks. The overall accuracy for the forest/non-forest cover reached 98.3% and declined slightly to 94.8% for the classification of the forest type. The use of the topographic information did not increase the accuracy of either result. The role of the topographic variables increased significantly in the process of tree species delineation. By combining the topographic information (in particular, digital elevation model) with the multi-temporal Sentinel-2 data, the classification of eight tree species improved from 75.6% to 81.7% (approach 1). A further increase in accuracy to 89.5% for broadleaf and 82% for coniferous species was observed following the stratified approach number 2. The highest overall accuracy (above 85%) was obtained for beech, oak, birch, alder, and larch. The study confirmed the potential of the multi-temporal Sentinel-2 data for accurate delineation of the forest cover, forest type, and tree species at the regional scale.

Fire plays an increasingly important role in deforestation and degradation of carbon-dense tropical peatlands in South-east Asia. In this study, analysis of a time-series of satellite images for the period 1973–2005 showed that repeated, extensive fires, following drainage and selective logging, played an important role in land-cover dynamics and forest loss in the peatlands of Central Kalimantan, Indonesia. A study of peatlands in the former Mega Rice Project area revealed a rising trend in the rate of deforestation and identified fire as the principal factor influencing subsequent vegetation succession. A step change in fire regime was identified, with an increase in burned area and fire frequency following peatland drainage. During the 23-year pre-Mega Rice Project period (1973–1996), peat swamp forest was the most extensive land-cover class and fires were of relatively limited extent, with very few repeated fires. During the 9-year post-Mega Rice Project period (1997–2005), there was a 72% fire-related loss in area of peat swamp forest, with most converted to non-woody vegetation, dominated by ferns or mosaics of trees and non-woody vegetation, rather than cultivated land.