Answers are grounded in this paper and cite their sources.
Tropical forest degradation inside protected areas (PAs) poses a subtle yet persistent threat to biodiversity, carbon storage, and ecosystem services. However, fine-scale patterns of degradation, particularly along PA boundaries, remain poorly understood. The present study provides a spatially explicit assessment of edge-driven forest loss in three PAs of Bangladesh—Bhawal National Park (BNP), Teknaf Wildlife Sanctuary (TWS), and Rema-Kalenga Wildlife Sanctuary (RKWS)—over 2001–2023. Using the Hansen Global Forest Change dataset (30 m resolution), we quantified tree cover loss across concentric buffer zones (0–500m, 500–1000m, 1000–1500m) and stratified results by canopy density to identify structurally vulnerable forest patches. Our study finds that degradation was concentrated along edges: RKWS and TWS showed the highest losses in the 0–500 m zone (12.4% and 11.7%). Proximity to roads intensified disturbance across all sites, although the spatial gradients differed among sites. BNP experienced the highest loss within the 50 m buffer, whereas RKWS and TWS showed major loss peaks within the broader 500–1,000 m buffers. Canopy-dependent loss patterns also varied by site- in RKWS, high-canopy forests (>75% cover) were most affected, accounting for 85% of the total forest-loss area, while BNP showed loss in 65-75% canopy class and TWS in >80% canopy classes. BNP exhibited diffuse degradation, whereas TWS and RKWS showed intensified edge fragmentation. Our findings highlight that legal protection alone is insufficient to maintain forest integrity in the country. Combining high-resolution remote sensing with spatial analysis can offer a scalable approach to detect degradation hotspots and guide restoration priorities. Such an approach can inform sustainable forest management and support progress toward global restoration commitments, including the Bonn Challenge and the UN Decade on Ecosystem Restoration.
No takes yet. Share an insight, caveat, or question.
Saimun et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: