New JAXA High-Resolution Land Use/Land Cover Map for Vietnam Aiming for Natural Forest and Plantation Forest Monitoring

摘要: Highly detailed and accurate forest maps are important for various applications including monitoring, forestry policy, climate change, biodiversity loss. This study demonstrates a comprehensive geographically transferable approach to produce 12 category high-resolution land use/land cover (LULC) map over mainland Vietnam in 2016 by remote sensing data. The included several natural categories (evergreen broadleaf, deciduous (mostly broadleaf), coniferous evergreen coniferous)) one representing all popular plantation forests such as acacia (Acacia mangium, Acacia auriculiformis, hybrid), eucalyptus (Eucalyptus globulus), rubber (Hevea brasiliensis), others. combined the advantages of sensor data integrating their posterior probabilities resulting from applying probabilistic classifier (comprised kernel density estimation Bayesian inference) each datum individually. By using different synthetic aperture radar (SAR) images (PALSAR-2/ScanSAR, PALSAR-2 mosaic, Sentinel-1), optical (Sentinel-2, Landsat-8) topography (AW3D30), resultant achieved 85.6% overall accuracy. major classes broadleaf had user’s accuracy producer’s ranging 86.0% 95.3%. Our identified 9.55 × 106 ha (±0.16 ha) 3.89 (±0.11 Vietnam, which were close Vietnamese government’s statistics (with differences less than 8%). study’s result provides reliable input/reference support policy sciences Vietnam.