Abstract
The United Nations sustainable development agenda emphasizes the importance of forests. China's forests cover 5% of the world's forest area, significantly influencing global climate and ecology. In recent decades, China's forests have undergone notable changes. Accurate forest cover maps are crucial for understanding forest distribution, conducting ecological research and sustainable management. However, there is a lack of forest cover maps satisfying the criteria. To this issue, this study focuses on developing a precise 16-m resolution forest cover map of China. For this purpose, we propose a forest classification framework based on weakly supervised deep learning and prior knowledge from open datasets. Utilizing this framework and GF-1 WFV satellite images, we generated China's forest cover map in 2020 named FCM16. The FCM16 is evaluated using 136,385 sample points, achieving an overall accuracy of 94.64 ± 0.12%, producer's accuracy of 91.12 ± 0.27% and user's accuracy of 87.31 ± 0.34%. Additionally, FCM16 was compared with existing forest-related datasets, demonstrating its reliability. In general, FCM16 effectively represents China's forest cover in 2020, providing a valuable resource for social and ecological analysis.