Distribution of mangrove community Kandelia obovata in China for 2020 based on separatable time period Sentinel-2 imagery

The protection, restoration, and management of China's mangroves are shifting from a focus on area to a focus on ecosystem functions. Different mangrove communities have different ecosystem functions; for example, communities with well-developed prop roots play a larger role in shoreline stability. However, national-scale mapping of mangrove communities is still lacking. Using fine-scale data sources can provide local-scale community distribution information, but the phenological changes make it challenging to generalize these methods to large latitude-span regions, such as the coastal China. To address the issue, we took Kandelia obovata community as an example, and explored an approach for identifying this mangrove community across coastal China using separatable time period Sentinel-2 imagery and a random forest algorithm. By analyzing the separable metric variations in southern and northern case areas, a time period with higher separability for identifying K. obovata was determined to train a random forest classifier. With the help of visual corrections using contemporaneous Google Earth imagery, we obtained the distribution of K. obovata community in China for the year 2020. Using an independently collected validation sample set, the overall accuracy of the produced map was 88.5%, indicating that the proposed approach is suitable for mangrove community identification across a large latitude-span. This research provides solid support for species-level mangrove conservation, accurate estimation of biomass and carbon storage, precise ecosystem service assessments, and sustainable mangrove management.