Key points are not available for this paper at this time.
Accurate mapping of mangrove distributions can provide useful information for the protection and management of mangrove ecosystems. In this study, by combining simple noniterative clustering (SNIC) and the object-oriented random forest classification method, the SNIC-RF algorithm for long-term annual mangrove remote sensing extraction was developed using the Google Earth Engine (GEE) platform. Based on Landsat and Sentinel time-series images, annual mangrove maps from 2000 to 2023 in the Guangxi Beibu Gulf (GBG) of China were generated, with an average overall accuracy of 97.64%. The result revealed that the mangrove area in the GBG recovered from 6,089.79 ha in 2000 to 9,747.46 ha in 2023. From 2000 to 2012, the mangrove area fluctuated, with a total growth rate of 8.82%. From 2012 to 2023, the mangrove area grew rapidly, with an average annual growth area of approximately 280.95 ha and a total growth rate of 47.09%. This study can provide methodological and data support for the continuous dynamic monitoring, protection and management of mangrove ecosystems. Our findings contribute to a deeper understanding of mangrove recovery dynamics in coastal urban agglomeration regions, offering insights for global mangrove conservation and restoration efforts.
Zhijie et al. (Thu,) studied this question.