Machine learning-based detection of dynamic changes in mangrove forest, Beilun Estuary

While mangrove forests in river-dominated estuaries have been extensively studied, wave-dominated estuarine mangroves have received relatively little attention. Here, a comprehensive analysis of mangrove forests in the Beilun Estuary, a typical wave-dominated estuary and the transboundary estuary between China and Vietnam, was conducted using a machine learning of continuous change detection and classification algorithm. The results revealed that the total area of mangrove forests in the Beilun estuary increased by 11.2 % during the period from 1986 to 2000, followed by a slight decline from 2000 to 2022. Over the past 37 years, a total of 568.4 ha of mangroves were lost, primarily on the landward side of the estuary's western part. However, this loss was offset by a gain of 688.1 ha, with newly established mangroves predominantly located behind the western barrier islands, as well as in the central and eastern parts of the estuary. Meanwhile, the shoreline of mangrove forests in the western part remained stable, while the central and eastern parts of the estuary presented seaward expansion at a slow rate of 2.1 ha/yr and 1.7 ha/yr, respectively. These findings suggest that the southeast waves limited the seaward expansion of mangrove forests in western estuary, while the combined action of tidal currents and waves resuspended and transported sediments to fill the tidal creeks, providing habitats for mangroves. Barrier islands formed by strong waves create a sheltered environment for mangrove growth. Land-use conversion to aquaculture ponds is the primary driver accounting for the loss and fragmentation of mangrove forests. However, the sea-level rise, at an increasing rate of 3.38 mm/yr, has not impeded the seaward expansion of mangroves in the Beilun Estuary. This research provides a profound perspective on dynamic changes in mangrove forests within a typical wave-dominated estuary, which can serve as a crucial reference for the future conservation and restoration strategies in a global scale. Copyright © 2025 Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies.