Time series satellite remote sensing reveals annual distribution, composition, and trajectory of tidal wetlands in the Yellow River Delta

Abstract

Tidal wetlands are vital ecological and economic bridges connecting terrestrial and marine ecosystems. Despite their importance, tidal wetlands face threats from human interference, biological invasion, unstable hydrology, and climate change, while remote sensing observations of them still face issues such as turbid water disturbance, limited datasets, and spectral similarity. As a result, detailed, long-term, and accurate annual mapping of tidal wetlands is often lacking globally, limiting the understanding and conservation of this important ecosystem. Taking the Yellow River Delta (YRD) as the studying site, we proposed an efficient and generalizable algorithm with Landsat images from 1985 to 2022 to systematically map the yearly tidal wetlands. The major components include tidal flats, dry flats, Suaeda salsa, Tamarix chinensis, Phragmites australis, artificial ponds and Spartina alterniflora. This algorithm effectively avoids interference from turbid water in tidal flat monitoring and spectral similarity in salt marsh monitoring while ensuring the reliability of image-based sample collection, achieving an average overall accuracy of over 92 %. The resultant 30-m spatial resolution map is the most comprehensive long-term observation of tidal wetlands to date, detailing their annual distribution and composition in the YRD. Furthermore, we investigated the evolution of the various components and subregions and found that the trajectory of tidal wetlands in the YRD exhibited six stages. This research aims to offer a robust solution for accurate tidal wetlands mapping and enhance our understanding of these ecosystems in turbid estuaries.