Monitoring dynamics of the Yalu River Estuary coastal wetland from 1990 to 2021 through the remote sensing approach

The Yalu River Estuary (YRE) coastal wetland, as a vital part of the Yellow Sea Large Marine Ecosystem, provides habitats for hundreds of thousands of wintering waterfowl and millions of migratory shorebirds along the East Asian-Australasian Flyway. Given the accelerated rate of wetland loss during the 20th and early 21st centuries worldwide, effectively monitoring the dynamic changes of the wetlands is critical for adaptive wetland management. This study utilized Landsat TM/OLI images in 1990, 2000, 2011, and 2021 as data sources, employing an object-oriented random forest model classification approach to analyze the dynamic changes of each wetland type in the YRE coastal wetland over the past 31 years. The results showed that: 1) the object-oriented random forest model achieved high accuracy in classifying wetland types in the study area, with an overall classification accuracy exceeding 88%; 2) shallow marine waters, tidal wetlands, aquaculture ponds, and paddy fields are the predominant wetland types over the years, distributed in bands successively along the coastline, from the sea to the land; 3) compared with 1990, natural wetlands in YRE were degraded by 2021 with a decrease of 1,880.93 hm 2 ; 4) the total area of natural and artificial wetlands remained relatively stable, due to the conversion of natural wetlands to artificial wetlands related to the development of agricultural, forestry, and fishery industries. The study indicates the reliability and efficiency of the application of the object-oriented random forest model classification approach on wetland monitoring. The trend of wetland change underscores the effectiveness of the reserve’s establishment.