National coastal wetland mapping over the last four decades: An annual classification with high accuracy.

As a critical component of global wetland ecosystems, coastal wetlands are facing dual pressures from climate change and human activities, highlighting the urgent need for a precise classification dataset with large spatial scale and high temporal resolution. The national coastal wetland mappings year by year over the last four decades are crucial for coastal integrated management. A robust and automated classification method was developed based on coastline delineation and integrations of multiple algorithms. By analyzing more than 60,000 Landsat images of China's coastal areas and incorporating field survey validation data, an annual coastal wetland spatial distribution dataset with 30-meter spatial resolution spanning from 1985 to 2024 was established. The dataset obtained an average Kappa coefficient of 0.957 and a range from 0.978 to 0.935, and the average overall accuracy was 97.6% with a range of 97.12% - 97.85%. The proposed method has achieved high-precision annual coastal wetland spatial datasets and effectively monitored the impacts of invasive species and ecological restoration projects on interannual changes in coastal wetlands.