Vegetation Dynamic in a Large Floodplain Wetland: The Effects of Hydroclimatic Regime

Floodplain wetlands are among the most dynamic ecosystems on Earth, featuring high biodiversity and productivity. They also sensitive to anthropogenic disturbances globally threatened. Understanding how flow regime drives spatiotemporal dynamics of wetland habitats is fundamental effective conservation practices. In this study, using Landsat imagery random forest (RF) machine learning algorithm, we mapped winter distribution four (i.e., Carex meadow, reedbed, mudflat, shallow water) in East Dongting Lake, a Ramsar middle lower Yangtze Basin China, for 34 years (1988–2021). The were explored through pixel-by-pixel comparisons. Further, response variations was investigated generalized additive mixed models (GAMM). Our results demonstrated constant expansion reedbeds shrinkage mudflats, that there three processes contributing reduction mudflat: (1) permanent replacement by reedbed; (2) irreversible loss water; (3) transitional swapping with meadow. These changes relative extent may degrade function wetland. Moreover, duration dry season date water level withdrawal identified as key parameters shaping size habitats. However, different vegetation showed distinct responses regime: while meadow increased earlier longer season, reedbed continuously expanded independent corresponding increase rainfall. findings suggested acts concert other factors, such climate change sand mining river channels, driving habitat transition floodplain landscape. Therefore, can only be achieved diverse restoration strategies addressing all drivers.