Identifying landscape functional states in polder heritage systems using remote sensing and self-organizing maps

The polder landscape heritage of the Taihu Basin in China represents a complex social–ecological system shaped by long-term human–water interactions. However, the functional organization of this landscape remains poorly understood. This study integrates remote sensing indicators, self-organizing maps, and landscape ecological theory to identify distinct functional states in polder landscape heritage systems. We first derive a landscape structural gradient from spatial configuration metrics and then classify functional states based on vegetation and hydrological remote sensing indicators. The results reveal three distinct functional states: highly disturbed artificial landscapes, high-productivity agro-composite systems, and hydrologically dominated pond-based polder ecosystems. These states exhibit nonlinear spatial distributions along the structural gradient, indicating that landscape functional organization is characterized by the coexistence of multiple configurations rather than by continuous variation. The findings highlight the importance of maintaining hydrological connectivity and landscape heterogeneity for the conservation and adaptive management of polder landscape heritage.