Landscape metric-enhanced cellular automaton for simulating land-use change in coastal wetlands

Coastal wetlands are key parts of blue carbon ecosystems, with unique structures and strong carbon sequestration abilities, playing vital roles in coastal protection. Traditional cellular automata (CA) models for land-use change often rely on pixel-level data, which fails to fully capture landscape structure. This study proposes a landscape structure-enhanced CA (LE-CA) model that integrates landscape features and cell transition coherence. The LE-CA model outperforms traditional pixel-based CA models in landscape similarity and prediction accuracy. It provides a solid scientific basis for sustainable coastal wetland planning and valuable insights for land-use change management.