Development of island-specific land-use classification system and spatiotemporal pattern evolution analysis using Landsat time series for coastal zone management: A case study of Zhoushan Archipelago, China

• By comprehensively consideration the unique characteristics of coastal management and island geographical environments, the island-specific land-use classification system is designed. • The long-term land-use mapping of Zhoushan Archipelago, China was carried out by integrating multi-feature parameters and employing the Random Forest Algorithm. • The characteristics and regularities of land-use and landscape pattern in Zhoushan Archipelago over the recent decades have been clarified. As a special geographical unit, islands serve as important platforms for expanding the blue economy. Existing land-use classification systems predominantly target terrestrial areas and often lack consideration of complex marine environments and periodic tidal effects on islands, which lead to insufficient applicability and lower precision in information extraction. To address these issues, this study adopted a land-sea coupling perspective and the long-term temporal analysis. By comprehensively consideration the unique characteristics of geographical environments and land-use modes, a hierarchical land-use classification framework was developed specifically for island environments through systematic integration of spectral, temporal, and geomorphological features derived from medium-resolution satellite imagery. Then, land use information was extracted by integrating multi-feature parameters and employing the Random Forest Algorithm (RFA), followed by an exploration on the spatiotemporal dynamics of it. The research results of Zhoushan Archipelago show that: (1) The developed classification system effectively consider the influences of periodic tidal inundation and land-sea interactions, more accurately reflecting the natural resource endowments and land-use patterns of island areas. (2) Integrating multi-feature parameters, the RFA demonstrated strong performance in land-use information extraction from remote sensing data of the Zhoushan Archipelago between 1985 and 2024, achieving an average overall accuracy (OA) of 93.56 % and a Kappa coefficient of 0.93. (3) The total land area of Zhoushan Archipelago exhibited a sustained expansion from 1,274.42 km 2 in 1985 to 1,456.57 km 2 in 2024. Urbanization triggered dramatic expansion of construction land, increasing from 47.62 km 2 to 426.50 km 2 (8.95 times). (4) Due to industrial transformation, the salt pans area declined by as much as 94.88 %. Driven by the fishery economy, aquaculture ponds areas experienced substantial expansion, from 8.08 km 2 to 35.09 km 2 , an increase of 4.34 times. Coastal reclamation primarily originates from shallow marine waters and bare intertidal flats, and the newly reclaimed land mainly allocated for construction land (119.12 km 2 ), cropland/grassland (29.04 km 2 ) and aquaculture ponds (17.66 km 2 ). (5) Landscape indices analysis indicated that, the Zhoushan Archipelago has undergone intensified landscape fragmentation, enhanced diversity, and decreased connectivity, although these trends have eased in recent years. This study can offer valuable guidance for the sustainable development and in-depth research of island land use.