multi-source fusion tidal flat dataset of China at 10m resolution (2019-2025)

This dataset is the China 10M Tidal Flat Dataset, which includes 10 folders representing 10 coastal provinces in China. Among them, Shanghai is included in Jiangsu, Tianjin is included in Hebei, and Hong Kong and Macao are included in Guangdong.The data format is tif, and each tif file has one band. A value of 1 represents tidal flat pixels, while a value of 0 represents non-tidal flat pixels. Abstract:The evolution of tidal flats is constrained by both natural factors and human activities. Addressing the issues of frequent cloud cover and insufficient quantitative analysis of driving mechanisms in coastal China, this study developed an automated extraction model on the GEE platform. By fusing Sentinel-1/2 imagery to construct high-density time series, the model effectively overcame cloud and tidal influences, achieving high-precision dynamic monitoring from 2019 to 2025. Validation using "Core Area Sampling" showed an average Overall Accuracy of 86.25% (Kappa=0.73), demonstrating excellent robustness.Further analysis revealed significant spatial heterogeneity in driving mechanisms: the Liaoning, Jiangsu, and Zhejiang-Fujian coasts are primarily dominated or constrained by thermal conditions (sea ice), hydrological and sediment processes (precipitation), and hydrodynamic environments (wind and waves), respectively; while human activities, such as Spartina alterniflora control, are the main cause of abrupt local area changes. The proposed multi-source fusion scheme and analysis results provide a scientific basis for all-weather tidal flat monitoring and differentiated coastal management.本数据集为中国1000万滩涂数据集（China 10M Tidal Flat Dataset），包含10个文件夹，分别对应中国10个沿海省份。其中上海市划入江苏省范畴，天津市划入河北省范畴，香港、澳门划入广东省范畴。数据格式为tif，每个tif文件仅包含单波段。波段值为1时代表滩涂像素，值为0时代表非滩涂像素。 【摘要】滩涂演化同时受自然因素与人类活动共同制约。针对中国沿海地区常见的云量覆盖频繁、驱动机制定量分析不足等问题，本研究在Google Earth Engine（GEE）平台上开发了自动化提取模型。通过融合Sentinel-1/2影像构建高密度时间序列，该模型有效克服了云层与潮汐的干扰，实现了2019年至2025年的高精度动态监测。采用"核心区域采样法（Core Area Sampling）"进行验证，结果显示总体精度均值达86.25%（Kappa系数=0.73），展现出优异的鲁棒性。进一步分析揭示了驱动机制存在显著的空间异质性：辽宁、江苏以及浙闽沿海分别主要受热力条件（海冰）、水文与沉积过程（降水）以及水动力环境（风浪）的主导或制约；而人类活动（如互花米草（Spartina alterniflora）治理）则是局部区域面积突变的主要诱因。本研究提出的多源融合方案与分析结果，为全天候滩涂监测与差异化沿海管理提供了科学依据。