海洋浮游植物群落结构信息对了解和研究碳循环和气候变化至关重要。浮游植物中含有的多种色素能够用于描述浮游植物群落结构特征和生理状态，浮游植物色素浓度的检测具有重要意义。本研究基于2016–2018年渤海、黄海和东海7次航次调查实验中采集的实测数据（包括浮游植物吸收系数、16种典型色素浓度和遥感反射率），利用高斯分解法开发了一种基于浮游植物吸收的色素浓度遥感模型，并使用原位观测数据集进行评估，得到的结果的误差可接受（例如，大部分色素模型的平均绝对百分比误差MAPE小于60%）。卫星匹配验证也表明反演结果与实测数据的一致性（大部分色素模型的MAPE范围在40%–60%之间）。将开发的模型应用于SeaWiFS和MODIS-Aqua遥感反射率月平均产品（1998–2020年），获得了渤海、黄海和东海区域16种色素浓度23年的时空变化数据记录。浮游植物色素浓度遥感数据集可以从https://doi.org/10.17632/bhcznf2m7v.1下载获得。本研究从遥感反射率卫星数据中得到复杂浑浊沿海水域中的浮游植物色素浓度，从而为渤海、黄海和东海精细化的海洋浮游植物群落结构研究提供数据支撑。

Marine phytoplankton community is essential for understanding the carbon cycle and climate change. Phytoplankton pigments can describe the composition and physiological state of phytoplankton communities. The detection of phytoplankton pigment concentrations is of great significance. Remote sensing technology permits macroscopic long-time series monitoring of phytoplankton pigment concentrations. Existing studies still have some limitations: 1) no remote sensing methods have been able to retrieve more types of pigments, and the existing studies mainly focus on a few pigments or pigment groups; 2) the existing pigment inversion algorithms are mostly based on oceanic water data, and the studies for optical Class II waters off China are not sufficient; 3) there is a lack of satellite remote sensing datasets of long time series of multiple phytoplankton pigment concentrations for phytoplankton-related fields to provide data support. This study collected phytoplankton absorption, 16 pigment concentrations, and remote sensing reflectance through seven cruise experiments in the Bohai Sea, Yellow Sea, and the East China Sea from 2016 to 2018. We developed a remote sensing model and further constructed a long-time series dataset of the spatiotemporal distribution of phytoplankton pigment concentration. In this study, remote sensing retrieval of concentrations was achieved by establishing the relationship between phytoplankton absorption and 16 pigments. The measured absorption coefficients were decomposed into Gaussian functions and the relationship between the Gaussian parameters and the measured pigment concentration was analyzed to construct inversion models. The two-component model of phytoplankton size classes was also used to achieve hyperspectral phytoplankton absorption. The models were evaluated for consistent performance and assessed using in situ datasets and leave-one-out cross-validation methods, yielding competitive and acceptable error results (e.g., mean absolute percentage errors (MAPEs) below ~60% for most pigments). Satellite-measured validation also produced promising prediction errors, yielding MAPEs in the range of 40%-60% for most pigments. The developed models were further applied to SeaWiFS and MODIS-Aqua remote sensing reflectance monthly mean products (1998-2020) to obtain a 23-year data record of the spatiotemporal patterns of 16 pigment concentrations in the Bohai Sea, Yellow Sea, and East China Sea. The satellite remote sensing dataset shows that 16 pigment distributions are similar, showing a decreasing trend from nearshore to offshore waters. In the Bohai Sea, the pigment concentration is high in winter and spring and low in summer. In summer, the pigment concentration peaks in the coastal areas from Jiangsu shallows to Zhejiang and Fujian. There is a triangular area of high concentration at the Yangtze River Estuary, with the triangular area extending from west to east in autumn and winter. The phytoplankton pigment concentration is relatively low for the outer deep water area, and the variation of concentration with the season is slight. The remote sensing datasets of 16 phytoplankton pigment concentrations can be downloaded from https://doi.org/10.17632/bhcznf2m7v.1. Scholars in related fields can research macroscopic and continuous phytoplankton community structure monitoring and physiological characteristics of phytoplankton in the Bohai, Yellow, and East China Seas based on the information from the pigment concentration remote sensing dataset. This dataset can enrich the understanding of marine phytoplankton pigment distributions and provide data support for satellite-based detection of phytoplankton community composition, which hopefully can inspire scholars.