IAP0.25度分辨率盐度观测格点数据集

本数据集使用前馈神经网络（FFNN）方法，将卫星遥感获得的全球尺度的高分辨率海表绝对动力地形（ADT）、海表温度（SST）及海表风场（SSW）等参数信息与海洋现场盐度观测数据、1° × 1°分辨率的IAP格点观测数据进行融合，构建了一套0.25° × 0.25°的覆盖海洋上层2000米的盐度观测格点数据集（1993-2018月平均）。该数据集相比较粗分辨率的数据能够更好的重现海洋中小尺度的信号，例如湾流、黑潮、南大洋等涡旋活动较为丰富区域的海洋变率。同时，在大尺度的信号上维持了IAP 1° × 1°分辨率格点数据的优势，能够准确重建大尺度变率。

This dataset employs the Feedforward Neural Network (FFNN) approach to fuse high-resolution global-scale sea surface parameters derived from satellite remote sensing, including absolute dynamic topography (ADT), sea surface temperature (SST), and sea surface wind (SSW), with in-situ ocean salinity observation data and 1° × 1° resolution IAP gridded observation data. A gridded salinity observation dataset with a resolution of 0.25° × 0.25° covering the upper 2000 meters of the ocean (monthly averages from 1993 to 2018) is thus constructed. Compared to coarser-resolution datasets, this product better reproduces mesoscale and small-scale ocean signals, such as ocean variability in regions with intense eddy activities like the Gulf Stream, Kuroshio Current, and Southern Ocean. Meanwhile, it retains the advantages of the original 1° × 1° IAP gridded data in representing large-scale signals, enabling accurate reconstruction of large-scale ocean variability.