基于图神经网络构建的全球溶解氧格点数据集(1960-2023)

该数据集为全球海洋的网格化溶解氧重建数据。重建数据基于WOD，Argo，CCHDO，Geotraces，GLODAP等海洋多变量观测数据，通过设计新型图神经网络算法，构建了1960年到2023年、1°×1°经纬度网格、0-5500米（33个标准深度层）的年平均格点数据。这一数据集为海洋科学研究提供了完整、精确的基础数据支持，有助于深入理解全球海洋缺氧环境的变化规律，并进一步研制了“精卫”海洋科学数据平台（https://jingwei.acemap.info）

This dataset is a gridded reconstructed dissolved oxygen dataset for the global ocean. Developed using a novel graph neural network algorithm and based on multi-variate in-situ ocean observational datasets including WOD, Argo, CCHDO, Geotraces and GLODAP, it provides annual mean gridded data spanning from 1960 to 2023, with a 1°×1° longitude-latitude grid resolution and 33 standard depth layers ranging from 0 to 5500 meters. This dataset delivers complete and accurate foundational data support for oceanographic research, enabling in-depth insights into the variation patterns of global ocean hypoxic environments. Furthermore, the "Jingwei" Marine Scientific Data Platform (https://jingwei.acemap.info) was developed based on this dataset.