基于深度学习海表风应力模型驱动的海气耦合模式用于ENSO预测（后处理数据）

原始数据主要包括ERA5再分析数据，时间分辨率为月，空间分辨率为1°×1°。此外，也包含由深度学习模型生成的纬向和经向风应力异常数据，主要用于进一步生成预测模式的初始状态。成果数据包括在预测时用于海气耦合模式初始化过程的海洋和大气变量，以及1996-2023年的预测结果。

The raw data mainly consists of ERA5 reanalysis data, with a monthly temporal resolution and a spatial resolution of 1°×1°. Additionally, it also includes zonal and meridional wind stress anomaly data generated by deep learning models, which are primarily used to further generate the initial states for prediction models. The resulting data covers oceanic and atmospheric variables utilized in the initialization procedure of the air-sea coupled model during forecasting, as well as the prediction outcomes spanning from 1996 to 2023.