Recent machine learning (ML)-based weather forecasting models have improved the accuracy and efficiency of forecasts while minimizing computational resources, yet still depend on traditional data assimilation (DA) systems to generate analysis fields. Four dimensional variational data assimilation (4DVar) enhances model states, relying on the prediction model to propagate observation to the initial field. Consequently, the initial fields from traditional DA are not optimal for ML-based models, necessitating a customized DA system. This paper introduces an ensemble 4DVar system integrated with the FuXi model (FuXi-En4DVar), which can independently generate accurate analysis fields. It utilizes automatic differentiation to compute gradients, and demonstrates the equivalence of these gradients with those derived from adjoint models. Experimental results indicate that this system preserves the physical balance of the analysis field and exhibits flow-dependent characteristics. These features enhance the propagation and assimilation of observation into the initial analysis field, thereby improving the accuracy of the analysis fields.

中文翻译：

---

FuXi-En4DVar： 一种基于机器学习天气预报模型的同化系统，确保物理约束


最近基于机器学习 （ML） 的天气预报模型提高了预报的准确性和效率，同时最大限度地减少了计算资源，但仍依赖传统的数据同化 （DA） 系统来生成分析字段。四维变分数据同化 （4DVar） 增强了模型状态，依靠预测模型将观察传播到初始场。因此，传统 DA 的初始字段对于基于 ML 的模型来说并不是最佳选择，因此需要定制的 DA 系统。本文介绍了一种集成了 FuXi 模型的集合 4DVar 系统（FuXi-En4DVar），该系统可以独立生成精确的分析字段。它利用自动微分来计算梯度，并证明这些梯度与来自伴随模型得出的梯度等效。实验结果表明，该系统保持了分析场的物理平衡，并表现出与流动相关的特性。这些功能增强了观测到初始分析字段的传播和同化，从而提高了分析字段的准确性。