基于理论模式GOLD天底数据同化研究热层大气可预报性

准确表征和预报热层大气对近地卫星轨道环境至关重要。先前研究认为热层大气变化缓慢，对过去的“记忆”可长达至数天，进而能够对其预报。然而，热层大气预报不仅依赖于准确的物理模型和初始场信息，而且取决于模型外部驱动参数。限于先前热层探测数据匮乏，学界对于热层在何种条件下可以准确预报，是否存在可预报的时间上限并未得到较好的认知。本项目将利用我们自主开发的基于理论模式同化系统同化GOLD中性温度和积分氧氮比天底探测数据，为理论模式预报提供准确初始场，并在同化过程中估计理论模型外部驱动参数。通过开展地磁平静日和不同空间天气事件期间（尤其是磁暴）热层大气预报研究，分析不同初始场和模型外部驱动参数估计对预报结果的影响，并探讨热层大气预报所需条件以及预报时长。该研究可以进一步加深我们对热层大气可预报性的认知，以期为日后近地卫星轨道环境预报提供关键依据。

Accurate specification and prediction of thermospheric state is critical to the low Earth orbit satellite environment. Previous studies suggest that the thermosphere changes so slowly, which means that it can “remember” the past for too long. However, the prediction of thermosphere depends on not only the initial field and first-principle model, but also the accurate estimation of model external drivers. Due to the lack of previous thermosphere observations, it is not clear that whether there is a needed conditions and limited time for thermosphere prediction in space community. In this program, based on our developed efficient ensemble Kalman filter assimilation system, the GOLD disk data (e.g., neutral temperature and neutral compositions) will be assimilated to provide an accurate initial field for theoretical model. Meanwhile, the model external drivers will be also estimated during the assimilation process. Through a series of researches on the thermosphere forecast during the geomagnetic quiet day and different space weather events (especially geomagnetic storm), the influence of the initial field and the model external drivers estimation on the forecast results will be analyzed. Meanwhile, the required conditions and forecast duration of the thermosphere will be also discussed. This study can further deepen our understanding of the forecast capability of thermosphere states, and provide the key evidence for the prediction of the low Earth orbit satellite environment in the future.