15:00 〜 15:15
[PEM14-18] Can deep learning more effectively reveal the link between the interplanetary environment and thermospheric mass density variations?
キーワード：thermospheric mass density, interplanetary environment, deep learning, modeling
The thermospheric mass density (TMD) is an important parameter in both space physics research and aerospace engineering applications. The state and variations of TMD are closely tied to solar-terrestrial coupling processes. We have long aimed to accurately describe TMD using modeling approaches. However, due to incomplete observations and the limitations of current modeling approaches, existing models each have their own shortcomings. It is particularly important to fully leverage the value of rare observational data and develop models that more effectively reflect the impact of the interplanetary environment on TMD. This report introduces some recent attempts we have made using deep learning (DL) techniques. With the aid of enhanced nonlinear fitting and feature mapping capabilities, we have developed a model that more accurately captures the influence of interplanetary environment conditions on TMD variations. With this DL model, we analyze the TMD disturbances during typical space weather conditions. We believe that adopting a more open-minded approach toward DL technologies will help us better discover and understand the link between TMD variations and the state of the interplanetary environment.