[M-GI24] Data assimilation: A fundamental approach in geosciences

[MGI24-P09] Assimilation of polar ionospheric data into a newly-developed emulator of global MHD simulation

*Shin ya Nakano^1,2,4, Ryuho Kataoka^3,4, Aoi Nakamizo⁵, Shigeru Fujita^2,1 (1.The Institute of Statistical Mathematics, 2.Joint Support Center for Data Science Research, 3.National Institute of Polar Research, 4.Graduate Institute for Advanced Studies, SOKENDAI, 5.National Institute of Information and Communications Technology)

Keywords:data assimilation, emulator, surrogate model, machine learning, polar ionosphere

We are developing a machine-learning-based emulator that mimics the outputs of the latest global MHD model of the magnetosphere-ionosphere system. In particular, the recently developed SMRAI2 (Surrogate Model for Aurora Ionosphere version 2) (Kataoka et al., 2024) is capable of reproducing realistic spatio-temporal patterns of the electric potential and current in the polar ionosphere by learning the results of long-term simulations obtained through the numerical space weather forecast carried out by the National Institute of Information and Communications Technology. The advantage of our emulator is its high computational efficiency. The time evolution scenarios under various solar wind conditions can instantaneously obtained by the emulator. This emulator can thus be used for ensemble forecast and ensemble-based data assimilation. In this study, we attempt to reproduce realistic polar ionosphere environments by data assimilation which incorporates actual ionospheric observations into the predictions with SMRAI2. We will report the current status and demonstrate some preliminary results of the data assimilation.

Reference
Kataoka, R., Nakamizo, A., Nakano, S., and Fujita, S. (2024): Machine learning-based emulator for the physics-based simulation of auroral current system. Space Weather, 22, e2023SW003720. https://doi.org/10.1029/2023SW003720