PRIME‐SH: A Data‐Driven Probabilistic Model of Earth's Magnetosheath

• Published in: Journal of geophysical research. Machine learning and computation Vol. 1; no. 3
• Main Authors: O’Brien, C., Walsh, B. M., Zou, Y., Qudsi, R., Tasnim, S., Zhang, H., Sibeck, D. G.
• Format: Journal Article
• Language: English
• Published: 01.09.2024
• Subjects: magnetosheath; model validation; plasma depletion layer; recurrent neural network; solar wind‐magnetosphere interaction; uncertainty
• ISSN: 2993-5210; 2993-5210
• DOI: 10.1029/2024JH000235

A data‐driven model of Earth's magnetosheath is developed by training a recurrent neural network (RNN) with probabilistic outputs to reproduce Magnetospheric MultiScale (MMS) measurements of the magnetosheath plasma and magnetic field using measurements from the Wind spacecraft upstream of Earth at the first Earth‐Sun Lagrange point (L1). This model, called Probabilistic Regressor for Input to the Magnetosphere Estimation‐magnetosheath (PRIME‐SH) in reference to its progenitor algorithm PRIME, is shown to predict spacecraft observations of magnetosheath conditions accurately in a statistical sense with a continuous rank probability score of 0.227σ (dimensionless standard deviation units). PRIME‐SH is shown to be more accurate than many current analytical models of the magnetosheath. Furthermore, PRIME‐SH is shown to reproduce physics not explicitly enforced during training, such as field line draping, the dayside plasma depletion layer, the magnetosheath flow stagnation point, and the Rankine‐Hugoniot MHD shock jump conditions. PRIME‐SH has the additional benefits of being computationally inexpensive relative to global MHD simulations, being capable of reproducing difficult‐to‐model physics such as temperature anisotropy, and being capable of reliably estimating its own uncertainty to within 3.5%. Plain Language Summary As the solar wind encounters Earth's magnetosphere and diverts around it, a shock is formed that heats and compresses the plasma and warps the magnetic field frozen into it. This shocked plasma and magnetic field, known as the magnetosheath, is what drives energy transfer at the magnetopause. Due to orbital constraints there is no continuous in‐situ monitor of magnetosheath conditions. Studies of solar wind magnetosphere interaction typically rely on solar wind conditions measured at L1 propagated to Earth by some algorithm, which are then either used directly or used to drive some model of the magnetosheath. This process has numerous points of uncertainty, from the choice of propagation algorithm to the choice of magnetosheath model (or lack thereof). To address these concerns with the traditional approach, this study develops a data‐driven model of the magnetosheath that uses data from L1 as its input. This new model, called Probabilistic Regressor for Input to the Magnetosphere Estimation‐magnetosheath, adapts a recurrent neural network architecture that is capable of estimating uncertainties for its predictions. This new model is verified to be accurate in a statistical sense, and is also capable of representing physics that is not explicitly incorporated in the model during training. Key Points Probabilistic Regressor for Input to the Magnetosphere Estimation‐magnetosheath (PRIME‐SH) is an algorithm that predicts plasma and magnetic field in Earth's magnetosheath using inputs from in‐situ monitors at L1 PRIME‐SH accurately predicts the magnetosheath conditions in a statistical sense and its predictions obey conservation laws at the shock PRIME‐SH can be used to easily assemble continuous maps of the magnetosheath, addressing spatial limitations of in situ data