Modeling Radiation Belt Electrons With Information Theory Informed Neural Networks. Issue 8 (29th August 2022)

• Record Type: Journal Article
• Authors: Wing, Simon; Turner, Drew L.; Ukhorskiy, Aleksandr Y.; Johnson, Jay R.; Sotirelis, Thomas; Nikoukar, Romina; Romeo, Giuseppe
• Is Part Of: Space weather. Volume 20:Issue 8(2022)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2022-08-29
• Subjects: radiation belt -- relativistic electrons -- solar wind drivers -- machine learning -- information theory -- empirical model -- phase space density; Space environment -- Periodicals; 551.509992
• DOI: 10.1029/2022SW003090
• Languages: English
• ISSNs: 1542-7390

Abstract

Abstract: An empirical model of radiation belt relativistic electrons ( μ = 560–875 MeV G −1 and I = 0.088–0.14 R E G 0.5 ) with average energy ∼1.3 MeV is developed. The model inputs solar wind parameters (velocity, density, interplanetary magnetic field (IMF) |B|, B z, and B y), magnetospheric state parameters (SYM‐H and AL), and L *. The model outputs the radiation belt electron phase space density (PSD). The model is operational from L * = 3 to 6.5. The model is constructed with neural networks assisted by information theory. Information theory is used to select the most effective and relevant solar wind and magnetospheric input parameters plus their lag times based on their information transfer to the PSD. Based on the test set, the model prediction efficiency (PE) increases with increasing L *, ranging from −0.043 at L * = 3 to 0.76 at L * = 6.5. The model PE is near 0 at L * = 3–4 because at this L * range, the solar wind and magnetospheric parameters transfer little information to the PSD. Using solar wind observations at L1 and magnetospheric index (AL and SYM‐H) models solely driven by solar wind, the radiation belt model can be used to forecast PSD 30–60 min ahead. This baseline model can potentially complement a class of empirical models that input data from low earth orbit (LEO). Plain Language Summary: An empirical model of radiation belt relativistic electrons with an energy of 1–2 MeV is developed. The model inputs solar wind parameters, magnetospheric state parameters, and L *. L * gives a measure of radial distance from the center of the Earth with a unit of R E (radius of the Earth = 6, 378 km). The model outputs the radiation belt electron phase space density (PSD). The model is operational from L * = 3 to L * 6.5. The model is constructed with an information theory informed neural networks. Information theory is used to select the relevant solar wind and magnetospheric parameters and their lag times based on the amount of information they provide to the radiation belt electrons. The model performance increases with increasing radial distance ( L *) because at distances close to Earth ( L * = 3–4), the solar wind and magnetospheric parameters provide little information about the radiation belt electron PSD. The model can be used to forecast radiation belt PSD 30–60 min ahead. Key Points: An empirical model to predict state of radiation belt relativistic electrons is developed The model prediction efficiency increases with increasing L * with a PE > 0.6 at L * > 5 The model can potentially complement a class of empirical models that input observations from low earth orbit