Automated Classification of Plasma Regions Using 3D Particle Energy Distributions. Issue 10 (29th September 2021)

• Record Type: Journal Article
• Authors: Olshevsky, Vyacheslav; Khotyaintsev, Yuri V.; Lalti, Ahmad; Divin, Andrey; Delzanno, Gian Luca; Anderzén, Sven; Herman, Pawel; Chien, Steven W. D.; Avanov, Levon; Dimmock, Andrew P.; Markidis, Stefano
• Is Part Of: Journal of geophysical research. Volume 126:Issue 10(2021)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2021-09-29
• Subjects: MMS -- machine learning -- bow shock; Magnetospheric physics -- Periodicals; Space environment -- Periodicals; Cosmic physics -- Periodicals; Planets -- Atmospheres -- Periodicals; Heliosphere (Astrophysics) -- Periodicals; Geophysics -- Periodicals; 523.01
• DOI: 10.1029/2021JA029620
• Languages: English
• ISSNs: 2169-9380

Abstract

Abstract: We investigate the properties of the ion sky maps produced by the Dual Ion Spectrometers (DIS) from the Fast Plasma Investigation (FPI). We have trained a convolutional neural network classifier to predict four regions crossed by the Magnetospheric Multiscale Mission (MMS) on the dayside magnetosphere: solar wind, ion foreshock, magnetosheath, and magnetopause using solely DIS spectrograms. The accuracy of the classifier is > 98 %. We use the classifier to detect mixed plasma regions, in particular to find the bow shock regions. A similar approach can be used to identify the magnetopause crossings and reveal regions prone to magnetic reconnection. Data processing through the trained classifier is fast and efficient and thus can be used for classification for the whole MMS database. Plain Language Summary: Magnetospheric Multiscale Mission (MMS) has been traversing the Earth's magnetosphere to help scientists understand how the tremendous amounts of energy are released through the phenomenon known as magnetic reconnection. The spacecraft can transfer to the Earth only 4% of its measurements due to link limitations. The success of the mission relies on the selection of the most relevant measurement intervals to be sent down to the science operation center. We have trained a small deep convolutional neural network which identifies the kind of plasma the spacecraft is traversing at each measurement interval with an excellent accuracy >98%. We have used our model to identify some of the most interesting regions, bow shocks. It took only a day for the model to process all observations collected by the MMS within 3 years. The model can save a substantial amount of time for the scientists in the loop whose role is to locate such regions manually. The proposed model is suitable for the hierarchy of models being built to fully automate the on‐ground data processing. Moreover, it is small enough to be embedded in the on‐board software of future missions. Key Points: We develop a technique for automated classification of plasma regions traversed by the Magnetospheric Multiscale Mission spacecraft Our model distinguishes solar wind, megnetosheath, magnetosphere, and ion foreshock with 98% accuracy It can be used to detect mixed plasma regions, bow shock and magnetopause crossings