Clustering of Global Magnetospheric Observations

• Published in: Journal of geophysical research. Machine learning and computation Vol. 1; no. 4
• Main Authors: Edmond, James, Raeder, Joachim, Ferdousi, Banafsheh, Argall, Matthew, Innocenti, Maria Elena
• Format: Journal Article
• Language: English
• Published: Wiley 01.12.2024
• ISSN: 2993-5210; 2993-5210
• DOI: 10.1029/2024JH000221

The use of supervised methods in space science have demonstrated powerful capability in classification tasks, but purely unsupervised methods have been less utilized for the classification of spacecraft observations. We use a combination of unsupervised methods, being principal component analysis, Self‐Organizing Maps, and hierarchical agglomerative clustering, to classify THEMIS and MMS observations as having occurred in the magnetosphere, magnetosheath, or the solar wind. The resulting classification are validated visually by analyzing the distribution of classifications and studying individual time series as well as by comparison to the labeled data set of a previous model, against which ours has an accuracy of 99.4% $\%$. The model has a variety of applications beyond region classification such as deeper hierarchical analysis, magnetopause and bow shock crossing identification, and identification of bursty bulk flows, hot flow anomalies, and foreshock bubbles. Plain Language Summary Machine learning in space science often uses supervised methods for classification, but we explore using unsupervised methods for classifying spacecraft observations. We combine principal component analysis, self‐organizing maps, and hierarchical clustering to classify whether observations occurred in the magnetosphere, magnetosheath, or solar wind for THEMIS and MMS. We verify classifications both visually and using a preexisting labeled data set, achieving 99.4% accuracy. Our model has additional applications such as the ability to analyze subgroups of clusters, identify boundary regions between the clusters, and flag important transient events related to the dynamics of the magnetosphere. Key Points MMS and THEMIS measurements are classified into magnetosphere, magnetosheath, and solar wind regions using unsupervised methods We created a data set of 5228 magnetopause and 3047 bow shock crossings inferred from the classifications The model is capable of detecting Hot Flow Anomalies, Foreshock Bubbles, and Bursty Bulk Flows