Using dimensionality reduction and clustering techniques to classify space plasma regimes: electron magnetotail populations

Abstract:

Collisionless space plasma environments such as the Earth’s magnetotail are primarily characterised by distinct particle populations. Although moments help us to distinguish different plasma regimes, the distribution functions provide more comprehensive information about the plasma state, especially at times when the distribution function includes non-thermal features. Unlike moments, distribution functions are not easily classified by a small number of parameters. We propose to distinguish between the different plasma regions by applying dimensionality reduction and clustering methods to electron distributions in pitch angle and energy space. We utilise three separate algorithms to achieve our classifications: autoencoders, principal component analysis, and agglomerative clustering.

We test our classification algorithms by applying our scheme to data from the Cluster-PEACE instrument measured in the Earth’s magnetotail. Traditionally, it’s thought that the Earth’s magnetotail is split into three different regions that are primarily defined by their plasma characteristics. The innermost plasma sheet typically contains a relatively hot, isotropic plasma with higher particle density, while the outermost lobes are characterised by a lower plasma density/temperature. The plasma sheet boundary layer forms the transition region in between the plasma sheet and the lobes. In contrast, we find it is optimal to have 8 distinct groups of distributions. By comparing the average distributions as well as the plasma and magnetic field parameters for each region, we can relate the groups to the canonical plasma sheet, the plasma sheet boundary layer, the current sheet, and the lobes. We compare our results to the ECLAT classifications based on plasma moments. Analysis of the different parameters shows clear distinctions between each of our classified regions and the ECLAT results.