We propose a novel machine learning algorithm for automatically detecting and classifying aurora in all–sky images (ASI). Our algorithm is largely trained without requiring ground–truth labels. By including a small number of labeled images, we are able to automatically label all of the approximately 700 million images in the Time History of Events and Macroscale Interactions during Substorms (THEMIS) ASI dataset from 2008 to 2022. We use a two–stage approach. In the first stage, we adapt the Simple framework for Contrastive Learning of Representations (SimCLR) algorithm to learn latent representations of THEMIS all–sky images. We then finetune a classifier network on the latent representations our model learns of the manually labeled Oslo aurora THEMIS (OATH) dataset. We demonstrate that this two–stage approach achieves excellent classification results on data for which there is no current benchmark. The outcome of this work is a set of annotations for all existing THEMIS ASI from 2008–2022. A statistical analysis of the occurrence rates of auroral labels with respect to solar wind parameters, interplanetary magnetic field vector, and geomagnetic indices shows agreement with the literature. We further investigate the occurrence rates of auroral phenomena in the annotated data set and their geoeffectiveness by utilizing the co–located THEMIS ground magnetometer data set. These annotations will facilitate efficient information retrieval for researchers interested in specific categories of aurora and will enable large scale statistical studies and machine learning analyses of THEMIS all–sky images that have not previously been possible.

Amy Keesee

and 2 more

Reconnection in the magnetotail transfers magnetic energy to thermal and kinetic energy in ions and electrons. These particles are injected both Earthward and tailward from the reconnection region. The Earthward particles are transported to the inner magnetosphere where they drive the ring current and radiation belts. The injections are observed in the plasma sheet in conjunction with dipolarizations of the magnetic field. The particles have been found to travel within narrow flow channels, rather than broadly across the magnetotail, in spatially and temporally localized events known as bursty bulk flows (BBF). Simulations of such events show these narrow flow channels moving from the reconnection region to the injection region. However, global observations are needed to understand how BBFs connect the reconnection region and the inner magnetosphere during storms and substorms. Ion heating has been observed with in situ measurements at the reconnection region and within the dipolarization fronts and BBFs. Using energetic neutral atom (ENA) imaging, ion temperature maps can be calculated to provide such global observations. Regions of ion heating have been observed in these maps and comparisons with in situ measurements demonstrate that they are associated with these phenomena. An automated identification algorithm has been developed and run on our database of storm-time ion temperatures. We will present the results of case and statistical studies of the characteristics of these features.