Featurizing an In-situ and Imagery Conjunction Database for Auroral Current Closure Studies: Implications for Magnetosphere-Ionosphere Coupling

The relationship between flow, Field Aligned Current (FAC), and conductance in an auroral arc is governed by the current continuity equation. In an idealized, sheetlike discrete arc, a single satellite measuring electric and magnetic field along a trajectory can reconstruct much of the detail of current closure in the arc. Even relaxing sheetlikeness assumptions, 2d (cross and along-track) E field data and/or multiple-point satellite crossings, along with magnetically conjugate ground based imagery, can still constrain some of the details of current closure. Other more complicated arc structures, including Alfvenic arcs poorly described by quasistatic models, can in principle be distinguished by time variation in GBO data or by the significant difference between typical Alfvenic and Pedersen B/E ratios. In all cases, the combination of in-situ and imagery data for simultaneous use in analyzing current closure provides a powerful tool for studying ionospheric system science.
We have developed several tools to help extract as much information as possible from conjugate spacecraft and auroral imagery data - these include a wavelet transform based featurization of the Swarm ion flow and B field cuts, a routine that uses Swarm B/E ratios to determine conductance in sufficiently sheetlike arcs, and a tool to automatically distinguish, track, and featurize arcs in imagery. We have used these tools, together with published databases of Swarm spacecraft data and THEMIS-GBO imagery queried with the Aurora-X tool, to construct a large, featurized database of conjunctions which will allow us to study in-situ and visible arc statistics and develop ML-based predictive tools that effectively use the available, heterogeneous data. The featurization routines are especially useful for ML, as the dimensionality reduction inherent in featurization allows training on smaller datasets than would otherwise be possible.
These tools and analyses allow us to address several open questions in the field. Straightforwardly, featurized conjugate in-situ and imagery data can be used to examine the statistics of visible discrete arc number, width, separation, proper motion, and lifetime; and how these relate to the width, strength, shape, and relative position of associated in-situ current sheet and flow structures. More abstractly, the relationship between flow and FAC (with boundary conditions partially informed by imagery) can give information on when the magnetosphere acts more as a current or voltage source, at the time and length scales of discrete arcs. The possibility of tearing mode instabilities, with known conditions for existence, being responsible for auroral filamentation can be examined. Finally, the problem of predicting in-situ ionospheric conditions from imagery data (“reading the aurora”) can be attacked.