Auroral current continuity - a machine learning study using available data

Auroral events are characterized by a channel of high conductance plasma in the ionosphere, caused by precipitating particles from Earth’s magnetosphere ionizing neutrals as they stream down field lines. Current continuity, however, requires this current to close, and so a return current flows back up in regions where conductance is much lower. Hall and Pedersen currents, proportional to an electric field times their respective conductances, allow current to flow horizontally within the ionosphere. This leads to a predictable (idealized) pattern often recognizable in in-situ, satellite based data: a region of precipitating Field Aligned Current (FAC) and high conductance, where electric field and therefore ExB flow is small, nearby to a region of return FAC where conductance is low and so ExB flow must be high to allow for current closure.

Separately, the relaxation of the particles ionized by precipitation often has a visible signature detectable by ground based cameras, and therefore the brightness observed by these cameras works as a proxy for precipitation and to some extent conductance. It is expected that the bright region seen by the imager be contained within the precipitating current region determined by in situ measurements, but the shape and distribution of the brightness and FAC do not in principle need to be the same. We are using machine learning to examine real examples of conjunctions between in-situ (SWARM) and ground based imager (THEMIS-GBO) data, and attempting to predict some of these quantities from the others. Over 3000 conjunctions were found in 2014 alone, but many of these do not contain auroral arcs, so the number of usable conjunctions is limited. Flow, current, and brightness are featurized as gaussians using an automated algorithm to help reduce the amount of redundant data passed to the neural network and allow training with fewer events. This project aims to better understand the relationship between current, flow, and precipitation in the auroral ionosphere, as governed by the system science of the auroral current continuity equation.