Disturbance Storm Time (Dst) index estimation using deep learning applied to Swarm satellite data

Computed from the intensity of the globally symmetrical equatorial electrojet (Ring Current) measured by a series of near-equatorial geomagnetic observatories, the Dst (Disturbance Storm Time) is an hourly index of magnetic activity. To give the estimation of the Dst index through the magnetic data measured by the Swarm three-satellite mission, we selected and trained an Artificial Neural Network (ANN). From November 2014 to December 2019, we collected a big Swarm magnetic dataset, confined in space to three very narrow belts of low-to-mid latitude, to better resemble the geographic distribution of the four geomagnetic observatories used to estimate at ground Dst. We also extended the analysis to mid latitude locations to increase the number of satellite samples. By using a Deep Learning architecture and based on its performance, we selected the best topology and trained the network testing its modelling capabilities. The outcomes show that the ANN is able to give a reliable fast estimation of the Dst index directly from Swarm satellite magnetic data, especially during magnetically disturbed periods.