Space Weather Applications of Machine Intelligence (SWAMI) Seminar Series: Next-generation magnetosphere-ionosphere coupling: Progress through machine learning and information representation

Hosted by: Space Weather Technology, Research, and Engineering Center
Presented by:  Ryan McGranaghan, ASTRA, LLC, Louisville, Colorado

Friday 19 March 2021, 10:00am – 11:00pm (MDT)
https://cuboulder.zoom.us/j/98581049160

Graphic image of Sun with grid superimposed on itThe connection between the Sun and the Earth is a complex one, involving interactions and variabilities across a dizzying spectrum of scales and systems. The result is a relationship between us and our star that is observable only through a fleet of instruments, methods, and technologies yet creates weather in the near-Earth space environment that is both life-sustaining as well as life-threatening. This relationship is colloquially known as space weather.

To unravel the critical complexities and variabilities and to evolve beyond current approaches to understand space weather, new data-driven approaches and data analysis technologies are required. These data‐driven methods are taking on new importance in light of the continuously changing data landscape of the space weather system, a challenge (and opportunity) shared by all scientific disciplines. Therefore, the scientific community faces both an exciting opportunity and an important imperative to create a new frontier built at the intersection of traditional approaches and state‐of‐the‐art data‐driven sciences and technologies.

In this talk we will first present a powerful use case for the use of data science to transform Heliophysics and Space Weather: the coupling between the magnetosphere and ionosphere via particle precipitation. We will describe a new model that better captures the dynamics of this precipitation from a large volume of data by increasing the expressive capacity afforded by machine learning [McGranaghan et al., [2021]].

With a more capable representation of the organizing parameters, the model achieves a >50%reduction in errors from a current state-of-the-art model, better captures the dynamic changes of the auroral flux, and provides evidence that it can capably reconstruct mesoscale phenomena. We share a new framework for space weather model evaluation that culminates previous guidance from across the solar-terrestrial research community.

We will conclude with an active extension of this research to build a data science ecosystem that enables data-driven progress. This talk is intended to provide insight to and spark discussion around the characteristics of convergence and antidisciplinary and the role of data science, namely creating new progress and a new frontier through radical interdisciplinarity and methodology transfer.

McGranaghan et al., 2021: Next generation particle precipitation: Mesoscale prediction through machine learning (a case study and framework for progress) https://arxiv.org/abs/2011.10117