ASI | Agenzia Spaziale Italiana

Next seminars in agenda

Speaker:
Dr. Manolis Georgoulis
(Accademy of Athens, Greece)

Title of the Talk
Solar Weather Forecasting: A Research-to-Operations-to-Research Approach

Scheduled on:
27^th September 2021, at 15:00 CET

Abstract:
The term Space Weather encompasses an array of phenomena that incur a short-term effect
(meaning hours to days) to all solar system bodies, each responding differently on the basis
of its inherent properties (mainly a magnetosphere / atmosphere, or lack thereof). An
overwhelming majority of Space Weather phenomena are driven by solar magnetic activity,
hence the primary Space Weather driver can be referred to as Solar Weather.
I will particularly focus on the impulsive Solar Weather component, relative to explosive or
eruptive solar activity, namely solar flares, coronal mass ejections (CMEs) and solar
energetic particle (SEP) events. Predicting this continuum of events and their repercussions
at geospace, the cis-lunar environment and deep space has obvious benefits for the
protection of our assets, human and technological, as the space age matures into an
ambitious future.
I will emphasize the importance of tackling Solar Weather events in tandem, rather than
artificially separating them, and will outline a possible prediction framework that relies on
data, model and verification tasks. The root of this framework is advances made by
heliophysics research while the envisioned outcome is tangibly operational. Given that we
tackle an intrinsically Big Data environment with an immense parametric space, machine
learning and, possibly, deep learning methods are coming into play. This is not without
challenges, however, as the rarity of major Solar Weather events, their solar-cycle
modulated climatology and the sensible use of time series may stray even perfectly capable
models into inferior performance. Even if these problems are tackled efficiently, we have a
major interpretability issue, particularly for unsupervised machine learning and deep
learning methods, that precludes us from determining what works best and invest on
research with an enhanced operational face value.
Concluding, I will double down on the importance of a research-to-operations-to-research
(R2O2R) feedback loop that, combined with a strategic investment on space-borne
hardware on and beyond the Sun-Earth line, could bring us forward and set the stage for a
continuous, hopefully transformative, progress in Solar (and Space) Weather forecasting.