Description
Solar eruptions (flares and coronal mass ejections) play a central role in shaping space weather, making their accurate prediction and modelling an imperative. These violent magnetic explosions release enormous amounts of energy and can drive solar storms travelling at hundreds of kilometres per second, capable of delivering severe geomagnetic disruptions. The source of this energy lies in the solar photosphere, where the footpoints of large-scale coronal magnetic arcs are continuously sheared and twisted by turbulent granular motions, injecting energy and helicity into the coronal magnetic field — a process known as magnetic helicity condensation. The accumulated magnetic free energy is then released via magnetic reconnection. Current numerical models of solar eruption often involve simulating the Sun from the base of the convection zone to the upper corona, and are hence computationally expensive and poorly constrained by sub-photospheric dynamics that cannot be directly observed. Data-driven simulation, by contrast, is a powerful technique for realistic modelling of the solar atmosphere that can address both challenges at once. As a dynamic data assimilation technique, data driving removes the convection zone and substitutes its effect through evolving boundary conditions derived from high-resolution observations of the photosphere. It thereby offers well-constrained, physics-grounded coupling between the lower and upper solar atmosphere, which is essential for modelling solar eruptions. In this talk, I will briefly review our implementation of data-driving techniques commonly used in solar eruption simulation, and present our findings and inherent model uncertainties for two classes of simulations. The results will pioneer the first ensemble model of solar eruptions, and form the bedrock for ensemble prediction of space weather drivers.
| I am the presenting author | Yes |
|---|