Two-stage Instability Model for Explaining and Predicting the Generation of Coronal Mass Ejections

Coronal mass ejections (CMEs) are explosive releases of large volumes of magnetized coronal plasma into interplanetary space, and they are a significant cause of space weather disturbances, such as geomagnetic storms. Therefore, predicting CME occurrence is a considerable challenge; however, due to limited understanding of their generation mechanisms, accurate predictions have not been achieved. Meanwhile, various studies have explored the magnetic characteristics of active regions that determine whether solar flares can erupt or not into CMEs. In particular, Muhamad and Kusano (2025) recently found that a new parameter, consisting of the critical height (hc) at which torus instability can grow and the ratio of the direct to return electric currents, can effectively distinguish the source active regions where solar flares erupt and do not erupt to CMEs, with unprecedented accuracy. Based on these results, we propose a new CME generation mechanism, a "two-stage instability model," and verify it using 3D MHD simulations. The two-stage instability model suggests that, in the first stage, small-scale magnetic reconnection triggers the growth of a double-arc instability (Ishiguro and Kusano, 2017), which raises the twisted magnetic flux to the critical height (hc). In the second stage, the torus instability grows and drives CMEs. Simulations using a shear-arcade magnetic field as the initial condition clearly demonstrate the validity of this model. Furthermore, the simulation results suggest that (1) the two-stage instability model can explain the cause of the slow-rise phase, which is considered a precursor to CMEs, and (2) the dependence of the torus instability on the initial magnetic field distribution can provide insight into the physics that determines the duration and spatial extent of solar flares. Based on these results, we propose a new method for predicting CME occurrence from magnetic field data in active regions and discuss its forecasting capability.