1. Observations –Morphology –Quantitative properties Underlying Physics –Aly-Sturrock limit Present Theories/Models Coronal Mass Ejections (CME) S. K. Antiochos, NASA/GSFC

2. Recap of CME Physics For some reason magnetic shear concentrates at PILs producing filament channels –Exact topology unclear (especially twist) Filament field held down by overlying non- sheared coronal field –Need some mechanism to disrupt force balance catastrophically –Simply continuing the shearing does not do it! As shown by many simulations –Agrees with Aly-Sturrock limit

3. Bipolar (one polarity inversion line) initial magnetic field Filament-field formation by shearing and reconnection See pronounced expansion & kinking – but no eruption Demonstration of Non-Eruption (from, DeVore et al, 2005; Aulanier et al, 2005) + -

4. Non-Eruption Underlying physics: Corona has no lid Magnetic field lines can stretch indefinitely without breaking –Free to open slowly in response to photospheric stress and gas pressure (rather than erupt as CME) Slow opening (not associated with filament channels) observed to occur continuously in large-scale corona

5. Aly – Sturrock Limit Force-free field in an infinite volume (Aly 1984) –From div·T = 0, derive that: ∫ (B r 2 - B t 2 ) dA= 0, at limiting bdy –Implies that transverse component cannot increase indefinitely –Virial eqtn: ∫ B 2 dV= ∫ r ( -2 B r 2 + B 2 ) dA –Energy in interior related to field at bdy –Implies upper bound on energy –Aly-Sturrock conjectured lub is fully open field –Agrees with many simulations

6. Test of Aly – Sturrock Limit Roumeliotis et al, 2.5D sheared dipolar force-free field in spherical geometry Beyond certain shear, field expands outward exponentially Energy saturates at open limit Only certain BC are physical

7. Present Theories for Eruption Non-ideal evolution – reconnection Partial opening – ideal instabilities Non-quasi-static evolution – flux emergence? All appear to “work” in numerical simulations

8. NUMERICAL SIMULATIONS Solve 3D or 2.5D ideal/dissipative MHD with variety of numerical schemes –Both explicit and implicit –Both fixed and fully amr grids –Both Cartesian and spherical grids Initial conditions: –Usually equilibrium with varying degree of complexity –Simple dipole to observed photospheric fields with solar wind Boundary Conditions: –Open conditions at outer boundaries –Photospheric conditions main discriminator between models –Simple shear to incomprehensible contortions

9. NUMERICAL SIMULATIONS Term “simulation” is misnomer Simply method for obtaining approximate solutions to standard equations Drastic change in theory techniques, but still comes down to physical insight Hopefully numerical simulation will turn into user-friendly community tools

10. ARMS NUMERICAL SIMULATIONS Ideal MHD eqtns. (but numerical resistivity) –Use non-conservative energy equation for low-beta systems –Spherical grid with adaptive mesh refinement

11. Reconnection-Driven CME Models Breakout: –Field erupts to a state that it cannot get to by any ideal evolution –Magnetic reconnection removes overlying field, decreasing downward pull –Need topologically complex field –More than 1 dipole –Generally present on Sun

12. Non-Dipole Coronal Topology Field of two dipoles – axi-symmetric –Large global at Sun center, weaker near surface –Must have 4-flux system with separatrix bdys, and null

13. Magnetic Reconnection Frozen-in condition: – B-field lines ~ constants of the motion –Produces topological complexity and all solar activity Even in corona have finite diffusion, t ~ L 2 /η >> 10 6 years, for L ~ 1 Mm –If L sufficiently small, field lines lose identity and can “reconnect” on short time scales, but only over localized region –Need to develop significant magnetic structure on small scale for reconnection to be effective –Magnetic topology plays critical role

14. Breakout Model 2D multi-polar initially potential field Create filament channel by simple footpoint motions Outward expansion drives breakout reconnection in corona

15. Breakout reconnection allows for explosive eruption Flare current sheet, flare reconnection, and twisted flux rope all consequences of ejection CME with no flare possible for slow eruptions Breakout Model

16. 3D simulation using 3D AMR code Lynch et al “Create” prominence by simple boundary flows Reproduces standard features of CMEs/flares

17. Breakout Model 3D simulation by Roussev et al (2008) of 04/21/02 event Complex topology with flux transfer prior to eruption Generalized breakout process

18. Cancellation Model Loss of equilibrium/ flux cancellation: Reconnection/emergence at photosphere converts downward to upward tension Produces twisted flux rope in corona, prior to eruption Rope loses equilibrium, jumps upward Subsequent flare reconnection accelerates ejection –Van Ballejooigen, Forbes, Mikic/Linker, Amari, …

19. Flux Cancellation Model Analytic model by Forbes et al Detailed simulation of 05/12/97 event (Titov et al)

20. Aneurism Model Ideal instability (kink-like) Part rather than remove overlying field Need twisted flux rope –Sturrock, Fan, Kliem, …

21. Aneurism Model 3D simulation by Fan et al. System driven only by flux “emergence” Kink or torus instability depending on overlying field So far only idealized configurations

22. –Apparently have three mechanisms that can produce explosive CMEs in 3D simulations: Reconnection (Breakout), loss-of-equilibrium (flux cancellation), ideal instability –All require sheared prominence field –All produce twisted flux rope as a result of eruption –Flux cancellation and ideal instability require twisted flux rope before eruption Models for CME Initiation

23. 64K Question What is the pre-eruption structure of the prominence field? –Clearly has strong shear –Does it have twist (twisted flux rope topology) NRL VAULT image of 06/16/02, 20K material, spatial resolution < 200 km Little evidence for twist in either structure or motions, but exact topology still unclear See Rob’s movies!