Three-dimensional MHD simulations of emerging flux and associated magnetic reconnection. Hiroaki Isobe (DAMTP, Cambridge / Tokyo) Takehiro Miyagoshi, Kazunari Shibata (Kyoto) Takaaki Yokoyama (Tokyo). 6th Solar-B Science Meeting, 8-11 November 2005, Kyoto. Outline. - PowerPoint PPT Presentation
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Three-dimensional MHD simulations of emerging flux and associated magnetic
6th Solar-B Science Meeting, 8-11 November 2005, Kyoto
Outline• Theories of magnetic reconnection and its difficulties
– Huge magnetic Reynolds number– Scale gap
• Key observations and its implication to theory– fractal nature of current sheet– plasmoid ejection
• 3D MHD simulations of emerging flux and its reconnection with overlying coronal field– formation of filamentary structure– patchy reconnection– Observations by Solar-B
Theories of magnetic reconnection
€ VinVa=lL=Rm−1/2≈10−7
Sweet-Parker reconnection
Reconnection rate:
Parker 1957, Sweet 1958
€ VinVa∝lnRm≈0.01−0.1Petschek reconnection
Petschek 1964•MHD simulations: if resistivity is localized, Petschek-like reconnection (i.e., with slow shocks) occurs.
•Magnetosphere obs. and Lab experriments suggest that fast reconnection occurs when current sheet become as thin as ion Larmor radius or ion innertial length.
Fundamental problem in fast reconnection: huge scale gap
Ion inertia length 2/1611010100⎟⎠⎞⎜⎝⎛⎟⎠⎞⎜⎝⎛≈=−KTGBcmeBvcmriLiIon Larmour radiuscmrflare910≈typical size of solar flare
•Laminar and steady reconnection with tiny diffusion region? Unlikely. •Mesoscale (MHD) structure?•Self-similar evolution in free space (Nitta P45)
2/1
310, 10300
−
− ⎟⎠
⎞⎜⎝
⎛≈=cmn
cmc
piionin ω
l
Fractal nature of current sheets
Hard X-ray emission(Ohki 1992)
Power spectrum of radio (610MHz) emission (Karlicky et al. 2005)
Fractal!
Fine spatial structure in reconnection events
Small kernels in flare ribbons(Fletcher, Pollock & Potts 2004)
aurora
Surges/jets
Supra-arcade downflows (McKenzie & Hudson 1999, Innes, McKenzie & Wang 2003, Asai et al. 2004 )
Plasmoid (flux rope) ejection
• Simultaneous acceleration of plasmoid and energy release • Slow rise and heating of plasmoid/flux rope prior to the hard X-ray burst
Ohyama & Shibata 1997
Zhang et al. 2001
Laboratory experiment
color: current densitysolid lines: magnetic field lines
Reconnection rate is enhanced when current sheet (plasmoid) is ejected (Ono et al. 1997).
•Upper convection zone - photosphere/chromosphere - corona
•Horizontal magnetic sheet in the convection zon+ uniform background field.•Grid: 800x400x620. Calculation was carried out using 160 processors of the Earth Simulator (about 8 hours for 50000 steps).
Result: overview
•Basically similar evolution to 2D simulation.
Magnetic field lines
Magnetic field lines + isosurface of |B| + temperature
Filamentary structure from the magnetic Rayleigh-Taylor instability
•The top of the emerging flux becomes top-heavy => unstable to the Rayleigh-Tyalor instability
•Bending of magnetic field is stabilized => Filamentary structure
x
zy
Color: mass density
Isosurface of mass density
Hαimage of arch filaments
Why top-heavy?Nonlinear evolution of Parker instability is approximately self-similar (Shibata et
al. 1990).
simulation self-similar solution
density at middle
Field lines at different time
•The outermost part deviates from self-similar solution.
Why top-heavy?
The outermost field lines undergo:
• Compression between coronal pressure above and magnetic pressure below.
• Larger curvature radius => smaller gravity along B => less evacuation.
divvpara
divvperp
Evolution of the Rayleigh-Taylor instabilityDensity at the y-z plane
t=70
t=81t=78
t=76
• Small wavelength modes grow first (larger linear growth rate)• Larger scales from inverse cascade• Scale (width of filaments) may change with the presence of shear€
γ= kyg(ρ 2 − ρ1) / (ρ 2 + ρ1)
Fourier modes of Vz
Linear growth rate
Observational signature
• Rayleigh-Taylor instability– Rising loops and (relatively) sinking loops exist alternatively.– Sinking loops are denser and probably colder– Spectroscopy in H-alpha and/or EUV
• When the emerging loops become coronal temperature?– time scale of indivisual loop emergence 〜 1000 s– For EIS observations, time cadence is more important
• Vortex excited by secondary Kelvin-Helmholtz instability– Tortional Alfven wave – Small scale twist in individual filaments– Chromospheric magnetic field measurement
Formation of filamentary current sheets
Mass density isosurface (gray)Current density distribution (color)
Mass density (contour) and current density (color) at the y-z cross section
•Deformation of magnetic field by the Rayleigh-Taylor instability =>current formation in the periphery of the dense filaments•Dissipation of these current sheets may result in intermittent heating, leading to the formation of the hot/cold loops system.
xz
y
Reconnectin in the interchanging current sheet
•Larger current density and smaller mass density in the rising part of the R-T instability
•anomalous resistivity sets in locally•fast reconnection occurs in spatially intermittent way
•Reconnection inflow enhance the nonlinear evolution of the R-T instability => nonlinear instability
Anomalous resistivityB×
・
Intermittent, patchy reconnection
Isosurfaces of velocity.
•Fast reconnection occurs in spatially intermittent way after the ejections of plasmoids.•Many narrow reconnection jets from initially laminar current sheet.
Isosurfaces of gas preasure+ magnetic field line
Many small plasmoids in the intialy laminar current sheet.
Conjecture•R-T instability can occur if there is density jump across the current sheet and effective acceleration (in suitable direction).•Effective acceleration is likely to exist in dynamically evolving system (like eruptive flares, CMEs, solarwind-magnetosphere) and in driven reconnections.
•RT instability is ideal instability, hence no restriction from large Reynolds number.
•Possible scenario may be...•Small scale turbulence grows by R-T instability (and couple with micro-scales) •tearing occurs in small scale•formation of large plasmoids (flux ropes) by coallescense => fast reconnection in global scale
Observation of reconnection signatures by Solar-B
• Few spectroscopic detection of reconnection inflows/outflows (Innes et. al. 1997, Lin et al. 1995).
=>EIS– Precise determination of the reconnection rate.– Slow shock? (Shiota et al. 2003)
• Turbulent broadening Vturb ≈VA in hottemperature lines such as FeXXIV(10MK) if reconnection is fractal, i.e., many small reconnection.
FeasibilityTemperature: T ≈ 5-20MK. FeXXIV(192.08) is suitable?.
Time scale: t ≈ 10-100 (s).
If turbulent-enhanced Sweet-Parker, the width of current sheet w is:w = MAL ≈ 10 - 1000 km.
Assuming the (pre-flare) density of 109 cm-3, EM ≈ 1024-26 cm-5.
For FeXXIV(192.028), several photons /s/pixel ... not easy but possible.(Thanks: Helen Mason)
Same figure with isosurface of current density
3D structure of reconnectionMagnetic field line + current density distribution
• Petschek-like slow shocks
Reconnection inflow/outflow in 3D
Isosurface of |V| Velocity and current density on the x-z cross section
Velocity on the y-z cross section in the outflow region. Contour is velocity perpendicular to the figure.Diverging
•Outflow in diverging = more effective in plasma expelling. => faster than 2D reconnection?
x
zy
Velocity on the y-z cross section in the outflow region.Converging
Comparison of the reconnection rate with 2D case
•Locally, 3D reconnection is faster and more bursty than 2D reconnection.•The spatial average is comparable with 2D.•Rayleigh-Taylor does not occur in 2D... so the local condition near the reconnection point is not the same.
Reconnection rate measured by maximumηJ in the y-z cross section.
Preliminaly.
Summary
• Filamentary structure spontaneously arises due to the magnetic Rayleigh-Taylor instability in the emerging flux.
• Current sheets are formed in the periphery of arch filaments due to the R-T instability. Intermittent heating.
• Magnetic reconnection becomes patchy, due to the interchanging of the current sheet.
• Fine structure and dynamics in EFR filaments/loops (SOT/EIS/XRT)
• Detection of reconnection inflows/outflows (EIS/XRT)• Turbulence in the current sheet (EIS)• EFR is suitable to target to catch the reconnection event (ev
en in solar minimum).
Why top-heavy?Density in 2D simulations
with coronal field
without coronal field
•Vertical distribution of density (color symbols) and Bx(solid) at the mid-point of the emerging flux.•Color indicates the Lagrangean trace of the fluid elements (i.e. same color indicates the same field lines.)
Divergence of the velocity components parallel and perpendicular to B
divV
divVpara
divVperp
•Top-heavy sheath locates at orange-magenta boundary.
•Divergence of V (especially Vperp) changes the sign at this point. => convergence.
Deviation from self-similar solution (Shibata et a. 1990)
•The outermost part of emerging flux deviate from self-similar solution.•Compressed beteen coronal pressure and upward magnetic pressure from below (divVperp<0)•Outerfield lines have larger curvature radius and hence smaller effective gravity along B. (effect of divVpara )
simulation self-similar solution Density at the midpoint of the emerging flux.
Density@midpoint t=10 t=80
t=84 t=88 t=92
t=96 t=104 t=112
Helical flux rope
Helical structure erupting after a prominence eruption (TRACE/EUV; Liu & Kurokawa 2004)
•With the presence of the guide field (By), helical flux rope is formed.
(This calculation is still preliminary)
The Earth Simulator•A parallel vector computer system installed at the Earth Simulator Centre, in Yokohama, Japan.•Fastest computer in the world since 2002 until September 2004.
↓Tokyo↑Yokohama
Kyoto↓
The Earth Simulator: hardware and software
•640 Processor Nodes (PNs)•One PN consists of 8 vector-type arithmetic processors (APs) and 16 GB shared memory..
•In total, 5120 APs and 10TB memory (distributed).•40Tflops at peak, 35.86Tflops for Linpack Benchmark