MURI meeting notes March 1, 2004 (sketchy, biased, and with no graphics) Motto: "H-alpha is the garbage heap of the solar atmosphere!" Comment #1: If anybody reads these and finds errors, please let me know at hhudson@ssl.berkeley.edu. Comment #2: My notes are more complete in places where I am (a) not so interested, but (b) not bored. I. MURI EVENTS OVERVIEWS George Fisher: May 1, 1998. Global field models are beautiful. May 12, 1997 - not much sign of shear in the region that erupts. Classic example of ribbon blithely crossing sunspot umbra and not reflecting the photospheric magnetic structure. Janet - "If you can't figure this one out, give up!" Also good candidate for heliospheric modeling. Nice impulsive particle event with shock signature (?) after about days. Yan Li: Halloween overview October 28, fast CME and X17 October 29, fast CME and X11 EIT difference images show huge brightenings in each case. AR 486 has amazing complexity. Sa"m's footpoint images. Janet Luhmann: October 29 prompt particle event (see Sam Freeland's "extremely useful" latest events page); canonical SEP connection a la Peggy Shea with IP connectivity and shock location. Soft spike at the time of shock arrival at Earth. Here it's complicated because of multiple events. The early particles are thought to be from a "foreshock" region involving particles running ahead. ACE did not perform so well as a sentry because of the intensity and complexity of the events. II. 8210 EVENTS Canfield: Is it large-scale (Moore-Sterling) or small-scale? IVM movie. Caution regarding empty areas in the vector-field measurements. Welsch: Introducing things with Amari model via Regnier. LCT tracking of magnetic or intensity features... Negative and positive field tracked separately. Need to understand vertical flows to avoid LCT confusion. Speeds are 0.2-0.3 km/s. The LCT has no physics, so how to get input consistent with the induction equation? MEF method has a dB_z/dt estimate => phi, one of the scalar potentials. Constraint that v be as small as possible consistent with data. Demoulin-Berger "ILCT". Canfield: Regnier movies of vector field including separatrix @ E with apparent flux transfer perhaps matching blue-shift events; also a similar structure in the emerging-flux region. The 15 VMGs during times of some flaring. "Energy rate" plot with differences of B model energy content. Also plot Poynting flux vs time based upon MEF ideology. Are things self-consistent and without arbitrary assumptions within this framework? McTiernan: Nonlinear FFF from the Roumeliotis formalism (Wheatland et al. ApJ 540, 1150, 2000): iteration from potential field e.g., minimize "objective function" L @ B constant and potential on boundary. L has jxB and div B terms. Test worked on Low-Lou model. With real data, L does not go to zero. Test on 8210 converges in a couple of hours in IDL, resulting in L about 0.2. Abbett et al. 2004 => view of "canyon" seen in the SXT images. Neat possible result that the "open" field lines come from larger B in the photosphere. "Non-force-freeness" is concentrated near the bottom layer. Abbett: Data-driving of modeling? How to emerge flux into a magnetized atmosphere, currently active research area. Now have methods to get horizontal derivatives of v and B needed for induction equation (LCT, MEF, VMG). Problems of self-consistency in putting the coronal boundary at the photosphere. Boundary cells below photosphere; mystery region below that which may contain emerging flux. How to relate these domains? One assumption would be to ignore forces exerted by the coronal component. "Ghost cell" specification in the mystery zone. Magara FFF simulations => initial-state estimation, comparison with potential, Wheatland, MHD simulations. Nice color-coded nonforcefreeness graphic illustrating the concentration at low altitudes. Discussion with Yang Liu about what is good and what is bad. Worry about using field lines without plasma loading to guide the eye. Barnes: MCT on 8210, KD in audience so blame Dana for problems. In general, this approach cannot reproduce the details of the field, but may provide a guide to the basic physics of its evolution. Connectivity plots for 8210 charges. Do the parameters change? On a statistical basis, "no", and these estimations include some sort of self-consistent error propagation. Submergence/emergence vs reconnection. Ignore the latter! Pairwise charge identification methods. In 8210 need some domains to have both emergence and submergence. How to reconcile coronal domains with MCT photospheric domains during flare evolution Manchester: CR 1935 as published in Roussev et al. APJ 595, L57, which uses a parametrization of gamma = (n+2)/n from 1.04 to 1.5 with radial distance. Zoran surprised by large amount of structure at 1AU, claimed by Zoran not to be reasonable in an MHD model. CME calculation based on Gibson-Low flux rope (out of equilibrium @ initial state). Roussev: May 2 eruption modeled again in April ApJL paper. Rotary boundary motions, then converging motions to create flux rope. Modeling time 441 computer hours for 7 real-time hours! Motions at 1% of local Alfven velocity for coronal lower boundary condition. Shock formation early and strong at low altitude, despite high Alfven speed? Numerical reconnection dissipates current sheet and allows eruption. Mach 4.5 at 2 solar radii. Not really consistent with observations e.g. of Zhang, but this is a first try. Sokolov: SEPs attacking astronauts behind 0.5 gm/cm^2 (2 mm) Al => 20 MeV. Diffusive shock acceleration (Dolginov-Toptygin 1966). Need assumptions about diffusion coefficient and resonance condition. Nice curve to be shown later! Foreshock thickness a function of particle energy => acceleration should be best below 10 R_sun. See Bulanov & Sokolov 1984). Conditions on the formation of the shock wave imposed by SEP requirements. OCTOBER 29 (Fisher's 49th!) Metcalf: Na D line observations from Mees from among the first routine such observations. The observations are only line-of-sight. SOLIS is doing this too, but with the Ca triplet instead. The pre-flare magnetogram at 18:46 is excellent, but things got worse. The flare is complicated, and the post-flare magnetograms are poor. Do flares cause bad seeing? The line goes into emission for a long time, which makes the post-flare view complicated. Magnetic virial theorem => total energy, done better here because the D-line is formed in a force-free domain. Get 3.6+-1 x 10^33 ergs. Find open configuration at 10x total virial energy for this AR (the IVM field of view); the errors in this kind of comparison make it difficult to do this with the photospheric field. Magnetograms during the flare show obvious flare effects and unusual Stokes profiles. Post-flare result does not establish an energy change at this yet at even the 1-sigma level. Leka: Now the photospheric observations for the same instrument for the same event, but the two channels are not simultaneous. There are the usual caveats about the inversion and disambiguation. At 17:12 get an excellent look at vertical currents and "alpha_best". Magnetic morphology shows an intrusion of opposite-polarity sunspots that is not so visible in the intensity (spots). The bottom line: the survey data are pretty good for a boundary condition, but not for direct observations during the flare itself. BBSO (Qiu et al) paper on magnetic effects (ApJ 601, L195, 2004). McTiernan: RHESSI data, mostly from Sa"m's Web page. There are particles in the late phase. The movie shows two sets of contours (eg 15-20, 50-100 keV) overlaid on TRACE images. Footpoint motion at higher energies is distinct and smooth. How does the morphology fit onto the magnetograms (or better yet, the 3D coronal B map)? Tune into RHESSI-Nessie 2 and/or the RST workshop in December (www.rhessi-soho-trace.org). Jackson: LASCO observations of October 28, 29; SMEI data. MURI graduate student Tamsen Dunn. Scintillation index work from IPS for background to general problem of tomographic reconstructions of ICME development. Bastille Day 2000 results. The IPS results for this period are limited in information. .... apologies, Bernie - no further notes here! Discussion of October flares: how to do before/after vector field. Global field modeling has been done by Janet. Ilia says "why not" when asked about modeling for these events. Yang Liu: May 12 1998 sigmoid => arcade and a moving ribbon crossing the umbra again. The prototypical morphology all around. Mitaka vector magnetic observations show different alpha between negative and positive domains. Either the data are not good, or it is not potential, or perhaps both. Have compared potential-field extrapolation based on LOS data with the vector transverse field suggest a factor 1.2 would be needed; for Mees (post-flare epoch) again compare alphas and LOS mapping, with better results. Suggest vertical current decreases after the flare by about 40% Also look at difference in synoptic maps and see a difference at 1.2 R_s level: the "average height" of the field gets larger at the time of the event. Discussion of this point; Zoran seems to disagree based upon the arbitrariness of field-line identificdation; Bernie wants currents circulating in the corona; Hugh suggests than in ideal MHD this behavior should occur if the CME represents the opening of coronal field. KD points out that photospheric magnetic signatures of filament channels are hard to see. Evolution of magnetic field, combining Mitaka, Huairou, Mees, MDI... movie. Are the Mitaka data trustworthy? Is there anything physical to discuss? Canfield points out that alpha distributions CAN be complex even in a sunspot umbra. "The data is unusable..." - Zoran's opinion, ie he would not like to model them. More discussion of what changes are believable and what aren't. K.D. Leka: "I think Zoran will like my talk a lot!" IVM data for 1997 May 9-15 make a "miserable picture". KD is making lots of technical comments, including such items as flat-fielding and seeing, to justify this position. For example, a vertical-currrent jump at the penumbral boundary is artifactual. Bottom line: for the 12th can provide a B_z boundary condition. There is nothing trustworth in the time series, though, except for gross changes. Yan Li: LCT results show converging flows for 1997 May 12, also pretty visible from the direct movie. Can't see rotation easily. [Yan, sorry, I ran off to fetch something! - HH] Yang Liu de-projected the images, but George points out that this would not cure LOS/transverse crosstalk. Zoran Mikic: Idealized models of May 12 event. Grumbling about the small spatial scale of the region, but working with a simplest representation of field that could mimic it. Find helmet-streamer equilibrium for active region plus large dipole => PB movie. CH notch results as PCH extension reaches into the AR. The orientation of the AR determines the behavior of the neutral line; according to Spiro some connectivities would allow eruption. The disconnection of the neutral line - a bubble within the helmet streamer - effects of orientation are really physical (Janet comment). Structural measures of the field, eg expansion factor ("K factor" of Demoulin, basically the mirror ratio and not the Jacobian of the map) that shows the location of a separatrix or QSL. Now add some shear flow, twisting on either side, hoping to get an eruption. The disconnected neutral line seems to be a clue to the eruption. What about Spiro's "canals" of open field? (BW question). The breakout model seems like a reasonable scenario, not "repugnant". Ilia comment - need higher resolution. Zoran comment - cancellation is more important than shear. Yan Li points out the "homologous" events before and after May 12. Q. Why is the simplification necessary? A. Magnification of structure and distillation of physics. Dusan Odstrcil: The rest of the heliosphere! Ambient solar wind from SAIC (Riley) and from PFSS (Arge). Zhao's "cone model" with e.g. ~650 km/s radial velocity. Structure develops with high speeds out of plane, and interaction with ambient SW. Connectivity can jump tens of degrees. Evolution at ACE rather well matched. There is no flux rope, though, so B at ACE does not look so great because of the "cone model" used. Can see effects of SW structure but it is not too well understood in detail because it's derived from synoptic models. [Dusan's presentations are hard to capture in small numbers of words - ed.] Marek Vandas 3D flux-rope model next, or input from simulations. Q. Is the concave-out shape of the shock surprising? A. No. In fact SMIE may be seeng this plausible effect. Discussion Is "magnification and distillation" as practiced by the MHD modelers the right thing to do? How can one use the VMG data? How much twist is needed to result in an eruption? How much cancellation? At 90% of open field, one of Zoran's calculations suggests 10% cancellation is needed. Piet Martens's comment about even small shear and plenty of convergence leading to eruption cited. How to know the axial field if the VMG doesn't show it? How soon before we can run the coronal/SW models in real time, like the magnetospheric people seem to be doing? A long, long, time, say the modelers, because of the range of scales. How to validate if the characterization is multidimensional? Pretty pictures useful for educational purposes are not really enough. How much work do modelers do to get something to work? How hard is it to simulate an eruption? How does one track the parameter-space surveys necessary to get somewhere?