MURI meeting notes March 2, 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. DISCUSSION OF DATA ISSUES Leka: [missed a bit of the beginning, sorry!] disambiguation of vector field measurements, basic principles and algorithms (maybe about 5 independent ones). Q: Do any of the methods involve using coronal EUV data for self-consistency? A: (KD) Maybe Alex Pevtsov did something like this, but no. Q: Can't Stephane and K.D. just collaborate and come up with a "best" 8210 magnetogram (and ditto for flow field)? A: (GB) What does "best" mean? Q: What are the precise uncertainties in a given magnetogram? A: (RCC) I don't care what they are! The best way for the modelers is to proceed with numerical experiments, as Zoran described yesterday. Q: Can't we terminate the discussion here and move on? A: (ZM) No. FLOWS Fisher: All of this work is an example of a development arising from having to look at real data (continuing the discussion above!) Focus on the magnetic induction equation as the essential physics. Correlation tracking.. Z component of induction equation involves full vector field and 3D flows. Assumption v.B=0 is OK - see Welsch et al. Q: how different is the horizontal velocity obtained this way? A: (interruption) Li: Regular LCT results. Application to several regions, May 12 shown yesterday. "Last best" AR from 1996 as another example. October 25 filament disappearance/CME. See converging flows in spite of apparent divergence of field from rotation to rotation. Discussion of systematic errors, coherence, relationship to differential rotation, sampling.. Q: If surface dynamos aren't real, why is this interesting? A: (GF) Probably we're seeing a trace of the subsurface flows that ARE important. Q: Where is the cancelling flux in this region? The white and black don't meet. A: It's a question of contrast. Note: Dave Bercik thinks there might be lots of intranetwork flux. Welsch: Flare Genesis data. Georgoulis reduction to Doppler, probably from Stokes I. LCT comparison shows morphology OK. MEF and Doppler contours in a 4-color presentation shows a "somewhat less than ideal" agreement. Parallel flows are handled differently (not at all in MEF) probably explain this. If one had truly vertical field and were looking along it, one should get agreement, but... Q: Consistent, inconsistent, or can't tell? A: Hunch is, inconsistent. CORONAL CONTEXT DATA Luhmann: PFSS modeling, though often bashed, is really the first thing that one should do to get an idea about structure and evolution. It is cheap and easy. CR1922 (April 1997) synoptic results, the homologous events and their PFSS field lines. Spherical view showing last closed field lines; SXT images showing illumination of N PCH boundary; Arge experiment with deleting AR from PFSS models. Q: what about holes punched in the heliosphere, ie unconnected neutral lines? A: (ZM) it's debated, but I think there are unconnected neutral lines. Hudson: Appeal to make use of the precise images of coronal field orientation obtained from TRACE in particular into the modeling. Discussion elicits murmurs of agreement but no concrete ideas for how to do this in a practical way. Luhmann: CR1935 (May 1998). TIL arcade shows up in CR1936 but not 1935 in the PFSS models. Q: Are the large-scale fields really important? Are the active-region fields really important? Q: Is the trans-equatorial arcade here analogous to a filament channel? The TILs are characterized by high temperatures, despite long field lines, and by being anchored outside the strongest-field parts of the AR (if any). A: (JL) Not sure I agree... Q: How about homologous eruptions? A: Some say sure, others puzzled. INTERPLANETARY Manchester: CME models extending to 1 AU with adaptive grids... Textbook case of ICME: shock, sheath, tangential discontinuity, cloud signatures, rarefaction behind - all of this seems to be simulated reasonably well. For the MURI events, find May 12 97 to have something like this canonical structure. [Extensive discussion of nature of foreshock, sheath region, etc.] May 98 is more complicated; several halo CMEs and the ICMEs presumably interact with each other. [Discussion of how these interactions work and how models might help to clarify them]. Luhmann: October event overview. There is not much on the Web, probably because of some special issue in preparation. ACE interplanetary observations show October 28 and 29 plus a grazing passage of the November 4 event. SW velocities went out of range, so infer from SWICS helium data at near 2000 km/s. Note that the Nov. 4 event did not appear to decelerate even out to the distance of Saturn. How can this be? Let us solve that problem over lunch. BOUNDARY CONDITIONS Well, the Saturn problem was not resolved, unfortunately. Abbett: Code coupling is interesting and possible, but it is still not clear how to physical feedback across the boundary. A case in point is flux emergence into a magnetized corona, rather than an empty one; note also the extensive discussion of the new methods for combining fields and flows in detail. Discussion of Roussev's boundary conditions: in the corona just above but with the photospheric (synoptic map) fields. Rotation assumed at 1% Alfven speed in a geometry derived from the Wilcox-based PFSS. New plan would be to graft the MDI data onto the Wilcox data currently used. At what level is one simulating an active region, or simulating a model of an active region? Discussion of 8210 in general. Do the Wilcox data imply (through PFSS) the opposite sense of rotation from the direct movie view or LCT? Are sign conventions in the observatories different? Meeting dissolves while people look at synoptic charts and argue locally. The bottom line might be that the lower-resolution Wilcox data have led to a model not based on an actual sunspot, but on an unresolved agglomeration of field. So, input data with higher resolution is needed. Discussion of initial conditions for force-free models; mixed results with IVM? See Montana MURI and the nuggets there, specifically http://solar.physics.montana.edu/muri/nuggets/2003_apr.html and Regnier's home page. McTiernan: A view of field lines being drawn... NLFF or "quasi-force-free" for 8210. Local potential field external boundary. Here one is minimizing the force via the Roumeliotis method. Q: Do these field lines agree with Stephane's? A: Probably not where they leave the box, but they should on smaller scales. Q: How well does the fitted field match the observations? A: (WA) 15-20% or so? ...three simultaneous discussions, all of them interesting... hope I can capture some conclusions but don't count on it... McTiernan is coding IDL before our very eyes... Welsch: Image comparison of Stephane's field with the observations shows drastic discrepancies (orthogonal or antiparallel red and blue arrows just where the field structure is interesting, ie where the flare occurred). This concludes the answer to Zoran's question just above. Abbett: The R-W-M technique will not produce the discrepancies seen in the Regnier calculation. Ideally the discrepancies will have been squashed down into the dreaded zone of non-force-freeness? Or somwhere else? Q: Should a Roussev-style model start from a force-free model, or is it more self-consistent to allow it to relax itself according to the calculation? A: Hmmm. Fisher: Review Lundquist calculation of SXT apparent signal in 8210. Speculation that it would not be so bad with a potential model either. Hudson: Plug for FASR as a means of characterizing field extrapolations. Discussion suggests that some MHD still will be necessary, at least to discriminate the harmonics.s MODEL COUPLING Sokolov: Discussion of models/components, infrastructure, superstructure. FORTRAN 90 vs C++. Infrastructure implies a "coupling toolkit" eg block-adaptive grid. Components include coupling to ionosphere, magnetosphere including kinetic treatments, radiation belt, module for particle acceleration, corona and Dusan-type model. Deadline June! Q: Is the MHD from BATSRUS? A: not all. The models coupled can even be of different dimensionality. Q: How about timestep matching? A: Brutal force was used originally; better coupling may get out of hand depending on the models. Q: Is the superstructure one single executable? A: Yes. Q: How does this relate to CISM? A: It is surely different technically. Odstrcil: Coupling to SAIC coronal model. If one is beyond the critical point the driving is one-way. The two-way coupling is more challenging, since the models need to share and interact. JASTP papers on magnetosphere/upper atmosphere. There is experience in the atmospheric community but we have harder problems. Q: Is solar rotation important in the corona at all? A: (GF) Probably one can ignore rotation if the process is fast, but in the SW of course. For differential rotation it is not necessary to have a rotating coordinate system. A: (DO) We assume a rigid rotation model. The choice of visualization tool for uniform use is Open DX rather than IDL-based Tekplot. Q: What are the biggest challenges for the coronal modelers to yield better realism? Where should efforts be focused? A: (GF) We need something that looks like an eruption from the real Sun. Ilia's work is really not at the active-region scale yet (see earlier discussion). A: (ZM) How to relate coronal model to in-situ data, ie the properties of the flux rope. A: (IR) SEPs. A: (HH) Active regions are vital. Canfield: Leamon paper on twist of magnetic clouds has 15 cases of clean clouds and a good AR/VMGs. Can we not compare these cases? A: How? Technical discussion of how MHD models relate to the potential field. Does one separate low-resolution and high-resolution structures, or does one just take 90 orders to get 1-degree resolution in a single synoptic description? JL asserts that 90 orders do a great job representing MDI data. How about 120 harmonics? "Easy."