ebook img

Reconstructing the Subsurface Three-Dimensional Magnetic Structure of A Solar Active Region Using SDO/HMI Observations PDF

0.7 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Reconstructing the Subsurface Three-Dimensional Magnetic Structure of A Solar Active Region Using SDO/HMI Observations

Reconstructing the Subsurface Three-Dimensional Magnetic Structure of A Solar Active Region Using SDO/HMI Observations Georgios Chintzoglou and Jie Zhang 3 1 School of Physics, Astronomy and Computational Sciences, George Mason University, 4400 University Dr., 0 MSN 6A2, Fairfax, VA 22030, USA 2 [email protected] n a J 0 2 ABSTRACT ] Asolaractiveregion(AR)isathree-dimensionalmagneticstructureformedintheconvection R zone, whose property is fundamentally important for determining the coronal structure and S solar activity when emerged. However, our knowledge on the detailed 3-D structure prior to . h its emergence is rather poor, largely limited by the low cadence and sensitivity of previous p instruments. Here, using the 45-second high-cadence observations from the Helioseismic and - Magnetic Imager (HMI) onboard the Solar Dynamics Observatory (SDO), we are able for the o r first time to reconstruct a 3-D datacube and infer the detailed subsurface magnetic structure t s of NOAA AR 11158 and to characterize its magnetic connectivity and topology. This task is a accomplished with the aid of the image-stacking method and advanced 3-D visualization. We [ find that the AR consists of two major bipoles, or four major polarities. Each polarity in 3-D 1 shows interesting tree-like structure, i.e. while the root of the polarity appears as a single tree- v trunk-like tube, the top of the polarity has multiple branches consisting of smaller and thinner 1 flux-tubes which connect to the branches of the opposite polarity that is similarly fragmented. 5 6 Therootsofthefourpolaritiesalignwellalongastraightline,whilethetopbranchesareslightly 4 non-coplanar. Ourobservationssuggestthatanactiveregion,evenappearinghighlycomplicated . on the surface, may originate from a simple straight flux-tube that undergoes both horizontal 1 0 and vertical bifurcation processes during its rise through the convection zone. 3 Subject headings: Sun: general — Sun: solar interior — Sun: surface magnetism 1 : v 1. Introduction neticpolarities,welldescribedbyHale’sandJoy’s i X laws (Hale et al. 1919). The subject of AR magnetic structure in the r There is limited information on the AR three- a Solar Convection Zone (SCZ, outer ∼ 220Mm dimensional (3-D) structure inferred from obser- of the solar radius) is one of the least under- vations of AR emergence. Zwaan (1987) provided stood topics but it is of crucial importance for a toy-model explanation of AR emergence by at- constraining solar dynamo models and explain- tributing it to the subsurface structure of an Ω- ing what drives solar activity and space weather. loop of magnetic flux with a frayed crest that It is widely believed that ARs seen on the sur- breaks through the surface giving rise to the ob- face are magnetic flux-tubes that are being cre- served appearance of bipolar ARs. The work by ated by the dynamo process at a depth in the Strous et al. (1996) and Strous & Zwaan (1999) SCZ(Charbonneau2005). Subsequently,theflux- extends this model to include the horizontal dy- tubes emerge through the photospheric surface namics in order to explain the fact that many givingbirthtoARsorsunspotsandmagneticloop flux-tubes emerge in multiple locations. Tanaka systems in the corona. On the surface, there is a (1991) studied complex (delta configuration)ARs highorderofregularityonthepatternofARmag- that exhibit rotational proper motions during 1 their evolution (starting from Hale’s-and-Joy’s- face (Fan 2009). laws-incompatible towards being compatible) and Inthisletter,wepresentthe implementationof attributed it to a knotted 3-D topology. Also, an image-stacking technique as a means to recon- Leka et al. (1996) found evidence that flux-tubes structandstudythe 3-Dstructureofanemerging emerge kink-deformed by the current they carry, AR directly from observations and with great de- thus containing a twisted 3-D structure differ- tail. The NOAA AR 11158presented in Figure 1, ent from an Ω-shaped flux-tube. However, it is isknownastheonethatproducedtheveryfirstX- generally difficult to determine the detailed 3-D classflare(X2.2)oftheSolarCycle24anditsener- structureofanAR,due tothe limitationofprevi- getics have been studied thoroughly by Sun et al. ous observations in terms of temporal and spatial (2012). The photospheric magnetogram images resolution. showacomplexARasseenfromthesurface. How- On the other hand, there has been a consider- ever, with our 3-D reconstruction method, it is able amount of theoretical work which have been rather evident that AR 11158has a much simpler developedoverthe pastfour decades trying to at- origin, i.e. the horizontal and vertical bifurcation tack the issue computationally (for a review, see of a single progenitor flux-tube. This is proba- Fan 2009). The models of emergence in the SCZ bly the first study of directly reconstructing the are (a) the Thin-Flux-Tube model (TFT, Spruit detailed subsurface 3-D structure and topology of 1981) and (b) the anelastic MHD model (Gough solar active regions. 1969) . While both models work well in the lower SCZ, they might not be valid at the top layers 2. Methodology of the SCZ (that is, 20 - 30 Mm below surface), In this study, we used high time-cadence and as the flux-tubes are not thin (TFT assumption high spatial resolution observations (0′.′5/pixel) breaks down) and the velocity field is not sub- taken by the HMI instrument on board the SDO sonic(anelasticapproximationbreaksdown). The spacecraft (Schou et al. 2012). The HMI instru- reasonfor making such differences lies behind the ment is able to take full-disk maps of the line-of- large pressure gradient at the layers close to the sight(LOS)B-fieldevery45seconds. Thestarting surface (Fan 2009; Stein 2012). Thus, theoretical works usually split the SCZ into two parts, the time, t0,ofourselectedobservationperiodwason 10-Feb-2011,00:00:28UTwhentheAR11158first lower SCZ and the upper SCZ (∼ 20Mm). ◦ ◦ emergedatheliographiccoordinatesE53 S20 and Withthe improvementofcomputationalpower the ending time of the period under study was and sophisticated algorithms, it has been possi- on 16-Feb-2011, 11:18:27 UT when the AR was bleformore“realistic”numericalexperiments(i.e. ◦ ◦ at W30 S20 , well passed the central meridian. fully compressible,radiative-convective3-D MHD During this six-day-longperiod, the AR had gone simulations)toexploretheformationofporesand through the emergence phase and fully developed sunspots (e.g. Cheung et al. 2010; also see re- into a mature region. For each of the observa- view by Stein 2012), although the spatial domain tion frames, we performed a geometrical correc- achieved so far is still very small (the “deepest” tion to get the radial, or normal field Bn, from simulations go down to a depth of 20Mm). Also, the LOS magnetograms. For further processing, radiative-convective MHD models provide no ex- ′′ ′′ weselectedacutoutof240 ×200 withaguiding planation on large scale characteristics of emerg- centerfollowingtheARattheCarringtonrotation ing ARs − such as the Joy’s law of AR tilts and rate. Lastly, we rotated and remapped the solar asymmetric foot-point separation − which are re- sphere fromthe centerof the cut-outs to the solar producible by global-scale models, like the TFT disk center, practically eliminating both the so- approximation (Caligari et al. 1995) and anelas- larrotationandthe projectioneffects aswell,and tic MHD simulations (Fan 2008). To this date, resulting in a good alignment of cut-out images. because of the computational restrictions of our By such pre-processing, we produced a uniform current era and the natural complexity of this dataset ready to be used in our stacking method. task, there’s no global, fully compressible MHD The image-stacking method works as follows. modelyetcapableinprobingtheevolutionofflux- Each of the frames is a map of the B-field at the emergence throughout the entire SCZ to the sur- 2 (thin) photospheric layer. We proceed onto mak- showsfinerstructuresthatenvelopethe“skeleton” ing a stack along the time dimension using a 7.5 structure of 1100 G. Apparently, the AR is com- minute cadence of the 2-D cutouts. This cadence posed of four major magnetic concentrations or effectively reduces the number of images by a fac- polarities, as indicated as P1, N1, P2,N2, respec- tor of 10, from 12330 images to 1233 images over tively in the bottom panel of Figure 2. The same the time period under study. The choice of the four polarities are marked in panel (c) and (f) in number of images accommodates the maximum Figure 1. As seen from the 3-D map, the four po- computer capacity in both hardware (in partic- laritiesoriginatefromtwobipoles,P1-N1andP2- ular, the memory) and the software. By starting N2. ThepositivepolarityP1connectstoitscorre- with t0 at the top of the stack (the X and Y di- spondingnegativepolarityN1,andP2connectsto mension) and adding images at later times con- N2. In other words, the pairs P1-N1, P2-N2 form secutively ata lowerheight (the Z dimension), we two neighboring flux-tube systems. It is easy to createa3-Ddatacube,whichcanbeusedtoinfer notethatP1andN1,andP2andN2arenotclos- the 3-D subsurface magnetic structure of the AR ingattheirapex,sinceatthislocation,thefieldis priortoits emergence. This technique is basedon weakandmostlytransverseto the LOS.However, theassumptionthatthesubsurfaceARemergesas ifwegodowntolowerB-fieldiso-surfaces,i.e. 400 a solid body, i.e. the observed flux on the surface G, we can clearly see an almost closed system of at each time instance corresponds to one partic- adjacent branch-like arches. ular height of the body. However, we know that Instead of coherent, solid flux-tubes, we ob- anemergingARissubjectedtostructuralchanges serve a very fragmented, branch-like appearance duetoturbulenceintheSCZandthenear-surface in all polarities of the bipoles. On the surface, processes. In particular, the near-surface pro- such fragmentation appeared as the continuous cesses may dominate the structure of weaker field emergence of individual small magnetic elements. (moreinthediscussionsection). Nevertheless,the However, these small magnetic elements exhibit AR selected in this paper is a strong AR with a remarkably ordered, “swarm”-like collective be- fast flux-emergence, thus making the surface ef- havior, separating in terms of polarity and coa- fect minimal. As a first-order approximation, the lescing in 3-D into big “tree-trunks” − i.e. the velocityoftheemergentstructureisassumedcon- four polarities of the quadrupolar AR. This tree- stant, thus each frame contributes equally to the branch-trunk feature signifies a deeper relation of height of the structure. For better showing the all the small magnetic features within a large- emergentstructure,weusedonly∼4.4daysworth scaleemergingstructure. Eachpolarityconsistsof of data (ending on 14-Feb-2011 04:25:17 UT), i.e. several branches, which are almost perfectly con- the first 800 images instead of the full processed nected to the branches in the opposite polarity dataset. The (x, y, z) final dimensions of the dat- along the flux-tubes. The branches are probably acube are 480 pix × 400 pix × 800 pix. The dat- caused by a bifurcation process which is further acube, which is produced in IDL, is imported to discussed below. PARAVIEW,avisualizationsoftwarepackage,for further processing and inspection. 3.2. Bifurcation in Height Both bipoles exhibit similar bifurcation along 3. Results the height (or time). An inspection on the 3- 3.1. 3-D Topology of the AR D data-cubes reveals that, for each bipole, there seems to be a two-phase evolution − or, equiv- InFigure2,wepresenttheresultedreconstruc- alently, a “grouping” of the individual small tion by showing the iso-surfaces in the 3-D data- branches to just two −but larger−groups we dub cubes for tworepresentativeconstantcontourval- “Mega-Branches”(“Mega-Branch-α” and “Mega- ues (see online material for a video fly-by around Branch-β” or for short “MBα” and “MBβ”, with the twoiso-surfaces). Onecontourlevelis at1100 “MBα” preceding “MBβ” in time, as illustrated G,whichshowsthe“skeleton”ofthestructure,i.e. in Figure 2). A more thorough study of the mag- the core structures of each sunspot/magnetic ele- netic topologyshouldinclude the time-flux profile mentoftheAR,andtheotheroneat400G,which to characterize the temporal evolution in a more 3 quantitative manner as discussed below. to the α-Episode rates, on average. Furthermore, In Figure3, we show the magnetic flux-versus- the onset times for the Episodes of the Bipoles time for each individual polarity of the AR, i.e. N1-P1/N2-P2 are only ∼ 6 hours apart, i.e. very P1,N1,P2andN2,andweoverploteachpolarity’s similarasitcanbealsoseeninthe3-Dreconstruc- unsigned profile in the same graph. According to tion of Fig 2. this plot, P1-N1 (solid lines) is the first one that 3.3. Bifurcation in Horizontal Direction emerges and it is followed by the second bipole, P2-N2, about only six hours later (dashed lines). At first sight, AR 11158’s magnetic topology From the time-flux profile in Fig.3, we see that seems rather complex. However, the tendency for for an individual bipole (i.e. N1-P1 and N2-P2) collinearity of the four polarities at the bottom the magnetic flux of its positive and negative po- of the 3-D cube, suggests that both flux-tubes larity is − to a first order − very similar; such might be related, even originating from the same similarity is in phase throughout the emergence parentalmagneticflux-tube. Thispictureisthere- period. Also, for both bipoles, two major flux- foresuggestingthatasinglesub-photosphericflux- emergencephases/episodesareidentified,withthe tube has been bifurcated into two tubes along the initial one being moderate (“α”-Episode) and the horizontal direction , as illustrated in Figure 4. later one (“β”-Episode) being a stronger “flux- The almost in-phase evolution of the fluxes with surge”. The major contributionto these two flux- time of the two bipoles further reinforces this in- emergence episodes is coming from the respective terpretation. adjacent “Mega-branches” of the emerging flux- Flux-tubes are non-coplanar when they first tube, as suggested by the 3-D reconstruction of emerge through the quiet sun (see Fig 1 and the Fig. 2. In each of the bipoles, the Mega-Branch online videos). However, right before the emer- that arrives first at the photosphere, “MBα”, ap- gence of the “flux-surge” in both of the bipoles, pears somewhat weaker and fragmented whereas thereisastrongpushing-asideofthealreadyemer- the branchthat arrives∼ 2 days later, “MBβ”, is gent Mega-Branches (“MBα’s”) in a manner like much stronger. Thus, for each bipole we have a they seemto “knowinadvance” aboutthe arrival bifurcation in height, as deduced from Figs. 2, 3. of the stronger “flux-surge” tubes (or “MBβ’s”), In Table 1, we provide the total unsigned flux suggestingthat the “MBβ’s” interact/collidesub- emergedforeachepisode’spolarityalongwiththe photospherically with the trunks of the “MBα” informationonthe durationofemergenceandthe tubes. At a later time, i.e. at the bottom of flux-emergence rate, selected manually from Fig- thedata-cube,thepolaritiesassumethe“correct”, ure 3. The duration of emergence is the time Joy’s law-compliant tilt. between the onset (annotated green lines) and Further,for eachbipole, the asymmetricpolar- the end-of-emergence (colored asterisks) for each ity separationsuggests flux-tubes with an oblique episode. The rate of emergence is defined as the Λ-shape instead of axisymmetric Ω-loops. The Λ- flux measured for each episode, divided by its re- shapehasitsleadinglegmoretiltedawayfromthe spective duration of emergence. For comparing verticaldirectionthanthetrailingleg. Thisasym- theindividualpolarities,wepresentmeasurements metry can be understood in terms of the Coriolis forα+β-Episodes,i.e. unifyingtheepisodesbyus- force acting on a rising flux-tube, as discussed by ing the onset time of α-Episode’s and for the end Caligari et al.(1995)withtheTFTapproximation time, the one of β-Episode’s (thus yielding infor- and also reproduced using anelastic MHD models mation on the emergence of the individual polari- (Abbett et al. 2001). ties). The AR 11158 is a strong AR where emer- gence lasted 110hours with a flux-emergence rate 4. Discussion and Conclusion of 5.99×1016Mx s−1, leading to a total emerged 22 unsignedflux of2.4×10 Mx. Going downto the In this letter, we presented a novel image- leveloftheindividualpolarities,weshouldbeable stacking technique for reconstructing the 3-D to quantify the similarities seen in the 3-D visu- structure ofbuoyantflux-tubes rising throughthe alization. For both bipoles, the β-Episode emer- solarsurfaceandformingobservedsolaractivere- gence rates are a factor of 3.2 larger as compared gions. Sequences of images have been used before 4 to infer the structure of ARs , e.g. Tanaka (1991) ing the lower SCZ, after they are born. However, and Leka et al. (1996). However, the previous at- the small arch-like magnetic “fibers” seen in the temptsarelimitedtotrackingthelocationsofAR 400Gisosurfacesmaybecausedbythestrongsur- centroids with time, but not the entire structure. face processes due to a large pressure gradient in To our best knowledge, this work is probably the the upper SCZ. first true implementation of the image-stacking This study also demonstrates that the image- techniquetoreconstructthedetailed3-Dstructure stacking technique is a promising method for of an AR, using advanced visualization software studying the 3-D structure of ARs prior to their and high-cadence high-resolution magnetogram emergence. In the future, we will study the mag- data. netic vector 3-D structure by fully using the mag- At least in the early stages of emergence, netic vector observations from the SDO/HMI in- the emerging magnetic structures are two non- strument. coplanar neighboring bipoles, but a more detailed picture reveals a bifurcated structure for both The authors wish to thank Drs. M. Linton, bipoles, in the horizontal direction and along the Y. Fan and M.K. Georgoulis for valuable discus- heightaswell. InthelowB-fieldiso-surfaces,mul- sions. G.ChintzogloualsothanksProf. C.E.Alis- tiple magnetic arches can be observed to emerge sandrakis for inspiring conversation and encour- in a continuous manner, with the like-polarities agement. We acknowledge the support from NSF coalescingwithtime. The3-Dreconstructionpro- ATM-0748003,NSFAGS-1156120. Oneoftheau- vided good evidence that Mega-branches could thors (G.C.) was supported by NASA Headquar- be originating from the same flux-tube below the tersundertheNASAEarthandSpaceScienceFel- photosphere. Last, we find that there’s a dual- lowship Program- Grant NNX12AL73H. phase evolution for both bipoles, as suggested by both the topology in 3-D and the time-flux pro- REFERENCES file of the AR, providing further evidence for a Abbett, W. P., Fisher, G. H., & Fan, Y. 2001, bifurcation in height. Observations also indicate ApJ, 546, 1194 that the two bipoles have a common origin. The twobipoleshaveasimilartopologyin3-D,similar Caligari, P., Moreno-Insertis, F., & Schussler, M. temporal evolution in flux-emergence, and most 1995,ApJ, 441, 886 significantly, appear almost collinear at the later stage of emergence. It is possible that the two Charbonneau, P. 2005, Living Reviews in Solar bipoles are the result of bifurcation of a single Physics, 2, 2 progenitor flux-tube early in the evolution. Cheung, M. C. M., Rempel, M., Title, A. M., & Itisinterestingtonotethatthe3-Dtopologyof Schu¨ssler, M. 2010, ApJ, 720, 233 the AR11158−foreachbipoleaswellasoverall− exhibits all the qualitative characteristics of the Fan, Y. 2008,ApJ, 676, 680 TFT approximation. The TFT model’s successes in reproducing observations qualitatively are well —. 2009, Living Reviews in Solar Physics, 6, 4 known (Caligari et al. 1995). The same qualita- Gough, D. O. 1969, Journal of Atmospheric Sci- tivecharacteristicsoftheTFTwerereproducedin ences, 26, 448 incompressible MHD simulations by Abbett et al. (2001) by including the Coriolis force due to the Hale,G.E.,Ellerman,F.,Nicholson,S.B.,&Joy, solar rotation, in order to study its effects on the A. H. 1919,ApJ, 49, 153 fragmentation of flux-tubes. This simulation re- produces the non-axisymmetrical topology that Leka, K. D., Canfield, R. C., McClymont, A. N., arises due to the Coriolis force. From our recon- & van Driel-Gesztelyi, L. 1996, ApJ, 462, 547 structionit is evident thatwe havesuchanasym- Schou, J.,Scherrer,P.H., Bush, R. I., et al.2012, metry. The fact that we observe it suggests that Sol. Phys., 275, 229 the upper SCZ hasnotasevereimpactinaltering the magnetictopologyof flux-tubes while travers- Spruit, H. C. 1981,A&A, 98, 155 5 Stein,R.F.2012,LivingReviewsinSolarPhysics, 9, 4 Strous, L. H., Scharmer,G., Tarbell, T. D., Title, A. M., & Zwaan, C. 1996, A&A, 306, 947 Strous, L. H., & Zwaan, C. 1999, ApJ, 527, 435 Sun, X., Hoeksema, J. T., Liu, Y., et al. 2012, ApJ, 748, 77 Tanaka, K. 1991, Sol. Phys., 136, 133 Zwaan, C. 1987, ARA&A, 25, 83 This2-columnpreprintwaspreparedwiththeAASLATEX macrosv5.2. 6 Fig. 1.— The first six *days of evolution of the AR 11158 as observed with the SDO/HMI LOS mag- netograph. The individual polarities are named after which bipole emerged first, e.g. bipole 1, hence we name its negative polarity N1 and its positive P1, and with N2 and P2 emerging at a later time. The ′′ ′′ white crossshows the position of the guiding center of the 240 ×200 FOVat a fixed heliographic latitude φ = −20◦. Note that the bipoles initially are non-collinear (dashed lines, panel (c)); at a later time they become quasi-collinear (panel (f)). 7 Fig. 2.— The 3-DreconstructionofAR11158usingthe image-stackingmethod onSDO/HMI LOSmagne- tograms. The flux-tubes shownhere areat |Bn|=400G(up) and 1100G (down). Ontopof eachdatacube is the last HMI LOS frame of the cube, i.e. the bottom frame, on 14-Feb-2011 04:25:57. The presented durationof the observationis 100.4hours. The positive X-axis directionis westwardand Y-axis northward. The length of the X-axis is roughly comparable with the height of the Solar Convection Zone (SCZ), i.e. about 200 Mm. The black loops are grouping the emergence episodes into Mega-Branches-α,β. 8 Fig. 3.— The time evolution of the flux for the bipoles N1-P1 (solid lines) and N2-P2 (dashed) suggests thatbothhaveanearlyemergencephase(i.e. “α”)or“front”offlux,followedbyastrongflux“surge”(“β”) as also can be seen in Fig 2. The Mega-Branches-α,β are the major contributors for each episode. Note the persistent lagging of the positive, i.e. leading polarities, P1 and P2 with respect to the following N1 and N2. The green lines denote the onset of emergence episodes. Also, the end of emergence is shown with an asterisk (*) in the respective color. 9 Fig. 4.— Model sketch of the emergence process of AR 11158 in the sub-photosphere. The plane is at the depth of the bottom of the SCZ that B-fields originally reside. The cylindrical structures shown at that plane in the SCZ are the flux-tubes of the toroidal field, generated by the solar dynamo process. For flux-tubes created in the South hemisphere during Solar Cycle 24 (like AR 11158), the B-field vector along the tubes is directed from West-to-East (here from right-to-left), as dictated by Hale’s law of polarity and the Babcock-Leighton dynamo theory. The flux-tubes also develop an asymmetric lambda-shape (“Λ”) as they rise. 10

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.