Ellerman bombs: fallacies, fads, usage
Robert J. Rutten, Gregal J. M. Vissers, Luc H. M. Rouppe van der Voort, Peter Sütterlin, Nikola Vitas
EEclipse on the Coral Sea: Cycle 24 Ascending
Editors P. S. Cally, R. Erd´elyi, A. A. NortonJournal of Physics Conference Series
Ellerman bombs: fallacies, fads, usage
Robert J. Rutten , , Gregal J. M. Vissers , Luc H. M. Rouppe vander Voort , Peter S¨utterlin and Nikola Vitas Lingezicht Astrophysics, ’t Oosteneind 9, 4158 CA Deil, The Netherlands Institute of Theoretical Astrophysics, University of Oslo, P.O. Box 1029, Blindern, N–0315Oslo, Norway Institute for Solar Physics, Albanova University Center, SE–106 91 Stockholm, Sweden Instituto de Astrof´ısica de Canarias, C/ Via Lactea S/N, E–38200 La Laguna, Tenerife, SpainE-mail:
Abstract.
Ellerman bombs are short-lived brightenings of the outer wings of H α thatoccur in active regions with much flux emergence. We point out fads and fallacies in theextensive Ellerman bomb literature , discuss their appearance in various spectral diagnostics,and advocate their use as indicators of field reconfiguration in active-region topography usingAIA 1700 ˚A images.
1. Introduction
Ellerman [1] described “solar hydrogen bombs” in 1917 as intense brightenings of the extendedwings of H α , H β and H γ , not visible in other lines and with the line cores unaffected. They lasta few minutes and occur repetitively in active regions with much flux emergence, preferentiallynear and especially between penumbrae. These properties define the Ellerman bomb (EB)phenomenon. The subsequent EB literature cannot be fully reviewed here but we point outsome fads and fallacies below. Bray and Loughhead [2] concluded in 1974 that “extensivemodern observations have added little to Ellerman’s original description” . This lack of progresschanged with high-resolution observing, first in the Flare Genesis flight [3] [4] [5] [6] [7] [8] andmore recently with the Swedish 1-m Solar Telescope (SST) [9] [10].
2. Magnetic concentrations as pseudo Ellerman bombs
Figure 1 shows snapshots from 3-hour multi-wavelength Dutch Open Telescope (DOT) movies .The H α wing image in the lower-left panel shows many bright grains that in the movie adhereto Ellerman’s description: “they seem to follow one another like the balls of a Roman candle” .About half exceed the mean intensity by over 30%, a few are brighter than 54% (4 σ ) excess.Traditional EB threshold criteria might classify these as EBs. However, movie inspectionincluding comparison with the parallel Ca II H and G-band movies shows that all the brightgrains are simply “network bright points” marking strong-field magnetic concentrations (MCs).They are not EBs but “moving magnetic features” described by [12] [13] [14] [15] [16] [17]. Norwas there flux emergence in this region evident in SOHO/MDI magnetograms. Clicking on citation numbers should open the corresponding ADS abstract page in a browser. Links: four-panel movie, wing-only movie. a r X i v : . [ a s t r o - ph . S R ] A p r llerman bombs: fallacies, fads, usage Figure 1.
Simultaneous DOT images of the main spot of decaying AR 10789 at view angle µ = 0 .
86. Clockwise: G band; blue Ca II H wing at ∆ λ = − .
35 ˚A; sum of the H α wings at∆ λ = ± . α core. Tick marks: arcseconds.MCs are bright in the continuum from hot-wall radiation [18]. The classic magnetostaticthin fluxtube/sheet paradigm [19] [20] [21] [22] [23] [24] [25] [26] explains their physics well andis well applicable [27], although actual MCs show complex morphology with rapid changes [28][29]. Since MCs are small and reside in dark intergranular lanes, their continuum brighteningis only seen at sub-arcsecond resolution [30]. MCs also appear as bright “line gaps” in thecores of neutral-metal lines due to ionization from evacuation, particularly in Mn I lines [31] [32]for which the contrast is not weakened by Dopplershifts in the surrounding granulation [11].They also appear markedly bright in the outer wings of H α [33], at larger contrast than inother diagnostics including the G band [34]. For the G band the contrast enhancement comesfrom dissociation of the CH molecules producing this feature, for the H α wings from smallercollisional damping. Both result also from MC evacuation.The H α core image in Fig. 1 shows the chromospheric fibril canopy that overlies and shieldsthe deep photosphere imaged in the far H α wings. H α does not sample the intermediatelayers (seen in Ca II H as shock-ridden clapotisphere [35]) due to its low-temperature opacitygap [36] [37] [38], so that at any wavelength its formation jumps between chromosphere anddeep photosphere depending on the chromospheric fibril opacity. The ∆ λ = ± . llerman bombs: fallacies, fads, usage Figure 2.
Synthesis of Ca II H wing and H α wing images for the numerical MHD simulationsnapshot of [11]. The scatter plots show brightness against magnetic field strength per pixel.Downdrafts make some MCs, including the brightest, brighter in the blue H α wing.because their brightness originates at about the same depth. The overlying fibrils block someMCs. Where they are sufficiently thin that the photospheric MC brightness shines through, theydegrade the MC image sharpness. This defocus, compared to the sharp LTE-formed G-bandand Ca II H-wing brightness features, results from ray spreading across the opacity gap andscattering through the effectively thin fibrils.Note in the H α wing image that the shielding by overlying fibrils appears to be less for theMCs in the moat than for MCs further away. Larger transparency of superpenumbral fibrils mayresult from repetitive shock heating by running penumbral waves [39] [40], ionizing hydrogentoo frequently to let it reach population equilibrium since hydrogen ionization/recombinationbalancing is exceptionally slow in cooling shock aftermaths due to the large n = 1–2 Ly α transition [41] [42].In the DOT movie some of the larger MCs do brighten momentarily, especially in the blue H α wing. Blue-wing brightening may result from MC downdrafts, as in the numerical simulationshown in Fig. 2. Downdrafts often occur in moat MCs that have opposite polarity to thesunspot [43]. MC downdrafts tend to produce shocks higher up [44] that are best seen inNa I D Dopplergrams [45]. Inspection of simultaneous H α and Na I D active-region Dopplergramsequences from the SST shows that such MC shock occurrence is often accompanied by H α blue-wing brightening.MC brightening may also result from field concentration by bathtub vorticity in granularswirls [46]. Such brightening also reflects increasing hot-wall radiation, and must not bemisinterpreted as MHD heating or reconnection [47]. However, swirl brightening seems a rarephenomenon. In the DOT data of Fig. 1 H α wing brightening shows no correlation with vorticityin the granular flows measured from the G-band movie. Rather, it happens in concert withCa II H-wing brightenings that seem mostly due to episodes of magnetic patch [48] compression llerman bombs: fallacies, fads, usage ✵ ✷ ✹ ✻ ✽ ✶✵① (cid:0)✁✂✄☎✆✄✝✵✷✹✻✽✶✵②✞✟✠✡☛☞✡✌ ✲✍✎✏ ✑ ✵ ✷ ✹ ✻ ✽ ✶✵① (cid:0)✁✂✄☎✆✄✝✵✷✹✻✽✶✵②✞✟✠✡☛☞✡✌ ✲✍✎✏ ✑ ✵ ✷ ✹ ✻ ✽ ✶✵① (cid:0)✁✂✄☎✆✄✝✵✷✹✻✽✶✵②✞✟✠✡☛☞✡✌ ✥✍✥ ✎ ✲(cid:0)✁✂ ✲✂✁✄ ✂✁✂ ✂✁✄ (cid:0)✁✂❉☎ ✥✆✝✂✄✂✂(cid:0)✂✂✂(cid:0)✄✂✂✷✂✂✂✐✞✟✠✞✡✐✟☛☞✌✍✎✐✟✍✌✍☛✏ ✲(cid:0)✁✂ ✲✂✁✄ ✂✁✂ ✂✁✄ (cid:0)✁✂❉☎ ✥✆✝✂✄✂✂(cid:0)✂✂✂(cid:0)✄✂✂✷✂✂✂✐✞✟✠✞✡✐✟☛☞✌✍✎✐✟✍✌✍☛✏ ✲(cid:0)✁✂ ✲✂✁✄ ✂✁✂ ✂✁✄ (cid:0)✁✂❉☎ ✥✆✝✂✄✂✂(cid:0)✂✂✂(cid:0)✄✂✂✷✂✂✂✐✞✟✠✞✡✐✟☛☞✌✍✎✐✟✍✌✍☛✏ Figure 3.
Upper row: SST EB images in H α at the specified wavelength separations from linecenter, taken from Fig. 1 of [9]. View angle µ = 0 .
67 with the limb direction to the top. Lower row:H α profile (solid) averaged over the full 67 ×
67 arcsec field of view of the observations and H α profile (dashed) for a pixel in the EB at ( x, y ) = (8 . , . α -wing contrast thresholds were applied. In our opinionthis warning should be heeded when reading [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60].
3. Ellerman bombs are not chromospheric
Figure 3 after Fig. 1 of [9] shows EBs at the unprecedented resolution of the SST. In thisslanted limbward high-resolution view EBs appear in the outer H α wings as tiny bright uprightflames that flicker rapidly while their footpoints travel along MC-filled network strands. Theensemble may be bright for many minutes but the subflames last only seconds. Inspection of thecorresponding SST movies demonstrates this behavior irrevocably. The elongated shape wasnoted before in lower-resolution data [61] [4] [62], as was their intermittent substructure [63].Figure 3 also demonstrates that EBs are purely photospheric phenomena. They are not seenat H α line center because the flames, even when tall (a few hundred to a thousand km [64] [61][65]), do not break through the overlying dense canopy of chromospheric H α fibrils that alwayscovers a growing active region. Comparison of the greyscale ranges (narrow rectangles in thelower panels of Fig. 3) shows that the EB wing emission is much brighter than the brightestline-core features. Some diffuse line-center brightening might again result from photosphericEB emission that passes through the opacity gap and scatters through the fibrils, but our SSTmovies suggest that such line-center brightening above EBs is uncommon. At ∆ λ = − . http://iopscience.iop.org/0004-637X/736/1/71/fulltext ; one example. llerman bombs: fallacies, fads, usage λ = − . λ = − . α fibril canopy is thicker in this flux-emergence region than around the little decaying spot ofFig. 1.Thus, an EB top may reach higher than the 400–500 km nominal height of the temperatureminimum in standard one-dimensional static-equilibrium models of the solar atmosphere [66][67] [68] [69]. The onset of the actual solar outward temperature rise likely varies between200 km in MCs and 2000 km in internetwork, varying temporally as well [42] [70], but in anycase, the jet-like EB flame protrusions originate from the deep photosphere, in the network atthe EB footpoint, and do not affect or poke through the overlying chromosphere defined [35] bythe H α fibrils.Apparent blue-brighter-than-red EB wing asymmetry [71] usually results from inverseEvershed flows [72] [73] [74] along the fibrils, the dark chromospheric line core shifting intothe red-wing emission. Thus, asymmetries and wavelengths of EB emission peaks do not defineDopplershifts of the EB emitting material [59] [6] but are set by the absorbing overlying fibrils[75].It has been suggested that EBs are accompanied by H α surges [76] [77] [78] [79] [80] [81].However, the high-resolution SST data of [9] contained only two tentative cases. Our newer SSTdata sampled in Fig. 4 contain at most one questionable case [10]. An EB–surge connection iscertainly not ubiquitous [4].Upshot regarding EB fallacies: most EB papers err in describing the EB phenomenon aschromospheric. It is not. EBs have no systematic counterpart in the overlying chromosphere,transition region or corona. Dark chromospheric H α cores are not EB ingredients, nor are theirDopplershifts.Comment regarding EB fads: EBs may combine strong downflows in the low photosphere [82]with outward flows higher up [81] [9]. Such bi-directional flows are reminiscent of “chromosphericanemone jets” seen in Hinode Ca II H data [83] [84] [85]. If such Ca II H jets actually are EBs,they are also confined to the photosphere. Just as for H α , not every bright Ca II H feature isnecessarily chromospheric.
4. Ellerman bomb visibility
Even though they are photospheric, proper EBs (those that adhere to Ellerman’s description)do not show up in neutral-metal lines nor in the continuum. At sufficient angular resolution theydo show continuum moustaches, G-band brightening [86], and narrow neutral-line gaps at theirfootpoint, sampling the MC from which they arise. Ellerman wrote: “they frequently appear inthe faculae so that their spectra are superposed on those of faculae, thus giving the appearance ofgreat extension to the bright “bomb” band ” which wasn’t heeded in most pseudo-EB literature.EB flames are also observable in Ca II 8542 ˚A, Ca II H & K, and in mid-UV continua [87][4] [6] [88] [89] [62] [90] [91] [86] [10]. They are not identically the same in these differentdiagnostics, but they often show up at the same space-time locations. We illustrate this forthe 1700 ˚A continuum in Fig. 4 by combining image cutouts from the new SST data of [10]with co-aligned image cutouts from SDO. Clearly, AIA’s 1700 ˚A shows the larger EBs also, atmuch lower resolution but with less interference by chromospheric fibrils. Since this continuumoriginates in the upper photosphere [67] [92], the bottom part of an H α EB is hidden in slanted1700 ˚A viewing but this is not noticeable at AIA’s angular resolution.AIA’s 1600 ˚A images (not shown) show the EBs too, at yet larger contrast, but these alsoshow a few more extended and yet brighter transition-region transients for this area and period.They have short-loop morphology and correspond to bright patches in H α line-center, are also llerman bombs: fallacies, fads, usage ✵ ✷ ✹ ✻ ✽ ✶✵① (cid:0)✁✂✄☎✆✄✝✵✷✹✻✽✶✵②✞✟✠✡☛☞✡✌ ✍✎✏✏ ✑ ✒✓✔✕✖✔✗✗✘ ✵ ✷ ✹ ✻ ✽ ✶✵① (cid:0)✁✂✄☎✆✄✝✵✷✹✻✽✶✵②✞✟✠✡☛☞✡✌ ❋✍ ✎ ✏✑✒✑ ✓ ✔✕✖✗✍✘ ✙ ✵ ✷ ✹ ✻ ✽ ✶✵① (cid:0)✁✂✄☎✆✄✝✵✷✹✻✽✶✵②✞✟✠✡☛☞✡✌ ❍❛ ✥✍✎ ❉✏ ✑ ➧ ✒ ✓ ✵ ✷ ✹ ✻ ✽ ✶✵① (cid:0)✁✂✄☎✆✄✝✵✷✹✻✽✶✵②✞✟✠✡☛☞✡✌ ❍❛ ✥✍✎✏✍✑✵ ✷ ✹ ✻ ✽ ✶✵① (cid:0)✁✂✄☎✆✄✝✵✷✹✻✽✶✵②✞✟✠✡☛☞✡✌ ❍✍✎ ✏✑✒✓✔✒✕✕✖ ✵ ✷ ✹ ✻ ✽ ✶✵① (cid:0)✁✂✄☎✆✄✝✵✷✹✻✽✶✵②✞✟✠✡☛☞✡✌ ❍✍✎ ✏✑✒✓✔✕✖✒✗✑✏ ✵ ✷ ✹ ✻ ✽ ✶✵① (cid:0)✁✂✄☎✆✄✝✵✷✹✻✽✶✵②✞✟✠✡☛☞✡✌ ✍✎✏✏ ✑ ✒✓✔✕✖✔✗✗✘ ✵ ✷ ✹ ✻ ✽ ✶✵① (cid:0)✁✂✄☎✆✄✝✵✷✹✻✽✶✵②✞✟✠✡☛☞✡✌ ❍✍ ✎ ✏✑✒ ✓ : U T ✵ ✷ ✹ ✻ ✽ ✶✵① (cid:0)✁✂✄☎✆✄✝✵✷✹✻✽✶✵②✞✟✠✡☛☞✡✌ ✍✎✏✏ ✑ ✒✓✔✕✖✔✗✗✘ ✵ ✷ ✹ ✻ ✽ ✶✵① (cid:0)✁✂✄☎✆✄✝✵✷✹✻✽✶✵②✞✟✠✡☛☞✡✌ ❋✍ ✎ ✏✑✒✑ ✓ ✔✕✖✗✍✘ ✙ ✵ ✷ ✹ ✻ ✽ ✶✵① (cid:0)✁✂✄☎✆✄✝✵✷✹✻✽✶✵②✞✟✠✡☛☞✡✌ ❍❛ ✥✍✎ ❉✏ ✑ ➧ ✒ ✓ ✵ ✷ ✹ ✻ ✽ ✶✵① (cid:0)✁✂✄☎✆✄✝✵✷✹✻✽✶✵②✞✟✠✡☛☞✡✌ ❍❛ ✥✍✎✏✍✑✵ ✷ ✹ ✻ ✽ ✶✵① (cid:0)✁✂✄☎✆✄✝✵✷✹✻✽✶✵②✞✟✠✡☛☞✡✌ ❍✍✎ ✏✑✒✓✔✒✕✕✖ ✵ ✷ ✹ ✻ ✽ ✶✵① (cid:0)✁✂✄☎✆✄✝✵✷✹✻✽✶✵②✞✟✠✡☛☞✡✌ ❍✍✎ ✏✑✒✓✔✕✖✒✗✑✏ ✵ ✷ ✹ ✻ ✽ ✶✵① (cid:0)✁✂✄☎✆✄✝✵✷✹✻✽✶✵②✞✟✠✡☛☞✡✌ ✍✎✏✏ ✑ ✒✓✔✕✖✔✗✗✘ ✵ ✷ ✹ ✻ ✽ ✶✵① (cid:0)✁✂✄☎✆✄✝✵✷✹✻✽✶✵②✞✟✠✡☛☞✡✌ ❍✍ ✎ ✏✑✒ ✓ : U T ✵ ✷ ✹ ✻ ✽ ✶✵① (cid:0)✁✂✄☎✆✄✝✵✷✹✻✽✶✵②✞✟✠✡☛☞✡✌ ✍✎✏✏ ✑ ✒✓✔✕✖✔✗✗✘ ✵ ✷ ✹ ✻ ✽ ✶✵① (cid:0)✁✂✄☎✆✄✝✵✷✹✻✽✶✵②✞✟✠✡☛☞✡✌ ❋✍ ✎ ✏✑✒✑ ✓ ✔✕✖✗✍✘ ✙ ✵ ✷ ✹ ✻ ✽ ✶✵① (cid:0)✁✂✄☎✆✄✝✵✷✹✻✽✶✵②✞✟✠✡☛☞✡✌ ❍❛ ✥✍✎ ❉✏ ✑ ➧ ✒ ✓ ✵ ✷ ✹ ✻ ✽ ✶✵① (cid:0)✁✂✄☎✆✄✝✵✷✹✻✽✶✵②✞✟✠✡☛☞✡✌ ❍❛ ✥✍✎✏✍✑ ✵ ✷ ✹ ✻ ✽ ✶✵① (cid:0)✁✂✄☎✆✄✝✵✷✹✻✽✶✵②✞✟✠✡☛☞✡✌ ❍✍✎ ✏✑✒✓✔✒✕✕✖ ✵ ✷ ✹ ✻ ✽ ✶✵① (cid:0)✁✂✄☎✆✄✝✵✷✹✻✽✶✵②✞✟✠✡☛☞✡✌ ❍✍✎ ✏✑✒✓✔✕✖✒✗✑✏ ✵ ✷ ✹ ✻ ✽ ✶✵① (cid:0)✁✂✄☎✆✄✝✵✷✹✻✽✶✵②✞✟✠✡☛☞✡✌ ✍✎✏✏ ✑ ✒✓✔✕✖✔✗✗✘ ✵ ✷ ✹ ✻ ✽ ✶✵① (cid:0)✁✂✄☎✆✄✝✵✷✹✻✽✶✵②✞✟✠✡☛☞✡✌ ❍✍ ✎ ✏✑✒ ✓ : U T Figure 4.
Two small-field cutouts of corresponding SST images (upper row of each set) andSDO images (lower rows) from the data of [10]. Magnification per pdf viewer is recommendedfor the SST images. View angle µ = 0 .
89; the panel tops are in the limb direction. The thirdset shows the same cutout as the second, but five minutes later. Each panel is bytescaledindependently. The arrows overlaid on the SST magnetograms show surface flows measuredfrom the continuum sequence by correlation tracking. llerman bombs: fallacies, fads, usage α line center.The He I 304 ˚A images in Fig. 4 indeed mimic the H α line-center morphology which appearsindifferent to EBs underneath.It is illustrative and recommended to play and blink SDO 1700 ˚A, 1600 ˚A, and He I 304 ˚Amovies of a very active region with much flux emergence, for example AR 11654 on January10, 2013 and the following days when it was crackling with short-lived brightenings in thesediagnostics. The plethora of transient brightness features in 1600 ˚A are the sum of different sorts,respectively seen in 1700 ˚A and in He I 304 ˚A. Very short-lived, pointlike 1700 ˚A brighteningswere likely photospheric H α -wing EBs, whereas more spread-out loop-like longer-durationbrightenings were likely flaring arch filaments and small flares, bright in the transition-regiondiagnostics and probably bright at H α line center. Inspection of such flux-emergence AIA moviesand of our high-resolution SST data suggests that distinction should be made between point-like(at SDO resolution) EBs and loop-like flaring arch filaments as different entities, respectivelyphotospheric and upper-atmosphere. AIA’s 1700 ˚A shows fewer of the latter type and is thereforethe best SDO diagnostic for EB studies. Figure 4 demonstrates that SDO’s resolution is poorcompared with the SST’s, but the obvious AIA advantage is that it samples the stronger EBsin any Earth-side active region at any time.Comment regarding EB fallacies: the low resolution of older data has led to claims of apparentcospatiality of higher-atmosphere brightenings with EBs. Examples: coincidence with EBs wasclaimed for 6 out of 16 microflares by [93] but the example pair in their Figs. 4 and 5 has8 arcsecond separation. The H α line-center brightening observed and modeled by [94] was notan EB. The H α ± .
35 ˚A feature claimed by [95] to be an EB at 1 MK doesn’t look like an EBto us. The long-lived EB of [96] at the foot of an arch filament looks like a large MC. Similarlyfor the Ca II H bright points bordering the flaring arch filament in [97], much like the brightpoints of [98]. [6] reported one coincidence of X-ray brightening at an exceptionally bright EBbut doubted ubiquitous co-spatiality. We have aligned our new SST data with SDO/AIA imagesequences and searched for but did not find any significant EB impact on the overlying transitionregion (He I 304 ˚A) and corona (Fe IX 171 ˚A) [10].Comment regarding EB fads: most EB studies are based on observations in the H α wings andtypically describe a single, a few, or at most some dozen EBs in a single active region duringa short period. This type of study should exploit IRIS in the near future since EBs will bewell visible in Mg II h & k, as in Ca II H & K. On the other hand, it seems time to exploit theAIA 1700 ˚A database to study very many more with respect to occurrence patterns.
5. Ellerman bomb detection
In [9] EBs were identified and selected manually on the basis of their flame morphology inthe SST H α -wing movies, meaning bright, narrow, tall, upright, flickering appearance in thelimbward view at this unprecedented resolution. More formal detection criteria are formulatedfor two SST data sets in [10]. One of the requirements defining an EB kernel is 55% excessintensity in the H α wings over the spatial average of the active region, while setting a thresholdof 5 σ or higher above the mean seems a good counterpart to select EBs in AIA 1700 ˚A images.Both thresholds are passed by the EBs in Fig. 4, whereas almost none of the bright H α -wingfeatures in the movie sampled in Fig. 1 passed a 55% H α -wing excess threshold, nor a 5 σ excessthreshold in simultaneous 1600 ˚A images from TRACE . Recently [60] claimed to detect over 3000 EBs during 90 minutes around a small spot rather like the one inFig. 1. Our application of the above threshold to the simultaneous SDO 1700 ˚A images suggests that they mostlydetected pseudo-EBs of which the H α -wing brightness has nothing to do with MHD heating. llerman bombs: fallacies, fads, usage “Onrarer occasions they [EBs] are superposed on bright reversals of H α over eruptive regions, but thisis an uncommon occurrence, and the distinction is easily made between the two phenomena bythe flickering of the “bomb” band compared to the H α reversal, due to the effect of seeing.” – asusing scintillation to distinguish stars from planets. Small size, fast variation, large brightness,and appearing in a region with much flux emergence together become the recipe to locate EBs inAIA 1700 ˚A image sequences. We are presently refining such automated detection for multipleviewing angles [99].
6. Ellerman bomb radiation mechanism
In traditional one-dimensional stratification modeling along a vertical line of sight the emergentH α profile maps the variation of the source function with depth, with the outermost wingssampling the deepest layers [100]. Postulating a suited temperature perturbation withappropriate structure and motion to a standard model atmosphere can then explain any H α excess emission profile [101] [88], but not unambiguously; different models may produce similarH α profiles [91].In the slanted perspective of Figs. 3 and 4 the EB flames appear as extended, dense, hotslabs that stick up from the network MCs into the otherwise H α -transparent upper photosphere,making cloud modeling more appropriate than one-dimensional plane-parallel modeling. The EBexcess emission and its profile including its moustache extent are properties of such a slab, whilethe core absorption and Dopplershift patterns are properties of the overlying fibrils. Furtheraway from line center the slab is optically thinner but not “deeper”.We attribute the non-visibilty of the EB slabs in neutral-metal lines to neutral-metalionization in the hot EB flames, and similarly their transparency in the continuum toH − ionization. Considerable ionization of neutral hydrogen is also likely, making cascaderecombination the main producer of the EB H α photons. Even if the ionization occurs onlybriefly during a momentary reconnection event, H α will so shine in a longer-duration afterglowbecause hydrogen recombination balancing is slow in the aftermath [41] [42].Following the suggestion of [102] we attribute the extended bright moustaches to subsequentthermal Thomson scattering of these photons, notwithstanding the small process crosssection,because the flames combine high temperature with very high (mid-photosphere) density. Whilethe line-center peak of the EB brightness profile remains hidden by overlying fibrils, the wingsmay gain brightness at large thermal electron Dopplershift whenever a line-core photon meetsan electron and is scattered our way. Since EBs are very optically thick at H α line center, beingnon-transparent already in the outer wings, we suggest that H α resonance scattering confinesline-core photons in a random walk within the EB until they are Dopplershifted into a wing andescape. This scattering mechanism will produce similar moustaches for other chromospheric lines(principally Mg II h & k, Ca II H & K, and Ca II 8542 ˚A) if an EB similarly confines scattering linephotons, which is likely unless these ions ionize too much. Obviously, numerical simulation withdetailed spectral synthesis including non-monochromatic Thomson scattering may vindicate thismoustache mechanism.In this view large moustache extent happens because EBs are photospheric and makes themvisible beyond the overlying fibrils in the spectrum. For H α such scattering occurs only withinthe flame, there being neither free electrons nor hydrogen atoms in the n = 2 level in thesurrounding atmosphere, and so it remains local. Flame images taken in the outer H α wingstherefore remain sharp, as demonstrated in Figs. 3 and 4. In Ca II H & K EBs get a diffuse halofrom additional resonance scattering in the surrounding upper photosphere [90].Comment regarding EB fallacies: Severny [49] [50], recognizing that the moustache extentcannot be explained by Stark broadening, invoked an EB scenario of nuclear explosion andrelativistic particle beams. This way of thinking became a school of thought in which flares llerman bombs: fallacies, fads, usage α line-core brightening by passing through andnot affecting the H α chromosphere. Linear line-core polarization attributed to the beam impactbecame the diagnostic for this notion [103] [104] [105] [106] [107] [108] [109] [110] [88] [111].In our opinion these authors failed to recognize that EBs are photospheric and occur withoutanything happening overhead. We note that in the example of [103] the polarization is notcospatial with the moustache, and that the excess emission profiles of [104] have no dip at linecenter. We suspect that the measured line core polarizations stem from flaring arch filaments,possibly through the mechanism of [112].
7. Ellerman bomb occurrence
The SST samples in Fig. 4 include magnetograms that show examples of what we find to becharacteristic behavior in these data. Small white opposite-polarity Stokes-V patches produceEB flames in the H α wing when colliding at relatively high speed with larger black patches,and then vanish. This is good evidence, similar to [90] but at higher resolution and withbetter statistics, that EBs mark strong-field cancelation and supports the notion that the brightEB flames are caused by photospheric reconnection [113] [114] [4] [80] [81] [115] [63] [9]. Thegasdynamical instability proposed by [116] seems a less likely explanation since it does not havefield emergence as necessary condition.Strong opposite-polarity field emergence may happen with small-scale ∩ loop shape inclassical Ω emergence [117] or with additional ∪ loop shape in serpentine emergence from theParker instability [118] [119], with connecting arch filaments [120] accompanied by dipped fieldsand bald patches [3] [7] [8] [65] [121] [122] [115] [123] (see cartoons in [65] [123] [124]). In thispicture chromospheric convergence of ∪ loop sides is thought to give reconnection in the line-tied regime [80] [124], but the EBs in our SST data occur at colliding photospheric flows in thefrozen-in regime (Fig. 4).Comment regarding EB fads: serpentine patterning into opposite-polarity pairs of the MCs inthe moat flow around decaying spots was already suggested by [14], but since EBs are emerging-flux phenomena it is tempting to assume serpentine emergence and ∪ loop patterning. Thisscenario may well be valid, but convincing proof requires long-duration wide-field imagingspectroscopy with spectral, spatial, and temporal resolution at least comparable to our SSTsequences. Obviously an IRIS topic. Photospheric feature tracking following [125] would be agood technique to analyze such data. Likewise, the notion of bi-directional anemone-jet flowswith reconnection a few hundred km up is attractive but needs further verification.
8. Conclusion: Ellerman bomb usage
EBs are the most spectacular solar photosphere phenomenon. They seem especially informativeas space-time markers of strong-field reconnection events in the photosphere, better thansearching for bipolar cancelation in magnetograms. Comparison of the SST and HMImagnetograms in Fig. 4 illustrates the need for superhigh resolution and sensitivity in thelatter approach and shows that even at the high SST quality one would not be able to identifyreconnection events from magnetogram sequences alone. Larger Stokes sensitivity is desirablebut longer integration would degrade the temporal and spatial resolution fatally. Thus, EBdetection seems a better way to locate small-scale reconnection events in emerging flux regions.Since the larger and brighter EBs are also visible in 1700 ˚A images even at the AIA resolution, theAIA database permits monitoring active-region field re-configuration through EB identificationin large data volumes [99]. EBs may so become useful photospheric telltales in studyingchromospheric active-region field topology evolution and energy loading.
EFERENCES Acknowledgments
We thank R. Rezaei for informative discussions and the referee for valuable suggestions. RJRacknowledges that the late C. Zwaan often suggested EBs as research topic, and thanks theLeids Kerkhoven-Bosscha Fonds for travel support. We made much use of NASA’s AstrophysicsData System Bibliographic Services. The macro to make the citations above link to ADS (atleast in the arXiv preprint) was contributed by EDP Sciences.
References [1] Ellerman F 1917
Astrophys. J. The solar chromosphere
Chapman & Hall, London[3] Bernasconi P N, Rust D M, Georgoulis M K and Labonte B J 2002
Solar Phys.
Astrophys. J.
Solar Variability:From Core to Outer Frontiers ( ESA Special Pub.
Astrophys. J.
Astrophys. J.
Adv. SpaceResearch Astrophys. J.
ApJ submitted[11] Vitas N, Viticchi`e B, Rutten R J and V¨ogler A 2009
Astron. Astrophys.
Solar Phys. Solar Magnetic Fields ( IAU Symposium vol 43) ed Howard R 329[14] Harvey K and Harvey J 1973
Solar Phys. Chromospheric Fine Structure ( IAU Symposium vol 56) ed Athay R G 201[16] Muller R and Mena B 1987
Solar Phys.
Astrophys. J.
Solar Phys. Magnetic flux tubes and transport of heat in the convection zone of the Sun.
Ph.D. ThesisUtrecht University[20] Spruit H C 1981 in
The Sun as a Star ed Jordan S
NASA Special Pub.
Astron. Astrophys.
Astron. Astrophys.
Solar Interior and Atmosphere
Univ Arizona Press 890[24] Solanki S K and Brigljevic V 1992
Astron. Astrophys.
L29[25] B¨unte M, Solanki S K and Steiner O 1993
Astron. Astrophys.
Space Science Rev. Astron. Astrophys.
Astron. Astrophys.
Astron. Astrophys.
Astrophys. J.
Astrophys. J.
Astrophys. J.
Astron. Astrophys.
Astron. Astrophys.
L15[35] Rutten R J 2012
Roy. Soc. London Phil. Trans. Series A
Solar Phys. Astron. Astrophys.
A86[38] Leenaarts J, Carlsson M and Rouppe van der Voort L 2012
Astrophys. J.
Astrophys. J. Lett.
L85[40] Christopoulou E B, Georgakilas A A and Koutchmy S 2001
Astron. Astrophys.
Astrophys. J.
EFERENCES [42] Leenaarts J, Carlsson M, Hansteen V and Rutten R J 2007 Astron. Astrophys.
Astron. Astrophys.
Astrophys. J. Lett.
L24[45] Rutten R J, Leenaarts J, Rouppe van der Voort L H M, De Wijn A G, Carlsson M and Hansteen V 2011
Astron. Astrophys.
A17[46] Bonet J A, M´arquez I, S´anchez Almeida J, Cabello I and Domingo V 2008
Astrophys. J. Lett.
L131[47] Jess D B, McAteer R T J, Mathioudakis M, Keenan F P, Andic A and Bloomfield D S 2007
Astron.Astrophys.
Astron. Astrophys.
The Observatory Soviet Astronomy Structure and Development of Solar Active Regions ( IAU Symposium vol 35) edKiepenheuer K O 293[52] Turon P J and L´ena P J 1970
Solar Phys. Solar Phys. Solar Phys.
Astron. Astrophys.
Astron. Astrophys.
Astron. Astrophys.
Solar Phys.
Magnetic Fields and Solar Processes ( ESASpecial Pub.
Solar Phys.
Solar Phys. Astron. Astrophys.
Pub. Astron. Soc.Japan The Observatory Astrophys. J.
Solar Phys. Astrophys. J. Suppl. Astrophys. J.
Astrophys. J. Suppl.
Astrophys. J.
Solar Phys. Austr. J. Physics Astron. Astrophys.
Astron. Astrophys.
L5[75] Dara H C, Alissandrakis C E, Zachariadis T G and Georgakilas A A 1997
Astron. Astrophys.
Solar Phys. Solar Phys. Solar Phys. Solar Phys. Astrophys. J. Lett.
L53[81] Matsumoto T, Kitai R, Shibata K, Otsuji K, Naruse T, Shiota D and Takasaki H 2008
Pub. Astron. Soc.Japan New Solar Physics with Solar-B Mission ( Astron. Soc. Pacific Conf. Series vol 369) ed Shibata K, NagataS and Sakurai T 113[83] Shibata K, Nakamura T, Matsumoto T, Otsuji K, Okamoto T J, Nishizuka N, Kawate T, Watanabe H,Nagata S, UeNo S, Kitai R, Nozawa S, Tsuneta S, Suematsu Y, Ichimoto K, Shimizu T, Katsukawa Y,Tarbell T D, Berger T E, Lites B W, Shine R A and Title A M 2007
Science
EFERENCES [84] Morita S, Shibata K, Ueno S, Ichimoto K, Kitai R and Otsuji K I 2010 Pub. Astron. Soc. Japan Astrophys. J.
Central European Astrophys. Bull. Astrophys. J. Lett.
L157[88] Fang C, Tang Y H, Xu Z, Ding M D and Chen P F 2006
Astrophys. J.
Solar Phys.
Pub.Astron. Soc. Japan Mem. Soc. Astron. Italiana Astrophys. J.
Astrophys. J.
Astrophys. J.
Astrophys. J.
Central European Astrophys. Bull. Astrophys. J. Lett.
L111[98] Kitai R and Muller R 1996
Solar Phys.
Radiative Transfer in Stellar Atmospheres
Lecture Notes Utrecht University[101] Kitai R 1983
Solar Phys. Mass Motions in Solar Flares and Related Phenomena (Procs 9th NobelSymposium) ed Oehman Y 109[103] Babin A N and Koval A N 1985
Solar Phys. Solar Phys.
Astron. Astrophys.
Astron. Astrophys.
Russian Phys. J. Astrophys. J. Lett.
L105[109] Kashapova L K 2002
Astron. Reports Astron. Astrophys.
Bulletin Crimean Astrophys. Obs.
Solar Phys. Nature
Astrophys. J.
Astron. Astrophys.
Solar Phys.
Solar Phys.
Astrophys. J.
Astrophys. J. Suppl. Astrophys. J.
Astrophys. J.
Astrophys. J. ( Astron. Soc. Pacific Conf. Series
EAS Publications Series
55) ed Faurobert M,Fang C and Corbard T 115[125] Strous L H, Scharmer G, Tarbell T D, Title A M and Zwaan C 1996
Astron. Astrophys.306