Magnetic and dynamical photospheric disturbances observed during an M3.2 solar flare

0.23 MB·English

by  C. Kuckein

Draftversion January20,2015
PreprinttypesetusingLATEXstyleemulateapjv.05/12/14
MAGNETIC AND DYNAMICAL PHOTOSPHERIC DISTURBANCES OBSERVED DURING AN M3.2 SOLAR
FLARE
C. Kuckein
Leibniz-Institutfu¨rAstrophysikPotsdam (AIP),AnderSternwarte16,14482,Potsdam,Germany
and
M. Collados and R. Manso Sainz
5 InstitutodeAstrof´ısicadeCanarias(IAC),V´ıaL´actea s/n,38205, LaLaguna,Tenerife,Spainand
Departamento deAstrof´ısica,UniversidaddeLaLaguna, 38206LaLaguna,Tenerife,Spain
1
Draft version January 20, 2015
0
2
ABSTRACT

n This letter reports on a set of full-Stokes spectropolarimetric observations in the near infrared Hei
a 10830 ˚A spectral region covering the pre-, flare, and post-flare phases of an M3.2 class solar flare.
J
The flareoriginatedon2013May17 andbelongedto activeregionNOAA11748. We detectedstrong
7 Hei 10830 ˚A emission in the flare. The red component of the He i triplet peaks at an intensity ratio
1
to the continuum of about 1.86. During the flare, Hei Stokes V is substantially larger and appears

reversed compared to the usually larger Sii Stokes V profile. The photospheric Sii inversions of the
]
R fourStokesprofilesrevealthefollowing: (1)themagneticfieldstrengthinthephotospheredecreasesor
isevenabsentduringtheflarephase,ascomparedtothepre-flarephase. However,thisdecreaseisnot
S
permanent. Aftertheflarethemagneticfieldrecoversitspre-flareconfigurationinashorttime(i.e.,in
.
h 30minutes afterthe flare). (2) Inthe photosphere,the line-of-sightvelocities showa regulargranular
p up- and down-flow pattern before the flare erupts. During the flare, upflows (blueshifts) dominate
o- the area where the flare is produced. Evaporation rates of ∼ 10−3 and ∼ 10−4 gcm−2s−1 have
r been derived in the deep and high photosphere, respectively, capable of increasing the chromospheric
t density by a factor of two in about 400 seconds.
s
a Keywords: Sun: chromosphere — Sun: flares — Sun: magnetic fields — Sun: photosphere — tech-
[ niques: polarimetric

1
v
1. INTRODUCTION the limb. In agreement with this, Hudson et al. (2008)
7
also predicted changes leading to more horizontal fields.
0 Understanding the magnetic field configuration in dy-
Recently, Wang et al. (2012)obtained first evidence of a
2 namic solar events such as flares is an unsolved problem
rapidly increasing horizontal magnetic field of ∼ 500 G
4 ofsolarphysics. Flaresarehard-to-predicteventsandare
alongthepolarityinversionline(PIL)inonly30minutes
0 the result of a globalchange in the magnetic field topol-
. ogy. These changes mainly happen in the chromosphere (startingatthebeginningoftheflare). Theauthorscon-
1 jecture that reconnection processes after the flare could
and corona. Therefore, full-Stokes polarimetric observa-
0
trigger the formation of low-lying horizontal fields.
5 tionsofflares,especiallyincludingthechromosphere,are
Here, we report on a set of full-Stokes spectropolari-
1 essential. They are, however, challenging. Many previ-
metric observations in the near infrared in a spectral in-
: ous studies have focused on the photospheric magnetic
v field underneath flare events. For instance, variations of terval covering the photospheric Sii 10827 ˚A line and
Xi the longitudinal magnetic field of the order of ∼ 100 G the chromospheric He i 10830 ˚A triplet. Using a scan-
wereobservedbyKosovichev&Zharkova(1999),andup ningspectrograph,severalmaps ofanactiveregionwere
ar to almost 300 G by Sudol & Harvey (2005) below X- obtained before, during and after the development of an
class flares. An enhancement of 400 G in the magnetic M3.2flare,makingitpossibletoderivethemagneticand
field strength was also detected by Kondrashova (2013) dynamicaldisturbances at photospheric layers. Interest-
during the onset of a microflare. ingly,somestudiesonflareshaverecentlybeenpublished
Itseemsclearthatchangeshappeninthephotospheric using data obtained in the same spectral region (e.g.,
magnetic field during flares. However, it is controversial Akimov et al. 2014; Judge et al. 2014; Sasso et al. 2014;
whetherthesevariationsareeitherassociatedwithanin- Zeng et al. 2014). With a data set with similar charac-
creaseoradecreaseofthemagneticfieldstrength. Petrie teristics, Judge et al. (2014)concentratedtheir study on
&Sudol(2010)studiedthephotosphericmagneticfieldin the seismic implications of the flare. These authors de-
77M-orX-classflares. Both,anincreaseanddecreaseof rivedanincreaseofthephotosphericmagneticfieldwhen
the magneticfield wasdetected indifferentflares. These comparing the maps before and after the flare. Unfor-
authors found absolute changes of up to ∼450 G in the tunately, they did not analyze the photospheric Stokes
longitudinalmagneticfieldandlinkedanincreaseofmag- spectra of their map taken during the flare. Sasso et al.
netic field strength to the presence of larger horizontal (2014) performed one scan of an active region after the
fields, as observed in their work concerning flares near onset of a flare and measured the Stokes profiles of the
Sii10827˚AlineandoftheHei10830˚Atriplet,deriving
themagneticstructureofafilamentpresentintheactive
[email protected]
2 Kuckein et al.
region. Other works (e.g., Akimov et al. 2014; Zeng et Inversions based on Response functions code (SIR; Ruiz
al. 2014) only used either He i 10830 ˚A chromospheric Cobo & del Toro Iniesta 1992). SIR solves the radia-
filtergrams or intensity spectra. In this paper, we com- tivetransferequationundertheassumptionoflocalther-
plement previous studies by analyzing the photospheric modynamicalequilibrium(LTE)andhydrostaticequilib-
magnetic and dynamic changes in the emission region rium. In this Letter, we focus on the inferred magnetic
before, during and after the flare. field strength and the LOS velocity patterns.
The original spatial resolution was preserved, with a
2. OBSERVATIONS pixelsizeof0.35×0.17arcsec2. The inversionswerecar-
ried out using the Harvard-Smithsonian reference atmo-
Observations of the active region (AR) NOAA 11748
spheremodel(HSRA; Gingerichetal.1971)asaninitial
were carriedout at the Vacuum Tower Telescope (VTT,
guess for the atmosphere. The model atmosphere covers
Tenerife,vonderLu¨he1998)on2013May17. Thesetup
included the Tenerife Infrared Polarimeter (TIP-II; Col- arangegivenasthelogarithmoftheLOScontinuumop-
lados et al. 2007) for full Stokes spectropolarimetry in ticaldepthτ at5000˚Aof1.4≤logτ ≤−4.0. Kuckeinet
the He i 10830 ˚A spectral region. al. (2012a) made an extensive analysis of inversion tests
using different initial model atmospheres. The output
NOAA 11748 sharply increased its activity towards
atmospheres were not much dependent on the starting
2013 May 13, when 14 flares, classes between C and
atmosphere,but aninitial magneticfield strengthof500
X, occurred. The number of flares and their activity
Gseemedtobeanoptimumchoice,sinceitreturnedthe
gradually decreasedin the next days. However,a strong
best fits while minimizing the number of iterations re-
M3.2 class flare was observed on 2013 May 17 while we
quired for the convergence. For this reason, we decided
were scanning with the slit along the PIL at coordinates
◦ ◦ to add this value to the initial HSRA guess model. The
(N 11 , E 36 ). The flare started at 8:43 UT, peaked at
stray-light profile in each map was computed averaging
around8:57UT, andendedat9:19UT.The firstspatial
the Stokes I profiles of pixels without polarization sig-
scan was carried out between 7:48 – 8:36 UT (pre-flare
nals. No more than four nodes in depth were necessary
phase), the second one between 8:36 – 9:06 UT (flare
for anyof the free parametersto achievegoodfits to the
phase), andthe third one between 9:06– 9:37UT (post-
four Stokes profiles.
flarephase),thus coveringthe wholeflareactivationand
relaxation process. An additional scan of the same area 4. RESULTS
wasperformedimmediatelyafterthethirdscan,between
4.1. Intensity and polarization profiles during the flare
9:38 – 10:08 UT.
The exposure time per slit position was 10 s and the The He i 10830˚A triplet originatesbetween the lower
scanning step 0′.′35. The pixel size along the slit was term 23S and the excited term 23P of orthohelium.
0′.′17. The Kiepenheuer-Institute Adaptive Optics Sys- It compr1ises three spectral lines w2h,1ic,0h are called the
tem (KAOS; Berkefeld et al. 2010) was locked on pho- “blue” component at 10829.09 ˚A (23S → 23P ), and
1 0
tospheric high-contrast structures, like pores and small the (blended) “red” component at 10830.30 ˚A (23S →
1
penumbrae, and was crucial to improve the fair-quality
23P ).
seeingconditions. TheobservedspectralrangewithTIP- 1,2
Photospherictemperatureenhancementsinflareshave
II spanned from 10824 to 10835 ˚A, with a spectral sam-
beenreportedby,e.g.,Chornogor&Kondrashova(2008)
pling of ∼11.1 m˚Apx−1. This spectralregioncomprises and Kondrashova (2013) using spectral lines in the visi-
thephotosphericSii10827˚Aline,thechromosphericHei ble. Likewise, Xu et al. (2004) detected a near-infrared
10830 ˚A triplet, and two telluric lines. intensity enhancement of the continuum inside an X10
white-light flare. During our M3.2 class flare, the Hei
3. DATAANALYSIS Stokes I profile showed a very strong emission, the red
Dark current, flat-field corrections and the standard component peaking at an intensity of ∼ 1.86, as seen
polarimetric calibration for the TIP-II instrument were in Figure 1. Though Hei 10830 ˚A emission against the
carried out (Collados 1999, 2003). The continuum was solar disk in flares has been previously seen by several
corrected for varying intensity due to changing air mass authors, such large values have not been reported yet.
andcenter-to-limbvariationsatdifferentpositionsonthe Previous studies showed intensity ratios to the contin-
solar disk while performing the scans. This was per- uum of ∼1.36 for a M2.0 flare (Li et al. 2007); .1.30 in
formed by second-order least-square polynomial fits and the decay phase of a C9.7 flare (Fig. 1; Penn & Kuhn
averages over quiet-Sun areas on the maps. All pixels 1995); .1.15 during a C2.0 flare (Fig. 2; Sasso et al.
alongthe slit within one scanstep were then normalized 2011). These works did not show any associated Hei
totheircorrespondingcontinuumvalue. Wavelengthcal- Stokes Q, U, and V profiles.
ibrationisbasedonthetwonearbytelluriclinesfollowing Figure 1 shows an illustrative example of the Stokes
theproceduredescribedinAppendicesAandBofKuck- profiles that corresponds to the maximum of the flare
ein et al. (2012b), which compares the wavelength sepa- (the exact position is marked by an arrow in Figure 2).
rationbetweentelluric linesina quiet-Sunareawiththe Remarkably, the amplitude of the circular polarization
Fourier Transform Spectrometer spectrum (FTS; Neckel Stokes V profile from Hei is larger than the one from
& Labs 1984) to obtain the spectral sampling. Finally, Sii. Furthermore, since the linear polarization signals
wavelengths were corrected for Earth’s orbital motions, (Stokes Q and U) in Si i are also very small, the mag-
solar rotation, and the solar gravity redshift (Martinez netic field is expected to be significantly stronger in the
Pilletet al.1997;Kuckeinetal. 2012b),thus the line-of- chromosphere than in the photosphere. Note that the
sight (LOS) velocities refer to an absolute scale. amplitude of the blue component of the He I Stokes V
ThephotosphericSiilinewasinvertedusingtheStokes profileissignificantlylargerthantheoneoftheredcom-
Magnetic and dynamical photospheric disturbances observed during an M3.2 solar flare 3
2.0 0.010
1.5 0.005
0.000
I 1.0 Q
−0.005
0.5
−0.010
0.010 0.04
0.005 0.02
0.000 0.00
U V
−0.005 −0.02
−0.010 −0.04
10826 10828 10830 10832 10834 10826 10828 10830 10832 10834
λ (Å) λ (Å)
Figure 1. ExampleofthefournormalizedStokesprofilesrecordedveryclosetothemaximumoftheflareat8:53UT.Theexactlocation
of the profiles is shown with an arrow in Figure 2. From left to right the followingspectral lines are seen: Sii line, Hei triplet, and two
telluriclines. TheverticaldashedlinesmarkthewavelengthatrestoftheSiilineandtheHeitriplet,respectively. Theaveragequiet-Sun
intensityspectrum isalsoplottedindashedlinesforcomparison.
ponent. Typically both profiles have similar amplitudes sented in the first row of Figure 2. These maps just
or even the blue component has a smaller amplitude. represent the part of the full scanned field of view cov-
In addition, while the V-profile of the blue component ered by the main part of the flare emission. From these
looks normal(i.e, rather antisymmetric), that of the red images,itcan be easily seenwhen the flare activates the
component seems distorted, especially its red lobe. This Hei emission (where I > 1). The second row shows the
might indicate that the Stokes V shown in the figure is photosphericmagneticfieldstrength,asderivedfromthe
a combination of at least two emission profiles. inversion of the Sii line. We will focus on the area in-
The observed reversal of the Stokes V profiles of Hei side the box which corresponds to the flare maximum,
with respect to Sii is due to the Hei spectral line ap- as indicated by the Hei emission. In the first column,
pearing in emission, whereas the Sii line appears in ab- i.e., before the flare started, the photospheric magnetic
sorption. Infact, the Sii line never goesinto emissionin field inside the box shows a patchy pattern of strong
ourdataseries,althoughitbecomesnoticeablyshallower fields. However, during the flare (second column), the
duringtheflare,asdemonstratedbythecomparisonwith field strength globally decreased being even absent in
theaveragequiet-SunspectrumplottedinFigure1. This some areas. Interestingly, in the post-flare panel (third
is the result of the heating of photospheric layers pro- column), the magnetic field has partially recovered its
duced by the flare. strengthandpattern. Hence,thereisclearevidencethat
The components of the Hei triplet appear redshifted the photospheric magnetic field changes during the flare
with respect to their wavelengthpositions at rest (verti- and afterwards tries to recover its initial configuration.
cal dotted lines in Figure 1), suggestingthat at the flare A further inspection of the Stokes profiles reveals that
peak, the plasma is slightly moving downwards towards the larger magnetic field strength in the pre- and post-
the photosphere. Two peaks in Stokes I can be distin- flare maps is mainly due to the larger amplitude of the
guishedintheusuallyblendedredcomponentofHei,one StokesV profiles. Thesecondmapaftertheflare(fourth
ofthemisslightlyredshifted. Duringtheflare,manypix- column) shows that the magnetic field is again losing
els show clear linear polarization signals in both the red strength and its pattern looks more like the one which
and blue components of He i 10830 ˚A (e.g., Stokes U in corresponded to the flare. These fluctuations may indi-
Figure 1), which cannotbe easily interpretedas Zeeman catethatthemagneticfieldmaybeundergoingtemporal
or scattering signatures. These anomalous polarization variations with a long relaxation time of hours.
patterns will be discussed in a forthcoming paper. As for the Doppler velocities presented in the last row
of Figure 2, a granular pattern, i.e., a mix of up and
4.2. Photospheric magnetic field suppression during the downflows is seen in the pre-flare panel. In response to
different flare phases the flare,the secondpanelshowsa largeareaofupflows.
Hence, photospheric material is predominantly pushed
The Hei emission pattern is used here as an indicator
upwardswheretheflareoccurs. Thefirstpost-flarepanel
of the flare eruption. To study the underlying photo-
still shows mainly upflows but a granular pattern is al-
sphere, we concentrate on the Sii line inversions. This
ready distinguishable. In the last panel the granulation
line is sensitive to a range of various heights within the
pattern is re-established.
photosphere. However, we will focus on a layer corre-
spondingtothegranularheightwhichiswellrepresented
by logτ =−1. 5. DISCUSSION
Slit-reconstructed monochromatic intensity images Wehaveshownaveryremarkableandunusualdataset
centered at the red component of the Hei line are pre- including pre-, flare, and post-flare slit scans in the in-
4 Kuckein et al.
7:49−8:03 UT 7:49−8:03 UT 8:51−9:05 UT 9:22−9:36 UT 9:53−10:08 U T
1.0
40 0.9
y (arcsec) 2300 000...678He I Stokes I
10
0.5
0 0.4
0 10 20
x (arcsec) 2000
40
1500
30
G)
20 1000B (
10 500
c)
se 0 0
c
ar −2
y ( 40
−1
30
0 −1m s)
20 v (k
1
10
0 2
0 10 20 0 10 20 0 10 20 0 10 20
x (arcsec)
Figure 2. Theisolatedupper lefthandframeshows acontinuum slit-reconstructed imageofthe firstmap(pre-flare). Next, fromtop to
bottom: Heimonochromaticred-componentslit-reconstructedimagessaturatedatanintensityofI=1(lineemission);totalphotospheric
magneticfieldstrengthB(inGauss)andLOSDopplervelocitiesinferredfromtheSiiinversions(clippedbetween±2kms−1). Fromleftto
right: thepre-flare,flare,andtwopost-flarephasesareshownwiththescanningtimesatthetop. Theboxesoutlinetheregion-of-interest
wheretheflareisseen,astracedbytheHeiemissionseeninthe8:51–9:05UTpanel. ThearrowmarksthepositionoftheStokesprofiles
showninFigure1.
There are roughly 30 minutes between the flare and the
Table 1
MeanDopplervelocitieshviandmassfluxhρvi, atthelayerof post-flareobservations. As seeninthe colorscaleinFig-
logτ =−1,insidetheboxofFigure2. ure 2, the magnetic field goes through changes of up to
1500 G in 30 minutes. These changes mainly happen in
Time Phase hvi hρvi
(h:mUT) (kms−1) (gcm−2 s−1) the box outlined in Figure 2, which lies between the up-
7:49–8:03 pre-flare −0.16±0.02 −2.2×10−3 per two pores and the lower PIL. The maximum of the
8:51–9:05 flare +0.11±0.01 +1.5×10−3 flare is neither located at the sunspot nor at the PIL of
9:22–9:36 post-flare +0.07±0.01 +1.1×10−3 the active region. In our data sets, it was found that 90
9:53–10:08 post-flare −0.08±0.01 −1.0×10−3 minutesseemedtobe anupperlimitforthetime needed
Note. —Thereare5368pixelsineachbox. Negative(positive) bythe magnetic fieldstrengthto recovermostofits pre-
valuesfordownflows(upflows) alongtheLOS. flare pattern and strength. However, the variations of
the magnetic field are not permanent. There is a strong
fraredHei 10830˚A spectralregionof anM3.2 class flare decreaseduring the flare,withareaswhere the magnetic
fieldwaspresentbeforetheflarebutiscompletelyabsent
withthespectropolarimeterTIP-II.Ourobservationsre-
vealcomplex StokesQ, U, andV profiles ofthe chromo- during the evolution of this phenomenon.
spheric Hei 10830 triplet during the flare. Furthermore, Table1 showssomeparametersinferredfromthe pho-
tospheric inversions in the flaring part of the region
unprecedented largeintensity ratios to the continuum of
the Hei red component where shown (I ∼ 1.86). How- (represented by the box of Figure 2), at the layer at
ever, the Hei emission is not associated with the whole logτ =−1. Beforetheflareoccurs,adownflowofaround
160 ms−1 is present in the average granular velocity.
areacoveredbytheflare,asalreadydetectedbyprevious
studies(Du &Li2008). Fromthe Siiinversionswehave During the flare and immediately after it, the granular
pattern appears very distorted and barely perceptible,
inferred the magnetic fields and the LOS velocities in
with an average upflow of some 100 ms−1. Later, in
thephotosphereforthepre-,flare,andpost-flarephases.
Magnetic and dynamical photospheric disturbances observed during an M3.2 solar flare 5
Onehourafterthe flare,mattercondensesagaindown
0.0020 to the photosphere. The magnetic field maps indicate
that while the flare is taking place, most of the pho-
tospheric magnetic field is below detectable limits. One
0.0015
−2−1m s) mcthraeeaymsseapogefnctuehtlaeictgesattsrhupacrtteustshrueesrhea,enatdhtiurnesgdpuprcrooidncugecstisnhhgeiarasnficeealxudpssaetdnrseainongntinho-f.
c 0.0010
g As the material evaporates, the gas pressure decreases
> ( gradually and the magnetic field tends to concentrate
v
ρ
< 0.0005 again trying to recover its original strength. However,
the mass downflow observed during the last map may
tend to increase again the gas pressure, leading again to
0.0000 an expansion of the magnetic field lines. This specula-
tive scenariomay explain at the same time the observed
−4 −3 −2 −1 0
log τ variations of the magnetic field strength and velocities
during the pre-, flare, and post-flare phases.
Figure 3. Stratification with logarithm of the optical depth
(logτ) of the average evaporation rate hρvi in the flaring region
indicatedbytherectangularboxofthesecondcolumninFigure2. The Vacuum Tower Telescope is operated by the
Kiepenheuer-InstituteforSolarPhysicsinFreiburg,Ger-
many, at the Spanish Observatorio del Teide, Tenerife,
Canary Islands. This project was supported in part by
thelastmap,thegranulationpatternisalmostrecovered
withanaveragedownflowslightlybelow100ms−1. This grant DE 787/3-1 of the German Science Foundation
(DFG). The authors would like to thank C. Denker and
evolution is compatible with a scenario where, once the
E. Khomenko for carefully reading the manuscript.
flare has started, the deep photospheric material heats
up and evaporates at a rate of ∼10−3 gcm−2s−1.
REFERENCES
Figure 3 shows the average dependence with depth of
the evaporation rate of the photosphere in the flaring
Akimov,L.A.,Belkina,I.L.,&Marchenko,G.P.2014,MNRAS,
region. From this stratification, the density change in 439,193
the chromosphere can be estimated from the continuity Berkefeld,T.,Soltau,D.,Schmidt,D.,&vonderLu¨he,O.2010,
equation: Appl.Opt.,49,G155
Chornogor,S.N.,&Kondrashova,N.N.2008,Sol.Phys.,250,
303
∂ρ d(ρv)
=− , (1) Collados,M.1999,inASPConf.Ser.184,ThirdAdvances in
∂t dz SolarPhysicsEuroconference: MagneticFieldsand
Oscillations,ed.B.Schmieder,A.Hofmann,&J.Staude, 3
where we have assumed that the material motion is es- Collados,M.,Lagg,A.,D´ıazGarc´ıa,J.,etal.2007,inASPConf.
sentiallyverticalandthemeasuredLOSvelocity,v,does Ser.368,ThePhysicsofChromosphericPlasmas,ed.P.
Heinzel,I.Dorotoviˇc,&R.J.Rutten(Coimbra,
notdiffersignificantlyfromthetrueverticalvelocity. In-
Portugal:ASP),611
tegrating equation (1) in height from the temperature
Collados,M.V.2003,inProc.SPIE4843, ed.S.Fineschi,55
minimum(zmin)upto the topofthe atmosphere(zmax), Du,Q.-S.,&Li,H.2008,ChineseJ.Astron.Astrophys.,8,723
one gets Gingerich,O.,Noyes,R.W.,Kalkofen,W.,&Cuny,Y.1971,
Sol.Phys.,18,347
∂σ Hudson,H.S.,Fisher,G.H.,&Welsch,B.T.2008,inASPConf.
=(ρv) −(ρv) , (2) Ser.383,SubsurfaceandAtmosphericInfluences onSolar
∂t zmin zmax Activity,ed.R.Howe,R.W.Komm,K.S.Balasubramaniam,
&G.J.D.Petrie(Sunspot, NM:ASP),221
where σ = Rzmaxρdz is the column mass density of the Kondrashova,N.N.2013,MNRAS,431,1417
zmin Kosovichev,A.G.,&Zharkova,V.V.1999,Sol.Phys.,190,459
chromosphere. Furtherassumingthatthe rate ofchange
Judge,P.G.,Kleint,L.,Donea,A.,SainzDalda,A.,&Fletcher,
of the density is constant during the flare and that no
L.2014, ApJ,796,85
massescapesaway((ρv)zmax =0),thenthetime interval Kuckein,C.,Mart´ınezPillet,V.,&Centeno,R.2012a,A&A,539,
required to produce a change ∆σ in the column density A131
is Kuckein,C.,Mart´ınezPillet,V.,&Centeno, R.2012b,A&A,
542,A112
Li,H.,You,J.,Yu,X.,&Du,Q.2007,Sol.Phys.,241,301
∆σ
∆t= . (3) MartinezPillet,V.,Lites,B.W.,&Skumanich,A.1997,ApJ,
(ρv) 474,810
zmin Neckel,H.,&Labs,D.1984,Sol.Phys.,90,205
We can evaluate (ρv) as the mass flux at the top of Penn,M.J.,&Kuhn,J.R.1995,ApJ,441,L51
our photospheric invezrmsiionns (i.e., ∼10−4 gcm−2s−1; see Petrie,G.J.D.,&Sudol,J.J.2010,ApJ,724,1218
RuizCobo,B.,&delToroIniesta, J.C1992,ApJ,398,375
Figure 3). For the aforementioned HSRA model atmo- Sasso,C.,Lagg,A.,&Solanki,S.K.2011,A&A,526,42
sphere, the chromospheric column density (i.e, integrat- Sasso,C.,Lagg,A.,&Solanki,S.K.2014,A&A,561,A98
ing the density from the temperature minimum up to Sudol,J.J.,&Harvey,J.W.2005,ApJ,635,647
z = 1850 km) is σ ∼ 0.042 gcm−2. Therefore, ac- vonderLu¨he,O.1998, NewARev.,42,493
max Wang,S.,Liu,C.,Liu,R.,etal.2012,ApJ,745,L17
cording to equation (3), the chromospheric density may
Xu,Y.,Cao,W.,Liu,C.,etal.2004, ApJ,607,L131
increase by a factor of two (∆σ = σ) in ∆t ∼ 400 sec- Zeng,Z.,Qiu,J.,Cao,W.,&Judge,P.G.2014,ApJ,793,87
onds.