The solar chromosphere at high resolution with IBIS. I. New insights from the Ca II 854.2 nm line

0.67 MB·English

by  G. Cauzzi

Astronomy&Astrophysicsmanuscriptno.ibis.caii.astroph (cid:13)c ESO2008
February1,2008
The solar chromosphere at high resolution with IBIS
I. New insights from the CaII 854.2 nm line
G.Cauzzi1,K.P.Reardon1,H.Uitenbroek2,F.Cavallini1,A.Falchi1,R.Falciani3,K.Janssen1,T.Rimmele2,A.
Vecchio1,andF.Wo¨ger2,4
8 1 INAF-OsservatorioAstrofisicodiArcetri,I-50125Firenze,Italy
0 2 NationalSolarObservatory,P.O.Box62,SunspotNM88349,USA
0 3 DipartimentodiAstronomia,Universita`diFirenze,I-50125Firenze,Italy
2 4 KipenheuerInstitutefu¨rSonnenphysics,D-79104Freiburg,Germany

n
February1,2008
a
J
ABSTRACT

3
Context.Thechromosphereremainsapoorlyunderstoodpartofthesolaratmosphere,ascurrentmodelingandobservingcapabilities

] arestillill-suitedtoinvestigateindepthitsfully3-dimensional nature.Inparticular,chromospheric observationsthatcanpreserve
h
highspatialandtemporalresolutionwhileprovidingspectralinformationoverextendedfieldsofviewarestillveryscarce.
p
Aims.Inthispaper,weseektoestablishthesuitabilityofimagingspectroscopyperformedintheCaII854.2nmlineasameansto
-
o investigatethesolarchromosphereathighresolution.
r Methods.WeutilizemonochromaticimagesobtainedwiththeInterferometricBIdimensionalSpectrometer(IBIS)atmultiplewave-
t lengths within the CaII 854.2 nm line and over several quiet areas. We analyze both the morphological properties derived from
s
a narrow-band monochromatic imagesandtheaveragespectral propertiesofdistinctsolarfeaturessuchasnetworkpoints,internet-
[ workareasandfibrils.
Results.Thespectralpropertiesderivedoverquiet-Suntargetsareinfullagreementwithearlierresultsobtainedwithfixed-slitspec-
2 trographicobservations,highlightingthereliabilityofthespectralinformationobtainedwithIBIS.Furthermore,theverynarrowband
v IBISimagingrevealswithmuchclaritythedualnatureoftheCaII854.2nmline:itsouterwingsgraduallysamplethesolarphoto-
7 sphere,whilethecoreisapurelychromosphericindicator.Thelatterdisplaysawealthoffinestructuresincludingbrightpoints,akin
1 totheCaIIH andK grains,aswellasfibrilsoriginatingfromeventhesmallestmagneticelements.Thefibrilsoccupyalarge
2V 2V
4 fractionof theobserved fieldof view eveninthequiet regions, andclearlyoutlineatmospheric volumes withdifferent dynamical
2 properties,stronglydependent onthelocalmagnetictopology.Thishighlightsthefactthat1-Dmodelsstratifiedalongthevertical
. directioncanprovideonlyaverylimitedrepresentationoftheactualchromosphericphysics.
9
Conclusions.ImagingspectroscopyintheCaII854.2nmlinecurrentlyrepresentsoneofthebestobservationaltoolstoinvestigate
0
thehighlystructuredandhighlydynamical chromospheric environment. Ahighperformanceinstrument suchasIBISiscrucialin
7
ordertoachievethenecessaryspectralpurityandstability,spatialresolution,andtemporalcadence.
0
v: Keywords.Sun:chromosphere—Sun:magneticfields—Instrumentation:highangularresolution—Instrumentation:interferom-
eters
i
X
r
a
1. Introduction Amongthechromosphericdiagnosticsaccessibletoground-
based observations, the CaII H and K resonance lines
have been used most extensively (see e.g. the review of
The chromosphereembodies the transition between the photo- Rutten&Uitenbroek1991).Theselinesarethebroadestlinesin
sphereandthecorona,tworegionsdominatedbyvastlydifferent thevisiblespectrum,samplingalargerangeinformationheight,
physicalregimes.Inparticular,itiswithinthechromospherethat andarethe onlyvisiblelinesthatprovidea directindicationof
the plasma β, the ratio of plasma kinetic pressure to magnetic thechromospherictemperaturerise withtheirH2 andK2 emis-
pressure,fallsbelowunity,signalingashiftfromhydrodynamic sionreversals.Becausethelinesoriginatefromthegroundstate
tomagneticforcesasthedominantagentinthestructuringofthe ofadominantionizationstage,theyaremostlycollisionallycon-
atmosphere.AsablydescribedinthereviewofJudge(2006),the trolled in the lower chromosphere, and sensitive to local tem-
combinedeffectsofmagneticfieldguidanceandsmallscalegas peraturetoamuchlargerdegreethanforinstancethehydrogen
thermodynamicsleadtotheimpressiveamountoffinestructures Balmer lines. Of these, the Hα line in particular has also been
(jets, spicules, fibrils, mottles, etc.) that uniquely characterize widely exploited, especially following the development of the
thispartofthesolaratmosphere.Suchfinestructurerepresents Lyotfilter (Lyot1933),whichgaverise tothe scienceof‘solar
a formidable challenge even to the most modern instrumenta- cinematography’,i.e.narrow-bandimagingobtainedatrapidca-
tion, as its study requireshighspectral resolution,necessary to dence.EventhoughHαspectraaredifficulttointerpretbecause
resolve line profilesencodinglarge gradientsor discontinuities ofthecomplicatedformationcharacteristicsofthisline,muchof
(e.g. the shocks of Carlsson&Stein 1997), combined with ex- whatweknowaboutthechromospherederivesfromthedetailed
tremelyhightemporalandspatialresolution(seee.g.therecent HαmorphologyobservedthroughLyot-stylefilters.
observationsbyvanNoort&RouppevanderVoort2006).
2 G.Cauzzietal.:ThesolarchromosphereathighresolutionwithIBIS
A somewhat neglected, but no less interesting, chromo- IFU proper, IBIS is a high performance instrument that com-
sphericdiagnosticsisrepresentedbytheCaIIinfraredtriplet(λ bines most of the advantages of a full spectroscopic analysis,
=849.8,854.2,866.2nm,hereafterCaIIIRT).Theselinesorig- usually obtained with single-slit spectrographs, with the high
inate fromtransitionsbetweenthe upper4p 2P levelsand spatialresolution,hightemporalcadenceandlargefieldofview
1/2,3/2
thelowermetastable3d2D levels.Transitionsbetweenthe typicaloffilterinstruments.Suchcharacteristicsarenecessaryto
3/2,5/2
groundstate4s2S ofCaIIandthesameupperlevelsgiverise obtainnewinsightsintothestructureanddynamicsofthechro-
1/2
totheHandKlines.Sincethebranchingratio,theratiobetween mosphere (see for example Vecchioetal. 2007b; Cauzzietal.
the spontaneousemission coefficientsof theH andK linesand 2007).AlthoughotherinstrumentssimilartoIBISarecurrently
those of the IRT lines, is about A /A ≡ 15, most photons operative, such as TESOS (Tritschleretal. 2002) and the new
HK IRT
emitted in the IRT result from excitation in the H and K lines, Go¨ttingenFabry-Perotsystem(Puschmannetal.2006),noother
leadingtoaverysimilartemperaturesensitivityforbothsetsof imagingspectrometerispresentlyabletoaccessthewavelengths
lines.Inaddition,the 3d2D aremetastable(i.e.,thereare oftheCaIIIRT.
3/2,5/2
no allowed electric dipole transitions with the 4s 2S ground
1/2
state),sothattheselevelscanonlybepopulatedfrombelowby
collisionalexcitation,strengtheningthesensitivityoftheIRTto 2. Instrumentalcharacteristics
localtemperatureevenmore.
First detected in the 1878 eclipse spectrum by C. Young
(Eddy 1973), the CaII IRT has remained essentially ignored
in solar observational work for nearly a century. From the
1960’sand up throughthe mid-1980’sit was occasionally em-
ployed,inparticularinmulti-linespectrographicstudies(among
others, Pasachoffetal. 1968; Linskyetal. 1970; Mein 1971;
Shine&Linsky 1972; Beckersetal. 1972; Shine&Linsky
1974;Lites1984).SpurredbytheavailabilityofCCDdetectors
with high sensitivity in this wavelength range, the last decade
hasinstead seen a rapid growthof observationalstudiesadopt-
ing these lines both for solar and cool star research; we refer
the reader to Socas-Navarroetal. (2006); Uitenbroek (2006b);
Tziotziouetal. (2006); Uitenbroeketal. (2006); Pietarilaetal.
(2007b)forrecentexamplesofsolarstudies,andtoChmielewski
(2000);Andrettaetal.(2005)forthestellarcase.
Despite the fact that near infrared wavelengths afford both
a higher photon flux and a reduced terrestrial atmospheric dis-
turbance (with respect to the wavelengths of H and K lines),
high resolution solar observations of the CaII IRT are still
scarce. The great majority of the previous observations were
in fact conductedwith fixed-slitspectrographs,either perform-
ing an area raster scan, requiring long repetition times and
thus compromising the temporal resolution (Flecketal. 1994;
Uitenbroek2006b; Pietarilaetal. 2007b), or keepingthe slit at
a fixed position (Deubner&Fleck 1990; Pietarilaetal. 2007b;
Langangenetal. 2007). The latter strategy allows for a high
temporalcadence, but makes it difficult to precisely follow the
small chromospheric structures, often of magnetic origin, that
are movedaway fromthe slit either by atmosphericturbulence
orsolarevolution.Amoreefficientobservationalstrategyispro-
vided by the Multichannel Subtractive Double Pass technique
(MSDP,Mein1991,2002).TheMSDPisinfactoneofthefew
examplesof anIntegralField Unit(IFU)devicebeingused for
opticalsolarobservations,withtheuniquecapabilityofacquir-
ing truly simultaneous spectra over an extended field of view.
However, for the case of the CaII 854.2 nm line, observations
are usually limited to a spectral range of about ± 40 pm from
line core (e.g. Tziotziou et al 2002), thus possibly neglecting Fig.1.Top panel:the solid line displaysthe CaII 854.2nm at-
importantinformationrelatedtostrongvariationsencodedinthe las profile, while the dashed line shows the effects of the 0.46
extendedlineprofile.Moreover,theusablefieldofview(FOV) nm FWHM prefilter utilized in IBIS observations. The actual
representsatrade-offwiththespectralcoverageandresolution, maximumtransmissionoftheprefilteris35%.Theasterisksin-
andisgenerallylimitedtoonlyafewarcsecinoneofthespatial dicate the wavelengths of the images in Fig. 2 and 3. Bottom
directions. panel:normalizedprefilterprofile(solid,uppercurve)andperi-
In this framework, we introduce here new high resolution odic transmission profile of the combinedFabry Perot’s (solid,
observations of the CaII 854.2 nm line obtained in a vari- multiple peaks curve). Note the logarithmic scale. The dashed
ety of solar structures with the Interferometric BIdimensional lineindicatestheeffectsoftheprefilterontheFPstransmission
Spectrometer(IBIS,Cavallini2006),installedattheDunnSolar profile,i.e.thelargesuppressionofthesecondarymaximaofthe
Telescope of the US National Solar Observatory.While not an curve,necessarytoreducethespectralparasiticlight.
G.Cauzzietal.:ThesolarchromosphereathighresolutionwithIBIS 3
Fig.2.CaII854.2nmdataacquiredonMay31,2004.Wavelengthrunsfromlefttoright,toptobottom,asindicatedinthepanels,
withoffsetsinpmfromnominallinecore.Thefigureshowsacutoutof64′′×41′′ofthefullcircularFOV,originally80”indiameter.
Eachpanelwasnormalizedforoptimumdisplay.Thetopmostleftpanelshowsthecospatial,cotemporalbroadbandimage(acquired
at710nm,FWHM=10nm).ThebottomrightpanelgivesthecotemporalHRMDImap,scaledbetween±500G.Theintensityin
thefarwingsoftheCaII854.2nmlineisstronglysensitivetothepresenceofmagneticstructures(cf.Leenaartsetal.2006, and
Sect.5.2)
.
IBISisatunablenarrowbandfilter,whosemaincomponents the filter to be tilted in the beam to avoid reflections. This also
are two air-spaced, 50 mm diameter Fabry-Perot interferom- permits the passband to be minimally tuned in a range of ap-
eters (Cavallini 2006; Reardon&Cavallini 2007). IBIS oper- proximately ±0.1nm from the central wavelength. We remark
ates in the spectral range 580–860nm, and can provide quasi- alsothatinthecaseofbroadlinessuchastheCaIIIRTthelim-
monochromaticimagesatanywavelengthinthatrangebysuit- ited prefilter passband effectively prevents the sampling of the
abletuningoftheinterferometers.However,duetotheperiodic lineallthewaytothecontinuum,asshownintheupperpanelof
nature of the spectral transmission profile, the analysis of any Fig.1.
givenspectrallinerequirestheuseofaprefilter,of0.3–0.5nm
TheinstrumentalspectraltransmissionofIBIShasbeenac-
FWHM,thatisolatesthecentraltransmissionorder.Currentlya
curately calculated and the FWHM at 854 nm was determined
prefilter for the CaII 854.2 nm line of the IRT is available for
tobe4.4pm(Reardon&Cavallini2007).Thespectralpurityof
IBIS.ThefilterwasprovidedbyBarrAssociates,withaFWHM
theinstrumentalprofileisquitehigh,withover95%ofthetotal
of0.46nm,35%transmission,centeredat854.25nm.Thecen-
transmissioncomingfromwithinarangeof±8pmaroundthe
tralwavelengthofthefilterisslightlyoffsettotheredfromthe
peak ofthe profileand spectralparasitic light, arising fromthe
nominalcentralwavelengthofthe CaII854.2nmlineto allow
repeating secondary orders of the transmission profile, of only
4 G.Cauzzietal.:ThesolarchromosphereathighresolutionwithIBIS
Fig.3.SameasFig.2,forJune2,2004data.Inthespeckle-reconstructedbroadbandimagemanybrightpointscanbeappreciatedin
thenetworkareas.TheMDImapshowsastrong,bipolarnetwork(imagescaledbetween±500G).Notehowthefibrillarstructures
originatingfromthenetworkelements,andwellvisibleintheinnerwingsandcoreimages,extendoverasubstantialfractionofthe
FOV,differentlytowhatoccursintheareadepictedinFig.2.Thetwodashedlinesintheupper-rightpanelindicatethepositions
wheretheslitspectraofFig.6havebeenobtained.
1.5%.ThelowerpanelofFig.1showshowthisisaccomplished locationofthestructuresremainstableduringthespectralscan-
with the combination of two interferometers and the prefilter, ning of the line. The high-order adaptive optics system of the
whichstronglysuppressesthesecondarymaximain theFabry- DST (Rimmele 2004) is routinely used with IBIS and greatly
Perotstransmission. aidsinstabilizingtheimage.Toremoveresidualimagemotion,
Piezoelectric tuning allows the instrumental profile to be especiallyatincreasingdistancesfromtheAOlockpoint,weob-
positioned at multiple wavelengths within a given line, mak- tainabroadband(“whitelight”)referenceimagestrictlysimul-
ingitpossible toacquirefullspectralinformationoverthe 80′′ taneouslywitheachnarrowbandimage.Thesereferenceimages
diameter circular field of view (FOV) of the instrument. The are obtained through 10 nm wide filters, and during the rela-
CaII854.2nmlineistypicallysampledat15–30wavelengthpo- tivelyshorttimeneededforaspectralscanshowessentiallythe
sitions,dependingonthescientificrequirements(seefollowing samescene(ascomparedtothespectralimageswherethestruc-
sections). Given the currentrate of acquisition (2.5 – 4 frames tures change dramatically during the scan). They are then nor-
s−1), this translates in an interval of 5 – 12 s to perform a full mallyusedtodestretchtheimagesinbothchannels,orformore
samplingoftheline.Aplannedupgradeforthecamerasystem advanced image reconstruction techniques, such as deconvolu-
inthenearfuturemayimprovethisratebyafactorof2–4. tionofthenarrowbandchannelimagesupportedbyMOMFBD
Inordertoobtainmeaningfulspectralinformationfromse- (vanNoortetal.2005)orspeckleimaging(Wo¨ger2006).
quentialimagesathighspatialresolution,itisnecessarythatthe
G.Cauzzietal.:ThesolarchromosphereathighresolutionwithIBIS 5
Finally, the spatial scale of the IBIS images is set at
0”.082/pixel. With the use of the procedures described above,
itisthuspossibletoexploittheperiodsofgoodtoexcellentsee-
ing occurringat the DST up to the nominaldiffractionlimit of
thetelescopeofλ/D=0”.23at854.2nm.
3. Observations
WepresentinthispapersomeexampleCaII854.2nmdataac-
quiredmostly in quiettargetareas, as listed in Table 1. We re-
markhoweverthatIBISobservationsinthisspectrallinearewell
suitedalsoforstudiesofsolaractivity.Ontheonehand,the80”
diameter FOV is able to accommodate large portions of active
regions(orevenfullsmallones),thusovercomingatypicalprob-
lem of spectrographicobservations.On the other,the ability to
obtainspectralprofilesinjustafewseconds(tenorless)allows
adetailedstudyofthedynamicsofrapidimpulsiveevents,such
as reconnection driven explosive events or small flares, occur-
ringinthesolarchromosphere.
Thesequenceofmonochromatic1imagesofFigure2shows
a scan acquired on May 31, 2004. The actual wavelengths of
observationare indicated in the panelsas offsets fromnominal
linecore.Thetargetwasaquietareaatdiskcenter,withmostly
unipolar,weaknetworkelementswithintheFOV.Thewholere-
gionwaspositionedattheedgeofanequatorialcoronalhole,as
seenfromEIT171Åfulldiskimages.Thefirstpanelshowsthe
corresponding white light image, while the last one displays a
cotemporal,highresolutionMDI mapofthe longitudinalmag-
neticflux. Fig.4. Speckle-reconstructed sample images from the weak
Fig. 3 shows the data acquired on June 02, 2004, with the plageregionofOct.1,2005.Upperpanel:imageobtainedat65
same setupasMay31,2004.Thetargetwas againa quietarea pminthebluewingoftheline.FOVisabout60”×40”.Bottom
atdiskcenter,thistimeroughlyencompassingafullsupergran- panel:correspondingimageat20pminthebluewing,obtained
ulesurroundedbysomeenhanced,bipolarnetwork.Thewhole 200slater.Imageshavebeenreconstructedapplyingthespeckle
regionappearedasafadingcoronalbrightpointintheEIT171 code of Wo¨ger (2006). See the online version for a movie of
thesedata.
Å full disk images, where occasional bouts of small scale ac-
tivitywereobserved.Numerousbrightpointsaswellasseveral
rapidlyevolvingsmallporescanbedistinguishedinthe broad-
4. Lineformation
bandimageintheregionsofthenetwork,testifyingtoastronger
magneticfluxasseenintheMDIpanel.Thisdatasethasbeenan- Themoststrikingcharacteristicsofthenarrowbandsequencesof
alyzedinVecchioetal.(2007b)andinpartinJanssen&Cauzzi Figs.2–4isthedramaticchangeinscenerywhentheinstrumen-
(2006). talpassbandmovesfromthewingstowardsthecoreoftheline
Fig.4showsthetargetregionofOct.01,2005,encompass- -thelatterdisplayingswarmsoffibrillarstructuresusuallyasso-
ingasmallpore(3”–4”diameter)withinaregionofweakplage ciatedonlywithHαcorediagnostics.The“criticalwavelength”,
neardiskcenter.Thisdatasetwasacquiredtotesttheprocedure at about30–40pm fromthe core,correspondsto the inflection
forperformingpost-factoimagereconstructionusingthespeckle pointin the line wherethe steep innerwingsleaveplace to the
code of Wo¨ger (2006), hence 50 repeated exposures were ob- muchshallowerouterwings(seetheprofilesinFig.1).
tainedateachwavelengthpoint,resultinginalongtotalacqui- Thesharpkneesin thelineprofilemarkthetransitionfrom
sition time. The top panel displays the reconstructed image at LTEabsorptionlineinthephotosphericouterwingstotheNon-
about 60 pm in the blue wing, while the bottom one gives the LTEchromosphericabsorptionlinecore.Thistransitionisare-
image at 20 pm, again in the blue wing. A movie displaying a flection of two effects, in the line source function and in the
cotemporalG-bandframe,togetherwith thetuningofthe IBIS line opacity, respectively. In the photosphere, where LTE ap-
passbandthroughthebluewingofthelineisavailableintheon- plies, the line source function decreases with height with the
lineversionofthispaper2.Thisdataclearlyillustratesthewealth electrontemperature,givingrisetoaslowdecreaseinlinewing
offinestructureinformationtheCaIIIRTcanprovidewhenob- intensity towardsline center. Indeed,because of their LTE for-
servedathighresolution(seealsoFig.11inRutten2007). mation,thewingsofthe854.2nmlineareusedasaccuratetem-
peraturediagnosticsinsolartypestars(e.g.,Linskyetal.1979;
Smith&Drake 1987). In the chromospheric layers collisional
1 Due to the classical mounting of the interferometers, the spectral
excitation and de-excitation loses dominance to radiative tran-
passbandofIBISexperiencesaradial,wavelength-dependent,blueshift
withintheFOV.Theeffecthasbeenremovedfromalltheimagesdis- sitions as the main population mechanisms in the line. Photon
playedinthispaper. losses to outer space propagate inward because of the scatter-
2 Movie 1, http://www.arcetri.astro.it/science/solar/IBIS/movies/ ing in the line thatresults. With these losses the radiation field
CaII.8542.scan.01Oct2005.movie.gif inthelinecannolongermaintaintheupperlevelpopulationof
6 G.Cauzzietal.:ThesolarchromosphereathighresolutionwithIBIS
Table1.SummaryoftheCaII854.2nmobservationsdescribedinthepaper.Dateisexpressedinyyyymmdd;∆xisthepixelspatial
scale;t istheexposuretime;δλisthespectralsamplingstep,while∆λisthetotalsampledrange;Nisthenumberofwavelength
exp
points;δtisthescantime,while∆tisthecadenceoftheobservations.ForthecaseoftheCaII854.2nmline,oftenthesamplingis
coarserintheextendedwingswithrespecttothecore.
Date Target Location ∆x(”) t (ms) δλ(pm) ∆λ(nm) N δt(s) ∆t(s) Duration(min)
exp
20040531 quietSun diskcenter 0.16 25 8–16 [−0.12,0.16] 27 7 19 55
20040602 quietSun diskcenter 0.16 25 8–16 [−0.12,0.16] 27 7 19 55
20051001 smallpore diskcenter 0.082 50 4 [−0.07,0.13] 35 700 – –
thelinetoLTElevels,andthelinesourcefunctiondropssteeply The outer wing images thus mostly display convective struc-
withheight. tures,onlyoccasionally“contaminated”byfibrilsthatshowup
as thin dark strikes. The core images instead display mostly
fibrillar, chromospheric structures, that appear most evident in
the blue or red inner wings depending on their line width and
Dopplershift.OnlyinveryquietportionsoftheFOVthefibrils
arenotthepredominantfeature,andsporadicbrightpointswith
sizeoftheorderof1arcsecondbecomevisibleinthecore(see
Sect.5.3).
5. Morphology
5.1.Outerwings–quietphotosphere
As just discussed, the high spatial resolution images obtained
withIBISintheouterwingsoftheCaII854.2nmlinearedom-
inated by photospheric structures. In particular, the pattern of
reversegranulation,i.e.apartialreversalofthecontrastbetween
granules and intergranules with respect to the continuum (see,
e.g. Ruttenetal. 2004, and referencestherein), stands out very
clearlywithinthequietportionsoftheFOV.Thisisvisible,for
example,bycomparingthewhite lightimagesof Figs.2 and3
vs.thepanelsat−100pm(seealsoJanssen&Cauzzi2006),or,
Fig.5. Contribution function for CaII 8542 computed in
evenmoreclearly,inthefirstimagesofMovie1.Lookingcare-
plane-parallel, hydro-static average quiet-Sun atmosphere
fullyatthewingimagesofFigs.2and3,onecanalsonoticea
(Fontenlaetal. 1993). Height zero refers to level at which op-
slight asymmetrybetween the blue and the red wings, with re-
ticaldepthinthecontinuumat500nmisunity.
versed granulation being more prominent in the blue, at equal
distancefromlinecore.Thisisduetotheeffectofcross-talkbe-
Thesharptransitionbetweenthesetworegimesresultsfrom tweenintensityandvelocitiesintheconvectivestructures,much
agapinthelineopacityaroundthetemperatureminimum.Since asthecaseformonochromaticimagesacquiredinthewingsof
the lower levels of the IRT are the metastable 3d 2D lev- typicalphotosphericlines(Janssen&Cauzzi2006).
3/2,5/2
els,theirpopulationinlowtemperatureregionsisreducedcom- As one movesfurther out in the wings, reverse granulation
pared to the groundlevel population.This is a small effect be- graduallydisappears,andthequietareasshowjustadiffusein-
cause the excitation energyof the D levelsis only 1.5 eV (e.g. tensitywithoutanyobviousstructure(comparee.g.thelastpanel
muchsmallerthanthe10eVofthehydrogenHαlineforwhich ofFig.3,at160pmfromlinecore).Giventhebroadersampling
the gap in line opacity is therefore much more pronounced). of the line in the red (cf. Table 1), this effect is mostly visible
Nevertheless,becauseofthisgapin line opacity,the formation in the red-mostwavelengthsdisplayed in the Figures. As men-
height quickly shifts from photospheric to chromospheric for tionedin Sect. 2, normalIBISsamplingdoesnotreachcontin-
smallwavelengthdifferencesaroundthekneesoftheline.This uum wavelengths (although technically feasible, the measures
quicktransitioninformationheightisobviousinthelineinten- would be much noisier and affected by a high levelof spectral
sitycontributionfunction,displayedinFigure5asafunctionof parasitic light). Hence, our observations do not go far enough
wavelength(onthehorizontalaxis)andheightintheatmosphere from the line core to see the normal granular structures of the
(on the vertical axis). At wavelengths for which the height of lowphotosphere.
τ ∼1correspondstotheheightofthelineopacitygap,thecon- Using 3-dimensional radiative hydrodynamics simulations,
λ
tribution function rises steeply with height, indicating that the Cheungetal. (2007) recently offered the most comprehensive
line intensity sensitivity quickly changes from photospheric to explanationofthehorizontalstructureofthephotospherictem-
chromospheric(seealsoUitenbroek1989;Qu&Xu2002). perature,and of reversegranulationin particular. The latter re-
TheverynarrowspectralpassbandofIBISmakesitpossible sults fromthe interplaybetweencoolingofgranulesrising and
toclearlydistinguishthesetworegimesdirectlyintheimaging, expanding within the optically thin layers of the photosphere,
with a rapid change with wavelength in observed morphology. and radiative heating acting against it. The (horizontally aver-
G.Cauzzietal.:ThesolarchromosphereathighresolutionwithIBIS 7
aged)layeratwhichthetemperaturecontrastbetweengranules analmostcompletesupergranularcell(∼28Mmindiameter)is
andintergranulesvanishesispositionedatabout150kmabove visibleinthecentral-rightportionoftheFOV.
the classical photospheric surface (τ = 1); photospheric di- The first feature that one discerns is the magnetic network,
500
agnostics originating above this height are bound to show evi- again outlined by clusters of small, bright elements that faith-
dence of the reversed temperature contrast with respect to the fullymapthepresenceofphotosphericmagneticelements.The
continuumgranulation.Thisisindeedconsistentwiththewing appearanceofthenetworkelementsinthecoreimagesismore
observationspresentedabove,wheninterpretedthroughthecon- diffusewithrespecttothewingsoftheline(Sect.5.2),owingto
tributionfunctionof Fig. 5: at about0.1 nm fromline core the thelateralspreadofthemagneticfieldwithheight.However,it
layerscontributingmosttotheintensityareabove∼150km,i.e. must be remarked that their spatial extension seems to be very
alevelwherereversedgranulationisalreadypresent.Instead,at dependenton the instantaneousseeing conditions,as well visi-
about0.15nmtheintensityformsslightlylower,inintermediate blefromthemovie.Aslongknown,thechromosphericnetwork
layerswherethegranularcontrastismuchreduced. elements evolve relatively slowly, with single features persist-
ingforseveralminutes,andtheypresentverydifferentspectral
propertiesanddynamicsthantherestoftheFOV(Sect.6).
5.2.Outerwings–themagneticphotosphere
In contrast, the internetwork portions of the FOV are
seething with smaller bright features, rapidly evolving, im-
Besidesreversegranulation,theouterwingimagesveryclearly
mersed in an otherwise very dark environment. These bright
display the location of magnetic structures, such as the plage
pointsreachthesameintensityasthenetworkpoints(about1.5–
or quiet network visible in Figs. 2, 3 and 4. Given their large
2timestheaveragecoreintensity),butaremoresharplydefined
temperature sensitivity, the intensity in the outer wings basi-
as roughlycircular areas with 1”–3” diameter,and evolve with
cally maps the lower opacity, higher temperature small scale
timescales of 1–2 minutes. At their darkest, they reach about
magneticelements.Theirvisibilityisfurtherenhancedatsome
0.75times the averagecore intensity.Many of the pointsseem
wavelengthsbythereducedcontrastofthesolargranulation.The
toreappearthroughoutthecourseoftheobservations,withaca-
co-temporalMDIhighresolutionmagneticmapsforthecaseof
denceofafewminutes,andabout10-20ofthemarepresentat
May31,2004andJune2,2004demonstratethatthereisaone-
any given time within the cell interior. These properties are in
to-onecorrespondencebetweenbrightwingstructuresandmag-
completeanalogytothepropertiesoftheH andK “grains”
neticpatches.Actually,thehigherresolutionIBISimagesclearly 2V 2V
(see the review of Rutten&Uitenbroek 1991). The latter have
reveal how single magnetic features visible in MDI are com-
been explained as due to shocks generated by acoustic waves,
posed of several distinct structures of sub-arcsec size, both in
withfrequenciesslightlyabovetheacousticcutoff,originatingin
thenetworkandpossibly,theinternetwork(Janssenetal.2003).
thephotosphericlayersandpropagatingupwardsintheabsence
The use of intensity proxiesas a diagnosticsof small-scale
ofmagneticfields(Carlsson&Stein1995,1997).Theradiative-
magnetic elements in the photospherehas been recently inves-
hydrodynamical model of Carlsson & Stein indicates that in-
tigated by Leenaartsetal. (2006), by numerically synthesizing
deedtheshocksshouldbewellvisibleinquietareasalsointhe
several spectral diagnostics in a snapshot of a 3D solar mag-
CaII854.2nmline(Pietarilaetal.2006)but,toourknowledge,
netoconvection simulation. They identify the outer wings of
theyhadneverbeenconvincinglyobservedin thisspectralsig-
CaII 854.2 nm as one of the best proxies (albeit at the price
nature.IBISobservationsthusopenthewaytoamorecomplete
of a reducedspatialresolutionwith respectto the wingsofHα
study of the phenomenon, as both the spectral and spatial do-
or Hβ), consistently with the results shown here. Somewhat at
mains are accessible at once, with much statistics provided by
oddswiththeirresults,however,wefindthattheintensityimages
large FOVs and the opportunity to analyze different magnetic
at −0.09nm fromline core alreadyshow the reversedgranula-
topologies. Results of such an analysis will be presented in a
tion pattern (as discussed in the previous Section), rather than
forthcomingpaper(Vecchioetal.2007a).
normal granulation. In our data magnetic structures are hence
Thethird,prominentfeatureintheCaII854.2nmcoreim-
better identified in intensity images obtained further out in the
agesisduetofibrils,thatseemtooriginatefromeventhesmall-
wings(uptothelimitofoursampling,i.e.0.16nmfromcore),
est magneticelements(compareMovie1). Theirappearanceis
where the granulation contrast almost vanishes. This discrep-
stronglyreminiscentofthestructuresobservedinHα(seethere-
ancyshouldbefurtherinvestigatedbyamoredetailedcompari-
centhighresolutionobservationsofRouppevanderVoortetal.
sonbetweensimulationsandhighqualityobservations.
2007), although we don’t have here any direct comparison be-
tween the two spectral signatures. Much as for the Hα fibrils,
5.3.Corestructures the structures observed in CaII 854.2 differ sensibly between
active regions and quiet Sun. More active regions harbor long,
Narrow band CaII 854.2 snapshots, acquired around the line andrelativelystablefibrilstogetherwithshorterdynamicfibrils
core wavelength in the quiet regions, display a clear segrega- (Fig.4),whilethequietSundisplaysbothlongmottles,connect-
tionoftheFOVinvariouscomponents(seeFigs.2and3).The ingthestrongermagneticconcentrationsinthenetwork(aclear
distinction is even clearer if one has access to the temporaldi- example is in the lower half of the FOV of Fig. 3) and shorter
mension:amovieintheonlineversion3 showstheevolutionof ones,presumablyoutliningfieldlinesthatreachoutwardtothe
the nominalCaII 854.2 nm line core as observedover the full coronaandinterplanetaryspace.ThestructuresvisibleinMovie
FOV and for the full duration of the observations of May 31, 2mostprobablybelongtothelattercategory,astheFOVliesat
2004 (55 minutes with cadence of 19 s). Although the seeing theedgeofacoronalhole.
conditionswere at times variable, and certainly worse than for Severalofthefibrilscharacteristicsareworthofnotice:first
thedataofJune02,2004,wechosethisdatasetasitprovidesa of all, they occupy a large fraction of the FOV, even in the
clearerpictureofthequietestinternetworkregions.Inparticular, quietest of instances. For example, in the region observed on
May 31, 2004, they occupy between 30 and 50% of the “in-
3 Movie 2: http://www.arcetri.astro.it/science/solar/IBIS/movies/ ternetwork” area, while they become the dominant component
31May2004.core.gif in the weak plage region of Fig. 4. Second, their dynamic is
8 G.Cauzzietal.:ThesolarchromosphereathighresolutionwithIBIS
completelydifferentfromtheinternetworkareas:lookingatthe
movie,oneisgiventheclearimpressionofmaterialflowingfrom
the network elements outwards, i.e. of predominant transver-
salmotionswith respectto thesolar surface,asopposedto the
mostly vertical dynamics of the bright points. We do not ob-
servesignificantlateralswayingofthefibrils,althoughthismay
be due to the lower spatial and temporal resolution compared
to RouppevanderVoortetal. (2007), who had a 1 s cadence
(comparedtothe19sinourdata)andaspatialresolutionnear-
ing the diffraction limit of the Swedish Solar Telescope (SST,
0.17” at the Hα wavelength). Finally, in some instances, small
brightpointsasthosedescribedabovearevisiblewithinthefib-
rils;whetherbecauseofphysicallateralmotions,orbecauseof
changesintheopacityofthestructure,attimesthefibrilsletus
glimpsethequietatmosphereunderneath.
5.4.FibrilsinCaII?
TheoverwhelmingpresenceoffibrilsintheCaII854.2nmcore
imageswasaninitialsurpriseoftheIBISobservations,asthey
hadneverbeenobservedatthislevelofdetailotherthaninthe
Hα core and inner wings. Earlier CaII 854.2 nm spectroheli-
ograms by Title (1966, Plates 16–18) as well as MSDP obser-
vationsprobablydidnotreachthenecessaryspectralandspatial Fig.6.“Slit”spectraobtainedaroundthemiddleofthesequence
resolution in order to well appreciate these characteristics, al- ofJune02,2004,fromtwohorizontalcutsofthedatadisplayed
though MSDP data have shown the presence of arch filament inFig.3.Thex-axisisthewavelengthdistancefromtheaverage
systemsinemergingfluxregions(Meinetal.2000). line core position. The left spectrum corresponds to the upper
A striking point is that fibrils are normally absent from slice, across network areas, while the right one correspondsto
CaIIHandKimaging,unlessobservedatveryhighspatialres- thelowerslice,andrepresentsquietinternetworkareas.
olutioninthecoreofthelines,whentheymightappearasbright,
uprightstalksinnetworkareas(e.g.Rutten2006,2007).Thishas
ledtothe(oftenunspoken)assumptionthatthelatterlinessam- 6. Spectralstructure
plea‘different’,possiblylowerchromospherethanHα,andthat
the enhancedvisibility of fibrils in Hα imagingis due to some The most attractive quality of two-dimensional spectrometers
peculiarities of the line formation, in particular its dependence lies in the possibility of obtaining sufficiently detailed spectral
onphotoionization-recombinationmechanisms. information over an extended FOV, in a period short with re-
specttotheevolutionarytimesofthestructures.Wedescribein
Given the strong interconnection between CaII H and K
this section some characteristicsof the CaII 854.2nm spectral
and the IRT formationdiscussed in the previousSections, why
thendoestheCaII854.2linedisplaysuchadifferent“chromo- profiles obtained from the IBIS observations presented above.
WeusethequietSundatasetobtainedonJune2,2004,thatcan
spheric” scenario? We believe that a large part of the problem
beeasilycomparedtoearlierresults.
lies not in the infrared line formation, but in the strong obser-
vationallimitationsthatstillaffectCaIIHandKdata.Inother
words,theissueisnotwhytheCaII854.2nmlinedisplaysthe
6.1.Spectraandaveragelineprofiles
chromosphericfibrils,butwhytheHandKdonotshowthem!
Essentially,thechromosphericsignaloftheCaIIHandKlines Fig. 6 shows two “slit spectra” derived from two cuts along
isconfinedtotherathernarrowcore(withDopplerwidthof10– the horizontal direction of the data displayed in Fig. 3, as in-
15 km s−1, i.e. ≤20 pm at 400nm), while mostof the imaging dicated in the Figure. The spectral axis is obtained interpolat-
isperformedwithbroadfilterswithpassbandsintherange30– ingthesequenceofmonochromaticvaluesforeachpixelalong
100 pm. These however strongly dilute the core signal, result- thecut.Giventheacquisitionsequence,thetimealsorunswith
ing in images that are heavily biased toward the wing, upper- thewavelength,andanintervalofabout7sseparatestheblue-
photospheric, signal. Scattered light, and general photometric most from the red-most wavelengths. The left spectrum inter-
noise further degrade the signal of interest. Finally, even with sects several network points (i.e. between positions 40 and 50
these relatively broad filters, the low photon flux near the core along the slit) and fibrils, while the right one covers mostly
oftheseverydeeplines(duealsotothelowtransmissionofthe quiet areas. The quality of the spectra is high, and appears
filters)requireslongexposures,whichcompoundstheincreased comparable to that of data acquired at high spatial resolution
seeing-induceddistortionsattheshorterwavelengths.Forthese withnormalspectrographs(seeforexamplethespectraobtained
reasonsonlyverybrightchromosphericstructures(forexample in the CaII 866.2 nm line at the Swedish Solar Telescope by
those related to network points) can survive the smearing and Langangenetal.2007).Numerousindicationsof thedynamics
stillbedistinguishableabovethephotosphericbackgroundinthe arevisibleinthetworegions,butwithverydifferentcharacteris-
“core”images(Rutten2006).Tobetterproveourpoint,adirect tics.Theleftpaneldisplaysstrongredshiftsofthelineinseveral
comparisonofsimultaneousCaIIKfiltergramsandCaII854.2 pixels (e.g. around positions 40 and 52, that become progres-
IBISdatawillbepresentedinaforthcomingwork.Preliminary sively strongerwithin a coupleofarcsec, and probablyinvolve
resultsarereportedinReardonetal.(2007). dynamic fibrils (DePontieuetal. 2007). In the right panel, in-
G.Cauzzietal.:ThesolarchromosphereathighresolutionwithIBIS 9
tudeareprobablyunavoidable,duebothtoslightresidualsinthe
datacalibrationandtotheintrinsicvariabilityofthesolarstruc-
tures in time and space. Still, all the general characteristics of
theatlasprofilearewellreproduced,includingtheasymmetryof
thelinecoretypicalofchromosphericlines(“inverseC-shape”,
Uitenbroek2006b).
Whatkindof individualprofilesenter the average?To gain
insight into this issue, we have attempted a spatial classifica-
tionforthethreefeaturesdiscussedbefore:network,fibrils,and
quiet internetwork. In this kind of analysis, often an intensity
thresholdisadoptedinthedefinitionofseparateclassesofpixels
(e.g.Deubner&Fleck1990).However,asclearfromSect.5.3,
thechromosphericsceneisprobablytoocomplextobeproperly
describedbyasingle,static parameter.Inparticular,thefibrils,
veryprominentatthecorewavelengths,appearherebothdarkor
brightwith respecttothe averagecontrast,makingtheirrecog-
nition quite difficult. As a first attempt we have identified the
threeclassesofpixelsbyadoptinganintensitythresholdapplied
tothetemporalaverageoftheobservations,butusing,foreach
feature,theintensityatthewavelengthpositionswheretheyare
mostclearlyseparatedfromthe background.First, thenetwork
hasbeendefinedasthebrightpixelsinthefarwingoftheline(at
+155pmfromlinecore,havingintensityaboveaverageplusone
time ther.m.s.).Thisthresholdclearlyidentifiesallof thepho-
tosphericmagneticfeatures,asseene.g.inthecotemporalMDI
magnetograms,thatalsohaveanobviouscorrespondenceinthe
chromospheric line core intensity. Then, the fibrils and the in-
ternetworkareaswererespectivelydefinedasthedarkandinter-
mediatepixelsintheinnerwingintensityimages(±20pmfrom
line core),usingthe intensityvaluesofthe averageplusor mi-
nus1/3ofther.m.s.asourlimits.Thesevaluesseparatedclearly
the two classes of pixels; in particular, they assured that there
werenoquietpixelsinthevicinityofnetworkpoints,wherefib-
rilssupposedlyoriginate.Fromourdefinition,thefractionalarea
occupiedwas9%(networkpoints),39%(fibrils)and35%(inter-
network),withtheremainingpixelsnotassignedtoanycategory.
Asexpected,theselatterunclassifiedpixelsliemostlyaroundthe
Fig.7. Average line profiles for the data acquired on June 02, magneticelements.Itisinterestingtonotethatinthisdatasetthe
2004. Panel a: the spectrum averaged over the whole dataset fractionalareacoveredbyfibrilsisactuallylargerthanthequiet
(thinsolidline)iscomparedtotheLiegeatlas(thicksolid).The Sun’s,reflectingtheclosedmagnetictopologyoftheregion.
effectof theprefiltertransmissionhavebeenremovedfromthe Theaveragedlineprofilesforthethreeclassesaredisplayed
observations.Theprofilesaveragedoverdifferentsolarfeatures in Fig. 7 as well. The network points average profile (dashed
definedbyusinganintensitythresholdarealsoshown:network line)ismuchbrighterthanthereferenceoneatallwavelengths
points (dashed); fibrils (dot-dashed) and internetwork (dotted). (however note that we don’t have any information on the con-
Panelb:Aspanela,butthistimefibrilsandinternetworkareas tinuumvalues),isslightlyredshifted(about150ms−1),andhas
weredefinedusingdynamicalproperties. acoreasymmetrylesspronouncedthantheaverageprofile.The
fibrils’profile(dash-dotted)isalmostcomplementarytothenet-
work one: it is slightly blueshifted with respect to the average
stead, stronger blueshifts are visible (e.g. position 12, 25, 33 overthewholeFOV(about100ms−1);itisdeeper,withacore
alongtheslit),thatarelessextendedinthespatialdirectionand intensity about3% lower (in units of the continuumintensity),
havemoreconstantamplitude.Thesearerelatedtothedevelop- whiletheoverallwingsremainclosetotheaverage.Thesesame
ment of acoustic shocks in non-magnetic areas (Pietarilaetal. properties have been reported by Harvey (2005) in a study of
2006;Vecchioetal.2007a). “calcium circumfaculae”using CaII 854.2 SOLIS data around
In Fig. 7 we show a comparison between the Liege atlas activeregions.He demonstratedhowthe areasof deeper,blue-
(thicksolidline)andtheaverageobservedlineprofile,obtained shiftedCaII854.2profilescoincidedwith theextendedfibrilar
summingoverthe wholedataset(morethan 2×107 individual regions observed in Hα core. Finally, the internetwork profile
profiles,thinsolidline).Theeffectsoftheprefiltertransmission (dottedline)isonlyslightlydeeperthantheaverage,whilemost
(cf. Fig. 1) have been accountedfor, and the wavelength offset oftheotherprofilecharacteristicsremainverysimilartotheav-
has been adjusted so that the minimum of the average profile erageoverthewholeFOV.
coincides with the atlas line core position. The general agree-
mentbetweentheprofilesisexcellent,showingdiscrepanciesof
6.2.Velocitydistribution
less than 2% of the continuum intensity around the inflection
pointin the wings(the largerdifferenceatthe bluelimitof the From the spectralprofiles described above,we have calculated
observationsisduetoatelluricline).Differencesofthisampli- line-of-sightvelocitiesforeachpixelintheFOVandeachtime
10 G.Cauzzietal.:ThesolarchromosphereathighresolutionwithIBIS
step.TothisendweusedtheDopplershiftsoftheintensitymin-
ima of the spectral profiles, derived from second degree poly-
nomial fits of the line core. The zero has been defined as the
spatio-temporal average position over the whole dataset. This
approachhasoftenbeenusedforthecaseofquietchromospheric
dynamics(e.g.Deubner&Fleck 1990), butone mustbe aware
that the derived values might not be fully representativeof the
actualplasmavelocities,especiallyinthepresenceofnonlinear
phenomenasuchasshocks.Inparticular,weexpectthatourve-
locitydeterminationmightfailforacertainfractionofthepixels
in the magnetic areas (2–3% of the total pixels), for which we
observeemissioninthebluewing,similarlytothecasereported
inPietarilaetal.(2007b).However,sinceweareonlyinterested
here in giving general properties, and given the overall small
numberof such pixels, we will keep thisdefinition asa conve-
nientparametricdescriptionofthedynamicsofthesystem.
TheupperpanelofFig.8displaysthedistributionofvelocity
valuesoverthe whole dataset(thick solid line).Positive values
indicate red-shifts, i.e. downward motions. The distribution is
skewedtowardspositivevalues,justifyingthepeakatthenega-
tive(upward)valueof−0.12kms−1sincewedefinedthezeroas
theaveragevalueoverthewholedataset.About97%ofthepix-
elsarecontainedinthe±3kms−1 range,butoutlierscanreach
highvelocityvalues,above10kms−1.Attheselargervaluesthe
distributionbecomesmarkedlyasymmetricinthepositivequad-
rant.
Decomposing the distribution into the three types of so-
lar features described above, i.e. network (dashed line), fibrils
(dash-dotted)andinternetworkareas(dotted),oneseesthatthe
various distributions have different widths, with the fibrils dis-
playing the least number of extreme vertical flows, especially
lackingstrongdownflows.Ther.m.s.valuesofthedistributions Fig.8. Upper panel: line-of-sight velocity distribution for the
are, 1.5, 1.4 and 1.25 km s−1 for, respectively, network, inter- datasetofJune2,2004.Solidline:overallFOVandtime(more
network areas, and fibrils. Further, the enhanced positive lobe than 2×107 values); dashed, dash-dotted and dotted: distribu-
in the total velocity distribution is largely due to the network tionsfor,respectively,networkpoints,fibrils,andinternetwork.
points, that accountfor about40%of this signal despite repre- Bottom panel: velocity power spectra for the same features
sentinglessthan10%ofthetotalpixels.Sinceneitherfibrilsnor (symbolsasupperpanel).Thecurvesrepresentanaverageover
quiet areas contribute much at the highest redshifts, the signal thecorrespondingpixels.
missingwithrespecttothewholeFOVdistributionmustbedue
tothepixelsleftunclassified,thatarepositionedaroundthenet-
work. Thisfurther underlinesthe highly dynamicnature of the
magnetic network elements, that for this enhanced flux region
throughthe use ofdestretchingtechniquesto correctfordiffer-
mightberelatedtothedynamicfibrilsoftenobservedinplages
entialimagemotion).
(Hansteenetal.2006;DePontieuetal.2007).
Thepowerspectrarelativetothethreedifferentatmospheric
components defined above are also displayed in Fig. 8, using
6.3.Velocitypowerspectra thesamesymbolsasFig.7.Thecurvesareanaverageoverthe
correspondingpixels,i.e.to recovertheglobalpowerspectrum
To address the presence and relevance of periodicities in the theymustbeweightedwiththeirrelativeoccurrence.Theycon-
CaII854.2nmdynamics,thetemporalevolutionofline-of-sight firmseveralofthepropertiesderivedbyearlierworks,inpartic-
velocities has been investigated via a standard Fourier analy- ular thatlow frequencyoscillationsare moreprominentin net-
sis,performedseparatelyoneachspatialpixel.TheNyquistfre- workelementsthanquietareas,whileperiodicitiesof3minutes
quencyoftheobservationsis26mHz,andthefrequencystepis orshorterareessentiallyduetothequietinternetwork(seee.g.
0.33mHz.ThethicksolidcurveofFig.8(bottompanel)shows theIntroductionofDeubner&Fleck1990).Thenetworkpoints
theresultingspatiallyaveragedvelocitypowerspectrum.Ithas displayaclearpeakat∼3.5mHz,butwecannotobviouslyiden-
abroadpeakbetween∼3and7mHz(periodicitiesbetween2.4 tifyanyfurthermaximaatlowerfrequencies(asclaimede.g.by
and 5 minutes), that indicates an almost equal contribution of Kalkofen 1997). The network power (per pixel element) at the
bothtypicalphotosphericandchromosphericoscillatorysignals 3.5 mHz frequencyis enhancedby abouta factor of 2 with re-
totheaverage.Thisisfullyinagreementwithearlierresultsob- spect to the rest of the FOV, as reported in Litesetal. (1993).
tainedusingsingle-slitspectrographicobservationsonquietso- Knownforsometime,theselatterpropertieshaverecentlybeen
lar regions(e.g.Noyes1967;Fleck&Deubner1989). We note related to the leakage of the dominant photospheric p-modes
thatthenoiselevels,determinedbytheaccuracyofthecorefit- intohigheratmosphericlayersduetotheloweringoftheacous-
tingprocedures,isverycomparablebetweentheseIBISdataand tic cutoff frequencyin magneticelements(Jefferiesetal. 2006;
previousspectroscopicstudies(thisaccuracywasonlyachieved Hansteenetal.2006;Vecchioetal.2007b).