A complete versionPreprint typeset using LATEX style emulateapj v. 08/22/09

MULTI-SIGHTLINE OBSERVATION OF NARROW ABSORPTION LINES IN LENSED QUASARSDSS J1029+26231,2

Toru Misawa3, Cristian Saez4,5, Jane C. Charlton6, Michael Eracleous6,7, George Chartas8, Franz E.Bauer9,10,11, Naohisa Inada12, and Hisakazu Uchiyama13

A complete version

ABSTRACTWe exploit the widely-separated images of the lensed quasar SDSS J1029+2623 (zem=2.197, θ =

22′′.5) to observe its outflowing wind through two different sightlines. We present an analysis of threeobservations, including two with the Subaru telescope in 2010 February (Misawa et al. 2013) and2014 April (Misawa et al. 2014b), separated by 4 years, and one with the Very Large Telescope,separated from the second Subaru observation by ∼2 months. We detect 66 narrow absorption lines(NALs), of which 24 are classified as intrinsic NALs that are physically associated with the quasarbased on partial coverage analysis. The velocities of intrinsic NALs appear to cluster around valuesof vej ∼ 59,000, 43,000, and 29,000 km s−1, which is reminiscent of filamentary structures obtainedby numerical simulations. There are no common intrinsic NALs at the same redshift along the twosightlines, implying that the transverse size of the NAL absorbers should be smaller than the sightlinedistance between two lensed images. In addition to the NALs with large ejection velocities of vej >1,000 km s−1, we also detect broader proximity absorption lines (PALs) at zabs ∼ zem. The PALs arelikely to arise in outflowing gas at a distance of r ≤ 620 pc from the central black hole with an electrondensity of ne ≥ 8.7×103 cm−3. These limits are based on the assumption that the variability of thelines is due to recombination. We discuss the implications of these results on the three-dimensionalstructure of the outflow.Subject headings: quasars: absorption lines – quasars: individual (SDSS J1029+2623)

1. INTRODUCTIONAGN outflows, potentially powered by one or more of

a variety of mechanisms (e.g., radiation force, magneticpressure, and magnetocentrifugal force), are importantingredients of quasar central engines and likely play a rolein quasar and galaxy formation/evolution because: 1)they extract angular momentum from accretion disks al-lowing gas accretion to proceed (e.g., Blandford & Payne1982; Emmering, Blandford, and Shlosman 1992; Konigl& Kartje 1994; Everett 2005), leading to the growth of

Electronic address: misawatr@shinshu-u.ac.jp1 Based on data collected at Subaru Telescope, which is operated

by the National Astronomical Observatory of Japan.2 Based on observations obtained at the European Southern

Observatory at La Silla, Chile in programs 092.B-0512(A)3 School of General Education, Shinshu University, 3-1-1 Asahi,

Matsumoto, Nagano 390-8621, Japan4 Korea Astronomy and Space Science Institute (KASI), 61-1,

Hwaam-dong, Yuseong-gu, Deajeon 305-348, Republic of Korea5 Department of Astronomy, University of Maryland, College

Park, MD 20742-24216 Department of Astronomy and Astrophysics, Pennsylvania

State University, University Park, PA 168027 Department of Astronomy & Astrophysics and Center for

Gravitational Wave Physics, The Pennsylvania State University,525 Davey Lab, University Park, PA 16802

8 Department of Physics and Astronomy, College of Charleston,SC 29424

9 Instituto de Astrof́ısica, Facultad de F́ısica, Pontificia Univer-sidad Católica de Chile, Casilla 306, Santiago 22, Chile

10 Millennium Institute of Astrophysics (MAS), NuncioMonseñor Sótero Sanz 100, Providencia, Santiago, Chile

11 Space Science Institute, 4750 Walnut Street, Suite 205,Boulder, Colorado 80301

12 Department of Physics, National Institute of Technology,Nara College, Yamatokohriyama, Nara 639-1080, Japan

13 Department of Astronomy, School of Science, GraduateUniversity for Advanced Studies, Mitaka, Tokyo 181-8588, Japan

black holes, 2) they also provide energy and momentumfeedback to the interstellar medium of host galaxies andto the intergalactic medium (IGM), and inhibit star for-mation (e.g., Springel, Di Matteo, & Hernquist 2005),and 3) they may promote metal enrichment of the in-tergalactic medium (IGM) (e.g., Hamann et al. 1997b;Gabel, Arav, & Kim 2006). These outflowing gases,which are difficult to observe directly, have been detectedas absorption lines in the spectra of about 50% of allquasars (e.g., Vestergaard 2003; Wise et al. 2004; Mis-awa et al. 2007a; Nestor et al. 2008; Muzahid et al. 2013).However, a limitation of these past studies is that theyobserve the outflowing gas only along a single sightline(i.e., one dimension) toward the nucleus of each quasar,although the absorber’s physical conditions probably de-pend on the location/orientation at which we observe it(e.g., Ganguly et al. 2001; Elvis 2000). Thus, the internalstructure of outflowing winds is still largely unknown.

Multiple quasar images, produced by gravitationallensing, provide a unique way to study the outflowing gasalong more than one sightline. Lensed quasars with largeimage separation angles have a higher chance of revealingstructural differences in the outflowing winds, especiallyin the vicinity of the continuum source. In this sense,the quasar images that are lensed by a cluster of galaxies(rather than a single massive galaxy) are very promis-ing targets. Among three such lensed quasars, SDSSJ1029+2623 at zem ∼2.197 (Inada et al. 2006; Oguri etal. 2008) is the best target because i) it has the largestlensed quasar image separation (θ ∼ 22′′.5) ever observed(see Figure 1 of Inada et al. 2006), and ii) it exhibitsabsorption features in the blue wings of the C IV, N V,and Lyα emission lines with ejection velocity of vej ≤

2 Misawa et al.

1,000 km s−1, which could be the result of outflowing gasmoving toward us from the central region. We call themProximity Absorption Lines (PALs) throughout this pa-per. We define PALs as a subcategory of Narrow Ab-sorption Lines (NALs) with ejection velocity of vej ≤1,000 km s−1. We use this terminology throughout thepaper to separate PALs from NALs with larger ejectionvelocities.

Misawa et al. (2013) obtained high-resolution spectraof the brighter two of the lensed images (images A andB) with the Subaru telescope in 2010 February and foundseveral clear signs that the origin of the PALs is indeedin the outflowing gas. First, they show the signature ofpartial coverage, which means the absorbers do not coverthe background flux source completely. There also existsa clear difference in the absorption profiles between thespectra of images A and B, which can be explained by ei-ther of the following two scenarios: (a) time variability ofthe absorption features over a time scale correspondingto the time delay between the two images (time varia-tion scenario; Chartas et al. 2007)14, or (b) a difference inthe absorption between the different sightlines of the out-flowing wind (multi-sightline scenario; Chelouche 2003;Green 2006). However, with a single-epoch observationwe cannot distinguish between these scenarios. Misawaet al. (2014b) performed a second observation about fouryears (1514 days in the observed frame) after the first ob-servation (which is longer than the time delay betweenimages A and B, ∆tAB ∼ 744 days), and found that thePALs were nearly stable and that most of the differencesbetween images A and B still remained. This evidencesuggests a multi-sightline scenario where the absorber’ssize should be smaller than the physical distance betweenthe sightlines of the lensed images, thus not covering bothsightlines. A possible explanation is that there are anumber of small clumpy clouds in the outflowing stream.Indeed, some of the outflowing gas is expected to consistof small gas clouds (dcloud ≤ 10−3 pc) with very largegas densities (ne ≥ 106 cm−3; Hamann et al. 2013; Joshiet al. 2014). Furthermore, recent radiation-MHD simula-tions by Takeuchi et al. (2013) reproduce variable clumpystructures with typical sizes of 20 times the gravita-tional radius (Rg), corresponding to dcloud ∼ 5×10−4 pc,assuming a black-hole mass for SDSS J1029+2623 ofMBH ∼ 108.72 M� (Misawa et al. 2013).

Such clumpy clouds can be examined more easilythrough narrow absorption lines (NALs, hereafter) withlarge offset velocities from the quasar because the cor-responding absorbers are (i) probably smaller than thePAL absorbers and (ii) not so crowded in velocity spaceas the PAL absorbers. Indeed, there are many NALsdetected in the spectra of both images A and B ofSDSS J1029+2623. Their origin is not only the out-flowing wind (intrinsic NALs, hereafter) but also cos-mologically intervening gas such as foreground galaxiesand the IGM (intervening NALs, hereafter). Althoughit has been traditionally believed that many NALs thatfall within 5,000 km s−1 of the quasar emission redshift(termed associated absorption lines or AALs) are physi-cally associated with the quasar (e.g., Weymann et al.

14 The time delay between images A and B is ∆tAB ∼ 774 daysin the sense of A leading B, while the time delay between images Band C ∆tBC is only a few days (Fohlmeister et al. 2013).

1979), we can separate intrinsic NALs from interven-ing ones more effectively by performing partial coverageanalysis. In order to form a global picture of the outflow-ing wind, we need to understand the physical conditionsof NAL absorbers (i.e., highly accelerated gas) as well asPAL absorbers (i.e., weakly accelerated gas).

In this paper, we present the results from our new spec-troscopic observation of SDSS J1029+2623 taken withthe Very Large Telescope (VLT), which enables us forthe first time to identify intrinsic NALs in a high qualityspectrum of this object. In §2, we describe the obser-vations and data reduction. The methods used for ab-sorption line detection and covering factor analysis areoutlined in §3. The results and discussion are presentedin §4 and §5. Finally, we summarize our results in §6. Weuse a cosmology with H0=70 km s−1 Mpc−1, Ωm=0.3,and ΩΛ=0.7 throughout the paper.

2. OBSERVATIONSWe acquired high resolution spectra of the brightest

two of the three lensed images of SDSS J1029+2623, Aand B with V = 18.72 and 18.67 mags, with the VLTusing the Ultraviolet and Visual Echelle Spectrograph(UVES) in queue mode (ESO program 092.B-0512(A)).The observations were performed from 2014 January 28to February 26 (epoch E2, hereafter), which is ∼4 yearsafter the first observation on 2010 February 10 (epochE1, hereafter; Misawa et al. 2013), and ∼2 months beforethe third observation on 2014 April 4 (epoch E3, here-after; Misawa et al. 2014b) with Subaru using the HighDispersion Spectrograph (HDS). We used a slit width of1.′′2, corresponding to R ∼ 33,000, while Misawa et al.(2013, 2014b) took R ∼ 30,000 and 36,000 spectra usingSubaru/HDS. The wavelength coverage is 3300–6600 Åin the 390/564 nm setting, which covers the O VI, N V,Si IV, and C IV doublets as well as the Lyα absorptionline at zabs ∼ zem. We also adopted 2×2 pixel binningin both the spatial and dispersion directions to increasethe S/N ratio. The total integration time is 26,670s andthe final S/N ratio is about 23 pix−1 around 4700Å forboth images.

We reduced the data to extract the one-dimensionalspectra in a standard manner using the UVES CommonPipeline Library (CPL release 6.6). We could not sepa-rate the third image (image C, V = 20.63) from image Bcompletely because the typical seeing of our observation(∼ 1.′′0–1.′′8) was comparable to the separation angle be-tween images B and C (θ ∼ 1.′′85)15.

Table 1 gives a log of the current observation withVLT/UVES as well as our past observations with Sub-aru/HDS, in which we list the target name, date of obser-vation, telescope/instrument used, spectral resolution,total exposure time, and signal-to-noise ratio (S/N). TheS/N is evaluated around 4700 Å, close to the C IV mini-BAL. In Figure 1, we show normalized spectra over thefull wavelength range of our observations for images Aand B. These spectra were binned every 0.5Å for displaypurposes, and the 1σ errors are also shown.

3. DATA ANALYSIS

15 On the other hand, a flux contamination from the image Cis almost negligible in the image B spectrum because the former ismuch fainter than the latter.

Multi-Sightline Observation of NALs 3

First, using the line detection code search, writ-ten by Chris Churchill, we detect all absorption fea-tures whose confidence level is greater than 5σ in thenormalized spectrum of each lensed image. We thenidentify N V, C IV, and Si IV doublets in the regionsfrom −1,000 km s−1 to 5,000 km s−1 (N V), from−1,000 km s−1 to 70,000 km s−1 (C IV), and from−1,000 km s−1 to 40,000 km s−1 (Si IV) around the corre-sponding emission lines, with the maximum velocity setin order to avoid the Lyα forest16. We also search for theMg II doublet in the whole range of the spectra. Absorp-tion troughs that are separated by nonabsorbed regionsare considered to be separate lines. In total, 4 Mg II, 19C IV, 2 N V, and 7 Si IV doublets are identified in the im-age A spectrum, while 3 Mg II, 22 C IV, 2 N V, and 7 Si IVdoublets are identified in the image B spectrum. Theequivalent widths of both blue and red members of dou-blets are measured for each line by integrating across theabsorption profile and these are listed in Table 2. We alsosearched for 10 single metal lines (O I λ1302, Si II λ1190,Si II λ1193, Si II λ1260, Si II λ1527, Al II λ1671,C II λ1036, C II λ1335, Si III λ1207, and C III λ1548)as well as Lyα and Lyβ and detected about 200 lines atthe same redshift as the doublet lines. These are sum-marized in Table ??. Other single lines or unidentifiedlines are not shown in Figure 1 and Tables 1–?? even ifthey are detected at a confidence greater than 5σ.

3.1. dN/dz AnalysisOne of the important properties of intrinsic NALs is

a number density excess of high-ionization doublets perunit redshift (i.e., dN/dz; Hamann et al. 1997a, and ref-erences therein). In order to compare the dN/dz fromour spectra with those from our previous study based on37 quasar spectra (Misawa et al. 2007a), we construct acomplete sample including only NALs whose blue dou-blet members would be detected even in the lowest S/Nspectrum in Misawa et al. (2007a). The correspondinglower limits of rest-frame equivalent widths (EWs) areEWrestmin = 0.056 Å, 0.038 Å, and 0.054 Å for C IV, N V,and Si IV, respectively. Here, the values of EWrestmin de-pend on the S/N ratio of the observed spectrum as

EW restmin =−U2 + U

√U2 + 4(S/N)2(M2LM

−1c + ML)

2(S/N)2(1 + zabs)×∆λ(Å),

(1)where U is the confidence level of the EW defined asEW/σ(EW), and ML and MC are the numbers of pixelsover which the equivalent width and the continuum levelare determined (Young et al. 1979; Tytler et al. 1987).Using equation (1), we confirm that the S/N of our VLTspectra is always larger than the required values exceptfor the region between λobs ∼ 4500 – 4525 Å (Figure 2).After removing weak NALs with EWrest < EWrestmin, wehave 12 C IV, 2 N V, and 3 Si IV doublets in image Aspectrum and 11 C IV, 2 N V, and 3 Si IV doublets in theimage B spectrum. We will call this the “homogeneous”NAL sample, hereafter. Following Misawa et al. (2007a),we also combined multiple NALs lying within 200 km s−1

16 Here, we define the velocity offset as positive for blueshiftedNALs from the quasar emission redshift that is determined fromMg II emission line (Inada et al. 2006).

of each other into a single NAL “system” because clus-tered lines are probably not independent even if theyhave a cosmologically intervening origin (e.g., Sargent,Steidel, & Boksenberg 1988).

All identified doublets (including both the homoge-neous and inhomogeneous NAL samples) are listed inTable 2, in which multiple NALs within 200 km s−1 ofeach other are separated by horizontal lines. The tablegives the ion name (ion), flux-weighted absorption red-shift (zabs), ejection velocity (vej) supposing they orig-inate in the outflow, rest frame EWs of blue and redmembers of the doublet (EWrestb , EW

restr ), identification

number in Figure 1 (ID), reliability class of intrinsic lines(described later), ionization class (described later), andvelocity difference from the first doublet at the lowestredshift in each absorption system (∆v). Table ?? sum-marizes other information including the flux-weightedline width of each system on a velocity scale, σ(v), asdefined in Misawa et al. (2007a), and other transitionsthat are detected at the same redshift, as well as the col-umn density (log N), Doppler parameter, b, and coveringfactor, Cf , for each absorption component in the system.

3.2. Partial Coverage AnalysisAmong several criteria, i) time variability, ii) par-

tial coverage, and iii) line locking are the most reliableproperties to distinguish intrinsic NALs from interveningNALs (Barlow & Sargent 1997; Hamann et al. 1997a, andreferences therein). When compared with broad intrin-sic absorption lines, intrinsic NALs are less likely to vary(Misawa et al. 2014a; Chen et al. 2015) and when theydo vary, their variation amplitude is small (Wise et al.2004; Misawa et al. 2014a). Therefore, the variabilitycriterion does not offer an efficient way of identifying in-trinsic NALs.

Partial coverage analysis is quite useful for our spectrabecause the resolving power is high enough to deblendNALs into multiple components. Using the Voigt pro-file fitting code minfit (Churchill 1997; Churchill et al.2003), we deblended NALs into 86 and 91 components inimages A and B, respectively. We do not include the Si IVNAL at zabs = 1.8909 because the blue and red membersof the doublet are both blended with other lines. Withminfit, we fit each NAL profile using the redshift (z),column density (log N in cm−2), Doppler parameter (bin km s−1), and covering factor (Cf) as free parameters.The covering factor (Cf) is the fraction of photons fromthe background source that pass through the absorber. Ifthe background source is uniformly bright, then Cf alsorepresents the fraction of the background source (i.e.,the continuum source and/or broad emission line regions,BELRs) that is occulted by foreground absorbers alongour sightline. If Cf is less than unity, it is likely thatthe absorbers are part of a quasar outflow because cos-mologically intervening absorbers like substructures inforeground galaxies and the IGM are less likely to haveinternal structures as small as a size of the backgroundflux sources (e.g., Wampler et al. 1995; Barlow & Sargent1997). The covering factor is evaluated in an unbiasedmanner as

Cf =(Rr − 1)2

1 + Rb − 2Rr, (2)

where Rb and Rr are the residual (i.e., unabsorbed) fluxes

4 Misawa et al.

of the blue and red members of a doublet in the nor-malized spectrum (c.f., Hamann et al. 1997b; Barlow &Sargent 1997; Crenshaw et al. 1999). If minfit givesunphysical covering factors for some components, e.g.,negative or greater than 1, we rerun the code assumingCf = 1 only for those components because the Cf valuesare very sensitive to continuum level errors, especiallyfor full coverage doublets (Misawa et al. 2005). In addi-tion to the fitting method above, we evaluate Cf valuesfor each pixel as Ganguly et al. (1999) did (pixel-by-pixelmethod). The fitting results by both methods are shownin Figure 3 and the fit parameters by the fitting methodare summarized in Table ??. We have confirmed thatthe Cf values provided by the two methods are in goodagreement with each other.

4. RESULTS4.1. Narrow Absorption Lines (NALs)

Using our NAL sample toward the lensed images ofSDSS J1029+2623, we perform several statistical anal-yses. In these analyses, in order to avoid any possiblebiases we do not include the two C IV NALs at zabs ∼1.9322 and 1.9788 that are covered in the Subaru spec-trum (Misawa et al. 2013) but not covered by our VLTdata.

4.1.1. dN/dz analysisIn Table 4, we summarize the number density of homo-

geneous NAL systems per unit redshift (dN/dz) or perunit velocity offset from the quasar (dN/dβ) for C IV,N IV, and Si IV, along with the Poisson noise in thesequantities (Gehrels 1986). All systems are also classifiedinto one of two categories according to the offset veloc-ity from the quasar; associated absorption lines (AALs)with vej ≤ 5,000 km s−1, and non-AALs with vej >5,000 km s−1, following Misawa et al. (2007a). We foundthat the dN/dz values for C IV, N V, and Si IV toward thisone quasar, SDSS J1029+2623, are larger than the aver-age values in the larger sample of Misawa et al. (2007a)by factors of ∼3, ∼6, and ∼2, respectively, although thisenhancement is not statistically significant because of thesmall number of NALs in our spectra. Because Misawa etal. (2007a) discovered that at least ∼20% of C IV NALsoriginate from winds based on partial coverage analysis,statistically we expect to detect two or more intrinsicC IV NALs among ∼10 homogeneous C IV NALs in thespectra of images A and B.

4.1.2. Intrinsic or Intervening NALsHigh-velocity NALs blueshifted with vej > 1000 km s−1

are expected to be outside of the range of velocities whereBELs are found, and thus, they absorb mostly continuumlight. The existence of partial coverage in high-velocityNALs suggests that the size of absorbers is comparableto or smaller than the continuum source (i.e., dcloud ≤Rcont). Following Misawa et al. (2007a), we separateall NALs (including those in the inhomogeneous sample)into three classes (classes-A, B, and C) based on partialcoverage analysis, where class-A includes NALs most re-liably classified as intrinsic while class-C includes NALsthat are consistent with full coverage or that cannot beclassified. Class-B contains NALs that show line-locking,which is a signature of a radiatively driven outflowing

wind and is only detectable if our sightline is approxi-mately parallel to the gas motion, as often seen in NALs(e.g., Benn et al. 2005; Bowler et al. 2014). Class-Balso contains systems that have tentative evidence forpartial coverage. As a result, 4 C IV NALs (includingPALs) are classified as intrinsic (two class-A and twoclass-B) among the 12 C IV NALs in the homogeneoussample from the spectrum of image A, while 3 C IV NALsare classified as intrinsic (two class-A and one class-B)among 11 NALs in the spectrum of image B. The frac-tion of intrinsic NALs (27 – 33%) is somewhat largerthan the average value of ∼20% (Misawa et al. 2007a),although our sample size is small. If we include the in-homogeneous sample, 5 C IV NALs are classified as 3class-A and 2 class-B among 19 NALs toward image A,while 9 out of 22 C IV NALs are classified into 4 class-Aand 5 class-B NALs toward image B. In addition to in-trinsic C IV NALs, we detected two class-B Si IV NALsonly toward image A in homogeneous sample, and threeclass-B Si IV NALs in each of the image A and B spectraafter including systems from the inhomogeneous sample.It is also noteworthy that we detect a large number ofline-locked NALs: 2 C IV and 3 Si IV NALs in image Aand 5 C IV and 3 Si IV NALs in image B, while only fivesystems are line-locked among 138 homogeneous NALstoward 37 quasars (Misawa et al. 2007a). This resultstrongly suggests that our sightline toward the centralsource is almost parallel to the outflowing streamline.

4.1.3. Ionization Conditions

The outflowing winds in the vicinity of the flux sourceare probably more highly ionized than most interveningabsorbers due to strong UV radiation from the contin-uum source, although it depends on the gas density ofthe absorbers. Indeed, a high ionization state has beenused as one indicator of the intrinsic properties of ab-sorbers (Hamann et al. 1997a, and references therein).Broad absorption lines (BALs) are often classified intothree categories according to their ionization level: high-ionization BALs (HiBALs), low-ionization BALs (LoB-ALs), and extremely low-ionization BALs showing Fe IIlines (FeLoBALs) (Weymann et al. 1991). A similar clas-sification has also been performed for NALs (Bergeron etal. 1994; Misawa et al. 2007a). Motivated by the liter-ature, we classify our NALs into three categories, basedon the detection of absorption lines in low (ionization po-tential; IP < 25 eV), intermediate (IP = 35 – 50 eV), andhigh (IP > 60 eV) ionization levels17. Interestingly, someclass-A NAL systems include low ionization lines (see Ta-ble 2), which suggests that intrinsic NAL absorbers havemultiple phases with different ionization states, as notedin Misawa et al. (2007a).

4.1.4. Similarities in NALs between Sightlines

Here, we present a new method of identifying intrin-sic NALs by exploiting our multi-sightline observation.If NALs with similar line profiles are detected at thesame redshift toward two sightlines, the corresponding

17 An important caveat here is that we do not necessarily de-tect all absorption lines because of the detection limits due to linestrength, observed wavelength coverage, and line blending (e.g., inLyα forest).

Multi-Sightline Observation of NALs 5

absorber must have a size larger than the physical dis-tance between the two sightlines. The absorber also can-not have any internal structures on the scale of the sight-line separation. Although foreground galaxies and IGMstructures also can cover both sight lines (whose physi-cal separation is ∼kpc or ∼Mpc scale), they usually havesignificant internal velocity structures on those scales, asoften seen in the spectra of lensed quasars (e.g., Ellisonet al. 2004). Only intrinsic absorbers can satisfy the re-quirement of having very similar profiles because theirsizes can be larger than the sightline separation (e.g.,sub-parsec scale) if they are at a small distance from theflux source (e.g., r < 1 kpc). Based on the ejection ve-locity distribution of class-A, B, and C NALs (Figure 4),in Figure 5 we summarize the distribution of velocity dif-ferences between NALs and PALs in the two sightlines(∆v) for systems within ∆v ≤ 200 km s−1 between thetwo sightlines. The distribution is almost uniform up to∆v ∼ 200 km s−1 with a clear peak near ∆v ∼ 0 km s−1.Among eight NAL pairs with ∆v ≤ 10 km s−1, four areC IV and N V PALs at zabs ∼ zem, which we will discusslater. The other four NAL pairs are C IV NALs at zabs∼ 1.8909, 2.1349, and 2.1819, and a Si IV pair at zabs∼ 1.8909. Although the velocity shift is very small forthese NAL pairs, their line profiles are clearly different ascompared in Figure 6. This means our two sightlines gothrough different absorbers or different regions of a sin-gle absorber. In either event, we cannot conclude thatthese are intrinsic NALs based only on this analysis. Onthe other hand, it is intriguing that no NAL pairs thatare classified into class-A/B have a common ejection ve-locity or line profile. For example, the C IV NALs atzabs = 1.7652 toward image A and at zabs = 1.7650 to-ward image B are both classified as class-A NALs. Al-though their ejection velocities are very close each other,their line centers and profiles are obviously different, asshown in Figure 7. We also note that the difference is re-markable in all three epochs (see Figure 8). This meansthe multi-sightline scenario (i.e., that two sightlines passthrough different regions of the outflow) is also applica-ble for intrinsic NALs with large ejection velocity as wellas for PALs as already confirmed in Misawa et al. (2013,2014b).

4.1.5. Time Variability of NALs

Because of the lower data quality (S/N ∼ 10 pixel−1)and narrower effective wavelength coverage of the Sub-aru/HDS spectra taken in epochs E1 and E3, we cannotmonitor NALs for time variability analysis with only afew exceptions. As an example, we compare spectra ob-tained at three epochs around the strong C IV NALs atzabs = 1.8909–1.9138 (ID = 35–44) in image A and C IVNALs at zabs = 1.8909–1.9119 (ID = 41–48) in image Bas shown in Figure 9. These NALs are obviously not vari-able, which is consistent with the past result that NALswith large ejection velocities are rarely variable (Chen etal. 2015).

4.2. Proximity Absorption Lines (PALs)The absorption line group at zabs ∼ zem with an ejec-

tion velocity of vej < 1000 km s−1 has already been ob-served twice in 2010 February (epoch E1) and 2014 April(epoch E3) with Subaru/HDS. Its origin is probably in

the outflowing gas because: i) there is evidence for par-tial coverage (e.g., there exists a clear residual flux at thebottom of the Lyα and N V absorption lines even thoughthey appear to be saturated.), ii) the profiles are vari-able, and iii) the profiles show signatures of line-locking(Misawa et al. 2013, 2014b). The most important resultfrom these past observations is that the absorption pro-files of the Lyα, N V, and C IV PALs in the spectra ofimages A and B are clearly different, especially for linesat vej < 0 km s−1 (see Figure 10). There are at least threepossible origins for the difference: a) a micro-lensing ef-fect, b) time delay between the images, and c) differentabsorber structure along the different sightlines. Amongthese, the first idea is immediately rejected because thelensed images show a common ratios between the radio,optical, and X-ray fluxes (Ota et al. 2012; Oguri et al.2012), which is not expected for micro-lensing. We canalso reject the time delay effect because the variability isalmost negligible between epochs E1 and E3 (Misawa etal. 2014b). A difference in column density between thetwo sightlines is the only acceptable explanation. To ourknowledge, this is the first time that an outflowing windhas been observed along multiple sightlines. Hereafter,we call the PALs at vej < 0 km s−1 (showing sightlinedifference) narrow PALs and those at vej > 0 km s−1(showing similar line profiles between sightlines) broadPALs (see Figure 10).

Our observation with VLT/UVES not only gives anadditional epoch for monitoring the PALs but enables usto study the detailed velocity structure of PALs usinghigher quality spectra with a S/N ratio double that ofthe past observations with Subaru. We first confirm theline-locking pattern in velocity plots of the C IV PAL asshown in Figure 11, which was already noted in Misawaet al. (2013). Like the other NAL absorbers, the PALabsorbers also appear to be outflowing almost parallelto our sightline. We also reconfirm that the C IV PALsbecame shallower (i.e., decreasing in equivalent width)over a large velocity range of the profile between epochsE1 and E2/E3 (see Figure 11) at ∆trest ∼1.3 years inthe quasar rest-frame. However, these lines are almoststable on the shorter time scales of ∆trest ∼0.05 yearsbetween epochs E2 and E3. The broad spectral coverageof the VLT/UVES spectra allows us to detect for thefirst time the O VI doublet corresponding to the PALabsorber (see Figure 10). There is also a hint of the Si IVdoublet detected, but it appears blended with other linesat lower redshift.

4.2.1. Comparison of Covering Factors

In addition to line profiles and strengths, we also com-pare covering factors for the clean part of the C IV andN V PALs in the three epochs. The fitting parameters tothe PALs in epoch E2 with minfit are summarized in Ta-ble ??. The total column densities of C IV and N V PALsafter summing all Voigt components are log(NNV/cm−2)= 15.62 and log(NCIV/cm−2) = 15.90 in image A and16.04 and 15.89 in image B spectrum. These values areall consistent with the corresponding values that we mea-sured in epoch E1 (Misawa et al. 2013) with differencesof a factor of ≤ 2. Because all PALs have small offsetvelocities and are located on the BELs in the spectra,we should consider the BELRs as well as the contin-

6 Misawa et al.

uum source as the background flux source. Because theC IV PAL is partially self-blended (i.e., blending of theblue members of one doublet with the red member ofanother doublet), we can compare covering factors effec-tively only for the narrow PALs at vej ∼ −100–0 km s−1(corresponding to ∆v ∼ 400–500 km s−1 in Figure 12),for which self-blending is negligible18.

We show the fitting results for the PALs in epoch E2in Figure 12 and Table ??. The average covering fac-tors of C IV and N V at vej ∼ −100–0 km s−1 (i.e., ∆v∼ 400–500 km s−1 in Figure 12) are Cf = 0.47±0.03and 0.58±0.05 toward image A, and Cf = 0.23±0.02 and0.39±0.08 toward image B. These values are all consis-tent with the corresponding values in epoch E1 towardimage A (Cf = 0.47±0.05 and 0.61±0.07) and image B(Cf ∼ 0.2 and 0.35±0.05)19 and in epoch E3 toward im-age A (Cf = 0.45±0.09 and 0.54±0.09) and image B (Cf= 0.20±0.05 and 0.34±0.19). Thus, covering factors arenot variable at least on a timescale of ∆trest ∼ 1.3 years.

The N V PALs all have larger Cf values than that forC IV, which is consistent with past results that higherionization transitions tend to have larger coverage frac-tions (e.g., Petitjean & Srianand 1999; Srianand & Pe-titjean 2000; Misawa et al. 2007a; Muzahid et al. 2015).This result suggests that i) the size (i.e., distance fromthe central black hole) of the N V BELR is smaller thanthat of the C IV BELR, and/or ii) the size of the N Vabsorbers is larger than that of the C IV absorbers. It isalso interesting that the PALs in the image A spectrumalways have larger Cf values than those in the image Bspectrum, which is additional evidence for variations inthe structure along multiple sightlines.

5. DISCUSSION5.1. Velocity anisotropies in NAL absorbers

Line-locking is seen in both NALs and PALs, thus weare likely to be observing the quasar almost along thedirection of the outflow. Nonetheless, we discovered avelocity difference between the two sightlines (see Ta-ble 2). Given the redshifts of the lens (zl = 0.58; Oguriet al. 2008) and the source (zs = 2.197) the separationangle of the light rays that form images A and B, as seenfrom the source20, is θ′ ∼14.′′6. This velocity gradientsuggests that the outflowing wind has internal velocityanisotropies on a scale of ∼1 km s−1 arcsec−1 on aver-age. Indeed, the hydrodynamic simulations of Proga &Kallman (2004) show this type of internal velocity vari-ations. A recent radiation-MHD simulation by Takeuchiet al. (2013) also reproduced such velocity variations over

18 Negative ejection velocity for these components could be dueto our underestimation of these values because the emission redshiftis determined from broad UV emission lines, as done for this quasarby Inada et al. (2006), which are systematically blueshifted fromthe systemic redshift that is measured by narrow, forbidden lines(see, e.g., Corbin 1990; Tytler & Fan 1992; Brotherton et al. 1994;Marziani et al. 1996) by about 260 km s−1.

19 Because no absorption component was used for C IV PAL atvej ≤ 0 km s−1 in the epoch E1 spectrum (Misawa et al. 2013), weadopt an average line depth for this region as a covering factor.

20 The separation angle as seen by the source is given by θ′

= (Dol/Dsl) × θ, where θ is the observed separation angle of theimages (22.′′5) and Dol and Dsl = ((1 + zs)/(1 + zl)) × Dls areangular diameter distances from the observer to the lens and fromthe source to the lens, respectively.

a typical spatial scale of ∼ 5×10−4 pc. This is consistentwith the size of NAL absorbers, which is also compara-ble to the size of the continuum source (∼ 2.5×10−4 pc;Misawa et al. 2013). However, we should note that thevelocity variations in the simulations are found in theinner part of the wind at distances of ∼50Rg from thecenter.

The velocity distribution of intrinsic NALs that areclassified into Class-A or B appears to cluster aroundvalues of vej ∼ 59,000, 43,000, and 29,000 km s−1, exceptfor two intrinsic NALs at vej ∼ 49,500 km s−1 (class-A)and ∼ 8,500 km s−1 (class-B) in the image B spectrum(Figure 4). These clustering patterns are reminiscent ofthe filamentary structures obtained by numerical simu-lations (e.g., Proga, Stone, & Kallman 2000). If thesepatterns are indeed due to filamentary structures, thereshould exist some velocity anisotropies within them. Ve-locity dispersions in the three intrinsic NAL clusters areδv ∼ 900, 260, and 1200 km s−1 respectively, which cor-respond to about 1.6%, 0.6%, and 4.0% of their averageejection velocity.

5.2. Ionization condition in PAL absorbersThe Lyα, C IV, and N V PALs have been monitored

at three epochs (E1, E2, and E3) between 2010 Febru-ary and 2014 April. Between epochs E1 and E2/E3,C IV PALs show a clear variation in their strength (i.e.,depth). There are several possible reasons for this timevariability: (a) gas motion across our line of sight (e.g.,Hamann et al. 2008; Gibson et al. 2008; Muzahid et al.2015), (b) changes in the ionization state of the absorber(e.g., Hamann et al. 2011; Misawa et al. 2007b), and (c)redirection of photons around the absorber by scatter-ing material (e.g., Lamy & Hutsemékers 2004; Misawaet al. 2010). Among these, the first scenario can be re-jected because all absorption components in the C IVPAL vary in concert, which requires the implausible sit-uation in which all clouds cross our sightline simulta-neously (c.f. Misawa et al. 2007b). The third scenariois also less likely because it requires a variation in thecovering factor while the Cf values remain almost sta-ble both in C IV and N V PALs between epochs E1 andE2/E3. Thus, only the scenario involving a change inionization state deserves further investigation. This sce-nario is further separated into two variants: i.e., C3+ions are ionized to C4+ or they recombine to C2+. Bothvariants of this scenario can explain the decreasing EWof the C IV PALs. Without knowing an absorber’s ion-ization parameter21, we cannot tell which one is morelikely. If the latter variant applies, we can place con-straints on the electron density and the distance fromthe ionizing photon source by the same prescription asused in Narayanan et al. (2004), taking the variabilitytime scale as an upper limit to the recombination time.Based on the observation that the C IV PALs vary be-tween epochs E1 and E3 over ∆tobs = 1514 days (i.e.,∆trest = 474 days)22 we can place a lower limit on the

21 The ionization parameter U is defined as the ratio of hydrogenionizing photon density (nγ) to the electron density (ne).

22 We compare spectra in epochs E1 and E3 instead of epochsE1 and E2 because the E2 observation spans the time range 2014January 28 to February 26, which gives an additional uncertaintyfor measuring the variation time scale.

Multi-Sightline Observation of NALs 7

electron density of the absorber as ne > 8.7× 103 cm−3,and an upper limit on the distance from the flux sourceas r < 620 pc, assuming U ∼ 0.02, the value at whichthe C IV and N V ions are close to the optimal ionizationstates for those elements (e.g., Hamann et al. 1995).

5.3. Global Picture of the Outflow fromSDSS J1029+2623

In Table 5, we summarize the physical properties ofbroad PALs, narrow PALs, and intrinsic NALs. We alsopresent a possible geometry of the outflow along our twosightlines in Figure 13 based on our previous constraints.

First, we see almost the same absorption profiles ofLyα, N V, and C IV PALs in the two lensed images ex-cept for a clear difference in the narrow PALs at vej ∼−100–0 km s−1. We also confirm that none of the in-trinsic NALs have common absorption profiles betweenthe images. These results suggest that the size of thebroad PAL absorbers are larger than the projected sepa-ration between sightlines at the distance of the absorbers,rθ, while the narrow PAL absorbers and NAL absorbershave sizes smaller than rθ. Such common absorption pro-files are, however, observable regardless of the absorber’ssize, if their distance from the flux source is smaller thanthe boundary radius rb at which the two sightlines oflensed images become fully separated with no overlap(Misawa et al. 2013). The boundary distance is ∼3.5 pcfor SDSS J1029+2623 if only the continuum source iscounted. If we also consider the BELR (whose size isestimated to be ∼0.09 pc)23 as the background source,rb would be ∼1200 pc (see Figure 13).

We can also place constraints on the absorber sizebased on partial covering analysis. Intrinsic NALs withlarge ejection velocities have partial coverage, althoughthey absorb only the continuum photons. This meanstheir physical scale is comparable to or smaller than thesize of the continuum source, dcloud ≤ 2.5×10−4 pc. Onthe other hand, broad PAL absorbers cover almost en-tirely both the BELR and the continuum source, whichmeans the size of the absorbers as a whole is compara-ble to or larger than the BELR size, ≥ 0.09 pc. NarrowPAL absorbers may consist of a number of small clumpyclouds because they show partial coverage.

In our VLT/UVES spectra, we detected high-ionization transitions like O VI and N V in the PALsystems, while the intrinsic NAL systems are in variousionization states, with or without high-ionization tran-sitions, as already noted in the literature (e.g., Misawaet al. 2007a; Ganguly et al. 2013). Some C IV NAL ab-sorbers also show N V transitions, while others do not.Because Lyα and N V in the PAL system are less vari-able than C IV and because N V has larger covering factorthan C IV, the cross-section of Lyα and N V should belarger than that of C IV. PAL absorbers with strong O VIlines are probably located much closer to the ionizing fluxsource than NAL absorbers, although this conclusion de-pends on the density of the absorber.

It is well known that both broad and narrow absorptionlines at zabs ∼ zem tend to vary (Wise et al. 2004; Misawaet al. 2014a), while NALs with large ejection velocitiesare rarely variable (e.g., Chen et al. 2015). Indeed, the

23 This is calculated by Misawa et al. (2013) based on the em-pirical equation in McLure & Dunlop (2004).

PALs in our spectra are variable, while most of the C IVNALs are probably stable between the three epochs, aswe observed for a few strong C IV NALs. Misawa etal. (2014a) suggest that broader absorption lines like thebroad PALs can vary mainly due to a change in theirionization state while narrow absorption lines like narrowPALs and NALs vary primarily due to the gas motiontransverse to our sightlines. The gas motion scenariocould explain why narrow PALs are variable but high-velocity NALs are not. Based on the dynamical model ofMurray et al. (1995) and more recent investigations (e.g.,Misawa et al. 2005; Hall et al. 2011), the absorbers atlarger distance from the center have small transverse (i.e.,orbital) velocities compared to their radial velocities. Ifintrinsic NALs lie at larger distances than narrow PALs,their small transverse velocities would rarely lead to timevariability, while the large transverse velocities of narrowPALs can lead to time variability more frequently.

6. SUMMARYIn this study, we performed a spectroscopic observa-

tion for images A and B of the gravitationally lensedquasar SDSS J1029+2623, and monitored the absorp-tion profiles in these spectra as well as in the previ-ous two observations. Using high quality spectra takenwith VLT/UVES, we detected intrinsic narrow absorp-tion lines (NALs) as well as broader, proximity absorp-tion lines (PALs), and fit models to the line profiles.Based on the results of our multi-sightline spectroscopy,we discuss a possible geometry and internal structure ofthe outflowing wind along our sightlines. Our main re-sults are as follows.

We detected 66 NALs, of which 24 are classifiedas intrinsic NALs (physically associated with thequasar) based on partial coverage analysis.

Class-A and B NALs cluster at vej ∼ 59,000,43,000, and 29,000 km s−1, which is reminiscent ofthe filamentary structures that are often obtainedin numerical simulations.

Our multiple sightline observation suggests thatthe size of the broad PAL absorbers are larger thanthe projected distance between sightlines (rθ) whilethe narrow PAL absorbers and intrinsic NAL ab-sorbers have sizes smaller than rθ if their radialdistances are greater than the boundary distance.

While PAL systems show only high ionization tran-sitions, including O VI, intrinsic NAL systems showa wide range of ionization conditions with andwithout low ionization transitions like O I, Al II,and Si II, as noted in the literature.

No class-A/B NALs (i.e., candidates for intrinsicNALs) have common absorption profiles in the twolensed images, which means that the outflow hasan internal velocity structure whose typical spatialscale is smaller than the physical distance betweenthe sightlines (i.e., ≤ rθ).

Short-time variation in the PALs is probably dueto a change in the ionization state of the gas. If thisis the case, we can place a lower limit on the gas

8 Misawa et al.

density as ne ≥ 8.7× 103 cm−3 and an upper limiton the absorber’s distance from the flux source asr ≤ 620 pc.

Based on our best knowledge, we present a possiblegeometry of the outflow along our two sightlines,in which we identify different structures in the out-flowing wind that can produce, respectively, broadPALs, narrow PALs, and NALs with large ejectionvelocities.

For further investigations of the outflow’s internalstructure, especially in the transverse direction, weshould perform the same observations for image C ofSDSS J1029+2623. We also aim to observe several lensedquasars with smaller separation angles (θ ∼ 2′′) by a sin-gle massive galaxy to examine ∼10 times finer internalstructure in an outflow, as already mentioned in Misawaet al. (2014b).

We thank the anonymous referee for comments thathelped us improve the paper. We would like to thank

Masamune Oguri, Poshak Gandhi, and Chris Cullitonfor their valuable comments. We also would like tothank Christopher Churchill for providing us with theminfit and search software packages. The researchwas supported by the Japan Society for the Promotionof Science through Grant-in-Aid for Scientific Research15K05020, JGC-S Scholarship Foundation, and partiallysupported by MEXT Grant-in-Aid for Scientific Researchon Innovative Areas (No. 15H05894). CS acknowledgessupport from CONICYT-Chile through Becas Chile74140006. FEB acknowledges support from CONICYT-Chile (Basal-CATA PFB-06/2007, FONDECYT Regular1141218, ”EMBIGGEN” Anillo ACT1101) and the Min-istry of Economy, Development, and Tourism’s Millen-nium Science Initiative through grant IC120009, awardedto The Millennium Institute of Astrophysics, MAS. JCCand ME acknowledge support from the National ScienceFoundation through award AST-1312686.

REFERENCES

Arav, N., Borguet, B., Chamberlain, C., Edmonds, D., &Danforth, C. 2013, MNRAS, 436, 3286

Benn, C. R., Carballo, R., Holt, J., et al. 2005, MNRAS, 360, 1455Barlow, T. A., & Sargent, W. L. W. 1997, AJ, 113, 136Bergeron, J., et al. 1994, ApJ, 436, 33Blandford, R.D. & Parne, D.G., 1982, MNRAS, 199, 883Bowler, R. A. A., Hewett, P. C., Allen, J. T., & Ferland, G. J.

2014, MNRAS, 445, 359Brotherton, M. S., Wills, B. J., Steidel, C. C., & Sargent,

W. L. W. 1994, ApJ, 423, 131Chartas, G., Eracleous, M., Dai, X., Agol, E., & Gallagher, S.

2007, ApJ, 661, 678Chen, Z.-F., Gu, Q.-S., Chen, Y.-M., & Cao, Y. 2015, MNRAS,

450, 3904Chelouche, D. 2003, ApJ, 596, L43Churchill, C. W., Vogt, S. S., & Charlton, J. C. 2003, AJ, 125, 98Churchill, C. W. 1997, Ph.D. Thesis, Univ. California, Santa CruzCorbin, M. R. 1990, ApJ, 357, 346Crenshaw, D. M., Kraemer, S. B., Boggess, A., Maran, S. P.,

Mushotzky, R. F., & Wu, C.-C. 1999, ApJ, 516, 750Gehrels, N. 1986, ApJ, 303, 336Dorodnitsyn, A. V. 2009, MNRAS, 393, 1433Dunn, J. P., Bautista, M., Arav, N., et al. 2010, ApJ, 709, 611Ellison, S. L., Ibata, R., Pettini, M., et al. 2004, A&A, 414, 79Elvis, M. 2000, ApJ, 545, 63Emmering, R.T., Bladford, R.D., & Shlosman, I., 1992, ApJ, 385,

460Everett, J. E., 2005, ApJ, 631, 689Fohlmeister, J., Kochanek, C. S., Falco, E. E., et al. 2013, ApJ,

764, 186Foltz, C., Wilkes, B., Weymann, R., & Turnshek, D. 1983, PASP,

95, 341Gabel, J.R., Arav, N., & Kim, T.-S. 2006, ApJ, 646, 742Ganguly, R., Lynch, R. S., Charlton, J. C., et al. 2013, MNRAS,

435, 1233Ganguly, R., Bond, N. A., Charlton, J. C., et al. 2001, ApJ, 549,

133Ganguly, R., Eracleous, M., Charlton, J. C., & Churchill, C. W.

1999, AJ, 117, 2594Gibson, R. R., Brandt, W. N., Schneider, D. P., & Gallagher,

S. C. 2008, ApJ, 675, 985Green, P. J. 2006, ApJ, 644, 733Hall, P. B., Brandt, W. N., Petitjean, P., et al. 2013, MNRAS,

434, 222Hall, P. B., Anosov, K., White, R. L., et al. 2011, MNRAS, 411,

2653Hamann, F., Chartas, G., McGraw, S., et al. 2013, MNRAS, 435,

133Hamann, F., Kanekar, N., Prochaska, J. X., et al. 2011, MNRAS,

410, 1957Hamann, F., Kaplan, K. F., Rodŕıguez Hidalgo, P., Prochaska,

J. X., & Herbert-Fort, S. 2008, MNRAS, 391, L39Hamann, F., & Sabra, B. 2004, AGN Physics with the Sloan

Digital Sky Survey, 311, 203

Hamann, F., Barlow, T. A., Junkkarinen, V., & Burbidge, E. M.1997a, ApJ, 478, 80

Hamann, F., Barlow, T. A., & Junkkarinen, V. 1997b, ApJ, 478,87

Hamann, F., Barlow, T. A., Beaver, E. A., et al. 1995, ApJ, 443,606

Inada, N., Oguri, M., Morokuma, T., et al. 2006, ApJ, 653, L97Jannuzi, B. T., Hartig, G. F., Kirhakos, S., et al. 1996, ApJ, 470,

L11Joshi, R., Chand, H., Srianand, R., & Majumdar, J. 2014,

MNRAS, 442, 862Konigl, A., & Kartje, J. F. 1994, ApJ, 434, 446Kurosawa, R. & Proga, D. 2009, ApJ, 693, 1929Lamy, H., & Hutsemékers, D. 2004, A&A, 427, 107Marziani, P., Sulentic, J. W., Dultzin-Hacyan, D., Calvani, M., &

Moles, M. 1996, ApJS, 104, 37McLure, R. J., & Dunlop, J. S. 2004, MNRAS, 352, 1390Misawa, T., Charlton, J. C., & Eracleous, M. 2014a, ApJ, 792, 77Misawa, T., Inada, N., Oguri, M., et al. 2014b, ApJ, 794, L20Misawa, T., Inada, N., Ohsuga, K., et al. 2013, AJ, 145, 48Misawa, T., Kawabata, K. S., Eracleous, M., Charlton, J. C., &

Kashikawa, N. 2010, ApJ, 719, 1890Misawa, T., Charlton, J. C., Eracleous, M., Ganguly, R., Tytler,

D., Kirkman, D., Suzuki, N., & Lubin, D. 2007a, ApJS, 171, 1Misawa, T., Eracleous, M., Charlton, J. C., & Kashikawa, N.

2007b, ApJ, 660, 152Misawa, T., Eracleous, M., Charlton, J.C., & Tajitsu, A., 2005,

ApJ, 629, 115Murray, N., Chiang, J., Grossman, S. A., & Voit, G. M., 1995,

ApJ, 451, 498Muzahid, S., Srianand, R., Charlton, J., & Eracleous, M. 2015,

arXiv:1509.07850Muzahid, S., Srianand, R., Arav, N., Savage, B. D., & Narayanan,

A. 2013, MNRAS, 431, 2885Narayanan, D., Hamann, F., Barlow, T., Burbidge, E.M., Cohen,

R.D., Junkkaribeb, V., & Lyons, R., 2004, ApJ, 601, 715Nestor, D., Hamann, F., & Rodriguez Hidalgo, P. 2008, MNRAS,

386, 2055Nomura, M., Ohsuga, K., Wada, K., Susa, H., & Misawa, T.

2013, PASJ, 65,Oguri, M., Schrabback, T., Jullo, E., et al. 2012, arXiv:1209.0458Oguri, M., Ofek, E. O., Inada, N., et al. 2008, ApJ, 676, L1Ota, N., Oguri, M., Dai, X., et al. 2012, arXiv:1202.1645Petitjean, P. & Srianand, R., 1999, A&A, 345, 73Proga, D., & Kallman, T. R. 2004, ApJ, 616, 688Proga, D., Stone, J. M., & Kallman, T. R., 2000, ApJ, 543, 686Sargent, W.L.W., Boksenberg, A., & Steidel, C.C., 1988, ApJS,

68, 539Springel, V., Di Matteo, T., & Hernquist, L. 2005, ApJ, 620, L79Srianand, R. & Petitjean, P., 2000, A&A, 357, 414Takeuchi, S., Ohsuga, K., & Mineshige, S. 2013, PASJ, 65, 88Tytler, D., & Fan, X.-M. 1992, ApJS, 79, 1Tytler, D., Boksenberg, A., Sargent, W. L. W., Young, P., &

Kunth, D., 1987, ApJS, 64, 667

Multi-Sightline Observation of NALs 9

Vestergaard, M., 2003, ApJ, 599, 116Wampler, E. J., Chugai, N. N., & Petitjean, P. 1995, ApJ, 443,

586Weymann, R. J., Morris, S. L., Foltz, C. B., & Hewett, P. C.

1991, ApJ, 373, 23Weymann, R. J., Williams, R. E., Peterson, B. M., & Turnshek,

D. A. 1979, ApJ, 234, 33

Wise, J. H., Eracleous, M., Charlton, J. C., & Ganguly, R. 2004,ApJ, 613, 129

Young, P. J., Sargent, W. L. W., Boksenberg, A., Carswell, R. F.,& Whelan, J. A. J., 1979, ApJ, 229, 891

10 Misawa et al.

TABLE 1Log of Observations

Target Obs Date Instrument R Texp S/Na Referenceb

(sec) (pix−1)

SDSS J1029+2623 A 2010 Feb 10 Subaru/HDS 30000 14400 13 12014 Jan 28 – Feb 3 VLT/UVES 33000 26670 23 2

2014 Apr 4 Subaru/HDS 36000 11400 14 3SDSS J1029+2623 B 2010 Feb 10 Subaru/HDS 30000 14200 13 1

2014 Feb 4 – 26 VLT/UVES 33000 26670 23 22014 Apr 4 Subaru/HDS 36000 11400 14 3

a Signal to noise ratio at λobs ∼ 4700Å.b References. (1) Misawa et al. 2013, (2) This paper, (3) Misawa et al. 2014b.

Multi-Sightline Observation of NALs 11TA

BLE

2A

bso

rptio

nSyst

ems

wit

hD

oublet

Lin

es

Image

AIm

age

BIo

nz a

bs

vej

EW

rest

ba

EW

rest

rb

IDC

lass

-1c

Cla

ss-2

dIo

nz a

bs

vej

EW

rest

ba

EW

rest

rb

IDC

lass

-1c

Cla

ss-2

d∆

ve

(km

s−1)

(Å)

(Å)

(km

s−1)

(Å)

(Å)

(km

s−1)

Mg

II0.5

111

190431

0.6

04±

0.0

08

0.4

95±

0.0

08

27,2

9C

1L

Mg

II0.5

124

190276

0.0

95±

0.0

04

0.0

52±

0.0

04

31,3

2C

3L

0.0

Mg

II0.5

125

190265

0.2

21±

0.0

05

0.1

45±

0.0

04

28,3

0C

1L

+19.8

Mg

II0.6

731

171003

1.6

89±

0.0

08

1.4

44±

0.0

08

49,5

0C

1L

Mg

II0.9

176

141245

0.1

93±

0.0

02

0.1

20±

0.0

02

61,6

3C

3L

0.0

Mg

II0.9

184

141157

0.1

86±

0.0

05

0.0

92±

0.0

05

67,6

8C

1L

+125.1

Mg

II0.9

187

141116

1.4

72±

0.0

06

1.3

66±

0.0

06

62,6

4C

1+

172.0

CIV

1.6

085

60206

0.0

35±

0.0

02

0.0

26±

0.0

02

7,1

0B

2H

CIV

1.6

149

59493

0.1

05±

0.0

04

≤0.1

32

12,1

3B

2IH

0.0

CIV

1.6

151

59472

0.3

08±

0.0

03

0.2

97±

0.0

03

8,1

1B

2,H

IH+

22.9

CIV

1.6

160

59378

0.0

47±

0.0

03

≤0.0

33

9,1

2C

2+

126.2

CIV

1.6

221

58699

≤0.2

84

0.1

44±

0.0

02

16,2

0C

2,H

H0.0

CIV

1.6

230

58601

≤0.5

12

0.1

90±

0.0

07

16,1

8B

2,H

H+

103.0

CIV

1.6

253

58350

0.0

21±

0.0

02

0.0

14±

0.0

02

18,2

2B

2H

CIV

1.6

556

55033

0.1

33±

0.0

04

0.0

81±

0.0

03

25,2

6C

1H

CIV

1.6

924

51029

0.0

84±

0.0

02

0.0

46±

0.0

02

27,2

8C

2H

0.0

CIV

1.6

930

50965

0.4

98±

0.0

05

0.1

85±

0.0

07

23,2

4C

2,H

LIH

+66.8

CIV

1.7

065

49509

0.0

42±

0.0

02

0.0

32±

0.0

02

29,3

0A

2H

CIV

1.7

212

47932

0.4

93±

0.0

05

0.0

53±

0.0

04

25,2

6C

2,H

HC

IV1.7

617

43596

0.0

10±

0.0

02

0.0

11±

0.0

02

33,3

6C

3LIH

0.0

CIV

1.7

627

43496

0.0

94±

0.0

02

0.0

66±

0.0

02

34,3

7B

1+

108.6

CIV

1.7

650

43250

0.0

41±

0.0

02

0.0

22±

0.0

02

35,3

8A

2H

0.0

CIV

1.7

652

43224

0.0

47±

0.0

02

0.0

35±

0.0

02

31,3

2A

2LIH

+21.7

CIV

1.8

909

30091

0.3

55±

0.0

04

0.3

09±

0.0

03

35,3

7C

1,H

LIH

0.0

CIV

1.8

909

30088

0.4

56±

0.0

06

≤0.5

01

41,4

3C

2,H

IH0.0

SiIV

1.8

909

30088

0.1

81±

0.0

03

≤0.2

68

5,1

0B

2,H

0.0

SiIV

1.8

909

30087

0.1

21±

0.0

04

0.1

19±

0.0

04

5,1

4B

20.0

SiIV

1.8

944

29737

0.3

80±

0.0

05

≤0.3

83

6,1

7C

2,H

LIH

0.0

CIV

1.8

945

29723

≤0.6

69

≤0.5

92

42,4

4C

2,H

+10.4

CIV

1.8

949

29686

0.4

13±

0.0

04

0.3

18±

0.0

04

36,3

9C

2,H

LIH

+51.8

SiIV

1.8

949

29685

0.1

37±

0.0

03

≤0.4

64

6,1

4B

2+

51.8

SiIV

1.8

975

29413

0.0

32±

0.0

02

0.0

19±

0.0

02

8,1

9B

2LIH

0.0

CIV

1.8

982

29343

0.1

93±

0.0

04

0.1

46±

0.0

05

38,4

0C

1,H

IH+

72.5

SiIV

1.8

982

29342

0.0

29±

0.0

03

0.0

13±

0.0

02

7,1

7C

2+

72.5

SiIV

1.9

016

28991

0.1

62±

0.0

03

0.1

42±

0.0

02

9,2

1C

1,H

LIH

0.0

CIV

1.9

019

28964

0.4

57±

0.0

05

0.3

63±

0.0

06

45,4

6C

1,H

+31.0

SiIV

1.9

024

28917

0.0

23±

0.0

02

0.0

13±

0.0

02

11,2

3B

2+

82.7

CIV

1.9

115

27981

0.5

72±

0.0

08

0.4

48±

0.0

07

41,4

3C

1,H

LIH

0.0

SiIV

1.9

115

27987

0.4

39±

0.0

04

0.3

18±

0.0

04

13,1

9C

1,H

0.0

SiIV

1.9

118

27949

≤0.2

48

≤0.1

97

15,2

4C

2,H

+30.9

CIV

1.9

119

27946

0.8

37±

0.0

09

0.6

79±

0.0

09

47,4

8C

1,H

LIH

+41.2

CIV

1.9

138

27745

0.4

84±

0.0

09

0.3

59±

0.0

09

42,4

4C

3,H

+236.9

SiIV

1.9

141

27718

≤0.4

61

0.2

05±

0.0

05

15,2

0B

2,H

LIH

+267.8

SiIV

1.9

420

24880

0.0

41±

0.0

02

0.0

19±

0.0

03

21,2

2C

2LI

CIV

2.1

081

8460

0.2

09±

0.0

04

0.1

10±

0.0

04

51,5

2B

1,H

IH0.0

CIV

2.1

084

8426

0.1

30±

0.0

02

0.0

74±

0.0

02

45,4

6C

2IH

+29.0

CIV

2.1

196

7348

≤0.0

65

0.0

24±

0.0

01

47,4

8C

2LH

0.0

CIV

2.1

209

7223

0.5

52±

0.0

03

0.4

70±

0.0

03

53,5

4C

1,H

LIH

+125.0

CIV

2.1

270

6643

0.1

45±

0.0

01

0.1

14±

0.0

01

49,5

0C

1H

0.0

CIV

2.1

276

6585

0.0

29±

0.0

02

0.0

14±

0.0

02

55,5

7C

2IH

+57.5

CIV

2.1

285

6500

0.1

83±

0.0

02

0.1

55±

0.0

02

56,5

8C

1,H

+143.9

CIV

2.1

349

5884

0.0

93±

0.0

03

0.0

46±

0.0

02

51,5

2C

1H

0.0

CIV

2.1

349

5886

0.0

73±

0.0

02

0.0

32±

0.0

02

59,6

0C

2H

0.0

12 Misawa et al.TA

BLE

2A

bso

rptio

nSyst

ems

wit

hD

oublet

Lin

es

CIV

2.1

800

1604

0.0

11±

0.0

01

0.0

08±

0.0

01

53,5

5C

2IH

0.0

SiIV

2.1

815

1458

0.1

15±

0.0

06

0.0

41±

0.0

05

33,3

4C

1+

141.5

CIV

2.1

818

1428

0.2

98±

0.0

02

0.2

09±

0.0

02

61,6

2C

1,H

LIH

+169.7

CIV

2.1

819

1416

0.4

25±

0.0

04

≤0.2

97

54,5

6C

1,H

+179.2

SiIV

2.1

821

1403

0.0

88±

0.0

02

0.0

62±

0.0

03

39,4

0C

1+

198.0

CIV

2.1

897

686

0.8

48±

0.0

03

≤1.4

05

63,6

4A

1,H

H0.0

CIV

2.1

898

676

0.9

02±

0.0

03

≤1.2

53

57,5

8A

1,H

H+

9.4

NV

2.1

900

657

0.9

45±

0.0

04

≤0.8

82

1,3

A1,H

+28.2

NV

2.1

900

658

0.8

88±

0.0

04

≤0.7

17

1,3

A1,H

+28.2

NV

2.1

955

141

≤1.4

94

1.3

41±

0.0

04

2,4

A1,H

H0.0

NV

2.1

955

141

≤1.2

31

1.2

01±

0.0

03

2,4

A1,H

H0.0

CIV

2.1

957

122

1.5

69±

0.0

03

1.2

24±

0.0

03

65,6

6A

1,H

+18.8

CIV

2.1

958

113

≤1.5

90

1.3

03±

0.0

03

59,6

0A

1,H

+28.2

aR

est-

fram

eeq

uiv

ale

nt

wid

thofblu

eco

mponen

tofdouble

tline.

bR

est-

fram

eeq

uiv

ale

nt

wid

thofre

dco

mponen

tofdouble

tline.

cR

elia

bility

class

ofin

trin

sic

lines

base

don

part

ialco

ver

age

analy

sis,

defi

ned

inM

isaw

aet

al.

(2007a).

Ifan

equiv

ale

nt

wid

this

larg

een

ough,w

ecl

ass

ify

the

double

tashom

ogen

eous

sam

ple

with

am

ark

of“H

”.

dIo

niz

ation

class

ofabso

rption

syst

emw

ith

hig

h(H

;IP

>60

eV),

inte

rmed

iate

(I;IP

=35

–50

eV),

and/or

low

(L;IP

<25

eV)

ioniz

ation

transi

tions.

eVel

oci

tydiff

eren

cefr

om

the

firs

tdouble

tin

each

abso

rption

syst

em.

Multi-Sightline Observation of NALs 13TA

BLE

3Lin

eParameters

of

Narrow

Abso

rptio

nC

omponents

λobsa

z absb

vejc

log

Nσ(v

)/bd

Oth

erIo

n(Å

)(k

ms−

1)

(cm

−2)

(km

s−1)

Cfe

Ionsf

Image

A

Mg

II4225.6

0.5

111

190431

47.3

Mg

Iλ2853,Fe

IIλ2600

(Fe

IIλ2383)

0.5

108

190464

12.9

6±

0.0

65.1±

0.2

0.9

3+

0.0

7−

0.0

6

0.5

112

190426

13.5

1±

0.0

510.5±

0.5

1.0

0+

0.0

6−

0.0

6

0.5

112

190423

13.6

9±

0.3

337.1±

7.2

0.1

9+

0.1

0−

0.0

90.5

113

190408

12.6

0±

0.0

27.6±

0.3

1.0

0

Mg

II4229.5

0.5

125

190265

17.9

Mg

Iλ2853,Fe

IIλ2600

(Fe

IIλ2344,Fe

IIλ2383)

0.5

125

190269

12.6

9±

0.0

26.0±

0.4

1.0

00.5

126

190261

12.7

0±

0.2

713.3±

4.2

0.8

3+

0.4

6−

0.2

8

Mg

II5376.1

0.9

176

141245

14.1

Mg

Iλ2853,Fe

IIλ2600

(AlII

Iλ1863)

0.9

176

141246

12.9

5±

0.0

24.7±

0.2

1.0

0

Mg

II5365.3

0.9

187

141116

68.1

Mg

Iλ2853,Fe

IIλ2344,2

600

(AlII

Iλ1863)

0.9

181

141190

12.5

0±

0.0

24.1±

0.3

1.0

00.9

185

141146

14.0

3±

0.0

216.9±

0.2

1.0

00.9

189

141094

14.9

4±

0.0

722.0±

0.5

0.9

9+

0.0

3−

0.0

3

CIV

4048.7

1.6

151

59472

26.0

SiIV

λ1394,1

403

1.6

151

59473

17.3

3±

0.0

68.7±

0.1

1.0

0+

0.0

4−

0.0

4

CIV

4050.1

1.6

160

59378

13.1

1.6

160

59379

13.6

6±

0.1

510.3±

1.1

0.4

2+

0.0

8−

0.0

8

CIV

4067.7

1.6

230

58601

38.3

1.6

229

58617

13.8

4±

0.0

213.0±

0.8

1.0

01.6

231

58590

13.9

3±

0.0

65.4±

0.6

1.0

01.6

236

58538

12.8

8±

0.0

99.7±

3.0

1.0

0

CIV

4169.3

1.6

930

50965

58.4

SiII

λ1527,A

lII

λ1671

(OI

λ1302,SiII

λ1260,C

IIλ1335,SiIV

λ1394)

1.6

928

50991

13.6

9±

0.0

239.0±

1.8

1.0

01.6

934

50925

13.5

2±

0.0

217.9±

0.7

1.0

0

CIV

4212.9

1.7

212

47932

41.9

Lyα

1.7

209

47955

13.2

5±

0.0

211.1±

0.7

1.0

01.7

214

47908

13.1

6±

0.0

29.2±

0.7

1.0

0

CIV

4281.1

1.7

652

43224

10.0

Lyα(O

Iλ1302,SiII

λ1207,SiIV

λ1394)

1.7

652

43223

13.7

1±

0.1

16.1±

0.5

0.5

8+

0.0

6−

0.0

6

CIV

4475.7

1.8

909

30091

30.8

Lyα,SiII

λ1527,A

lII

λ1671,(S

iII

λ1190,1

192,1

260,C

IIλ1335,SiII

Iλ1207)

1.8

909

30090

14.8

3±

0.0

520.9±

0.5

0.9

6+

0.0

5−

0.0

5

SiIV

4029.3

1.8

909

30088

27.1

Lyα,SiII

λ1527,A

lII

λ1671,(S

iII

λ1190,1

192,1

260,C

IIλ1335,SiII

Iλ1207)

1.8

906

30121

12.4

8±

0.1

12.8±

0.9

1.0

01.8

907

30111

12.4

9±

0.1

48.8±

3.1

1.0

01.8

910

30081

13.5

6±

0.0

211.2±

0.4

1.0

0

CIV

4481.8

1.8

949

29686

36.5

Lyα(C

IIλ1335,SiII

Iλ1207,N

Vλ1239,1

243)

1.8

945

29728

12.8

5±

0.2

67.7±

3.0

1.0

0

14 Misawa et al.TA

BLE

3Lin

eParameters

of

Narrow

Abso

rptio

nC

omponents

1.8

947

29703

14.1

4±

0.2

212.5±

4.2

1.0

01.8

949

29684

14.5

7±

0.2

39.0±

1.3

1.0

01.8

952

29654

13.7

4±

0.0

28.7±

0.4

1.0

0

SiIV

4034.7

1.8

949

29685

38.2

Lyα(C

IIλ1335,SiII

Iλ1207,N

Vλ1239,1

243)

1.8

949

29684

16.1

7±

0.0

52.2±

0.4

1.0

0

CIV

4487.0

1.8

982

29343

22.5

Lyα

1.8

981

29353

13.8

7±

0.0

711.9±

1.0

0.8

8+

0.0

7−

0.0

71.8

983

29332

13.4

5±

0.0

212.8±

0.6

1.0

0

SiIV

4039.4

1.8

982

29342

19.9

Lyα

1.8

980

29360

11.9

5±

0.0

94.8±

1.6

1.0

01.8

983

29337

12.4

0±

0.0

49.3±

1.2

1.0

0

CIV

4507.6

1.9

115

27981

48.8

Lyα,SiII

λ1527,A

lII

λ1671

(OI

λ1302,SiII

λ1190,1

193,1

260,SiII

Iλ1207)

1.9

115

27984

14.7

9±

0.0

436.0±

0.9

0.9

8+

0.0

8−

0.0

8

SiIV

4057.9

1.9

115

27987

44.5

Lyα,SiII

λ1527,A

lII

λ1671

(OI

λ1302,SiII

λ1190,1

193,1

260,SiII

Iλ1207)

1.9

112

28017

13.4

3±

0.0

413.7±

1.2

1.0

01.9

114

27995

13.3

5±

0.0

47.5±

0.7

1.0

01.9

115

27979

13.4

6±

0.0

346.9±

2.4

1.0

01.9

116

27973

13.8

4±

0.0

712.3±

1.4

0.9

6+

0.0

6−

0.0

6

CIV

4511.2

1.9

138

27745

64.0

Lyα,SiII

λ1527,A

lII

λ1671,C

IIλ1335

(SiII

λ1190,1

193,1

260,SiII

Iλ1207,N

Vλ1239,1

243)

1.9

133

27797

14.1

0±

0.0

823.1±

1.4

0.6

9+

0.0

9−

0.0

81.9

141

27716

14.3

4±

0.0

221.2±

0.6

1.0

0

SiIV

4087.8

1.9

141

27718

17.2

Lyα,SiII

λ1527,A

lII

λ1671,C

IIλ1335

(SiII

λ1190,1

193,1

260,SiII

Iλ1207,N

Vλ1239,1

243)

1.9

141

27720

13.6

1±

0.0

210.9±

0.4

1.0

01.9

143

27694

12.3

8±

0.0

94.9±

1.7

1.0

0

SiIV

4100.4

1.9

420

24880

13.2

Lyα,SiII

λ1527,A

lII

λ1671,C

IIλ1335

(OI

λ1302,SiII

λ1190,1

193,1

260,SiII

Iλ1207)

1.9

420

24879

12.7

0±

0.0

210.2±

0.7

1.0

0

CIV

4812.5

2.1

084

8426

21.2

Lyα,SiIV

λ1394

(SiII

Iλ1207)

2.1

083

8435

13.4

6±

0.0

311.8±

0.7

1.0

02.1

086

8415

13.1

9±

0.0

610.8±

1.0

1.0

0

CIV

4829.8

2.1

196

7348

17.1

Lyα,C

IIλ1335

2.1

196

7349

13.6

3±

0.1

413.9±

1.0

0.4

3+

0.1

2−

0.0

9

CIV

4841.2

2.1

270

6643

16.4

Lyα(N

Vλ1239,1

243)

2.1

270

6644

13.6

8±

0.1

014.5±

0.6

0.9

0+

0.1

2−

0.1

22.1

270

6638

13.8

6±

0.0

25.2±

0.1

1.0

0

CIV

4853.4

2.1

349

5884

35.3

Lyα

2.1

347

5903

13.0

3±

0.0

220.9±

1.5

1.0

02.1

351

5866

13.3

4±

0.1

811.3±

0.8

0.7

0+

0.4

5−

0.2

4

CIV

4923.2

2.1

800

1604

14.3

Lyα

2.1

799

1606

13.4

8±

0.3

510.2±

2.3

0.1

5+∞

−0.0

8

SiIV

4434.2

2.1

815

1458

73.3

Lyα(S

iII

Iλ1207)

2.1

810

1502

12.7

2±

0.0

35.8±

0.8

1.0

0

Multi-Sightline Observation of NALs 15TA

BLE

3Lin

eParameters

of

Narrow

Abso

rptio

nC

omponents

2.1

818

1427

12.5

4±

0.0

719.5±

4.4

1.0

02.1

822

1395

12.9

4±

0.2

53.0±

1.1

0.7

2+

0.2

3−

0.2

3

CIV

4926.3

2.1

819

1416

34.0

Lyα(S

iII

Iλ1207)

2.1

817

1436

13.6

9±

0.0

916.5±

1.7

0.8

8+

0.1

0−

0.1

02.1

820

1409

13.8

9±

0.0

129.1±

0.3

1.0

02.1

822

1394

13.6

1±

0.0

16.0±

0.3

1.0

0

CIV

g4944.5

2.1

937

347

400.1

Lyα,SiIV

λ1394,O

VI

λ1032,1

038

2.1

881

878

13.2

4±

0.0

125.6±

0.8

1.0

02.1

886

831

14.4

9±

0.0

711.6±

0.6

0.4

3+

0.0

2−

0.0

2

2.1

889

795

14.2

7±

0.2

38.6±

1.3

0.3

8+

0.0

2−

0.0

2

2.1

892

773

13.7

7±

0.2

011.1±

1.4

0.5

2+

0.0

8−

0.0

8

2.1

894

748

14.3

0±

0.1

77.8±

1.0

0.4

4+

0.0

2−

0.0

2

2.1

900

693

14.4

9±

0.0

225.3±

0.6

0.8

4+

0.0

2−

0.0

2

2.1

909

613

14.0

4±

0.0

320.3±

0.5

0.9

8+

0.0

3−

0.0

32.1

914

567

13.3

1±

0.0

116.6±

0.3

1.0

02.1

943

287

14.5

3±

0.0

217.8±

0.2

0.9

1+

0.0

2−

0.0

2

2.1

947

254

14.8

4±

0.1

810.0±

1.0

0.5

2+

0.0

2−

0.0

2

2.1

953

201

14.6

3±

0.0

222.5±

0.7

0.8

9+

0.0

2−

0.0

2

2.1

961

124

14.8

8±

0.0

329.4±

0.9

0.8

1+

0.0

2−

0.0

2

2.1

968

61

14.4

8±

0.0

225.9±

0.6

0.8

4+

0.0

2−

0.0

2

2.1

975

−11

14.2

8±

0.0

219.6±

0.5

0.4

6+

0.0

2−

0.0

2

2.1

981

−63

13.9

1±

0.0

314.9±

0.6

0.4

7+

0.0

3−

0.0

3

2.1

985

−100

14.0

9±

0.0

911.5±

1.0

0.1

8+

0.0

2−

0.0

2

NV

h3956.5

2.1

937

343

428.4

Lyα,SiIV

λ1394,O

VI

λ1032,1

038

2.1

879

889

13.3

8±

0.0

223.8±

1.2

1.0

02.1

885

834

13.8

1±

0.1

317.4±

1.0

0.8

5+

0.1

8−

0.1

7

2.1

893

762

14.2

6±

0.0

442.5±

1.7

0.9

7+

0.0

6−

0.0

6

2.1

900

691

14.8

8±

0.0

320.6±

0.5

0.8

8+

0.0

3−

0.0

3

2.1

909

614

14.5

5±

0.0

223.5±

0.5

0.9

7+

0.0

3−

0.0

32.1

914

567

13.9

0±

0.0

115.0±

0.2

1.0

02.1

944

285

14.8

3±

0.0

222.1±

0.4

0.9

7+

0.0

3−

0.0

3

2.1

953

201

14.9

2±

0.0

141.1±

1.6

0.9

7+

0.0

3−

0.0

3

2.1

961

125

15.4

5±

0.3

314.9±

1.8

0.8

4+

0.0

4−

0.0

4

2.1

966

75

15.2

1±

0.0

526.7±

0.8

0.9

2+

0.0

3−

0.0

3

2.1

976

−14

14.3

8±

0.0

425.6±

1.4

0.6

2+

0.0

3−

0.0

3

2.1

981

−61

14.2

1±

0.0

910.9±

1.2

0.5

3+

0.0

6−

0.0

62.1

984

−91

13.6

6±

0.0

126.3±

0.7

1.0

0

Image

B

Mg

II4229.3

0.5

124

190276

11.7

Fe

IIλ2600

0.5

124

190275

12.6

8±

0.1

811.3±

0.9

0.6

9+

0.2

9−

0.2

0

Mg

II4678.6

0.6

731

171003

78.2

Mg

Iλ2853,Fe

IIλ2344,2

383,2

600

0.6

725

171070

13.0

2±

1.0

32.4±

1.9

0.3

1+

0.1

0−

0.0

9

0.6

726

171059

13.4

9±

0.1

47.5±

1.7

0.9

2+

0.0

8−

0.0

8

16 Misawa et al.TA

BLE

3Lin

eParameters

of

Narrow

Abso

rptio

nC

omponents

0.6

727

171047

12.8

3±

0.5

216.5±

18.7

1.0

0+

0.2

1−

0.1

8

0.6

729

171032

13.0

3±

0.2

48.1±

2.4

1.0

0+

0.0

6−

0.0

60.6

730

171018

13.2

7±

0.0

112.0±

0.3

1.0

00.6

731

171008

14.2

9±

9.7

93.2±

9.0

0.9

8+

0.0

6−

0.0

60.6

732

170994

13.2

9±

0.0

116.5±

0.4

1.0

00.6

733

170980

13.6

7±

0.1

620.7±

5.0

1.0

0+

0.0

5−

0.0

50.6

735

170955

13.2

9±

0.0

117.6±

0.3

1.0

00.6

736

170947

13.3

8±

0.4

46.0±

2.0

0.5

5+

0.1

4−

0.1

4

Mg

II5364.4

0.9

184

141157

33.1

0.9

181

141186

12.4

5±

0.5

99.1±

2.0

0.2

9+∞

−0.3

2

0.9

184

141158

12.6

5±

0.1

29.6±

1.0

0.6

8+

0.1

8−

0.1

30.9

185

141142

12.2

2±

0.0

211.8±

0.7

1.0

0

CIV

4038.4

1.6

085

60206

11.2

1.6

085

60206

13.5

7±

0.1

88.0±

0.9

0.4

4+

0.1

0−

0.1

0

CIV

4048.4

1.6

149

59493

20.6

(SiIV

λ1394,1

403)

1.6

150

59491

12.3

8±

0.0

82.5±

4.5

1.0

01.6

150

59487

14.0

4±

0.0

69.2±

0.4

0.7

1+

0.0

6−

0.0

6

CIV

4066.3

1.6

221

58699

19.0

1.6

222

58699

14.1

2±

0.0

213.4±

0.4

1.0

0

CIV

4064.5

1.6

253

58350

13.7

1.6

253

58348

13.5

7±

0.2

46.0±

0.9

0.2

8+

0.0

8−

0.0

8

CIV

4111.4

1.6

556

55033

22.1

1.6

556

55033

13.8

2±

0.0

614.7±

0.5

0.8

5+

0.0

7−

0.0

6

CIV

4168.4

1.6

924

51029

14.7

1.6

925

51027

13.4

2±

0.0

111.4±

0.5

1.0

0

CIV

4190.2

1.7

065

49509

11.7

1.7

065

49508

13.6

8±

0.1

47.8±

0.8

0.4

6+

0.0

7−

0.0

6

CIV

4275.7

1.7

617

43596

11.1

Lyα

1.7

617

43596

14.1

7±

9.9

94.5±

25.6

0.1

5+

0.1

1−

0.1

1

CIV

4277.1

1.7

627

43496

20.9

Lyα,SiII

λ1527

(SiII

λ1260,C

IIλ1335,SiII

Iλ1207,SiIV

λ1394,1

403)

1.7

626

43503

13.7

8±

0.0

77.6±

0.7

0.8

3+

0.0

7−

0.0

6

1.7

628

43479

13.9

2±

0.3

722.0±

16.2

0.1

7+∞

−0.0

7

CIV

4280.7

1.7

650

43250

15.8

Lyα

1.7

650

43249

13.6

4±

0.2

011.7±

1.4

0.3

6+

0.1

2−

0.1

0

CIV

4475.7

1.8

909

30088

48.6

Lyα(S

iII

Iλ1207)

1.8

904

30140

13.4

2±

0.0

311.4±

0.8

1.0

01.8

910

30084

14.3

1±

0.0

625.0±

2.3

1.0

01.8

913

30055

13.1

5±

0.5

910.4±

6.0

1.0

01.8

915

30027

13.0

0±

0.0

811.3±

2.2

1.0

0

SiIV

i4029.3

1.8

909

30087

38.2

Lyα(S

iII

Iλ1207)

Multi-Sightline Observation of NALs 17TA

BLE

3Lin

eParameters

of

Narrow

Abso

rptio

nC

omponents

SiIV

4034.0

1.8

944

29737

76.4

Lyα,SiII

λ1527,A

lII

λ1671

(SiII

λ1190,1

193,1

260,C

IIλ1335,SiII

Iλ1207)

1.8

940

29771

13.8

7±

0.0

222.5±

0.4

1.0

01.8

947

29700

12.3

3±

0.0

810.1±

2.1

1.0

01.8

950

29675

12.9

2±

0.0

211.3±

0.7

1.0

01.8

954

29629

12.7

0±

0.0

312.6±

0.9

1.0

0

CIV

4488.7

1.8

945

29723

75.2

Lyα,SiII

λ1527,A

lII

λ1671

(SiII

λ1190,1

193,1

260,C

IIλ1335,SiII

Iλ1207)

1.8

939

29785

14.1

4±

0.0

916.9±

1.9

1.0

01.8

942

29753

14.5

1±

0.0

518.7±

1.4

1.0

01.8

946

29708

13.2

3±

0.1

47.7±

1.7

1.0

01.8

949

29682

14.2

5±

0.0

316.6±

1.2

1.0

01.8

953

29638

13.7

6±

0.0

731.2±

5.3

1.0

0

SiIV

4038.4

1.8

975

29413

11.2

Lyα,C

IVλ1548,1

551

(SiII

λ1260,SiII

Iλ1207)

1.8

975

29414

13.0

7±

0.2

07.4±

0.9

0.5

0+

0.1

6−

0.1

3

SiIV

4044.2

1.9

016

28991

18.5

Lyα,SiII

λ1527,A

lII

λ1671

(SiII

λ1190,1

193,1

260,C

IIλ1335,SiII

Iλ1207)

1.9

016

28992

13.8

7±

0.0

412.8±

0.3

0.9

2+

0.0

5−

0.0

5

CIV

4492.7

1.9

019

28964

52.4

Lyα,SiII

λ1527,A

lII

λ1671

(SiII

λ1190,1

193,1

260,C

IIλ1335,SiII

Iλ1207)

1.9

017

28988

14.5

1±

0.0

321.6±

0.5

0.9

6+

0.0

7−

0.0

7

1.9

024

28918

13.8

7±

0.0

512.5±

0.4

0.9

7+

0.0

8−

0.0

7

SiIV

4045.2

1.9

024

28917

11.1

Lyα,C

IVλ1548,1

551

1.9

023

28918

13.1

9±

0.2

56.8±

1.2

0.3

1+

0.1

0−

0.0

9

SiIV

4058.4

1.9

118

27949

26.7

Lyα