The Next Generation Virgo Cluster Survey. XIX. Tomography of Milky Way Substructures in the NGVS Footprint

2.6 MB·

by  Deborah Lokhorst

To appear in The Astrophysical Journal
PreprinttypesetusingLATEXstyleemulateapjv.01/23/15
THE NEXT GENERATION VIRGO CLUSTER SURVEY. XIX. TOMOGRAPHY OF MILKY WAY
SUBSTRUCTURES IN THE NGVS FOOTPRINT
Deborah Lokhorst1,10, Else Starkenburg1,2,11, Alan W. McConnachie3, Julio F. Navarro1,12, Laura Ferrarese3,
Patrick Coˆte´3, Chengze Liu4,5, Eric W. Peng6,7, Stephen D.J. Gwyn3, Jean-Charles Cuillandre8, Puragra
Guhathakurta9
To appear in The Astrophysical Journal
ABSTRACT
TheNextGenerationVirgoClusterSurvey(NGVS)isadeepu∗giz surveytargetingtheVirgocluster
6
1 of galaxies at 16.5 Mpc. This survey provides high-quality photometry over an ∼ 100 deg2 region
0 straddling the constellations of Virgo and Coma Berenices. This sightline through the Milky Way is
2 noteworthy in that it intersects two of the most prominent substructures in the Galactic halo: the
VirgoOver-Density(VOD)andSagittariusstellarstream(closetoitsbifurcationpoint). Inthispaper,
n
weusedeepu∗giimagingfromtheNGVStoperformtomographyoftheVODandSagittariusstream
a
using main-sequence turnoff (MSTO) stars as a halo tracer population. The VOD, whose centroid
J
is known to lie at somewhat lower declinations (α ∼ 190◦, δ ∼ −5◦) than is covered by the NGVS,
0
is nevertheless clearly detected in the NGVS footprint at distances between ∼ 8 and 25 kpc. By
2
contrast, the Sagittarius stream is found to slice directly across the NGVS field at distances between

25 and 40 kpc, with a density maximum at (cid:39) 35 kpc. No evidence is found for new substructures
]
A beyond the Sagittarius stream, at least out to a distance of ∼ 90 kpc — the largest distance to which
we can reliably trace the halo using MSTO stars. We find clear evidence for a distance gradient in
G
the Sagittarius stream across the ∼ 30 deg of sky covered by the NGVS and its flanking fields. We
h. compare our distance measurements along the stream to those predicted by leading stream models.
p Keywords: Galaxy: halo – Galaxy: stellar content – Galaxy: structure – Local Group
-
o
r
t 1. INTRODUCTION facie evidence that the assembly of the halo is an ongo-
s ingprocess. Whilesomeofthesestreamsareclearlystill
a Large-scale stellar mapping of the Milky Way by
attached to their progenitor systems (e.g. the Sagittar-
[ panoramic optical and infrared imaging surveys — most
ius tidal stream), others are seemingly isolated and left
notably the Sloan Digital Sky Survey (SDSS) and the
1 adriftawayfromtheirancestralhomes(e.g.,theOrphan
Two Micron All Sky Survey (2MASS) — has revealed a
v stream; Belokurov et al. 2006; Grillmair 2006). Needless
rich and complex panoply of faint, remote substructures
5 to say, the properties of these stellar streams provide us
(e.g., Belokurov et al. 2006; Majewski et al. 2003; New-
8 withinvaluableinformationonboththeaccretionhistory
berg et al. 2007, 2002; Ibata et al. 2001a,b, 2002; Martin
3 of the Milky Way and its gravitational potential.
et al. 2014). Indeed, the many stellar streams that have
5 The archetypal stellar stream — that of the tidally
now been discovered in the outer Galaxy provide prima
0
disrupting Sagittarius dwarf galaxy (discovered by Ibata
.
1 1DepartmentofPhysics&Astronomy,UniversityofVictoria, et al. 1994, 1995) — has now been mapped extensively
0 Victoria,B.C.,V8P1A1Canada across both the northern and southern hemispheres of
6 2Leibniz-Institut fu¨r Astrophysik Potsdam, An der Stern- the Galaxy (e.g., Mateo et al. 1996, 1998; Totten & Ir-
1 warte16,14482Potsdam,Germany win 1998; Majewski et al. 1999, 2003; Ivezi´c et al. 2000;
3National Research Council, Herzberg Astronomy & Astro-
: Yanny et al. 2000; Dohm-Palmer et al. 2001; Mart´ınez-
v physics, 5071 West Saanich Road, Victoria, B.C., V9E 2E7
i Canada Delgado et al. 2001a; Vivas et al. 2001; Ruhland et al.
X 4Center for Astronomy & Astrophysics, Department of 2011;Ibataetal.2001a,2002;Majewskietal.2003;New-
Physics&Astronomy,ShanghaiJiaoTongUniversity,Shanghai
r 200240,China berg et al. 2002; Bellazzini et al. 2003; Newberg et al.
a 5Shanghai Key Lab for Particle Physics and Cosmology, 2003; Belokurov et al. 2006; Yanny et al. 2009; Correnti
ShanghaiJiaoTongUniversity,Shanghai200240,China et al. 2010; Niederste-Ostholt et al. 2010; Koposov et al.
6Department of Astronomy, Peking University, Beijing 2012; Carlin et al. 2012b; Jerjen et al. 2013; Pila-D´ıez
100871,China
et al. 2014; Belokurov et al. 2014; Koposov et al. 2015;
7Kavli Institute for Astronomy & Astrophysics, Peking
University,Beijing100871,China Hyde et al. 2015; Huxor & Grebel 2015).
8CEA/IRFU/SAP, Laboratoire AIM Paris-Saclay, As the quality and quantity of data for the stream
CNRS/INSU, Universit´e Paris Diderot, Observatoire de has improved, a wide range of models for its forma-
Paris, PSL Research University, F-91191 Gif-sur-Yvette Cedex,
tion have been developed and refined (e.g., Velazquez
France
9Department of Astronomy and Astrophysics, University of & White 1995; Mart´ınez-Delgado et al. 2001b; Helmi &
CaliforniaSantaCruz,1156HighStreet,SantaCruz,CA95064, White 2001; Ibata et al. 2001b; Helmi 2004; Law et al.
USA 2005; Johnston et al. 2005; Fellhauer et al. 2006; Law
10Emailaddress: [email protected]
11CanadianInstituteforAdvancedResearch(CIFAR)Global & Majewski 2010; Pen˜arrubia et al. 2010; Casey et al.
Scholar 2012; Deg & Widrow 2013; Ibata et al. 2013; Gibbons
12CanadianInstituteforAdvancedResearch(CIFAR)Senior et al. 2014). The latest data now suggest that it is diffi-
Fellow
2 Lokhorst et al.
cult to accommodate both the kinematic and photomet- seems obviously associated with the Sagittarius leading
ric measurements of the leading and trailing arm within arm,thebifurcationfeature,oranyofthepredictedtrail-
one single prolate, spherical or oblate Milky Way poten- ing arm features or older wraps in this region. It has
tial (although it is possible if the density profile of the beensuggestedthatmany, andperhapsall, ofthesesub-
haloisallowedgreaterfreedom,Ibataetal.2013). From structures could be explained as the remnants of a now
a modeling standpoint, the discovery of an apparent bi- disrupted satellite galaxy (e.g., Carlin et al. 2012 and
furcation in the stream (Belokurov et al. 2006; Koposov references therein) with stellar proper motion measure-
et al. 2012; Slater et al. 2013), a recently discovered off- ments for some of the substructures seeming to point to
setbetweenthetrailingarmandthepredictionsofmany anorbitthathasjustpassedpericenter(Casetti-Dinescu
models (Belokurov et al. 2014; Koposov et al. 2015) and etal.2006;Carlinetal.2012). Whateverthetrueexpla-
the possible existence of faint, associated streams (Ko- nation, it is clear that the sightline towards the constel-
posov et al. 2013) have further complicated matters. At lation of Virgo is a complex and intriguing “crossroad”
the present time, two widely used models for the Sagit- in the Milky Way — and one for which a consensus has
tarius tidal stream are those of Law & Majewski (2010, yet to emerge.
hereafter LM10) and Pen˜arrubia et al. (2010, hereafter By a lucky coincidence, this region of the sky is also
Pen10). To lessen the tension between the leading and home to the rich cluster of galaxies nearest to the Milky
trailing arm data, these models both use a triaxial — Way: the Virgo cluster, at a distance of 16.5 Mpc (Mei
thoughnearlyprolate—halopotentialthathastheawk- et al. 2007; Blakeslee et al. 2009). The cluster has been
ward property of not being aligned with the plane of the the subject of numerous past imaging studies (see, e.g.,
Milky Way disk. The Pen10 model was specifically fo- Richter & Binggeli 1985; Binggeli et al. 1985; Cˆot´e et al.
cused on modelling the stream bifurcation, which they 2004 and references therein), and is the target of a new,
attributetoamoderateamountofrotationinthestream deep, panoramic, multi-band survey with the Canada-
progenitor. However, no evidence for such rotation has France-Hawaii Telescope (CFHT): the Next Generation
been found in the core of the Sagittarius dwarf galaxy Virgo Cluster (NGVS; Ferrarese et al. 2012). Although
(Pen˜arrubia et al. 2011). theprimarygoaloftheNGVSisacensusandcharacter-
More recently, Vera-Ciro & Helmi (2013, hereafter izationofbaryonicsubstructurewithintheVirgocluster,
VC13) have calculated the orbit of the Sagittarius dwarf the survey is also ideally suited for studying the struc-
galaxy in a Galactic potential that includes both a vary- ture of the Milky Way halo along this direction. For
ing shape with radius — as is seen in cosmological simu- instance, in their NGVS study of Virgo’s globular clus-
lations(Vera-Ciroetal.2011)—andtheinfluenceofthe ter populations, (Durrell et al. 2014) clearly identified
LargeMagellanicCloud(LMC).WhilenotafullN-body the Sagittarius stellar stream as a prominent foreground
simulation,acomparisonoftheirtestparticleorbittothe feature (see their Figure 5). Our study thus builds upon
latest data showed that this new potential can alleviate the tradition of using wide-field photometric data sets
some of the earlier tensions. It also showed that the in- acquired for background galaxies or clusters to study
clusionoftheLMCpotentialcansignificantlychangethe the intervening halo: those of, e.g., Martin et al. (2014),
orbitoftheSagittariusanditstidalstream. Inaddition, who mapped the highly structured Galactic foreground
G´omez et al. (2015) have shown that in such a scenario within the deep Pan-Andromeda Archaeological Survey
the Milky Way’s response to the orbit of the LMC must (McConnachie et al. 2009), and Pila-D´ıez et al. (2014),
also be taken into account. As we shall argue below, who determined distances along the Sagittarius stream
new simulations of the Sagittarius stream including all andvariousothersubstructuresfrompencil-beamsurvey
these effects are now needed, while deeper imaging at data targeting galaxy clusters.
selected positions along the stream, including direct dis- In this work, we focus on the two main stellar over-
tancemeasurements,areneededtoconstrainthevarious densitiesin,andaround,theNGVSregion: theSagittar-
model parameters. ius stellar stream and the VOD. Using old, metal-poor
A second prominent substructure in this region of the starslocatednearthemainsequenceturn-off(MSTO)re-
sky, but one lying at a closer distance, is the Virgo gion, we are able to measure accurate distances to these
Stellar Stream or Virgo Over-Density (hereafter VOD). over-densities and explore their three dimensional struc-
The VOD is now recognized to span more than 1000 ture within the Galaxy. These measurements allow us
square degrees (Juri´c et al. 2008; Jerjen et al. 2013) not only to probe the geometry of these substructures
and contain several dense clumps embedded within it buttotestthepredictionsofnumericalmodelsthathinge
(Vivas et al. 2001; Newberg et al. 2002; Vivas & Zinn on the assumed properties of their progenitors and the
2003,2006;Newbergetal.2007;Kelleretal.2008,2009, Galactic halo potential.
2010). In addition to the large-scale excess of halo stars This is the first NGVS paper to focus specifically on
that was originally used to identify the VOD, several theMilkyWayforegroundstarpopulation. Otherpapers
distinct, kinematically-grouped substructures have now in this series have examined the distribution of globular
been found in this region (Duffau et al. 2006; Newberg clusters within the Virgo cluster (Durrell et al. 2014),
et al. 2007; Vivas et al. 2008; Prior et al. 2009; Starken- the properties of star clusters, UCDs and galaxies in
burg et al. 2009; Brink et al. 2010; Duffau et al. 2010; the cluster core (Zhu et al. 2014; Zhang et al. 2015; Liu
Casey et al. 2012; Duffau et al. 2014). Although most et al. 2015), the internal dynamics of low-mass galaxies
of these kinematical structures show high positive galac- (Gu´erou et al. 2015, Toloba et al., submitted), abun-
tocentric velocities, their spread in velocity can be more dance matching of low-mass galaxies in the cluster core
than 100 km s−1. The nature of these over-densities — (Grossaueretal. 2015,inpress),interactionswithinpos-
andtheirrelationshiptotheVODitselfandtoeachother sibleinfallinggalaxygroups(Paudeletal.2013),optical-
— has yet to be conclusively established. None of them IR source classification methods (Mun˜oz et al. 2014), a
Milky Way Substructures in the NGVS Footprint 3
30
)
s
e 20
e
r
g
e NGVS
d
( 10

n
o
i
t
a
n 0
i
l
c
e
D
−10
−20
220 200 180 160 140
RA (degrees)
Figure 1. A stellar density map made using SDSS stars selected to have colors of 0.2 < (g −r ) < 0.3 and (u −g ) >
SDSS SDSS SDSS SDSS
0.4, and magnitudes of 20.4 < g < 21.4. This selection showcases the various large-scale Milky Way substructures in the northern
SDSS
hemisphere that happen to fall along the line of sight to the constellation of Virgo. The location of the 100 deg2 NGVS survey footprint
anditsfourbackgroundregionsareshown. TheNGVSlandssquarelyonthenortherntidalstreamoftheSagittariusdwarfgalaxy, with
thetwolowerbackgroundfieldsalsofallingonthetidalstream.
new member of the inner Oort cloud (Chen et al. 2013) depths of u∗ =24.8, g =25.9, i=25.1 and z =23.3 AB
and a catalogue of photometric redshifts for background mag. Thesurveyhassub-arcsecseeinginallbands,with
sources (Raichoor et al. 2014). a median seeing of 0(cid:48).(cid:48)54 in the i−band. The survey also
The paper is organized as follows. We begin with a includes partial coverage of the NGVS footprint in the r
short introduction to the NGVS in §2. In §3, we give band (see, e.g., Raichoor et al. 2014).
anoverviewofthemainsubstructuresvisibleinthevery In Figure 1, we show the location of the NGVS foot-
deepNGVSdata,includingtheVODandtheSagittarius printwithinthelarge-scaleMilkyWaystellarhalointhe
tidal stream. In §4, we determine accurate distances to Northern hemisphere, as seen by SDSS. In this figure,
the Sagittarius tidal stream using the NGVS data and SDSS DR7 was used to create a density map of MSTO
compare thosedistances to thepredictions ofthree lead- stars selected by the following color criteria:
ing stream models. We summarize and conclude in §5.
0.2 < (g −r ) <0.3
SDSS SDSS (1)
2. OBSERVATIONSANDDATA (uSDSS−gSDSS) >0.4.
The NGVS is a multi-band, panoramic imaging sur- A magnitude cut of 20.4 < g < 21.4 was also ap-
SDSS
vey of the Virgo cluster carried out with the MegaCam plied in order to select SDSS stars having approximate
instrumentmountedatprimefocusonthe3.6mCanada- distances of 15 (cid:46) d (cid:46) 35 kpc. In this distance range,
(cid:12)
France-HawaiiTelescope(CFHT).Fulldetailsonthesur- the two over-densities that we focus on in this paper —
vey design, reduction procedures, data products and sci- namely, the VOD and the brighter branch of the Sagit-
encegoalshavebeenpresentedinFerrareseetal.(2012). tariustidalstreambifurcation—areunmistakable. The
Inbrief,thesurveycoversanareaof104deg2 inscribed Sagittarius tidal tail is running diagonally from (α,δ)∼
within the virial radii of Virgo’s two main subclusters: (120◦,20◦)to(200◦,10◦). Thefainterbifurcationfeature
the A subcluster to the north, centered on Virgo’s cD isseenabovethemainstream. Bycontrast,theVODap-
M87, and the B subcluster to the south, centred on pearsasanimmenseover-densitywhichpeaksat(α,δ)∼
Virgo’s optically brightest galaxy, M49. Observations (190◦, –5◦). Although its center lies well south of our
were carried out in four optical bands — u∗, g, i, and z fields, the VOD also extends northward into the NGVS
— in the MegaCam filter system13 to 10σ point-source footprint. Note that the NGVS also includes four out-
13 Note that the MegaCam bands have similar names to those length range nor do they have the same response function: see
of the SDSS, but the filters do not cover exactly the same wave- http://www.cadc.hia.nrc.gc.ca/en/megapipe/docs/filt.html.
4 Lokhorst et al.
lying background fields (see the four squares in Figure 1
and Ferrarese et al. 2012) located ∼ 16◦ from the cen- 8
ter of the footprint. Among the four NGVS background
fields, two lie well off the main portion of the Sagittar- 18
ius tidal stream, but two others fall squarely along the
6
stream. As we show below, these two fields turn out to
provideimportantconstraintsontheoverallgeometryof 20
the stream.
Our analysis of these substructures obviously requires i 4
a catalog of stellar sources selected from the NGVS.
WithintheNGVSfootprint,acatalogoptimizedforcom- 22
pact and unresolved sources was created as described
2
in Liu et al. (2015) and summarized below. Note
that the NGVS has implemented several data process- 24
ing pipelines, each optimized for a specific goal; for
the purpose of this work, a source catalog was gener- 0
-0.5 0.0 0.5 1.0 1.5 2.0 -0.5 0.0 0.5 1.0 1.5 2.0
atedusingNGVSstacksprocessedthroughtheMegaPipe g - i g - i
(Gwyn 2008) pipeline which adopts a global estimation
of the sky background. For compact and unresolved Figure 2. Color-magnitude diagrams (CMDs) for compact
sources,thesestacksprovidethehighestphotometricac- sources in the NGVS. (Left Panel) In this CMD, the location of
halo main sequence turnoff (MSTO) stars, disk dwarf stars and
curacy(seeFerrareseetal.2012fordetails). Photometry
compactbackgroundgalaxiesarelabelled. Anover-densemainse-
wasthenperformedusingSExtractor(Bertin&Arnouts quence with a turnoff at i ∼ 21 is clearly visible, corresponding
1996)indual-imagemodewiththeg−band—thedeep- to the tidal stream of the Sagittarius dwarf galaxy. The best-fit
estoftheNGVSbands—asthedetectionimage. Aper- theoretical isochrone for the tidal stream is shown as a reference.
Typicalphotometricerrorbarsareshownalongtheleftsideofthe
turemagnitudesweremeasuredinaperturesofdiameter
diagram. (Rightpanel)InthisCMD,threerepresentativepolygons
3, 4, 5, 6, 7, 8, 16 and 32 pixels (each pixel correspond- used to select mainsequenceturnoff stars are shown, correspond-
ing to 0(cid:48).(cid:48)187) and then corrected to an infinite aperture. ingtodistances(frombrighttofaint)of∼10,20and30kpc. Each
This last step was performed by applying an aperture mainsequenceselectionregioncorrespondstothesimilarlylabelled
densitymapinFigure6. Themainlocusofredglobularclustersin
correctionmeasuredbymatchingtheMegaCam16-pixel
the Virgo cluster is also indicated, as the bluest globular clusters
aperture magnitudes to SDSS PSF magnitudes (trans- havethesame(g−i)colorashalomainsequencestars.
posed to the MegaCam photometric system as given in
Colour vs Colour Plot of NGVS Data
equation4ofFerrareseetal. 2012)foranumberofbright 2.0 3.5
but unsaturated stars. All other aperture magnitudes
(within 3, 4, 5, 6, 7, and 8 pixels) were then corrected to 3.0
a16-pixelaperture. BecauseofPSFvariationsfromfield 1.5
to field, aperture corrections were calculated separately 2.5
for each field and, of course, for each filter.
1.0
Point sources were then selected as having −0.1 ≤ − i) 2.0
(m4 − m8) ≤ +0.15 mag, where m4 and m8 are the (g 1.5
corrected aperture magnitudes measured within 4 and 0.5
8pixels,respectively. Thesmallerofthesetwoapertures
1.0
is well matched to the average seeing of the MegaCam
0.0
imagesgiventhepixelscale. ThePSFissufficientlyuni- 0.5
formovertheimagesthatallpointsourcesclustertightly
in (m −m ) (the same is generally true for the differ- −0.5 0.0
4 8
ence in any two aperture magnitudes). Although point- 0.0 0.5 1.0 1.5 2.0
(u* − g)
source selections were carried out separately in each of
the u∗, g and i bandpasses, the overall results are not Figure 3. Acolour-colourdiagramoftheNGVSstarsscaledwith
affected by the choice of bandpass (or a combination of densityinsidebinsof0.1mag. TheMSTOcolorselectionisshown
inbluewhilethegreencurveshowsaPARSECisochronewithan
them). Therefore, for the remainder of this work, we
age of 9 Gyr and a metallicity of [M/H] = –0.7 dex. The main
use the catalog of stellar sources derived in the g band. region occupied by background galaxies is also shown. Red con-
As a final step in the catalog preparation, the photom- tourssurroundtheirlocationandthelocusofwhitedwarfsinthe
etry for each object was de-reddened following Schlegel color-colorplane. A(u∗−g)MSTOcolorcutwasusedtoseparate
whitedwarfsstarsfromtheMSTOstars,whichoverlapin(g−i)
et al. (1998). Note that the reddening along this line
color.
of sight is quite low (see, e.g., Boissier et al. 2015), with
M87,whichmarksthecenteroftheVirgocluster,having
E(B−V)=0.022 mag. disk dwarfs in the solar neighbourhood. At the faintest
Figure 2 shows the (Hess) colour-magnitude diagram magnitudes, one sees residual contamination from com-
(CMD) of all NGVS stellar (point-like) sources identi- pact, background galaxies. At brighter magnitudes, still
fied in this way. The average photometric error in i and another source of contamination is discernible: globu-
(g−i)asafunctionofi-bandmagnitudeisshownonthe lar clusters belonging to the Virgo cluster which appear
left-handsideofthefigure. Thisfigureillustratesthelo- as nearly point-like at this distance (Jord´an et al. 2005;
cation of Milky Way stars along this line of sight, which Durrell et al. 2014). We will discuss our treatment of
is dominated by faint MS halo stars, plus some very red these various contaminants below.
Milky Way Substructures in the NGVS Footprint 5
is centered on an isochrone that we henceforth adopt as
5500
the fiducial for this analysis: a PARSEC isochrone hav-
ing an age of 9 Gyr and a metallicity of [M/H] = –0.695
4400 dex (Bressan et al. 2012), which has MSTO i-band ab-
c] solute magnitude = 3.8. These values have been chosen
p
s [k tomatchtheSagittariusstreampropertiesinthisregion
e of the sky (Chou et al. 2007). In the left panel of Fig-
c 3300
an ure2,thePARSECisochroneisoverplottedontopofthe
Dist mostvisibleturn-offfeaturethatweseewithintheNGVS
d CMD: that of the Sagittarius stream. Grafted onto this
e 2200
at isochrone, our MS boxes are wide enough to accommo-
ul
c date MS stars having a dispersion in metallicity and age
al
C centered on these mean values.
1100
TheMSboxeshavebeenlimitedin(g−i)colourspace
MMSSTTOO sseelleeccttiioonn:: 00..11 << ((gg -- ii)) << 00..3355 by requiring:
00
0.20 < (g−i) <0.55. (2)
00 1100 2200 3300 4400 5500
Actual Distances [kpc]
HeretheblueboundarycorrespondstotheMSTOcolour
and the red boundary has been chosen to minimize con-
5500 taminationbyVirgoglobularclusters(whosepositionin
the CMD is indicated in the right panel of Figure 2; see
also Figure 1 of Durrell et al. 2014). By moving this se-
4400
c] lection box vertically in the CMD, we can identify MS
kp stars at a range of heliocentric distances. When doing
s [ so, the vertical widths of the boxes are scaled smoothly
e
nc 3300 as a function of magnitude to ensure that the physical
a
st depthofeachsampleofstarsselectedwithintheboxesis
d Di 5kpc. ExamplesofMSboxesatthreedifferentdistances
ate 2200 are shown in the right panel of Figure 2.
ul For stars selected inside any of the MS boxes, we im-
c
al pose a second constraint: a colour selection of
C
1100
0.58 < (u∗−g) <0.90 (3)
MMSSTTOO sseelleeccttiioonn:: 00..11 << ((gg -- ii)) << 00..3355
00..5588 << ((uu -- gg)) << 00..99
00 that eliminates disk white dwarfs from our sample of
00 1100 2200 3300 4400 5500 MSTO halo stars. In Figure 3, this additional colour
Actual Distances [kpc] selection in shown in the (u∗ −g)–(g−i) colour-colour
diagram. The location of white dwarf stars and back-
Figure 4. Distances derived from a mock catalog of stars taken ground galaxies is shown, demonstrating that, although
fromtheTRILEGALmodeloftheGalaxyinthedirectionofNGVS thewhitedwarfsandMSstarshavesimilar(g−i)colours,
(using the NGVS color band system and magnitude range). The they are cleanly separated using the (u∗−g) index. Es-
distances were derived assuming a MSTO magnitude of i = 4 —
the approximate MSTO magnitude of halo stars (age ∼ 11 Gyrs, peciallyatfaintermagnitudes,wefindthatupto∼20%
[Fe/H]∼-1.0)—andplottedagainsttheinputdistancesfromthe of our MS samples would actually be (disk) white dwarf
mockcatalog. Intheupperpanel,a(g−i)cuthasbeenappliedto candidates if this extra color selection criterion were ig-
themocksample;inthelowerpanel,anadditional(u∗−g)cuthas
nored. Such a contamination rate would heavily skew
also been applied. The solid black line is the average calculated
distancevs. actualdistance,displayingthetrend’sagreementwith the distance estimates for these stars, as demonstrated
thedashedredone-to-oneline. Itisclearfromacomparisonofthe in Figure 4. In this figure, we calculate MSTO distances
running averages that, without a (u∗−g) color cut to weed out to stars in the TRILEGAL model mock catalog (Girardi
white dwarfs in the sample, the derived distance profile will be
etal.2005,2012)fromwhichwetookaselectionofstars
highlyskewed.
comparable in galactic latitude and magnitude range to
our NGVS sample. The distance is calculated from the
3. EXPLORINGMILKYWAYHALOSUBSTRUCTURE
difference between the apparent magnitude of MSTO
3.1. Selecting a Tracer Population
stars in the model (selected by 0.1 < (g − i) < 0.35)
In this work, we use main sequence (MS) stars located and an absolute magnitude of i=4, in accordance with
near the MSTO as our stellar tracers. MSTO stars are theapproximateMSTOabsolutemagnitudeofhalostars
not as accurate distance indicators as some other stellar with age ∼ 11 Gyrs and [Fe/H] ∼ -1.0. In order to de-
tracers (e.g., blue horizontal branch = BHB stars), but rive the needed accurate distances to individual stars -
they are more numerous than other stellar types and, rather than a characterisation of the distance distribu-
as such, are attractive targets for mapping large-scale tion of the population - we focus for the purpose of this
structure. test solely on the main-sequence turn-off stars and thus
TomaphalofeaturesusingMSstars,webeginbycon- use a (g −i) range smaller than that used for the MS
structingapolygonintheCMDcenteredontheMSofa boxes. The two panels show the derived and model in-
particular theoretical isochrone (an “MS box”) in order putdistancesbothwithandwithoutanextra(u∗−g)cut
toselectstarslocatedneartheturn-offregion. Eachbox appliedtothemodelstars. Clearly,fordeepphotometric
6 Lokhorst et al.
# of point sources per deg2 # of stars per deg2
Virgo Cluster Galaxies
0 2.0•103 4.0•103 6.0•103 8.0•103 1.0•104 1.2•104 0 500 1000 1500 2000
18
16
14
g)
e
d
n ( 12
o
ati
n
cli 10
e
D
8
6
192 190 188 186 184 182 192 190 188 186 184 182 192 190 188 186 184 182
RA (deg) RA (deg) RA (deg)
Figure 5. (Left panel) ThelocationoftheVirgoclustergalaxiesinsidetheNGVSfootprint. Thesymbolsizehasbeenscaledbygalaxy
B-bandluminosity. TheredcrossesmarkthelocationsofM49andM87,whicharelocatednearthecentersofsubclustersAandB.(Middle
panel) A density map of all NGVS compact sources with i < 23.5. M49 and M87 are again marked by red crosses. The obvious density
enhancementsatthepositionsofthebrightestVirgogalaxiesareduetocontaminationofthestellarcataloguebyVirgoglobularclusters.
(Rightpanel)AdensitymapofmainsequencestarsfromtheNGVS,selectedusingthecolorcutsgiveninEquations2and3. Inthismap,
there are no longer any significant over-densities at the locations of M49, M87 (red crosses) or any other Virgo galaxies: i.e., the MSTO
selectionishighlyeffectiveinremovingglobularclustersfromthestellarsample.
studiesofhalostarpopulations,theadditionofasecond SDSSphotometryalonehasmainlyfocussedondistances
color index that includes a blue bandpass is absolutely <20 kpc (e.g., Juri´c et al. 2008).
essential.
BackgroundgalaxiesinoursampleofMSTOhalostars 3.2. Blue Horizontal Branch stars as tracers
arealsoreducedbythecombinedcolourcuts,butextend WeexaminedthepossibilityofadditionallyusingBHB
intotheMSTOregionshowninFigure3withincreasing stars as stellar tracers. SDSS studies of MS stars have
numberatfaintermagnitudes. Significantcontamination often been accompanied by BHB star studies. Due to
from these misidentified background galaxies occurs at their relatively high luminosities, BHB stars allow one
i∼23.5(seeFig2),thereforeweadoptthisasthedepth to probe larger distances than is possible with MS stars.
limit to our analysis. Unfortunately, the Megacam u∗ filter extends to redder
The effectiveness of this approach in eliminating con- wavelengths than the SDSS u filter, compromising our
tamination from background sources, including globular ability to include BHB stars in our analysis: i.e., the
clusters in the Virgo cluster, is illustrated in Figure 5. additional red coverage includes a significant region red-
In the left panel of this figure, we show the location of ward of the Balmer break, whereas in SDSS most of the
Virgo galaxies within the NGVS footprint, with sym- u-band transmission is dominated by light blueward of
bol sizes scaled according to galaxy B-band luminosity. the break. This difference — though small in terms of
The supergiant elliptical galaxies M49 and M87 are de- wavelengthcoverage—isimportantforBHBstudiesbe-
notedbytheredcrossesineachpanelofthisfigure. The cause it means that the filter loses its gravity sensitivity
middle panel shows a density map of the NGVS point for these hot stars. We are therefore unable to sepa-
sourcesbrighterthani=23.5mag,aselectionthatelimi- rate BHB stars from contaminating blue straggler stars
natesmostofthecontaminationfrombackgroundgalax- (which have a similar temperature but a much higher
ies. However, it is clear that significant contamination gravity) and therefore cannot use BHB stars as a com-
frommisidentifiedVirgoglobularclustersremains, man- plementary tracer population.
ifesting as strong density peaks at the location of M49,
M87, and many other Virgo galaxies. In the right panel 3.3. Halo Tomography
of Figure 5, we show a map of MS stars selected using
In Figure 6, we use our selection criteria for MS boxes
both a selection on i-band magnitude and on location
placed at different distances to perform tomography of
in the (u∗ −g)–(g −i) colour-colour diagram. Clearly,
the Milky Way halo. Density maps for stars within each
the contamination from globular clusters has now been
distance slice are shown in 3(cid:48) by 3(cid:48) cells, smoothed with
largely eliminated by the colour-colour selection and no
a Gaussian filter of FWHM = 12(cid:48). The nine panels in
longer affects our view of the intervening halo.
Figure6showMSstarscenteredonourfiducialisochrone
In the remainder of the paper, we adopt the MS star
at mean heliocentric distances of d = 5, 10, 15, 20, 25,
selection process described above, which allows the halo (cid:12)
30, 40, 50 and 60 kpc. These distances are labelled in
to be traced out to distances of d ∼ 90 kpc. Because
(cid:12) each panel, along with the mean MSTO magnitude in
NGVS is several magnitudes deeper than SDSS – the
each bin.
95% completeness limit of SDSS is reached at i = 21.3
Already at distances of 5–12 kpc, an over-density be-
(Abazajianetal.2009)–itallowsustoprobealotdeeper
gins to appear in the southern half of the survey foot-
into the outer halo. Main-sequence mapping based on
print,atalocationanddistanceconsistentwiththeVOD
Milky Way Substructures in the NGVS Footprint 7
18
i. ii. iii.
16
g) 14
e
d
n ( 12
o
ati 200
n VOD
cli 10
e
D
8
i ~17.1 i ~18.7 i ~19.7
TO TO TO
6 5.0 kpc 10.0 kpc 15.0 kpc
18
iv. v. vi.
RA (deg) RA (deg) RA (deg) 150
eg) 1146 S gr tid al stre a m 2g
d e
on ( 12 er d
eclinati 10 Stars p
D 100 of
8 #
i ~20.3 i ~20.8 i ~21.2
TO TO TO
6 20.0 kpc 25.0 kpc 30.0 kpc
18
vii. viii. ix.
RA (deg) RA (deg) RA (deg)
16
g) 14 50
e
d
n ( 12
o
ati
n
cli 10
e
D
8
i ~21.8 i ~22.3 i ~22.7 0
TO TO TO
6 40.0 kpc 50.0 kpc 60.0 kpc
192 190 188 186 184 182 192 190 188 186 184 182 192 190 188 186 184 182
RA (deg) RA (deg) RA (deg)
Figure 6. DensitymapsofMSTOstarsintheNGVSfootprint. TheninepanelsstepoutwardsinheliocentricdistancethroughtheMilky
Wayhalo,displayingMSTOstarsselectedwithinboxestailoredaroundthemainsequenceofaPARSECisochrone. Acolorcutin(u−g)
has been used to eliminate white dwarfs from the MSTO star sample (see Figure 3). The boxes used to select stars in panels ii, iv, and
viareshowninFig2. Thewidthoftheboxesinmagnitudearescaledwithdistancesuchthattheapproximaterangeindistancewithin
each box is 5 kpc. The mean distance and MSTO magnitude of the stars within each density map are labelled in the lower left of each
panel. Thestarshavebeenbinnedwithin3(cid:48) ×3(cid:48) pixelsandsmoothedwithaGaussianfilterhavingFWHM=12(cid:48). Asthestarsincrease
indistancefromthedisk,thefeaturereferredtoastheVirgoOver-Density(VOD)appearsatd(cid:12)∼8kpc(panel ii). TheVODincreases
indensityandspatialextentatslightlylargerdistances(d(cid:12) ∼15kpc), andthendiminishesbeginningaroundd(cid:12) ∼20kpc. Inpanel vi,
weseetheSagittariustidalstreamclearlyflowingfromlowtohighdeclinationasRAdecreases. Inordertoinvestigatethechangeinthe
Sagittarius tidal stream properties across the NGVS, the two spatial boxes shown in panel vi are used to determine an overall trend in
stream distance. At distances well beyond the stream, we see a smooth spheroid halo with no other significant substructure (panels viii
andix).
(see §1). At distances of 25–40 kpc, the Sagittarius tidal lumpystellardistributionthatcouldbeduetoamixture
stream is clearly seen slicing across the NGVS footprint. ofstarsbelongingtothesetwosubstructures. Wedonot
At farther distances (up to 50 kpc), the Sagittarius tidal clearly detect the bifurcation feature that runs parallel
stream still appears at a low level. Although, a certain to the Sagittarius stream at slightly higher declinations
distance spread within the stream is seen, as well as ex- (e.g., Belokurov et al. 2006; Koposov et al. 2012; Slater
pected (see also Sections 4.2.2 and 4.4), the tails of this et al. 2013; de Boer et al. 2015). This is because, at the
distribution can be mostly ascribed to remnant contam- highestdeclinationsprobedbytheNGVS,thesecondary
ination from white dwarf populations and photometric stream is too faint and no clear background region is
uncertainties. Atintermediatedistances, weseearather available within the NGVS footprint for it to stand out
8 Lokhorst et al.
against. hints of a significant distance gradient in the Sagittarius
The VOD, which has previously been observed at dis- stream across the NGVS field, with the largest right as-
tancesrangingfrom5to30kpc,isknowntospreadover censionregionbeingsystematicallyfurtherawaythanthe
thousands of square degrees (Juri´c et al. 2008; Jerjen otherextreme. WeexplorethisfurtherinFigure7,where
et al. 2013). The VOD stellar density reaches its max- we show the CMDs of halo MS stars selected within the
imum at lower declinations than are accessible by the red and blue boxes drawn in panel vi of Figure 6. The
NGVS and, indeed, most previous studies have targeted two NGVS background fields at δ ≥ 20◦, which do not
areas of declinations lower than those examined here. intersectanyobvioussubstructuresinthisdistancerange
We note that several formation models for the Sagittar- (seeFigure1),wereusedtocreateareferenceCMDthat
ius stream suggest that some trailing arm debris and/or was subtracted from the CMDs of the tidal stream re-
older wraps could be present withinthe NGVS footprint gions. These background-subtracted CMDs are shown
atsmallerdistancesthantheleadingarm. Basedonpho- in the left and middle panels of Figure 7. Our fiducial
tometryalone(i.e.,withoutanyradialvelocityorproper isochrone is shown as the dashed green curve in each
motion information), it is difficult to conclusively rule panel,shiftedtoacommondistancemodulusof(m−M)
out the possibility that some of the substructure we at- = 17.8 mag (d ≈ 36.3 kpc).
(cid:12)
tribute to the VOD may instead be associated with the AdifferentialCMDfortheSagittariusstreamisshown
Sagittarius tidal debris. However, according to current in the right panel of Figure 7, in which the CMD in the
models,anysecondarystreamsarenotexpectedtobeas middle panel (composed of stars within the blue spatial
strong as the prominent over-density we see in our data. boxinpanelviofFigure6)wassubtractedfromtheCMD
For instance, in the LM10 model, no debris features in intheleftpanel(composedofstarsintheredspatialbox
the NGVS fields are predicted at ∼12-18 kpc, which is inpanelviofFigure6). Theresultsofthissubtractionin
whereweseeourpeakdensity. Henceforth,wewillthere- therightpanelofFigure7arecolourcodedaccordingto
forerefertothesoutherlyfeaturedetectedbetween5and the differential over- or underabundance of the number
25 kpc as the VOD and consider it separately from the of stars in the red or blue spatial box respectively. This
Sagittarius stream. differential analysis illustrates that the mean distance to
We see the VOD first appear at d ∼ 8 kpc and dis- thestreamchangeswithpositiononthesky. Thisismost
(cid:12)
appear completely by d ∼ 25 kpc. In this distance clearly seen to the left and right of the main sequence
(cid:12)
range many smaller “hotspot” regions can be observed of the overplotted fiducial isochrone. Whereas left of
within the broader over-density feature itself. Spatially, the isochrone main sequence the pixels are mostly blue
the VOD is separated from the Sagittarius tidal stream (correspondingtoarelativeoverabundanceofstarsinthe
in distance as well as location on the sky. In the density blue spatial box at larger distance), they are mostly red
maps shown in Figure 6, the stream appears clearly in ontherightside,inlinewitharelativeoverabundanceof
the mid-upper region of the footprint, whereas the VOD starsintheredspatialboxatcloserdistances. Byfitting
is strongest at the very bottom of the survey, at the de- anisochronetobothspatialfieldsseparately,wefindthat
clinations (cid:46) 8◦. A sharp cut-off is used to indicate the the vertical shift between the two best fits corresponds
strongestpartoftheVODinpanel iiofFigure6,though to a distance difference of ∼ 13 kpc.
it appears across the entire region with density decreas-
ing towards the top-right (see panels ii and iii). 4.2. Main Sequence Turn-off Star Analysis
As shown in, e.g., Helmi et al. (2011), galactic sub-
4.2.1. Refining our CMD analysis
structures tend to be distributed very anisotropically
on the sky in halo simulations carried out within a The deep photometry and excellent image quality
ΛCDM framework. One should therefore be careful availableintheNGVSallowustoidentifyMShalostars
about na¨ıvely linking two over-densities on the sky to in this region of the sky — with minimal contamina-
a common physical origin. Indeed, it may be prudent to tion — to larger distances than has been possible in the
look more deeply into regions that are already known to past (e.g., Juri´c et al. 2008; Sesar et al. 2011). Addi-
show an abundance of substructures. Although this is tionally, this excellent dataset enables the measurement
exactly what we have done in our analysis, it is worth of distances to a higher level of precision than we have
notingthatwefindnoconvincingevidencefornewover- demonstrated thus far using broad MS boxes. We there-
densities lying behind the Sagittarius leading arm. fore present a more refined approach in this section.
We begin by fine tuning the box dimensions so that
4. THEDISTANCEGRADIENTOFTHESAGITTARIUS their width at each luminosity is proportional to the
TIDALSTREAM
uncertainty of the photometric measurements (resulting
In this section, we use our halo star density maps to in a trapezoid-like shape in the CMD; see the left and
examine the three-dimensional structure of the Sagittar- middle panels of Figure 7). This allows us to compen-
ius tidal stream, specifically its distance gradient along satenaturallyforthelargerphotometricuncertaintiesof
the line of sight. fainterstarswhenassessingtheirlikelihoodasSagittarius
streammembers. Theaimofthisexerciseistomaximize
4.1. CMD Analysis of the Sagittarius Tidal Stream
our selection of Sagittarius stream stars, so we use again
The Sagittarius stream first appears in Figure 6 at a our fiducial isochrone as the ridge line of this MS box.
distance of d ∼ 20 kpc and reaches its greatest promi- Subsequently, we step this spatial box down the CMD
(cid:12)
nence in panel vi, at a distance of ∼ 30 kpc. The stream to represent populations at increasing distance, and at
is clearly visible as a broad swath extending from higher eachstepapplytheweightingschemetoeffectivelycross-
to lower declination as right ascension increases. correlate the observed stellar density on the CMD with
Examination of the panels vi-vii of Figure 6 show thatofthemodelisochrone. AteachstepoftheMSbox,
Milky Way Substructures in the NGVS Footprint 9
18 600 18 600 18 150
19 19 19
500 500 100
20 20 20
400 el 400 el 50 el
x x x
g 21 300 per pi g 21 300 per pi g 21 0 per pi
s s s
22 ar 22 ar 22 ar
St St St
200 # 200 # -50 #
23 23 23
100 100 -100
24 24 24
25 0 25 large0 region2s5 -150
0.0 0.5 1.0 1.5 2.0 0.0 0.5 1.0 1.5 2.0 0.0 0.5 1.0 1.5 2.0
225500
g - i g - i g - i
y
Figure 7. (Left and middle panels) Cositlour22-00m00agnitude diagrams (CMDs) of the stars in two locations at either end of the Sagittarius
n
tNidGaVlSstrreeafemrenwciethfiineldthse(aNtGdVecSlin(saeteiopnasnoeflDeδvi∼o11f552F001igaunrde∼6)2.7Bdaecgk)g.roAunPdAsRoSuErcCesihsoacvherobneeen(wsuitbhtraanctaegdefroofm9tGhyerCaMndDasfmoretcallalircitityyuosfin[Mgt/wHo]
= –0.7 dex; dashed green line) is plotteder in both diagrams at a distance modulus of 17.8, along with the selection box used in the main
sequence cross-correlation algorithm descmbribe11d0000in §4.2. The theoretical isochrone and the main-sequence selection box have been placed
CatMtDheslhoocwatiinogntohfetrheesiSdaugailttoafrituhsetCidMalDsstNurienamth55,e00wphreicvhiohuasspaancelelsa.rlRyeodvecro-rdreensspeonmdasintoseaqugreenacteerstnaurmpboeprulsattairosn.in(tRhieghfitrsptanCeMl)DD,iffanerdenbtliuael
correspondstoagreaternumberofstarsinthe00secondCMD.ThecolorshavebeenscaledwiththenumberofexcessstarsineitherCMD
followingthetwo-tonedcolorbar(withnegativenumbersreferringtoanexcessofstarsinthesecondCMD).Thereisaclearverticalshift
1144 1155 1166 1177 1188 1199
between the two populations around the main sequence of the theoretical CMD, indicating a distance gradient to the Sagittarius tidal
Distance Modulus (m - M)
streamacrosstheNGVSfootprint.
ra: 182-184 and dec: 10-11
18 250
g) 16 y
(de 14 nsit 200
on 12 De 150
nati 10 ber 100
Decli 8 Num 50
6
0
190 186 182 14 15 16 17 18 19
RA (deg) Distance Modulus (m - M)
Figure 8. (Left panel) Schematic diagram showing the 2 deg2 regions used to tile the Sagittarius tidal stream, each of which was used
to calculate a mean stream distance. (Right panel) Number density plotted against distance modulus for Sagittarius tidal stream main
sequencestarsderivedfromstellarcountsintheredandbluetilesshownintheprecedingpanel(redandbluecurves,respectively). These
luminosity functions were found via the main sequence cross-correlation algorithm outlined in Section 4.2. Their maxima correspond to
the distances of the tidal stream. This plot shows that the Sagittarius tidal stream is located at a larger heliocentric distance in the
blue (eastern) tile than in the red (western) tile, a result consistent with the CMD analysis: see Figure 7. The histograms result from a
“bootstrapping”methodtodeterminedistanceuncertainties: thecross-correlationalgorithmwasperformedonrandom90%completeness
subsets (250 for each tile), yielding a distribution of peak distances from each sample. The red (blue) histogram corresponds to the red
(blue)luminosityfunction.
as the region shifts down the CMD, the highest weights distancesatvariouspositionsalongthestream,wedivide
are assigned to stars lying along the ridge line of the MS theNGVSfootprintinto2deg2 cellsandapplythealgo-
region (i.e., those most likely to be at the exact distance rithmtoeachcellseparately. Figure8illustratesthispro-
of the MS region). The assigned weights decrease for cess,emphasizingthedistancegradientalongthestream
stars that are more displaced from the ridge line accord- within the NGVS footprint. In the left panel, two cells
ing to a gaussian function whereby the width is set such ateitherendofthestreamhavebeenhighlighted. Inthe
that the edges correspond to a 3σ level, following Equa- right panel, we show the corresponding MS star density
tion1ofPila-D´ıezetal.(2014). Whenimplementingthis distributions for these regions. The two peaks are sepa-
algorithm, we apply the same selections on (g−i) and rated in mean distance by 6 kpc. This is slightly smaller
(u∗−g)colordescribedin§3(i.e.,Equations2and3). At compared to the distance separation inferred from the
eachpositionoftheMSbox,theintegraloftheweighted CMD analysis described in Section 4.1, but note that
stars inside the box represents the MS star density. the spatial boxes are different as well.
Thereareseveralsourcesofuncertaintyintrinsictothis
4.2.2. Sagittarius Stream Distances method. The first is the choice of theoretical isochrone
used for the ridge line of the MS region (which we have
In the above approach, a peak in the MS star den-
chosen specifically to match what is known about Sagit-
sityprofilecorrespondstoanover-densityatthedistance
tariusstreamstarsinthisregionofthesky,seee.g. Chou
modulus applied to the fiducial isochrone. To measure
10 Lokhorst et al.
60 4.3. Comparison with Other Measurements
Majewski et al. (2004)
Shi et al. (2012) Acomparisonofthedistancesderivedinthisworkwith
Belokurov et al. (2014)
50 Pila−Diez et al. (2014) those from papers in the literature is shown in Figure 9.
Forconvenience, thedistanceshavebeenplottedagainst
kpc] 40 the Sagittarius stream coordinate, Λ(cid:12), which is a coor-
e [ dinate system defined along the Sagittarius stream (for
c details, see Majewski et al. 2003).14 Since only one of
n
a 30 the Sagittarius stream coordinates is given in this fig-
st
di ure on the x-axis, a naive plotting of literature results
will give a large range caused by targets that are associ-
20
ated with leading or trailing arm debris respectively (as
well as older wraps). Here we show only stars associated
10 withleadingarmdebrisbythevariousauthors,asthisis
240 250 260 270 280 the component of the Sagittarius stream that we see in
Λ [°] the NGVS footprint. There is generally good agreement
O ·
among the different studies, especially when one consid-
Figure 9. ComparisonofdistancestotheSagittariustidalstream
ers that different tracers have been used in the various
calculated in this work to those for leading arm of the Sagittar-
ius stream inferred from blue horizontal branch stars (Belokurov studies, each with their own typical distance uncertain-
etal.2014),redhorizontalbranchstars(Shietal.2012),Mgiants ties. Moreover, Majewski et al. (2004) and Shi et al.
(Majewskietal.2004)andmainsequencenearturn-offstars(Pila-
(2012) use individual stars as tracers, whereas our mea-
D´ıez et al. 2014). Our measurements within the NGVS footprint
surements,aswellasthepointsshownforPila-D´ıezetal.
areshownascoloredtriangles,whilethetwoblacktrianglesshow
measurementsmadeintheNGVSbackgroundfields. NGVSdata (2014)andBelokurovetal.(2014), measuredistancesto
pointshavingthesamerightascensionareplottedusingthesame a population of tracer stars, which explains the reduced
color.
dispersion in these latter measurements. The larger dis-
persion in Majewski et al. (2004) relative to the work of
et al. 2007). Other sources of error include sampling er-
Shietal.(2012)canfurtherbeexplainedbythefactthat
rors caused by the limited number of stars available in
Shi et al. (2012) measured only distances to stars that
a given box, photometric errors on the individual stellar
were spatially overlapping with the LM10 model of the
magnitudesandcolors,andcontaminationbyforeground
Sagittarius stream, whereas no such restriction was used
and/or background sources.
in the earlier work of Majewski et al. (2004).
We have accounted for photometric uncertainties in
Ourobserveddistancestothestreamareshownasthe
our distance measurements by implementing a “boot-
colored triangles in Figure 9, as well as in each panel of
strapping” method. Random 90% completeness stellar
Figure 10. The symbol sizes of each distance measure-
subsetswereselectedandthecross-correlationalgorithm
menthavebeenscaledbytheareaunderthecorrespond-
was run on each of them to investigate the robustness
ing peak in the bootstrapping method histogram. To
of the original distance estimate. This procedure has
supplement measurements within the main body of the
been repeated 250 times for each region. The red and
NGVS footprint, we also include the two NGVS back-
blue histograms shown in the right panel of Figure 8 are
ground fields situated along the Sagittarius stream (see
the90%completenesspeakdistancedistributionsforthe
Figure 1). These are shown as black triangles offset by
sametwoexampleregionsusedtocreatethecorrespond-
∼ 15◦ from the other NGVS data points.
ingMSdensitydistributions. Thepercentilesofthe90%
completeness histograms are taken as the uncertainties
4.4. Comparison with Numerical Models
on our distance measurements throughout this work.
Figure 10 compares our measured distances to two N-
Incaseswherethebootstrappingmethodyieldedmore
bodymodelsfortheSagittariusstream—byLaw&Ma-
than one distinct peak in MS star density distribu-
jewski (2010) and Pen˜arrubia et al. (2010) — as well as
tion, the individual peaks were treated as separate over-
theSagittariusorbitmodelofVera-Ciro&Helmi(2013).
densities — each with its own distance estimate and ac-
The two N-body models, in the upper and middle pan-
companying error distribution. This approach acknowl-
els,areshownasdensitycontours(wherethedarkergray
edgesthepossibilitythatanygivenlineofsightmightcut
corresponds to a higher density of model particles). In
through multiple over-dense regions, rather than forcing
the lower panel, the heavy black line corresponds to the
the method to find only a single peak or, worse, defining
median orbit of a test particle in the potential consid-
a final distance as the mean of two distinct peaks. How-
ered by VC13. The 1σ and 2σ orbits from the median
ever, we do define a threshold above which a distance
are shown as the dashed dark and light grey lines.
peak is considered to be a real feature. In what follows,
Our comparison between the model predictions and
we consider only those over-densities that were identi-
data points shows that the measured distances to the
fied by the bootstrapping method as a main feature in
Sagittarius stream are consistently smaller than those
morethan40%ofthecases. Becausetheyshouldprovide
predicted by the LM10 model. In fact, the densest part
a useful test of future simulations, we have summarized
of the stream predicted by the model does not signifi-
ourfindingsinTable1. Fromlefttoright,thecolumnsof
cantlyoverlapwiththedata. ThePen10model—which
thistablerecordthecellrightascensionanddeclination,
includes a rotation in the main body of the satellite —
measured distance modulus, upper and lower 1σ errors,
and adopted weights. These weights correspond to the
14 Throughoutthiswork,weusethecoordinatesystemdefined
fraction of times the overdensity peak at this distance is
byMajewskietal.(2003),ratherthanthetransformationgivenby
recovered in our bootstrapping tests. Belokurovetal.(2014).