Cross-Roads in Studies of Galaxy Evolution

5 Pages·0.1 MB·English

by  Richard S Ellis

Cross-Roads in Studies of Galaxy Evolution
RICHARD S. ELLIS1
Astronomy Department, MS 105-24, Caltech, Pasadena CA 91125; [email protected]
1
0
0 ABSTRACT
2
Studies of galaxy evolution at optical and near-infrared wavelengths have reached an in-

n
teresting point in their historical development and the arrival of a new millenium provides an
a
appropriate occasion to review the overalldirection in which this subject is moving. The aim of
J
this short essay is to emphasize the qualitatively different datasets now arriving and the more
9
1 detailed and challenging scientific analyses needed for future progress.

1
v
1
5 1. Introduction integral field unit spectrographs on ground-based
3 telescopes with adaptive optic feeds, or using the
1 Progress made in studying faint galaxies is no Next Generation Space Telescope.
0 longer measured solely in terms of the depth at
How is the community responding to the chal-
1 which samples are located but also in terms of
0 lenge of interpreting these increasingly-detailed
the ever-increasing detail provided for objects at
/ datasets and to what extent does the emerging
h fixed redshifts and flux limits. Consider a typi-
picture of how galaxies evolve indicate new di-
p cal 23rd magnitude galaxy at a redshift z ≃0.7.
- rections which define future instrumental require-
o In the 1970’s our knowledge about such an ob-
ments? The question is not merely of sociologi-
r ject would be very limited; it would have a pho-
t cal interest. Strategic changes in direction, when
s tographic apparentmagnitude and possibly a sin-
necessary, represent a natural outcome of the sci-
a
gle broad-bandcolor. By the mid-1980’s,it might
: entific process and thus it is important to try and
v have some CCD photometry and a low dispersion
anticipate them. This is particularly so for faint
Xi spectrum from a 4-meter telescope yielding a red- galaxystudieswhichcontinuetomotivatethecon-
shift,k-correctedluminosityandsomeemissionor
r struction of future large telescopes.
a absorptionline characteristics. By the mid-1990’s
Until recently, observers of faint galaxies were
a WFPC-2 image from Hubble Space Telescope
largely preoccupied with conducting ‘census sur-
(HST) would provide morphological and struc-
veys’, whose aim is to define the epochs within
tural information, and ground-based photometry
which galaxies of various photometric selections
overarangeofwavelengthsfromtheUVtotheIR
lie, as a precursor to evolutionary tests. We may
mightcharacteriseits spectralenergydistribution
be approaching the time when such studies will
(SED) fairly well. Late in the 1990’s, intermedi-
be complete and thus face something of a “cross-
ate dispersion spectroscopy with the new genera-
roads” in the subject. In this essay I discuss
tion of 8-10 meter telescopes has provided some
whether both observersandtheorists arereadyto
dynamicaldata, in terms of line widths or, excep-
interpretthemuchmoredetaileddatasetsthatwe
tionally, rotation curves. This trend of increasing
can expect fairly soon.
detail on high redshift galaxies is sure to continue
as we anticipate obtaining resolved spectra from
2. Earlier Cross-Roads
1This Essay is one of a series of invited contributions
which will appear in the PASP throughout 2000-2001 to We have encountered cross-roads before in our
marktheupcomingmillenium(Eds.) studies of galaxy evolution. A major one oc-
1
curred when it was realised that main sequence dependentevolutionarytrendswithsomeevidence
brightening arising from stellar evolution in first- for an increase in number density with look-back
ranked cluster galaxies could significantly com- time.
plicate their use in tracing the history of the A less obvious cross-roads appeared after the
cosmic expansion. That distant galaxies might completion of the 10-meter Keck telescopes. It
be useful as probes of galactic history did not, mightseemobviousthatmorepowerfultelescopes
of course, escape the attention of early pioneers should be used to extend the earlier magnitude-
(Sandage 1961). However, the (still-familiar) ea- limited redshift surveys faintward. However, the
gerness to constrain the cosmological parameters mean redshift depth is only marginally increased
clearly took priority. Tinsley was perhaps the with a larger telescope in this way. Most fainter
most ardent advocate in recognizing this cross- galaxies are simply less luminous sources occu-
roads (Tinsley 1976) convincingly inspiring ob- pying the same redshift space as those which
serverstoreturntostudiesofthecountsandcolors dominate the 4-meter surveys. Photometric se-
of faint field galaxies with the new generation of lection techniques, the most successful of which
wide-field4-metertelescopes(Peterson et al 1979; are based on Lyman and Balmer discontinuities,
Tyson & Jarvis 1979;Kron 1980;Koo 1985). The more efficiently yield high redshift samples with
emerging data were comparedwith her evolution- surface densities well-matched to Keck’s LRIS
arysynthesismodels (Tinsley 1977;Tinsley 1980) (Steidel et al 1996; Steidel 2000). Such tech-
and laid the foundation for an entire industry of niques provided the first glimpse at the abun-
surveying the faint skies which has grown apace dance of early galaxies as well as their likely con-
since 1980. tribution to the star formationandmetal produc-
In present-day parlance, Tinsley assumed tion history of the Universe (Steidel et al 1999;
galaxies underwent “pure luminosity evolution”. Madau & Pozzetti 1999).
Theyevolvedinisolationandthuswereconserved
innumber. Differentstarformationhistorieswere 3. NewDirectionsforHighRedshiftStud-
chosen to reproduce the colors of the present-day ies
Hubble sequence in an admirably simple scheme
So what can we learn from this brief his-
(Struck-Marcell & Tinsley 1978). As noted by
tory in terms of how observational strategies
earlierworkers,thepresentcolorsofspheroidscan
might change in the future? Whilst determin-
be fit with model galaxies with a high formation
ing the redshift distribution of flux-limited sam-
redshift and a rapidly declining star formation
ples, N(m,color,z), with survey spectrographs
rate; such objects would be spectacularly lumi-
was successful in the past, it is less likely to be
nousandblueclosetotheirformationepoch. The
a productive technique in the future. Just as mi-
faint blue excess seen in the counts (Ellis 1997)
crowavebackgroundenthusiastsarguethat“soon”
could originate from such a primordial popula-
we will know the gamut of cosmological param-
tion. Finding primeval galaxies and discovering
the epoch of galaxy formation spectroscopically eters (and so can concentrate on understanding
structure formationand the nature of Λ and dark
thus became a new ‘Holy Grail’ for observers.
matter) so, many argue, we may soon have the
For many reasons, these searches were largely in-
statistical redshift distribution N(z) for any pop-
conclusive (Pritchet 1994) but they did serve to
ulation we desire from photometric redshift tech-
emphasize the importance of faint object spec-
niques (Connolly et al 1996) renderingtraditional
troscopy.
redshift surveys unnecessary, at least for the pur-
Efficient spectrographs equipped with multi-
pose discussed above.
objectfeedsenabledsystematicsurveysofmagnitude-
Thisisnottounderestimatethechallengesthat
limited samples heralding importance advances.
remain in convincing skeptics of this optimistic
Evolution of galaxy populations could be ad-
viewpoint. Photometric redshift techniques suf-
dressedby comparingwell-defined samples within
ferfromanumberofpotentiallydangerouspitfalls
differentredshiftintervals(Lilly et al 1995;Ellis et al 1996;
not least the reliance on template SEDs based ei-
Cowie et al 1996; Cohen et al 2000). These sur-
ther on (i) local galaxydata (which surely cannot
veyshaverevealedacomplexpictureofluminosity-
2
be representativeofthe bewildering variety of ob- 4. The Challenges of Interpreting Re-
jectsweseeatallepochs)or(ii)evolutionarysyn- solved Galaxy Data
thesismodels whicharelargelyuntestedexceptin
Ibelievewefaceanothersignificantcross-roads
the brightregimewherenormalHubble types still
in the subject of galaxy evolution. Resolved data
dominate. Onemust alsorememberthatthe pho-
on distant galaxies may soon be arriving from a
tometric precision appropriate in the oft-quoted
varietyoffacilities,ground-basedandspace-based,
HubbleDeepFieldcomparisons(Hogg et al 1998)
but are we ready to deal with it?
is considerably better than that available in typi-
cal ground-based imaging surveys. Wemightbeginaddressingthisproblembycon-
sidering what has been learnt already from the
Two important new scientific directions have
community’s experience so far with HST images
emerged in the past few years. First, photomet-
of distant galaxies. The most significant conclu-
ric redshift selection can already isolate samples
sions here havecome fromthe verylimited subset
sufficient to permit their angular clustering to in-
of HST imaging for which ground-based redshift
dicate the spatial correlation function at early
data is available. The mismatch in the field of
times (Adelberger et al 1998; Daddi et al 2000;
view between ground-based survey spectrographs
Brunner et al 2000; McCarthy et al 2000). This
andWFPC-2hasseriouslylimitedprogressinthis
provides a valuable measure of large scale struc-
regard and the arrival of the Advanced Camera
tureintermediateinepochbetweenthemicrowave
for Surveys will not make a big impact unless
background and the present, crucial for a direct
thereisaconcertedcampaignforpanoramicimag-
test of gravitational instability. Spectroscopic
ingwithHST.Disregardingrichclusters,onlythe
data is still highly desirable in these imaging sur-
Groth strip, the CFRS/LDSS/Hawaii survey and
veys because if, ultimately, galaxy clustering can
HDFfieldshavebeenmosaicedwithHSTtomatch
becharacterisedadditionallybydynamicalorstel-
ground-based data. Collectively these represent a
larmass,ratherthanviamerepopulationcolorsor
pitifully small coverage considering the potential
star formationrates, the connectionwith contem-
impact of HST imaging in this subject area.
porary model predictions will be much sounder.
These deep surveys differ significantly from their The primary scientific results on galaxy evolu-
local magnitude-limited equivalents (2dF, SDSS) tionfromthe limitedHSTdatasetsinclude claims
inthatredshift-dependentselectioneffectscompli- for (i) physically smaller sources at high red-
cate the analysis necessitating forward modelling shift (Pascarelle et al 1995; Bouwens et al 1998),
of the various biases (Kauffmann et al 1999). (ii) an increasing abundance of peculiar, pos-
sibly merging, blue star-forming systems by
The other direction of importance is one that
z≃1 (Glazebrook et al 1995; Driver et al 1995;
was established via the arrival of WFPC-2. The
Brinchmann & Ellis 2000),and(iii)aslowgrowth
first resolved images of distant galaxies caused
in spiral disks (Lilly et al 1998) and field ellipti-
much optimism in the subject because of the po-
cals (Menanteau et al 2000; Im et al 2000) over
tential for examining evolution of galaxy mor-
0<z <1. Each is a purely empirical result whose
phologies(Griffiths et al 1994). Mostwouldagree
placeinaquantativeunderstandingofgalaxyevo-
thatthe recognitionoffamiliar-lookingspheroidal
and spiral galaxies at redshifts z ≃1 was an im- lution is hindered by, at best, a loose interface
with contemporary theories of structure forma-
portant development. However, as we anticipate
tion. Whereas uncertain assumptions concern-
building ground-based instruments that will de-
liver resolved spectroscopy of faint galaxies, it is ing gas cooling, feedback and related star for-
mation physics are necessary to use structure
relevant to ask what, in terms of quantitative as-
formation models to predict cosmic star forma-
trophysics, we can learn about galaxy evolution
tion histories, luminosities and color distributions
from resolved data.
(Baugh et al 1998; Cole et al 2000), a more pro-
found challenge is now faced in extending theo-
retical models to predict detailed morphological
and structural information. In this respect, rely-
ing solely ona foundation ofgravitationalphysics
3
may be quite inadequate. One suspects the sub- 2<z <5 likewise forms a major driver for multi-
ject would benefit considerably from the expe- conjugate adaptive optics on current and future
rience of those studying the detailed dynamics large telescopes as well as for NGST instruments.
and interstellar media of nearby systems. Some Resolved data for samples of high redshift
progress is possible with dynamical data such as galaxies is not only an inevitable next step in
rotation curves and internal velocity dispersions the developmentofastronomicalinstrumentation,
(Vogt et al 1997; Mo et al 1998; Treu et al 2000) but also a logical scientific direction if we wish to
but only a small subset of the faint galaxy pop- truly understand the variety of galaxy types we
ulation can currently be studied in this way. In- see today. However, observers and theorists need
deed, where attempts have been made to mimic to invest far more intellectual effort to reap the
the hydrodynamical evolution of galaxies, serious benefits of the data arriving soon.
discrepancies arise, for example in understanding
howstellarangularmomentumissustainedindisk REFERENCES
galaxies (Weil et al 1998). Clearly, there is a lot
Abraham, R.G., Valdes, F., Yee, H.K.C. & van den
of work for theorists to do!
Bergh, S. 1994 ApJ, 432, 75.
Moreover, the interpretational deficiency does
not lie entirely at the hands of the theorists. Abraham, R.G., Tanvir, N.R., Santiago, B.X.
We can ask if observers have thought hard et al.MNRAS,279, L47.
enough how to optimally condense their resolved
Adelberger, K.L., Steidel, C.C., Giavalisco, M.
data into useful astrophysical information. In-
et al.1998 ApJ,505, 18.
genious “objective” criteria have been intro-
duced to classify distant systems taking into ac- Bacon, R. 2000 in NGST Science & Technology, eds.
count instrumental, surface brightness and other, Smith,E. & Long, K., ASP Conf. Ser., 207, 333.
redshift-dependent, biases (Abraham et al 1994;
Baugh,C.M.,Cole,S.,Frenk,C.S.&Lacey,C.G.1998
Odewahn 1995;Abraham et al 1996;Conselice et al 2000).
ApJ498, 504.
HST galaxy images can also be ‘decomposed’
structurally,yieldingparameterssuchasdisk/bulge Bouwens, R., Broadhurst, T.J. & Silk, J. 1998 ApJ,
ratiosanddiskscale-lengths(Marleau & Simard 1998; 506, 557.
Ratnatunga et al 1999), although the practicality
Brunner, R.J., Szalay, A.S. & Connolly, A.J. 2000
of doing this for anything other than familiar reg-
ApJ,541, 527.
ular systems remains unclear. Nonetheless, a gap
remains between any objective taxonomic classi- Brinchmann, J. & Ellis, R.S.2000 ApJ,536, L77.
fication and astrophysicalconclusions if, as seems
Cohen, J.G., Hogg, D.W. & Blandford, R. et al.2000
likely, galaxies transform from one class to an-
ApJ,538, 29.
otherasaresultofenvironmentalormergingpro-
cesses (Dressler et al 1997). Perhaps tools appli- Cole, S.M., Lacey, C.G., Baugh, C.M. & Frenk, C.S.
cable to the entire population, equivalent to the 2000 MNRAS,319, 168.
stellar Hertzsprung-Russell diagram, are needed
Connolly,A.J.,Csabai,I.&Szalay,A.S.1996AJ,110,
to make progress in terms of a bigger picture
2655.
(Brinchmann & Ellis 2000).
The observationalmomentumtowardsthe pro- Conselice, C., Bershady, M & Gallagher, J 2000
A&A354, L21.
vision of detailed information on high redshift
galaxies is increasing, which serves only to em-
Cowie, L.L., Songaila, A., Hu, E.M. & Cohen, J.G.
phasize the need to grasp the challenge of its in- 1996 AJ, 112, 839.
terpretation. On the future instrumentation web
sitesforthecurrentsuiteof8-10metertelescopes, Daddi, E., Cimatti, A., Pozzetti, L. et al.2000 A&A,
361, 535.
integralfieldunitspectrographsrepresentadomi-
2
nantfeature . Providingresolvedgalaxydataover
object so that each pixel yields an independent spectrum
(Bacon2000)
2An integral field unit remaps the 2-D image of a resolved
4
Dressler,A.,Oemler,A.,Couch,W.J.et al.1997ApJ, Sandage, A.R. 1961 ApJ,134, 916.
490, 577.
Steidel, C.C., Giavalisco, M., Pettini, M. et al.1996
Driver, S.P., Windhorst, R.A. & Griffiths, R.E. 1995 ApJ,462, L17.
ApJ,453, 48.
Steidel, C.C., Adelberger, K.L., Giavalisco, M.
Ellis, R.S., Colless, M., Broadhurst, T.J. et al.1996 et al.1999 ApJ,519, 1.
MNRAS,280, 235.
Steidel, C.C. 2000 S.P.I.E., 4005, 22.
Ellis, R.S. 1997 ARA&A,35, 389.
Struck-Marcell, C. & Tinsley, B.M. 1978 ApJ, 221,
Glazebrook,K.,Ellis,R.S.,Colless,M.et al.1995MN- 562.
RAS,275, L19.
Tinsley, B.M. 1976 ApJ,203, 63.
Griffiths,R.,Casertano,S.,Ratnatunga,K.et al.1994
ApJ,435, L19. Tinsley, B.M. 1977 ApJ,211, 621.
Hogg, D.W., Cohen, J.L., Blandford, R.D. et al.1998 Tinsley, B.M. 1980 ApJ,241, 41.
AJ, 115, 1418.
Tyson, A.J. & Jarvis, J.F. 1979 ApJ,230, L153.
Im, M., Simard, L., Faber, S.M. et al.2000 ApJ, in
Treu, T., Stiavelli, M., Casertano, S. et al.2000 MN-
press (astro-ph/0011092).
RAS,308, 1037.
Kauffmann, G., Colberg, J.M., Diaferio, A. & White,
Weil,M.,Eke,V.R.&Efstathiou,G.P.1998MNRAS,
S.D.M. 1999 MNRAS,307, 529.
300, 773.
Koo, D.C. 1985 AJ, 90, 418.
Vogt,N.,Phillipps, A.C., Faber,S.M.et al.1997 ApJ,
Kron, R.G. 1980 ApJS,43, 305. 479, L121.
Lilly,S.J.,Tresse,L.,Hammer,F.et al.1995ApJ,455,
108.
Lilly, S.J., Schade, D.J., Ellis, R.S. et al.1999 ApJ,
500, 75.
McCarthy, P., Carlberg, R., Marzke, R. et al.2000
in Deep Fields, ESO Publications in press (astro-
ph/0011499).
Madau, P. & Pozzetti, L. 1999 MNRAS,312, L9.
Marleau, F. & Simard, L. 1998 ApJ,507, 585.
Menanteau,F.,Abraham,R.G.&Ellis,R.S.2000MN-
RAS,in press (astro-ph/0007114).
Mo,H.J.,Mao,S.&White,S.D.M.1998MNRAS295,
319.
Odewahn, S 1995 PASP,107, 7700.
Pascarelle,S.,Windhorst,R.A.,Keel,W.C.et al.1995
Nature,383, 45.
Peterson, B.A., Ellis, R.S., Kibblewhite, E.J.
et al.1979 ApJ,233, L109.
Pritchet, C. 1994 PASP,106, 1052.
Ratnatunga, K.U., Griffiths, R.E. & Ostrander, E.J. This2-columnpreprintwaspreparedwiththeAASLATEX
1999 AJ, 118, 86. macrosv5.0.
5