Results using a halo-based group finder in the SDSS and

Results using a halohalo-based group
finder in the SDSS and 2dFGRS
Weinmann et al., astroastro-ph/0509147
Yang, Mo and van den Bosch,
astro--ph/0509626
astro
Wh a haloWhy
halo-based group
g o p finder?
finde ?
„
„
„
„
Attempts to identify galaxies residing in the
same dark matter halo.
More direct comparisons with models via halo
model or halo occupation distribution formalism.
Overcomes problems with definitions of
environment
i
incorporating
i
i a fixed
fi d spatial
i l or
density scale.
Gives three separate measures of environment:
halo mass; halohalo-centric radius; and large
large--scale
environment.
Problems measuring
g environment
on a fixed scale.
r5
r5
Group
G o p finding algorithm
algo ithm
1.
2
2.
3.
4.
5.
Use FOF algorithm with small linking length to
identify potential group centres.
Estimate the total group luminosity
luminosity.
Estimate group mass (using assumed M/L),
and thence radius and velocity dispersion.
Reassign group memberships to all galaxies
using threethree-dimensional density contrast in
redshift space.
Recompute a luminosityluminosity-weighted group
centre and iterate steps 2
2--5 until convergence.
„
„
„
„
„
Tested using 2dFGRS
mock catalogues.
~90% completeness.
p
~20% interloper fraction.
Better group mass
estimates by using
estimated Lgroup than by
using velocity dispersion.
Estimate Lgroup at higher
g
redshift from low
low--redshift
calibration, not using the
global luminosity
f
function.
ti
Late-type fraction
Tests of g
group
o p finding algorithm
algo ithm
SDSS sample
„
„
New York University Value Added Galaxy Catalogue,
based on SDSS DR2.
‘Main’
Main galaxy sample:
„
„
„
„
„
„
r<18
0.01≤z≤0.2
0.01
≤z≤0.2
Redshift completeness >0
>0.7
7
Stellar masses obtained from 4000
4000Å
Å break and H
Hδ
δA as
described by Kauffmann et al. (2003).
SFR from emission lines as in Brinchmann et al. (2004);
divide by stellar mass to obtain specific SFR.
179197 galaxies.
SDSS sample
‘L t type’
‘Late
t
’
‘I t
‘Intermediate’
di t ’
‘E l type’
‘Early
t
’
„
„
„
„
„
Almost a one
one--parameter
family of galaxies.
48 1% blue
48.1%
bl & active
ti
(‘late’).
30.7% red & passive
(‘early’)
20.1% red & active
(‘intermediate)
(‘i
di )
1.1% blue & passive –
unclassified.
More acttive Æ
SDSS sample
Redder Æ
T pe ffraction
Type
action by
b halo mass
Use Lgroup, not σ !
„
„
-20
20≥
≥0.1Mr-5logh≥
5logh≥g -22
Using
Us
g velocity
e oc ty
dispersion as a mass
estimator washes out
mass dependence of
type fraction.
Type
yp fraction of satellite galaxies
g
by halohalo-centric radius
Galactic confo
conformity
mit
S mma of SDSS results.
Summary
es lts
„
„
„
„
„
Late type fraction decreases smoothly with halo mass,
and decreases (weakly) with luminosity at fixed halo
mass.
Late type fraction increases with increasing halohalo-centric
radius.
~20%
20% intermediate
inte mediate ttype
pe gala
galaxies
ies at e
every
e halo mass,
mass
luminosity and halohalo-centric radius.
Satellite galaxies seem to know about the properties of
the halo central galaxy, even for haloes of a given mass.
Median galaxy properties depend only weakly on halo
mass.
A theoretical
theo etical inte
interlude
l de
‘The Gao Effect’: for
haloes of a given
mass (below about
1013h-1Msun),
) those
with more recent
formation times are
l
less
clustered.
l t d
Oldest 20%
Y
Youngest
t 20%
Gao, Springel & White (2005)
O in other
Or
othe words…
o ds
Haloes of a given
mass in denser
regions form earlier,
earlier
on average.
Wide dispersion
Wid
di
i in
i
formation times at a
given mass and
g
overdensity.
Looking
g for the Gao Effect in the
2dFGRS
„
„
„
Given the halohalo-based
groups (constructed as
for the SDSS), calculate
the galaxygalaxy-group cross
cross-correlation function.
Use η parameter for
central galaxies to
distinguish between early
types and late types
types.
Works for halo masses
>1013h-1Msun.
Relative bias of halo central
galaxies
More active Æ
Relati e bias split by
Relative
b ηc
„
„
Haloes with a more
passive central galaxy
have a higher relative
bias.
M i groups with
Massive
ih
passive but relatively
faint central galaxies
have the highest bias.
Consequences
q
for galaxy
g
y formation
models
„
Ram pressure stripping?
„
„
„
Strangulation?
„
„
Can it explain enhancement of spheroidals? Seems OK for S0s…
Harassment?
„
„
Scales inversely with mass, which seems to give wrong
dependence on luminosity.
H d to explain
Hard
l i same radial
di l dependence
d
d
in
i haloes
h l
off allll mass.
More trouble with luminosity dependence (LSBs more
vulnerable).
Merger history?
„
Drives differences between haloes, but via other processes;
large--scale bias must come from this to some extent.
large
Conclusions: Gao Effect + galactic
g
conformity
„
„
„
„
„
„
Clustering of central halo galaxy at given halo mass
depends on galaxy properties.
Properties
p
of satellite galaxies
g
correlate with properties
p p
of central galaxies at given halo mass.
This implies a bias dependence on the entire galaxy
population
p
p
of a group.
g p
Older, more clustered halo had galaxies fall in earlier and
become late types by some process.
If AGN feedback is active in the central g
galaxy,
y, it’s had
more time to act if the halo is older (hence consistent
with explanations for the bright end of the galaxy
luminosity function).
Lc dependence of group clustering strength is harder to
explain.
Overall type
yp fraction as a function
of magnitude
Colour
Colo vs.
s concentration
concent ation
Bimodalit of galaxy
Bimodality
gala properties
p ope ties
Type
T pe ffraction
action by
b magnitude
magnit de
Halo mass and magnitude
g
byy
galaxy type
Type
yp fraction of satellite galaxies
g
by halohalo-centric radius
Median galaxy
gala properties
p ope ties
P(M|L t pe)
P(M|L,type)
0 1(g
P(0.1
(g--r)|L,M)
)|L M)
P(c|L M)
P(c|L,M)