Title: Galaxies in the UKIDSS Large Area Survey
1Galaxies in the UKIDSS Large Area Survey
- Jon Loveday
- Anthony Smith
- Celine Eminian
- University of Sussex
2Outline
- UK Infrared Deep Sky Survey overview and status
- Near-IR luminosity function
- Photometric redshifts
- Physical Interpretation of near-IR Colours
- Conclusions/future prospects
3Goals
- Large-scale clustering to z 0.6 (BAO, neutrino
mass) - Evolution of galaxy properties (LF, SFR) and
clustering since z 0.6 - Try out techniques on real data before future
surveys such as DES, PanSTARRS, LSST etc begin
4UKIDSS
- UK Infrared Deep Sky Survey
- UKIRT 3.8m telescope plus WFCAM (4x20482
Hawaii-II arrays, 0.21 deg2) - Étendue of 2.38 m2 deg2 largest of any IR camera
until VISTA - zYJHK (1 2.5 ?) near-IR filters
- 5 surveys, 3 extragalactic
- Significantly deeper than 2MASS
5UKIDSS
- Observing started May 2005
- 7 year observing plan (50 of UKIRT time)
- Pipeline processing in Cambridge, archive in
Edinburgh - No consortium proprietary data period
- Data immediately available to ESO members once
verified - Rest of world 18 months later
6(No Transcript)
7UKIDSS Surveys
8Comparison with 2MASS
Lawrence et al 2007
9UKIDSS survey progress
10Near-IR Luminosity Function(Smith, Cross,
Loveday, in prep)
- UKIDSS-LAS DR2 K-band photometry SDSS DR5
redshifts - Need to allow for selection effects in
- r-band flux (SDSS spectro limit)
- K-band flux (UKIDSS completeness limit)
- UKIDSS angular size
- Surface brightness
11LAS Klt16 Vega (17.9 AB)
12SDSS 5740 deg2453,349 galaxies with
redshifts LAS-K 476 deg2
1319,105 galaxies to K16 over 195 deg2 (400,000
over 4000 deg2by end of 2009)
14Multivariate ? 1/Vmax method
16,452 galaxies within selection limits
15K-band BBD (1/Vmax)
(Bivariate Brightness Distribution)
16K-band BBD (SWML)
17Red core (u-r) gt 2.35 (SWML)
18Blue core (u-r) lt 2.35 (SWML)
19K-band luminosity function
20LF Summary
- UKIDSS K-band LF broadly consistent with previous
results - Some discrepancies between 1/Vmax and SWML
estimates - Low-luminosity discrepancy partly due to
large-scale structure? - UKIDSS will be competitive with 2MASS in terms of
volume/galaxy numbers with DR3 onwards (expected
December 2007) - Extend analysis to DXS, UDS and VISTA surveys
with photo-z to probe evolution
21Photometric RedshiftsCeline Eminian
- Use SDSS ugriz and UKIDSS-LAS YJHK magnitudes in
ANNz (Collister Lahav 2004) - Network architecture 510101 (5 bands) or
912121 (9 bands) - Committee of five networks
- For each sample, use SDSS spectroscopy
- 3/8 for training
- 1/8 for verification
- 1/2 for testing (numbers shown on plots)
22SDSS Main
23SDSS Main UKIDSS
24SDSS Main LRGs
25SDSS Main LRGs UKIDSS
26SDSS Main Adding near-IR photometry helps to
reduce outliers
27Photo-z Summary
- At low redshifts (z 0.6) addition of near-IR
photometry helps to improve errors by reducing
outliers - Lack of improvement for LRGs with UKIDSS data due
to - Small training set cf. network size?
- Uniformity of LRG SED?
- Severe lack of spectroscopic training data for
ordinary galaxies at redshifts between 0.2 and 1 - Cannot use LRG-trained network to predict
redshifts of non-LRGs - AAOmega service proposal in queue to obtain
spectroscopic redshifts of wide range of
galaxies out to z 0.6 from coadded data in
SDSS southern stripe
28(No Transcript)
29Physical Interpretation of near-IR Colours
- Eminian et al, 2007, MNRAS in press
- Compare 3-arcsec aperture photometry from SDSS
and UKIDSS-LAS with physical galaxy properties
deduced from SDSS spectra (SDSS-MPA database
Brinchmann et al 2004) and with stellar
population synthesis models - Pair matching technique to remove correlations
with mass, redshift and concentration
30- Increasing star-formation rate correlates with
bluer optical colours but redder near-IR colours
- Due to dominance of TP-AGB stars in HK bands
(Marraston 2005) - These stars also responsible for correlation of
HK with dust?
31Comparsion with BC03
Stars constant SFR Squares ? 3Gyr Ages 5,
10, 15 Gyr bot-top
32Comparsion with CB07 (prelim)
33Conclusions/Future Prospects
- Goal is to measure evolution in stellar mass and
clustering of a wide range of galaxy masses to z
0.6 - Well-calibrated photometric redshifts of
representative galaxies will be vital to do this - UKIDSS DR3 (December 2007) will probe volume
competitive with 2MASS and provide far cleaner
window function for clustering statistics - Immediate goal how well can large-scale
clustering be measured using photo-z (eg. w(?) in
photo-z slices) compared with using spectroscopic
redshifts? - Techniques can then be applied to UKIDSS DXS
UDS, VISTA
34(No Transcript)