[Paper Review] The VLT Survey Telescope ATLAS

T. Shanks, N. Metcalfe+10 more|arXiv (Cornell University)|Feb 18, 2015

Astronomy and Astrophysical Research2 references4 citations

TL;DR

This paper presents the VLT Survey Telescope ATLAS survey, an optical ugriz survey of 4700 deg² in the Southern Hemisphere with sub-arcsecond seeing (0."8–1."0 FWHM), deeper than SDSS, and photometric zeropoints accurate to ±0.02 mag. It demonstrates excellent photometric consistency with SDSS, validates star-galaxy separation, and confirms completeness limits at 5σ depths, establishing ATLAS as a high-precision, cosmologically valuable southern sky survey.

ABSTRACT

The VLT Survey Telescope (VST) ATLAS is an optical ugriz survey aiming to cover ~4700deg^2 of the Southern sky to similar depths as the Sloan Digital Sky Survey (SDSS). From reduced images and object catalogues provided by the Cambridge Astronomical Surveys Unit we first find that the median seeing ranges from 0.8 arcsec FWHM in i to 1.0 arcsec in u, significantly better than the 1.2-1.5 arcsec seeing for SDSS. The 5 sigma magnitude limit for stellar sources is r_AB=22.7 and in all bands these limits are at least as faint as SDSS. SDSS and ATLAS are more equivalent for galaxy photometry except in the z band where ATLAS has significantly higher throughput. We have improved the original ESO magnitude zeropoints by comparing m<16 star magnitudes with APASS in gri, also extrapolating into u and z, resulting in zeropoints accurate to ~+-0.02 mag. We finally compare star and galaxy number counts in a 250deg^2 area with SDSS and other count data and find good agreement. ATLAS data products can be retrieved from the ESO Science Archive, while support for survey science analyses is provided by the OmegaCAM Science Archive (OSA), operated by the Wide-Field Astronomy Unit in Edinburgh.

Motivation & Objective

• To create a deep, wide-area optical survey of the Southern Hemisphere comparable to SDSS in depth but with superior spatial resolution.
• To calibrate ATLAS photometry using APASS and SDSS overlap data, achieving photometric zeropoints accurate to ±0.02 mag.
• To validate photometric consistency, star-galaxy separation, and completeness limits across the survey.
• To enable cosmological science, including quasar clustering, ISW effect testing, and galaxy cluster detection, using ATLAS data.
• To provide public access to ATLAS data via ESO and OSA archives for broader astronomical use.

Proposed method

• Conducting a deep ugriz imaging survey using the VST and OmegaCAM, covering 4711 deg² with 0.21 arcsec pixels and 1 deg² field-of-view.
• Reducing raw images and producing source catalogues through the Cambridge Astronomical Surveys Unit (CASU).
• Calibrating photometry by comparing ATLAS star magnitudes (m < 16) with APASS in g, r, i, and extrapolating to u and z bands.
• Using 120 deg² NGC overlap with SDSS and 250 deg² SGP field to validate photometric linearity and magnitude consistency.
• Applying aperture photometry (1" radius) and Kron magnitudes for stars and galaxies, respectively, to ensure photometric reliability.
• Employing the OmegaCAM Science Archive (OSA) and ESO Science Archive to store, query, and distribute data via VO-compliant protocols and integration with SDSS, 2MASS, VHS, and WISE.

Experimental results

Research questions

• RQ1 How does the median seeing of ATLAS compare to SDSS, and what impact does this have on photometric depth and resolution?
• RQ2 To what extent is ATLAS photometry consistent with SDSS, and what are the color terms and zeropoint uncertainties?
• RQ3 How effective is the star-galaxy separation in ATLAS, and what are the misclassification rates in overlap regions?
• RQ4 What are the completeness limits for stars and galaxies in ATLAS, and how do they relate to the 5σ magnitude depth?
• RQ5 Can ATLAS data be used to detect large-scale structures such as the 'Local Hole' or superclusters in the Southern Hemisphere?

Key findings

• ATLAS achieves median seeing of 0."8 FWHM in i-band and 1."0 in u-band, significantly better than SDSS’s 1."2–1."5 FWHM.
• The 5σ magnitude limit for stellar sources is r_AB = 22.7, matching or exceeding SDSS depth across all bands.
• Photometric zeropoints are calibrated to ±0.02 mag accuracy via comparison with APASS and SDSS in overlapping fields.
• Star-galaxy separation yields 10% galaxy misclassification as stars and 15% star misclassification as galaxies, primarily due to interchip gaps, but can be mitigated with a conservative separation threshold.
• Galaxy number counts turn over at approximately the 5σ depth limit, indicating effective completeness for galaxies at these magnitudes.
• Star counts turn over at similar limits, but this is primarily driven by the star-galaxy separation algorithm rather than source density.