CF-HiZELS, a 10 deg$^2$ emission-line survey with spectroscopic follow-up: H\alpha, [OIII]+H\beta\ and [OII] luminosity functions at z=0.8, 1.4 and 2.2

TL;DR

This study presents a large-scale near-infrared emission-line galaxy survey covering 10 square degrees, providing detailed luminosity functions at redshifts 0.8, 1.4, and 2.2, with implications for future cosmological surveys.

Contribution

It offers the largest contiguous volume survey in the near-infrared, combining photometric and spectroscopic data to derive emission-line luminosity functions and analyze cosmic variance effects.

Findings

Detected over 6000 emission-line galaxy candidates.

Derived luminosity functions at z=0.8, 1.4, 2.2.

Quantified cosmic variance impacts on survey volume requirements.

Abstract

We present results from the largest contiguous narrow-band survey in the near-infrared. We have used WIRCam/CFHT and the lowOH2 filter (1.187$\pm$0.005 $\mu$m) to survey ~10 deg$^2$ of contiguous extragalactic sky in the SA22 field. A total of ~6000 candidate emission-line galaxies are found. We use deep ugrizJK data to obtain robust photometric redshifts. We combine our data with the High-redshift Emission Line Survey (HiZELS), explore spectroscopic surveys (VVDS, VIPERS) and obtain our own spectroscopic follow-up with KMOS, FMOS and MOSFIRE to derive large samples of high-redshift emission-line selected galaxies: 3471 H$\alpha$ emitters at z=0.8, 1343 [OIII]+H$\beta$ emitters at z=1.4 and 572 [OII] emitters at z=2.2. We probe co-moving volumes of >10$^6$ Mpc$^3$ and find significant over-densities, including an 8.5$\sigma$ (spectroscopically confirmed) over-density of H$\alpha$ emitters at z=0.81. We derive H$\alpha$, [OIII]+H$\beta$ and [OII] luminosity functions at z=0.8,1.4,2.2, respectively, and present implications for future surveys such as Euclid. Our uniquely large volumes/areas allow us to sub-divide the samples in thousands of randomised combinations of areas and provide a robust empirical measurement of sample/cosmic variance. We show that surveys for star-forming/emission-line galaxies at a depth similar to ours can only overcome cosmic-variance (errors <10%) if they are based on volumes >5x10$^{5}$ Mpc$^{3}$; errors on $L^*$ and $\phi^*$ due to sample (cosmic) variance on surveys probing ~10$^4$ Mpc$^{3}$ and ~10$^5$ Mpc$^{3}$ are typically very high: ~300% and ~40-60%, respectively.