Spitzer Wide-Area Infrared Extragalactic Research Articles

Preliminary Results of Using k-nearest-neighbor Regression to Estimate the Redshift of Radio-selected Data Sets**Released on 2018 June 1.

In the near future, all-sky radio surveys are set to produce catalogues of tens of millions of sources with limited multiwavelength photometry. Spectroscopic redshifts will only be possible for a small fraction of these new-found sources. In this paper, we provide the first in-depth investigation into the use of k-nearest-neighbor (kNN) regression for the estimation of redshift of these sources. We use Australia Telescope Large Area Survey (ATLAS) radio data, combined with Spitzer Wide-Area Infrared Extragalactic Survey infrared, Dark Energy Survey optical, and Australian Dark Energy Survey spectroscopic survey data. We then reduce the depth of photometry to match what is expected from the upcoming Evolutionary Map of the Universe survey, testing against both data sets. To examine the generalization of our methods, we test one of the subfields of ATLAS against the other. We achieve an outlier rate of ∼10% across all tests, showing that the kNN regression algorithm is an acceptable method of estimating redshift, and would perform better given a sample training set with uniform redshift coverage.

A Lognormal Luminosity Function for SWIRE in Flat Cosmology

The evaluation of the physical parameters or effects such as the luminosity function (LF) or the photometric maximum (PM) for galaxies is routinely modeled by their spectroscopic redshift. Here, we model LF and PM for galaxies by the photometric redshift as given by the Spitzer Wide-Area Infrared Extragalactic (SWIRE) catalog in the framework of a lognormal LF. In addition, we compare our model with the Schechter LF for galaxies. The adopted cosmological framework is that of the flat cosmology.

Radio Galaxy Zoo: machine learning for radio source host galaxy cross-identification

We consider the problem of determining the host galaxies of radio sources by cross-identification. This has traditionally been done manually, which will be intractable for wide-area radio surveys like the Evolutionary Map of the Universe (EMU). Automated cross-identification will be critical for these future surveys, and machine learning may provide the tools to develop such methods. We apply a standard approach from computer vision to cross-identification, introducing one possible way of automating this problem, and explore the pros and cons of this approach. We apply our method to the 1.4 GHz Australian Telescope Large Area Survey (ATLAS) observations of the Chandra Deep Field South (CDFS) and the ESO Large Area ISO Survey South 1 (ELAIS-S1) fields by cross-identifying them with the Spitzer Wide-area Infrared Extragalactic (SWIRE) survey. We train our method with two sets of data: expert cross-identifications of CDFS from the initial ATLAS data release and crowdsourced cross-identifications of CDFS from Radio Galaxy Zoo. We found that a simple strategy of cross-identifying a radio component with the nearest galaxy performs comparably to our more complex methods, though our estimated best-case performance is near 100 per cent. ATLAS contains 87 complex radio sources that have been cross-identified by experts, so there are not enough complex examples to learn how to cross-identify them accurately. Much larger datasets are therefore required for training methods like ours. We also show that training our method on Radio Galaxy Zoo cross-identifications gives comparable results to training on expert cross-identifications, demonstrating the value of crowdsourced training data.

Evolution of the far-infrared luminosity functions in the Spitzer Wide-area Infrared Extragalactic Legacy Survey

We present new observational determination of the evolution of the rest-frame 70 and 160 micron and total infrared (TIR) galaxy luminosity functions (LFs) using 70 micron data from the Spitzer Wide-area Infrared Extragalactic Legacy Survey (SWIRE). The LFs were constructed for sources with spectroscopic redshifts only in the XMM-LSS and Lockman Hole fields from the SWIRE photometric redshift catalogue. The 70 micron and TIR LFs were constructed in the redshift range 0<z<1.2 and the 160 micron LF was constructed in the redshift range 0<z<0.5 using a parametric Bayesian and the vmax methods. We assume in our models, that the faint-end power-law index of the LF does not evolve with redshifts. We find the the double power-law model is a better representation of the IR LF than the more commonly used power-law and Gaussian model. We model the evolution of the FIR LFs as a function of redshift where where the characteristic luminosity, $L^\ast$ evolve as $\propto(1+z)^{\alpha_\textsc{l}}$. The rest-frame 70 micron LF shows a strong luminosity evolution out to z=1.2 with alpha_l=3.41^{+0.18}_{-0.25}. The rest-frame 160 micron LF also showed rapid luminosity evolution with alpha_l=5.53^{+0.28}_{-0.23} out to z=0.5. The rate of evolution in luminosity is consistent with values estimated from previous studies using data from IRAS, ISO and Spitzer. The TIR LF evolves in luminosity with alpha_l=3.82^{+0.28}_{-0.16} which is in agreement with previous results from Spitzer 24 micron which find strong luminosity evolution. By integrating the LF we calculated the co-moving IR luminosity density out to z=1.2, which confirm the rapid evolution in number density of LIRGs and ULIRGs which contribute ~68^{+10}_{-07} % to the co-moving star formation rate density at z=1.2. Our results based on 70 micron data confirms that the bulk of the star formation at z=1 takes place in dust obscured objects.

A DEEP 1.2 mm MAP OF THE LOCKMAN HOLE NORTH FIELD

We present deep 1.2 mm continuum mapping of a 566 arcmin^2 area within the Lockman Hole North field, previously a target of the Spitzer Wide-area Infrared Extragalactic (SWIRE) survey and extremely deep 20 cm mapping with the Very Large Array, which we have obtained using the Max-Planck millimeter bolometer (MAMBO) array on the IRAM 30 m telescope. After filtering, our full map has an RMS sensitivity ranging from 0.45 to 1.5 mJy/beam, with an average of 0.75 mJy/beam. Using the pixel flux distribution in a map made from our best data, we determine the shape, normalization, and approximate flux density cutoff for 1.2 mm number counts well below our nominal sensitivity and confusion limits. After validating our full dataset through comparison with this map, we successfully detect 41 1.2 mm sources with S/N > 4.0 and S(1.2 mm)\simeq 2-5 mJy. We use the most significant of these detections to directly determine the integral number counts down to 1.8 mJy, which are consistent with the results of the pixel flux distribution analysis. 93% of our 41 individual detections have 20 cm counterparts, 49% have Spitzer/MIPS 24 micron counterparts, and one may have a significant Chandra X-ray counterpart. We resolve \simeq 3% of the cosmic infrared background (CIB) at 1.2 mm into significant detections, and directly estimate a 0.05 mJy faint-end cutoff for the counts that is consistent with the full intensity of the 1.2 mm CIB. The median redshift of our 17 detections with spectroscopic or robust photometric redshifts is z(median)=2.3, and rises to z(median)=2.9 when we include redshifts estimated from the radio/far-infrared spectral index. By using a nearest neighbor and angular correlation function analysis, we find evidence that our S/N>4.0 detections are clustered at the 95% confidence level.

Wide-field optical imaging on ELAIS N1, ELAIS N2, First Look Survey and Lockman Hole: observations and source catalogues

We present u-, g-, r-, i- and z-band optical images and associated catalogues taken primarily with the Isaac Newton Telescope Wide Field Camera on the European Large Area ISO Survey (ELAIS) N1 and N2, First Look Survey and Lockman Hole fields comprising a total of 1000 h of integration time over 80 deg2 and approximately 4.3 million objects. In this paper we outline the observations and data processing and characterize the completeness, reliability, photometric and astrometric accuracy of this data set. All images have been photometrically calibrated using the Sloan Digital Sky Survey and a uniform and homogeneous data set is composed over all the observed fields. Magnitude limits are u, g, r, i, z of 23.9, 24.5, 24.0, 23.3, 22.0 (AB, 5σ). These data have been used for optical identification of past and ongoing projects including the surveys ELAIS, Spitzer Wide-Area Infrared Extragalactic Survey, Spitzer Extragalactic Representative Volume Survey and Herschel Multi-tiered Extragalactic Survey.

Spectroscopic follow-up of 70-μm sources in Spitzer Wide-area Infrared Extragalactic Legacy Survey

We present spectroscopic follow-up observations of 70-μm selected galaxies from the Spitzer Wide-area InfraRed Extragalactic Survey XMM-LSS and Lockman Hole fields. We have measured spectroscopic redshifts for 293 new sources down to a 70 μm flux limit of 9 mJy and r < 22 mag. The redshift distribution peaks at z ∼ 0.3 and has a high redshift tail out to z = 3.5. We perform emission line diagnostics for 91 sources where [O iii], Hβ, [N ii], Hα and [S ii] emission lines are available to determine their power source. We find in our sample 13 quasi-stellar objects, one type 2 Seyfert galaxy, 33 star-forming galaxies, 30 composite galaxies, five LINERs and 21 ambiguous galaxies. We fit single temperature dust spectral energy distributions (SEDs) to 81 70-μm sources with 160-μm photometry to estimate dust temperatures and masses. Assuming the dust emissivity factor (β) as 1.5, we determine dust temperatures in the range ∼20–60 K and dust masses with a range of 106–109 M⊙. Plotting these objects in the luminosity–temperature diagram suggests that these objects have lower dust temperatures than local IR luminous galaxies. The Herschel Space Observatory will be crucial in understanding the nature of these sources and to accurately determine the shape of the Rayleigh–Jeans tail of the dust SED. We then model SEDs from optical to far-IR (FIR) for each source using a set of galaxy and quasar templates in the optical and near-IR and with a set of dust emission templates [cirrus, M82 starburst, Arp 220 starburst and active galactic nucleus (AGN) dust torus] in the mid-IR to FIR. The numbers of objects fitted with each dust template are 57 Arp 220, 127 M82, nine cirrus, one AGN dust torus, 70 M82 and cirrus, 26 M82 and AGN dust torus, and three Arp 220 and AGN dust torus. We determine the total IR luminosity (LIR) in range 108–1015 L⊙ by integrating the SED models from 8 to 1000 μm.

MORPHOLOGICALLY IDENTIFIED MERGING GALAXIES IN THE SWIRE FIELDS

We investigate the evolutional and environmental effects on star formation efficiency for more than 400 merging galaxies. The ~400 merging systems, with photometric redshifts smaller than 0.7, are obtained from a catalog of ~15000 morphologically identified merging galaxies derived from observations of the Canada-France-Hawaii Telescope. We also obtained the IR data of the merging galaxies from the Spitzer Wide-area InfraRed Extragalactic Survey (SWIRE). The redshift differences \Delta z between the member galaxies of these merging pairs show a large distribution with 0 < \Delta z < 0.4. We divide our merging pairs into two sub-samples with \Delta z < 0.05 and > 0.05 for further analyses. We find a statistically significant anti-correlation between the specific star formation rate (SSFR) and the separation of the merging galaxies for both sub-samples. Our analyses also show that although most of the merging systems do have enhanced star formation activity, only very rare ones display extremely high SFRs. Additionally, the SSFR of the merging galaxies also decreases when the magnitude difference between two member galaxies becomes large. However, we find that for the merging pairs with large luminosity contrast, the fainter components show higher SSFR than the brighter ones. Finally, there is a higher fraction of gas-poor mergers in galaxy clusters, and the SSFR of gas-rich mergers is reduced in cluster environments.

POLARIZED RADIO SOURCES: A STUDY OF LUMINOSITY, REDSHIFT, AND INFRARED COLORS

The Dominion Radio Astrophysical Observatory Deep Field polarization study has been matched with the Spitzer Wide-Area Infrared Extragalactic survey of the European Large Area Infrared Space Observatory Survey North 1 field. We have used VLA observations with a total intensity rms of 87 microJy beam^-1 to match SWIRE counterparts to the radio sources. Infrared color analysis of our radio sample shows that the majority of polarized sources are elliptical galaxies with an embedded active galactic nucleus. Using available redshift catalogs, we found 429 radio sources of which 69 are polarized with redshifts in the range of 0.04 < z <3.2. We find no correlation between redshift and percentage polarization for our sample. However, for polarized radio sources, we find a weak correlation between increasing percentage polarization and decreasing luminosity.

THE WYOMING SURVEY FOR Hα. III. A MULTI-WAVELENGTH LOOK AT ATTENUATION BY DUST IN GALAXIES OUT TOz∼ 0.4

We report results from the Wyoming Survey for H-alpha (WySH), a comprehensive four-square degree survey to probe the evolution of star-forming galaxies over the latter half of the age of the Universe. We have supplemented the H-alpha data from WySH with infrared data from the Spitzer Wide-area Infrared Extragalactic (SWIRE) Survey and ultraviolet data from the Galaxy Evolution Explorer (GALEX) Deep Imaging Survey. This dataset provides a multi-wavelength look at the evolution of the attenuation by dust, and here we compare a traditional measure of dust attenuation (L(TIR)/L(FUV)) to a diagnostic based on a recently-developed robust star formation rate (SFR) indicator, [H-alpha_obs+24-micron]/H-alpha_obs. With such data over multiple epochs, the evolution in the attenuation by dust with redshift can be assessed. We present results from the ELAIS-N1 and Lockman Hole regions at z~0.16, 0.24, 0.32 and 0.40. While the ensemble averages of both diagnostics are relatively constant from epoch to epoch, each epoch individually exhibits a larger attenuation by dust for higher star formation rates. Hence, an epoch to epoch comparison at a fixed star formation rate suggests a mild decrease in dust attenuation with redshift.

Specific star formation and the relation to stellar mass from 0 &lt;z&lt; 2 as seen in the far-infrared at 70 and 160μm

We use the Spitzer Wide-area InfraRed Extragalactic Legacy Survey (SWIRE) to explore the specific star-formation activity of galaxies and their evolution near the peak of the cosmic far-infrared (FIR) background at 70 and 160um. We use a stacking analysis to determine the mean FIR properties of well defined subsets of galaxies at flux levels well below the FIR catalogue detection limits of SWIRE and other Spitzer surveys. We tabulate the contribution of different subsets of galaxies to the FIR background at 70um and 160um. These long wavelengths provide a good constraint on the bolometric, obscured emission. The large area provides good constraints at low z and in finer redshift bins than previous work. At all redshifts we find that the specific FIR Luminosity (sLFIR) decreases with increasing mass, following a trend L_FIR/M* propto M_* ^beta with beta =-0.38\pm0.14. This is a more continuous change than expected from the {Delucia2007} semi-analytic model suggesting modifications to the feedback prescriptions. We see an increase in the sLFIR by about a factor of ~100 from 0<z<2 and find that the sLFIR evolves as (1+z)^alpha with alpha=4.4\pm0.3 for galaxies with 10.5 < log M*/Msun < 12. This is considerably steeper than the {Delucia2007} semi-analytic model (alpha \sim 2.5). When separating galaxies into early and late types on the basis of the optical/IR spectral energy distributions we find that the decrease in sLFIR with stellar mass is stronger in early type galaxies (beta ~ -0.46), while late type galaxies exhibit a flatter trend (beta \sim -0.15). The evolution is strong for both classes but stronger for the early type galaxies. The early types show a trend of decreasing strength of evolution as we move from lower to higher masses while the evolution of the late type galaxies has little dependence on stellar mass. We suggest that in late-type galaxies we are seeing a consistently declining sSFR..

EVIDENCE FOR INFRARED-FAINT RADIO SOURCES ASz&gt; 1 RADIO-LOUD ACTIVE GALACTIC NUCLEI

Infrared-Faint Radio Sources (IFRSs) are a class of radio objects found in the Australia Telescope Large Area Survey (ATLAS) which have no observable mid-infrared counterpart in the Spitzer Wide-area Infrared Extragalactic (SWIRE) survey. The extended Chandra Deep Field South now has even deeper Spitzer imaging (3.6 to 70 micron) from a number of Legacy surveys. We report the detections of two IFRS sources in IRAC images. The non-detection of two other IFRSs allows us to constrain the source type. Detailed modeling of the SED of these objects shows that they are consistent with high redshift (z > 1) AGN.

The discovery of a typical radio galaxy at z = 4.88

Abstract In this Letter, we report the discovery of a z= 4.88 radio galaxy discovered with a new technique which does not rely on pre-selection of a sample based on radio properties such as steep-spectral index or small angular size. This radio galaxy was discovered in the Elais-N2 field and has a spectral index of α= 0.75, i.e. not ultra-steep spectrum. It also has a luminosity consistent with being drawn from the break of the radio luminosity function and can therefore be considered as a typical radio galaxy. Using the Spitzer Wide-Area Infrared Extragalactic Survey (SWIRE) data over this field, we find that the host galaxy is consistent with being similarly massive to the lower redshift powerful radio galaxies (∼1–3L★). However, we note that at z= 4.88, the Hα line is redshifted into the IRAC 3.6 μm filter, and some of the flux in this band may be due to this fact rather than the stellar continuum emission. The discovery of such a distant radio source from our initial spectroscopic observations demonstrates the promise of our survey for finding the most distant radio sources.

The relationship between star formation rate and radio synchrotron luminosity at0 &lt; z &lt; 2

We probe the relationship between star formation rate (SFR) and radio synchrotron luminosity in galaxies at 0 < z < 2 within the northern Spitzer Wide-area Infrared Extragalactic survey (SWIRE) fields, in order to investigate some of the assumptions that go into calculating the star formation history of the Universe from deep radio observations. We present new 610-MHz Giant Metrewave Radio Telescope (GMRT) observations of the European Large-Area ISO Survey (ELAIS)-N2 field, and using this data, along with previous GMRT surveys carried out in the ELAIS-N1 and Lockman Hole regions, we construct a sample of galaxies which have redshift and SFR information available from the SWIRE survey. We test whether the local relationship between SFR and radio luminosity is applicable to z = 2 galaxies, and look for evolution in this relationship with both redshift and SFR in order to examine whether the physical processes which lead to synchrotron radiation have remained the same since the peak of star formation in the Universe. We find that the local calibration between radio luminosity and star formation can be successfully applied to radio-selected high-redshift, high-SFR galaxies, although we identify a small number of sources where this may not be the case; these sources show evidence for inaccurate estimations of their SFR, but there may also be some contribution from physical effects such as the recent onset of starburst activity, or suppression of the radio luminosity within these galaxies.

EVOLUTION OF THE RATE AND MODE OF STAR FORMATION IN GALAXIES SINCE z = 0.7

We present the star formation rate (SFR) and starburst fraction (SBF) for a sample of field galaxies from the ICBS intermediate-redshift cluster survey. We use [O II] and Spitzer 24 micron fluxes to measure SFRs, and 24 micron fluxes and H-delta absorption to measure of SBFs, for both our sample and a present-epoch field sample from the Sloan Digital Sky Survey (SDSS) and Spitzer Wide-area Infrared Extragalactic (SWIRE) survey. We find a precipitous decline in the SFR since z=1, in agreement with other studies, as well as a corresponding rapid decline in the fraction of galaxies undergoing long-duration moderate-amplitude starbursts. We suggest that the change in both the rate and mode of star formation could result from the strong decrease since z=1 of gas available for star formation.

The evolution of star formation in quasar host galaxies

We have used far-infrared data from IRAS, ISO, SWIRE, SCUBA and MAMBO to constrain statistically the mean far-infrared luminosities of quasars. Our quasar compilation at redshifts 0<z<6.5 and I-band luminosities -20<I(AB)<-32 is the first to distinguish evolution from quasar luminosity dependence in such a study. We carefully cross-calibrate IRAS against Spitzer and ISO, finding evidence that IRAS 100um fluxes at <1Jy are overestimated by ~30%. We find evidence for a correlation between star formation in quasar hosts and the quasar optical luminosities, varying as SFR proportional to L_opt^(0.44+/-0.07) at any fixed redshift below z=2. We also find evidence for evolution of the mean star formation rate in quasar host galaxies, scaling as (1+z)^(1.6+/-0.3) at z<2 for any fixed quasar I-band absolute magnitude fainter than -28. We find no evidence for any correlation between star formation rate and black hole mass at 0.5<z<4. Our data are consistent with feedback from black hole accretion regulating stellar mass assembly at all redshifts.

Mid-infrared spectroscopy of infrared-luminous galaxies atz∼ 0.5-3

We present results on low-resolution mid-infrared (MIR) spectra of 70 IR-luminous galaxies obtained with the infrared spectrograph (IRS) onboard Spitzer. We selected sources from the European Large Area Infrared Survey with S15 > 0.8 mJy and photometric or spectroscopic z > 1. About half of the samples are quasi-stellar objects (QSOs) in the optical, while the remaining sources are galaxies, comprising both obscured active galactic nuclei (AGN) and starbursts. Redshifts were obtained from optical spectroscopy, photometric redshifts and the IRS spectra. The later turn out to be reliable for obscured and/or star-forming sources, thus becoming an ideal complement to optical spectroscopy for redshift estimation. We estimate monochromatic luminosities at several rest-frame wavelengths, equivalent widths and luminosities for the polycyclic aromatic hydrocarbon (PAH) features, and strength of the silicate feature in individual spectra. We also estimate integrated 8–1000 μ mI R luminosities via spectral energy distribution fitting to MIR and far-IR (FIR) photometry from the Spitzer Wide-Area Infrared Extragalactic survey and the MIR spectrum. Based on these measurements, we classify the spectra using well-known IR diagnostics, as well as a new one that we propose, into three types of source: those dominated by an unobscured AGN, mostly corresponding to optical quasars (QSOs), those dominated by an obscured AGN and starburst-dominated sources. Starbursts concentrate at z ∼ 0.6–1.0 favoured by the shift of the 7.7-μm PAH band into the selection 15-μm band, while AGN spread over the 0.5 < z < 3.1 range. Star formation rates (SFR) are estimated for individual sources from the luminosity of the PAH features. An estimate of the average PAH luminosity in QSOs and obscured AGN is

THE DEEP SWIRE FIELD. I. 20 cm CONTINUUM RADIO OBSERVATIONS: A CROWDED SKY

We present results from deep radio observations taken with the Very Large Array at a center frequency of 1400 MHz covering a region of the Spitzer Wide-area InfraRed Extragalactic Survey of the Spitzer Legacy survey, centered at 10h46m00s, 59°01'00''(J2000). The reduction and cataloging of radio sources are described. An electronic catalog of the sources detected above 5σ is also presented. The survey presented is the deepest so far in terms of the radio source density on the sky. Perhaps surprisingly, the sources down to the bottom of the catalog appear to have median angular sizes still greater than 1 arcsec, like their cousins 10-100 times stronger. The shape of the differential log N-log S counts also seems to require a correction for the finite sizes of the sources to be self-consistent. If the log N-log S normalization remains constant at the lowest flux densities, there are about six sources per square arcminute at 15 μJy at 20 cm. Given the finite-source size this implies that we may reach the natural confusion limit near 1 μJy.

Properties of dusty tori in active galactic nuclei – I. The case of SWIRE/SDSS quasars

We derive the properties of dusty tori in active galactic nuclei from the comparison of observed spectral energy distributions (SEDs) of SDSS quasars and a precomputed grid of torus models. The observed SEDs comprise SDSS photometry, Two-Micron All-Sky Survey J, H and K data, whenever available, and mid-infrared (mid-IR) data from the Spitzer Wide-area InfraRed Extragalactic Survey. The adopted model is that of Fritz, Franceschini & Hatziminaoglou. The fit is performed by standard chi(2)-minimization; the model, however, can be a multicomponent comprising a stellar and a starburst component, whenever necessary. Models with low equatorial optical depth, tau(9.7), were allowed as well as 'traditional' models with tau(9.7) >= 1.0, corresponding to A(V) >= 22 and the results were compared. Fits using high optical depth tori models only produced dust more compactly distributed than in the configuration where all tau(9.7) models were permitted. Tori with decreasing dust density with the distance from the centre were favoured while there was no clear preference for models with or without angular variation of the dust density. The computed outer radii of the tori are of some tens of parsecs large but can reach, in a few cases, a few hundreds of parsecs. The mass of dust, M-Dust, and IR luminosity, L-IR, integrated in the wavelength range between 1 and 1000 mu m, do not show significant variations with redshift, once the observational biases are taken into account. Objects with 70-mu m detections, representing 25 per cent of the sample, are studied separately and the starburst contribution (whenever present) to the IR luminosity can reach, in the most extreme but very few cases, 80 per cent.

High‐Redshift QSOs in the SWIRE Survey and thez∼3 QSO Luminosity Function

We use a simple optical/infrared (IR) photometric selection of high-redshift QSOs that identifies a Lyman break in the optical photometry and requires a red IR color to distinguish QSOs from common interlopers. The search yields 100 z ∼ 3 (U-dropout) QSO candidates with 19 < r' < 22 over 11.7 deg2 in the ELAIS-N1 (EN1) and ELAIS-N2 (EN2) fields of the Spitzer Wide-area Infrared Extragalactic (SWIRE) Legacy Survey. The z ∼ 3 selection is reliable, with spectroscopic follow-up of 10 candidates confirming that they are all QSOs at 2.83 < z < 3.44. We find that our z ∼ 4 (g'-dropout) sample suffers from both unreliability and incompleteness but present seven previously unidentified QSOs at 3.50 < z < 3.89. Detailed simulations show our z ∼ 3 completeness to be ~80%-90% from 3.0 < z < 3.5, significantly better than the ~30%-80% completeness of the SDSS at these redshifts. The resulting luminosity function extends 2 mag fainter than SDSS and has a faint-end slope of β = − 1.42 ± 0.15, consistent with values measured at lower redshift. Therefore, we see no evidence for evolution of the faint-end slope of the QSO luminosity function. Including the SDSS QSO sample, we have now directly measured the space density of QSOs responsible for ~70% of the QSO UV luminosity density at z ∼ 3. We derive a maximum rate of H I photoionization from QSOs at z ∼ 3.2, Γ = 4.8 × 10−13 s−1, about half of the total rate inferred through studies of the Lyα forest. Therefore, star-forming galaxies and QSOs must contribute comparably to the photoionization of H I in the intergalactic medium at z ∼ 3.