arXiv subjects

S. Heinis

Publications and source records attributed to S. Heinis.

At least 19 recordsLinked to original sources

arXiv2016

GAMA/H-ATLAS: Common star-formation rate indicators and their dependence on galaxy physical parameters

We compare common star-formation rate (SFR) indicators in the local Universe in the GAMA equatorial fields (around 160 sq. deg.), using ultraviolet (UV) photometry from GALEX, far-infrared (FIR) and sub-millimetre (sub-mm) photometry from H-ATLAS, and Halpha spectroscopy from the GAMA survey. With a high-quality sample of 745 galaxies (median redshift 0.08), we consider three SFR tracers: UV luminosity corrected for dust attenuation using the UV spectral slope beta (SFRUV,corr), Halpha line luminosity corrected for dust using the Balmer decrement (BD) (SFRHalpha,corr), and the combination of UV and IR emission (SFRUV+IR). We demonstrate that SFRUV,corr can be reconciled with the other two tracers after applying attenuation corrections by calibrating IRX (i.e. the IR to UV luminosity ratio) and attenuation in the Halpha (derived from BD) against beta. However, beta on its own is very unlikely to be a reliable attenuation indicator. We find that attenuation correction factors depend on parameters such as stellar mass, z and dust temperature (Tdust), but not on Halpha equivalent width (EW) or Sersic index. Due to the large scatter in the IRX vs beta correlation, when compared to SFRUV+IR, the beta-corrected SFRUV,corr exhibits systematic deviations as a function of IRX, BD and Tdust.

astro-ph.GA↗

arXiv2016

The host galaxy properties of variability selected AGN in the Pan-STARRS1 Medium-Deep Survey

We study the properties of 975 active galactic nuclei (AGN) selected by variability in the Pan-STARRS1 Medium-Deep Survey. Using complementary multi wavelength data from the ultraviolet to the far-infrared, we use SED fitting to determine the AGN and host properties at $z<1$, and compare to a well-matched control sample. We confirm the trend previously observed that the variability amplitude decreases with AGN luminosity, but on the other hand, we observe that the slope of this relation steepens with wavelength resulting in a "redder when brighter" trend at low luminosities. Our results show that AGN are hosted by more massive hosts than control sample galaxies, while the restframe, dust-corrected $NUV-r$ color distribution of AGN hosts is similar to control galaxies. We find a positive correlation between the AGN luminosity and star formation rate (SFR), independent of redshift. AGN hosts populate the whole range of SFRs within and outside the Main Sequence of star forming galaxies. Comparing the distribution of AGN hosts and control galaxies, we show that AGN hosts are less likely to be hosted by quiescent galaxies, but more likely to be hosted by Main Sequence or starburst galaxies.

astro-ph.GA↗

arXiv2016

Of Genes and Machines: application of a combination of machine learning tools to astronomy datasets

We apply a combination of a Genetic Algorithms (GA) and Support Vector Machines (SVM) machine learning algorithm to solve two important problems faced by the astronomical community: star/galaxy separation, and photometric redshift estimation of galaxies in survey catalogs. We use the GA to select the relevant features in the first step, followed by optimization of SVM parameters in the second step to obtain an optimal set of parameters to classify or regress, in process of which we avoid over-fitting. We apply our method to star/galaxy separation in Pan-STARRS1 data. We show that our method correctly classifies 98% of objects down to i_P1= 24.5, with a completeness (or true positive rate) of 99% for galaxies, and 88% for stars. By combining colors with morphology, our star/classification method yields better results than the new SExtractor classifier spread_model in particular at the faint end (i_P1>22). We also use our method to derive photometric redshifts for galaxies in the COSMOS bright multi-wavelength dataset down to an error in (1+z) of sigma=0.013, which compares well with estimates from SED fitting on the same data (sigma=0.007) while making a significantly smaller number of assumptions.

astro-ph.IM↗

arXiv2015

Dust properties of Lyman break galaxies at $z\sim3$

We explore from a statistical point of view the far-infrared (far-IR) and sub-millimeter (sub-mm) properties of a large sample of LBGs (22,000) at z~3 in the COSMOS field. The large number of galaxies allows us to split it in several bins as a function of UV luminosity, UV slope, and stellar mass to better sample their variety. We perform stacking analysis in PACS (100 and 160 um), SPIRE (250, 350 and 500 um) and AzTEC (1.1 mm) images. Our stacking procedure corrects the biases induced by galaxy clustering and incompleteness of our input catalogue in dense regions. We obtain the full IR spectral energy distributions (SED) of subsamples of LBGs and derive the mean IR luminosity as a function of UV luminosity, UV slope, and stellar mass. The average IRX is roughly constant over the UV luminosity range, with a mean of 7.9 (1.8 mag). However, it is correlated with UV slope, and stellar mass. We investigate using a statistically-controlled stacking analysis as a function of (stellar mass, UV slope) the dispersion of the IRX-UVslope and IRX-M* plane. Our results enable us to study the average relation between star-formation rate (SFR) and stellar mass, and we show that our LBG sample lies on the main sequence of star formation at z~3.

astro-ph.GA↗

arXiv2015

The HerMES sub-millimetre local and low-redshift luminosity functions

We used wide area surveys over 39 deg$^2$ by the HerMES collaboration, performed with the Herschel Observatory SPIRE multi-wavelength camera, to estimate the low-redshift, $0.02<z<0.5$, monochromatic luminosity functions (LFs) of galaxies at 250, 350 and 500$\,μ$m. SPIRE flux densities were also combined with Spitzer photometry and multi-wavelength archival data to perform a complete SED fitting analysis of SPIRE detected sources to calculate precise k-corrections, as well as the bolometric infrared (8-1000$\,μ$m) luminosity functions and their low-$z$ evolution from a combination of statistical estimators. Integration of the latter prompted us to also compute the local luminosity density (LLD) and the comoving star formation rate density (SFRD) for our sources, and to compare them with theoretical predictions of galaxy formation models. The luminosity functions show significant and rapid luminosity evolution already at low redshifts, $0.02<z<0.2$, with L$_{IR}^* \propto (1+z)^{6.0\pm0.4}$ and $Φ_{IR}^* \propto (1+z)^{-2.1\pm0.4}$, L$_{250}^* \propto (1+z)^{5.3\pm0.2}$ and $Φ_{250}^* \propto (1+z)^{-0.6\pm0.4}$ estimated using the IR bolometric and the 250$\,μ$m LFs respectively. Converting our IR LD estimate into an SFRD assuming a standard Salpeter IMF and including the unobscured contribution based on the UV dust-uncorrected emission from local galaxies, we estimate a SFRD scaling of SFRD$_0+0.08 z$, where SFRD$_0\simeq (1.9\pm 0.03)\times 10^{-2} [\mathrm{M}_\odot\,\mathrm{Mpc}^{-3}]$ is our total SFRD estimate at $z\sim0.02$.

astro-ph.GA↗

arXiv2015

The GALEX Ultraviolet Virgo Cluster Survey (GUViCS). VI: The UV luminosity function of the Virgo cluster and its surrounding regions

We use the GALEX data of the GUViCS survey to construct the NUV luminosity function of the Virgo cluster over ~ 300 deg.2, an area covering the cluster and its surrounding regions up to ~ 1.8 virial radii. The NUV luminosity function is also determined for galaxies of different morphological type and NUV-i colour, and for the different substructures within the cluster. These luminosity functions are robust vs. statistical corrections since based on a sample of 833 galaxies mainly identified as cluster members with spectroscopic redshift (808) or high-quality optical scaling relations (10). We fit these luminosity functions with a Schechter function, and compare the fitted parameters with those determined for other nearby clusters and for the field. The faint end slope of the Virgo NUV luminosity function (alpha = -1.19), here sampled down to ~ NUV = -11.5 mag, is significantly flatter than the one measured in other nearby clusters and similar to the field one. Similarly M* = -17.56 is one-to-two magnitudes fainter than measured in Coma, A1367, the Shapley supercluster, and the field. These differences seem due to the quite uncertain statistical corrections and the small range in absolute magnitude sampled in these clusters. We do not observe strong systematic differences in the overall NUV luminosity function of the core of the cluster with respect to that of its periphery. We notice, however, that the relative contribution of red-to-blue galaxies at the faint end is inverted, with red quiescent objects dominating the core of the cluster and star forming galaxies dominating beyond one virial radius. These observational evidences are discussed in the framework of galaxy evolution in dense environments.

astro-ph.GA↗

arXiv2015

The spectral energy distribution of galaxies at z > 2.5: Implications from the Herschel/SPIRE color-color diagram

We use the Herschel SPIRE color-color diagram to study the spectral energy distribution (SED) and the redshift estimation of high-z galaxies. We compiled a sample of 57 galaxies with spectroscopically confirmed redshifts and SPIRE detections in all three bands at $z=2.5-6.4$, and compared their average SPIRE colors with SED templates from local and high-z libraries. We find that local SEDs are inconsistent with high-z observations. The local calibrations of the parameters need to be adjusted to describe the average colors of high-z galaxies. For high-z libraries, the templates with an evolution from z=0 to 3 can well describe the average colors of the observations at high redshift. Using these templates, we defined color cuts to divide the SPIRE color-color diagram into different regions with different mean redshifts. We tested this method and two other color cut methods using a large sample of 783 Herschel-selected galaxies, and find that although these methods can separate the sample into populations with different mean redshifts, the dispersion of redshifts in each population is considerably large. Additional information is needed for better sampling.

astro-ph.GA↗

arXiv2015

The GALEX Ultraviolet Virgo Cluster Survey (GUViCS). V. Ultraviolet diffuse emission and cirrus properties in the Virgo cluster direction

CONTEXT: The Virgo direction has been observed at many wavelengths in the recent years, in particular in the ultraviolet with GALEX. The far ultraviolet (FUV) diffuse light detected by GALEX bears interesting information on the large scale distribution of Galactic dust, owing to the GALEX FUV band sensitivity and resolution. AIMS: We aim to characterise the ultraviolet large scale distribution of diffuse emission in the Virgo direction. A map of this emission may become useful for various studies by identifying regions where dust affects observations by either scattering light or absorbing radiation. METHODS: We construct mosaics of the FUV and near ultraviolet diffuse emission over a large sky region (RA 12 to 13 hours, DEC 0 to 20 degrees) surrounding the Virgo cluster, using all the GALEX available data in the area. We test for the first time the utilisation of the FUV diffuse light as a Galactic extinction E(B-V) tracer. RESULTS: The FUV diffuse light scattered on cirrus reveals details in their geometry. Despite a large dispersion, the FUV diffuse light correlates roughly with other Galactic dust tracers (coming from IRAS, Herschel, Planck), offering an opportunity to use the FUV emission to locate them in future studies with a better resolution (about 5 arcsec native resolution, 20 arcsec pixels maps presented in this paper) than several usual tracers. Estimating the Galactic dust extinction on the basis of this emission allows us to find a smaller dispersion in the NUV-i colour of background galaxies at a given E(B-V)than with other tracers. The diffuse light mosaics obtained in this work are made publicly available.

astro-ph.GA↗

arXiv2015

Dust attenuation up to z~2 in the AKARI North Ecliptic Pole Deep Field

(Abridged) We aim to study the evolution of dust attenuation in galaxies selected in the IR in the redshift range in which they are known to dominate the star formation activity in the universe. The comparison with other measurements of dust attenuation in samples selected using different criteria will give us a global picture of the attenuation at work in star-forming galaxies and its evolution with redshift. Using multiple filters of IRC instrument, we selected more than 4000 galaxies from their rest-frame emission at 8 microns, from z~0.2 to 2$. We built SEDs from the rest-frame UV to the far-IR by adding data in the optical-NIR and from GALEX and Herschel surveys. We fit SEDs with the physically-motivated code CIGALE. We test different templates for AGNs and recipes for dust attenuation and estimate stellar masses, SFRs, amount of dust attenuation, and AGN contribution to the total IR luminosity. The AGN contribution to the total IR luminosity is found to be on average approximately 10% with a slight increase with redshift. Dust attenuation in galaxies dominating the IR luminosity function is found to increase from z=0 to z=1 and to remain almost constant from z=1 to z=1.5. Conversely, when galaxies are selected at a fixed IR luminosity, their dust attenuation slightly decreases as redshift increases but with a large dispersion. The attenuation in our mid-IR selected sample is found ~ 2 mag higher than that found globally in the universe or in UV and Halpha line selections in the same redshift range. This difference is well explained by an increase of dust attenuation with the stellar mass, in global agreement with other recent studies. Starbursting galaxies do not systematically exhibit a high attenuation

astro-ph.GA↗

arXiv2015

GALEX Detection of Shock Breakout in Type II-P Supernova PS1-13arp: Implications for the Progenitor Star Wind

We present the GALEX detection of a UV burst at the time of explosion of an optically normal Type II-P supernova (PS1-13arp) from the Pan-STARRS1 survey at z=0.1665. The temperature and luminosity of the UV burst match the theoretical predictions for shock breakout in a red supergiant, but with a duration a factor of ~50 longer than expected. We compare the $NUV$ light curve of PS1-13arp to previous GALEX detections of Type IIP SNe, and find clear distinctions that indicate that the UV emission is powered by shock breakout, and not by the subsequent cooling envelope emission previously detected in these systems. We interpret the ~ 1 d duration of the UV signal with a shock breakout in the wind of a red supergiant with a pre-explosion mass-loss rate of ~ 10^-3 Msun yr^-1. This mass-loss rate is enough to prolong the duration of the shock breakout signal, but not enough to produce an excess in the optical plateau light curve or narrow emission lines powered by circumstellar interaction. This detection of non-standard, potentially episodic high mass-loss in a RSG SN progenitor has favorable consequences for the prospects of future wide-field UV surveys to detect shock breakout directly in these systems, and provide a sensitive probe of the pre-explosion conditions of SN progenitors.

astro-ph.HE↗

arXiv2015

Selection of Burst-like Transients and Stochastic Variables Using Multi-Band Image Differencing in the Pan-STARRS1 Medium-Deep Survey

We present a novel method for the light-curve characterization of Pan-STARRS1 Medium Deep Survey (PS1 MDS) extragalactic sources into stochastic variables (SV) and burst-like (BL) transients, using multi-band image-differencing time-series data. We select detections in difference images associated with galaxy hosts using a star/galaxy catalog extracted from the deep PS1 MDS stacked images, and adopt a maximum a posteriori formulation to model their difference-flux time-series in four Pan-STARRS1 photometric bands g,r,i, and z. We use three deterministic light-curve models to fit burst-like transients and one stochastic light curve model, the Ornstein-Uhlenbeck process, in order to fit variability that is characteristic of active galactic nuclei (AGN). We assess the quality of fit of the models band-wise source-wise, using their estimated leave-out-one cross-validation likelihoods and corrected Akaike information criteria. We then apply a K-means clustering algorithm on these statistics, to determine the source classification in each band. The final source classification is derived as a combination of the individual filter classifications. We use our clustering method to characterize 4361 extragalactic image difference detected sources in the first 2.5 years of the PS1 MDS, into 1529 BL, and 2262 SV, with a purity of 95.00% for AGN, and 90.97% for SN based on our verification sets. We combine our light-curve classifications with their nuclear or off-nuclear host galaxy offsets, to define a robust photometric sample of 1233 active galactic nuclei and 812 supernovae. We use these samples to identify simple photometric priors that would enable their real-time identification in future wide-field synoptic surveys.

astro-ph.IM↗

arXiv2014

The GALEX Ultraviolet Virgo Cluster Survey (GUViCS). IV: The role of the cluster environment on galaxy evolution

We study the role of the environment on galaxy evolution using a sample of 868 galaxies in the Virgo cluster and in its surrounding regions selected from the GUViCS Survey with the purpose of understanding the origin of the red sequence in dense environments. We collected multifrequency data covering the whole electromagnetic spectrum for most of the galaxies. We identify the different dynamical substructures composing the Virgo cluster and we calculate the local density of galaxies using different methods. We then study the distribution of galaxies belonging to the red sequence, the green valley, and the blue cloud within the different cluster substructures. Our analysis indicates that all the most massive galaxies are slow rotators and are the dominant galaxies of the different cluster substructures generally associated with a diffuse X-ray emission. They are probably the result of major merging events that occurred at early epochs. Slow rotators of lower stellar mass are also preferentially located within the different high-density substructures of the cluster. They are virialised within the cluster, thus Virgo members since its formation. They have been shaped by gravitational perturbations occurring within the infalling groups that later formed the cluster. On the contrary, low-mass star-forming systems are extremely rare in the inner regions of the Virgo cluster A, where the density of the intergalactic medium is at its maximum. Our ram pressure stripping models consistently indicate that these star-forming systems can be rapidly deprived of their interstellar medium during their interaction with the intergalactic medium. The lack of gas quenches their star formation activity transforming them into quiescent dwarf ellipticals. This mild transformation does not perturb the kinematic properties of these galaxies which still have rotation curves typical of star-forming systems.

astro-ph.GA↗

arXiv2014

Modeling the connection between ultraviolet and infrared galaxy populations across cosmic times

Using a phenomenological approach, we self-consistently model the redshift evolution of the ultraviolet (UV) and infrared (IR) luminosity functions across cosmic time, as well as a range of observed IR properties of UV-selected galaxy population. This model is an extension of the 2SFM (2 star-formation modes) formalism, which is based on the observed "main-sequence" of star-forming galaxies, i.e. a strong correlation between their stellar mass and their star formation rate (SFR), and a secondary population of starbursts with an excess of star formation. The balance between the UV light from young, massive stars and the dust-reprocessed IR emission is modeled following the empirical relation between the attenuation (IRX for IR excess hereafter) and the stellar mass, assuming a scatter of 0.4\,dex around this relation. We obtain a good overall agreement with the measurements of the IR luminosity function up to z~3 and the UV luminosity functions up to z~6, and show that a scatter on the IRX-M relation is mandatory to reproduce these observables. We also naturally reproduce the observed, flat relation between the mean IRX and the UV luminosity at L$_{\rm UV}>$10$^{9.5}$ L$_\odot$. Finally, we perform predictions of the UV properties and detectability of IR-selected samples and the vice versa, and discuss the results in the context of the UV-rest-frame and sub-millimeter surveys of the next decade.

astro-ph.GA↗

arXiv2014

Constraining the Lyman Alpha Escape Fraction with Far-Infrared Observations of Lyman Alpha Emitters

We study the far-infrared properties of 498 Lyman Alpha Emitters (LAEs) at z=2.8, 3.1 and 4.5 in the Extended Chandra Deep Field-South, using 250, 350 and 500 micron data from the Herschel Multi-tiered Extragalactic Survey (HerMES) and 870 micron data from the LABOCA ECDFS Submillimeter Survey (LESS). None of the 126, 280 or 92 LAEs at z=2.8, 3.1 and 4.5, respectively, are individually detected in the far-infrared data. We use stacking to probe the average emission to deeper flux limits, reaching $1σ$ depths of ~0.1 to 0.4 mJy. The LAEs are also undetected at $\ge3σ$ in the stacks, although a $2.5σ$ signal is observed at 870 micron for the z=2.8 sources. We consider a wide range of far-infrared spectral energy distributions (SEDs), including a M82 and an Sd galaxy template, to determine upper limits on the far-infrared luminosities and far-infrared-derived star-formation rates of the LAEs. These star-formation rates are then combined with those inferred from the Ly$α$ and UV emission to determine lower limits on the LAEs Ly$α$ escape fraction ($f_{\rm esc}($Ly$α$)). For the Sd SED template, the inferred LAEs $f_{\rm esc}($Ly$α$) are $\gtrsim30%$ ($1σ$) at z=2.8, 3.1 and 4.5, which are all significantly higher than the global $f_{\rm esc}($Ly$α$) at these redshifts. Thus, if the LAEs $f_{\rm esc}($Ly$α$) follows the global evolution then they have warmer far-infrared SEDs than the Sd galaxy template. The average and M82 SEDs produce lower limits on the LAE $f_{\rm esc}($Ly$α$) of ~10 to 20% ($1σ$), all of which are slightly higher than the global evolution of $f_{\rm esc}($Ly$α$) but consistent with it at the 2 to 3$σ$ level.

astro-ph.CO↗

arXiv2014

Dust Spectral Energy Distributions of Nearby Galaxies: an Insight from the Herschel Reference Survey

We gather infrared (IR) photometric data from 8 to 500 microns (Spitzer, WISE, IRAS and Herschel) for all of the HRS galaxies. Draine & Li (2007) models are fit to the data from which the stellar contribution has been carefully removed. We find that our photometric coverage is sufficient to constrain all of the models parameters and that a strong constraint on the 20-60 microns range is mandatory to estimate the relative contribution of the photo-dissociation regions to the IR SED. The SED models tend to systematically under-estimate the observed 500 microns flux densities, especially for low mass systems. We provide the output parameters for all of the galaxies: the minimum intensity of the interstellar radiation field (ISRF), the fraction of PAH, the relative contribution of PDR and evolved stellar population to the dust heating, the $M_{dust}$ and the $L_{IR}$. For a subsample of gas-rich galaxies, we analyze the relations between these parameters and the integrated properties of galaxies, such as $M_*$, SFR, metallicity, H$α$ and H-band surface brightness, and the FUV attenuation. A good correlation between the fraction of PAH and the metallicity is found implying a weakening of the PAH emission in galaxies with low metallicities. The intensity of the IRSF and the H-band and H$α$ surface brightnesses are correlated, suggesting that the diffuse dust component is heated by both the young stars in star forming regions and the diffuse evolved population. We use these results to provide a new set of IR templates calibrated with Herschel observations on nearby galaxies and a mean SED template to provide the z=0 reference for cosmological studies. For the same purpose, we put our sample on the SFR-$M_*$ diagram. The templates are compared to the most popular IR SED libraries, enlightening a large discrepancy between all of them in the 20-100 microns range.

astro-ph.GA↗

arXiv2013

HerMES: Point source catalogues from Herschel-SPIRE observations II

The Herschel Multi-tiered Extragalactic Survey (HerMES) is the largest Guaranteed Time Key Programme on the Herschel Space Observatory. With a wedding cake survey strategy, it consists of nested fields with varying depth and area totalling ~380 deg^2. In this paper, we present deep point source catalogues extracted from Herschel-SPIRE observations of all HerMES fields, except for the later addition of the 270 deg^2 HeLMS field. These catalogues constitute the second Data Release (DR2) made in October 2013. A subset of these catalogues, which consists of bright sources extracted from Herschel-SPIRE observations completed by May 1, 2010 (covering ~ 74 deg^2) were released earlier in the first extensive Data Release (DR1) in March 2012. Two different methods are used to generate the point source catalogues, the SUSSEXtractor (SXT) point source extractor used in two earlier data releases (EDR and EDR2) and a new source detection and photometry method. The latter combines an iterative source detection algorithm, StarFinder (SF), and a De-blended SPIRE Photometry (DESPHOT) algorithm. We use end-to-end Herschel-SPIRE simulations with realistic number counts and clustering properties to characterise basic properties of the point source catalogues, such as the completeness, reliability, photometric and positional accuracy. Over 500, 000 catalogue entries in HerMES fields (except HeLMS) are released to the public through the HeDAM website (http://hedam.oamp.fr/herMES).

astro-ph.GA↗

arXiv2013

Ultraviolet to infrared emission of z>1 galaxies: Can we derive reliable star formation rates and stellar masses?

We seek to derive star formation rates (SFR) and stellar masses (M_star) in distant galaxies and to quantify the main uncertainties affecting their measurement. We explore the impact of the assumptions made in their derivation with standard calibrations or through a fitting process, as well as the impact of the available data, focusing on the role of IR emission originating from dust. We build a sample of galaxies with z>1, all observed from the UV to the IR (rest frame). The data are fitted with the code CIGALE, which is also used to build and analyse a catalogue of mock galaxies. Models with different SFHs are introduced. We define different set of data, with or without a good sampling of the UV range, NIR, and thermal IR data. The impact of these different cases on the determination of M_star and SFR are analysed. Exponentially decreasing models with a redshift formation of the stellar population z ~8 cannot fit the data correctly. The other models fit the data correctly at the price of unrealistically young ages when the age of the single stellar population is taken to be a free parameter. The best fits are obtained with two stellar populations. As long as one measurement of the dust emission continuum is available, SFR are robustly estimated whatever the chosen model is, including standard recipes. M_star measurement is more subject to uncertainty, depending on the chosen model and the presence of NIR data, with an impact on the SFR-M_star scatter plot. Conversely, when thermal IR data from dust emission are missing, the uncertainty on SFR measurements largely exceeds that of stellar mass. Among all physical properties investigated here, the stellar ages are found to be the most difficult to constrain and this uncertainty acts as a second parameter in SFR measurements and as the most important parameter for M_star measurements.

astro-ph.CO↗

arXiv2013

HerMES: Candidate High-Redshift Galaxies Discovered with Herschel/SPIRE

We present a method for selecting $z>4$ dusty, star forming galaxies (DSFGs) using Herschel/SPIRE 250/350/500 $μm$ flux densities to search for red sources. We apply this method to 21 deg$^2$ of data from the HerMES survey to produce a catalog of 38 high-$z$ candidates. Follow-up of the first 5 of these sources confirms that this method is efficient at selecting high-$z$ DSFGs, with 4/5 at $z=4.3$ to $6.3$ (and the remaining source at $z=3.4$), and that they are some of the most luminous dusty sources known. Comparison with previous DSFG samples, mostly selected at longer wavelengths (e.g., 850 $μm$) and in single-band surveys, shows that our method is much more efficient at selecting high-$z$ DSFGs, in the sense that a much larger fraction are at $z>3$. Correcting for the selection completeness and purity, we find that the number of bright ($S_{500\,μm} \ge 30$ mJy), red Herschel sources is $3.3 \pm 0.8$ deg$^{-2}$. This is much higher than the number predicted by current models, suggesting that the DSFG population extends to higher redshifts than previously believed. If the shape of the luminosity function for high-$z$ DSFGs is similar to that at $z\sim2$, rest-frame UV based studies may be missing a significant component of the star formation density at $z=4$ to $6$, even after correction for extinction.

astro-ph.CO↗