OzDES multi-object fibre spectroscopy for the Dark Energy Survey: Results and second data release

Over the next ten years, surveys using the next generation of multi-object spectroscopic facilities fed by rapidly configurable fibre positioning systems will obtain the spectra of tens of millions of sources. Facilities such as the Subaru Prime Focus Spectrograph (PFS, Tamura et al., 2018) on the 8.2-m Subaru telescope, the Dark Energy Spectroscopic Instrument (DESI, DESI Collaboration, 2016; Vargas-Magana et al., 2019) on the 4-m Mayall telescope, and the 4-metre Multi-Object Spectroscopic Telescope (4MOST, de Jong et al., 2019) on the 4-m VISTA telescope are either being built or on the cusp of entering operation. Other facilities, such as Mauna Kea Spectroscopic Explorer (MSE, The MSE Science Team et al., 2019) and MANIFEST (Lawrence et al., 2018) for the Giant Magellan Telescope, are being planned. Over the same time frame, the Legacy Survey of Space and Time (LSST; LSST Science Collaboration, 2017) at the Rubin Observatory will start imaging the entire southern sky multiple times in multiple pass-bands over 10 years. The advent of the LSST in combination with these spectroscopic facilities will allow contemporaneous imaging and spectroscopic studies of tens of thousands of targets.

As a herald to these future surveys, OzDES¹¹1Australian (aka “Oz”) Dark Energy Survey. has used the 2dF fibre positioner on the Anglo-Australian Telescope (AAT) to obtain spectra of thousands of sources in the 10 deep fields of the Dark Energy Survey (DES) over the 6 years than DES ran. OzDES has two main scientific goals: i) constraining the dark energy equation of state parameter using type Ia supernovae (SNe Ia), and ii) measuring the mass of super-massive black holes over a broad range of redshifts using AGN reverberation mapping. Additionally, OzDES has obtained spectra for a number of ancillary projects, such as improving and quantifying the accuracy of redshifts obtained from broad-band photometry. In total, about 40% of the fibre hours not used for active supernovae, host galaxies and AGN were used for these projects.

This paper presents results from the full six years of OzDES, and describes the second OzDES data release²²2Available from Data Central: https://datacentral.org.au. Future releases will be announced on Data Central. After giving a brief summary of the principal scientific aims of OzDES in Sec. 2, we provide an overview of OzDES operations for the last three years of the survey. Throughout the paper we will use Y4, Y5, and Y6 to denote the fourth, fifth and sixth years of OzDES. The observing strategy for the first three years of OzDES (Y1, Y2 and Y3) and the precursor survey are discussed in two papers: Yuan et al. (2015) gives an overview up the end of Y1 (2013), including the OzDES precursor survey, and Childress et al. (2017) gives an overview up until the end of Y3 (2015). We will refer to these papers as Yu15 and Ch17 henceforth. The reliability and completeness of OzDES redshifts are analysed in Sec. 4, and the data release is described in Sec. 5. In the sections that follow, we examine various aspects of the survey, such as the frequency at which we were able to target AGN, and how the signal-to-noise ratio behaves with time for objects that have the longest integration times: the host galaxies of transients. Before giving the summary of the paper in the final section, we compare OzDES to TiDES, a OzDES-like survey using the 4MOST facility (Swann et al., 2019). All magnitudes listed in this paper are measured using 2″ diameter apertures and are on the AB magnitude system.

The Dark Energy Survey (DES) was a six-year programme, ending in 2019, using the DECam imager (Flaugher et al., 2015) on the CTIO Blanco 4m telescope in Chile. DES consisted of two surveys: a wide-field survey covering 5,000 square degrees, and a deep survey consisting of 10 DECam fields located in several well-studied extra-galactic fields. These 10 fields, the coordinates of which can be found in Table 2 of D’Andrea et al. (2018), were targeted repeatedly in griz over the course of the first five years of DES with a cadence of approximately 6 days ,(Diehl et al., 2018), and consist of 8 relatively shallow fields and 2 extra deep fields (known as X3 and C3). The regular cadence facilitated the discovery and follow-up of transient sources, such as supernovae (SNe), and the monitoring of variable objects, such as AGN.

DES aims to place constraints on the dark energy equation of state parameter and deviations from General Relativity using four astronomical probes: weak gravitational lensing, galaxy clusters, the large scale distribution of galaxies, and type Ia supernovae. Results from small subset of the DES data set have been published in a number of papers, the most recent of which are constraints that combine supernovae, gravitational lensing, and galaxy clustering (Dark Energy Survey, 2019a).

DES data have been used in many studies covering a diverse range of fields (Dark Energy Survey, 2016). Notable examples include studies of bodies in the outer solar system (Khain et al., 2018, for example), AGN accretion disk sizes (Mudd et al., 2018; Yu et al., 2020), dwarf galaxies and stellar streams in the halo of the Milky Way (Marshall et al., 2019; Li et al., 2019, for example) and rare transients such as merging neutron stars (Palmese et al., 2017), super-luminous supernovae (Smith et al., 2018; Angus et al., 2019), and rapidly evolving transients (Pursiainen et al., 2018).

Since 2012, OzDES has been targeting sources in the 10 DES deep fields with the 2dF fibre positioner on the AAT. As shown in Fig. 1 of Yuan et al. (2015), the fields of view of 2dF and DECam are very similar. In addition to the 10 DES deep fields, OzDES targeted the DES MaxVis field during one of the OzDES observing runs. Centred at RA=$97.5\degree$ and Dec=$-58.75\degree$, it is observable during the entire DES observing season. It is one of the DES standard star fields that was observed during twilight. Yu et al. (2020) used quasars in this field to study continuum reverberation mapping.

The scientific aims of OzDES focus on two areas: i) using redshifts and distances of type Ia supernovae to measure the expansion history of the Universe and thereby constrain the dark energy equation of state parameter; and ii) measuring the time lags between the continuum and the broad lines in a sample of AGN to measure their central black hole masses and to investigate the possibility of using AGN as a luminosity distance indicator that would explore a redshift range that is currently beyond the reach of supernovae.

Over the six years that OzDES ran, OzDES obtained redshifts for 7,000 candidate supernova hosts (D’Andrea et al., 2018) and spectroscopically confirmed several hundred supernovae (SNe). These redshifts will be used with distances to the supernovae inferred from broad-band lightcurves obtained by DES to measure the expansion history of the Universe and to constrain the dark energy equation of state parameter. Using an initial sample of 207 spectroscopically confirmed type Ia supernovae from the first three years of data in combination with 122 low redshift supernova from the literature and constraints from the Cosmic Microwave Background (CMB), the Dark Energy Survey (2019b) obtained a 7% constraint on a constant dark energy equation-of-state parameter. Analysis of the full DES sample is ongoing, and the final cosmological results will include approximately an order of magnitude more supernovae, the majority of which will be classified photometrically.

At the same time, OzDES has monitored a sample of 771 AGN up to $z\sim 4$ with the goal of measuring black hole masses using the time lag between variations in the continuum, measured from the broad-band photometry obtained by DES, and the broad lines, measured from the spectra obtained by OzDES. The first time lags, based on the first four years of data, were published in Hoormann et al. (2019).

In addition to these two main projects, OzDES is facilitating the studies of a number of areas, both within DES and outside of DES, by obtaining redshifts of targets of interest in these 10 fields using the “spare" fibres that are not allocated to an AGN, an active transient, or a host galaxy. The most notable examples are: radio galaxies from the ATLAS survey (Franzen et al., 2015); brightest cluster galaxies (Webb et al., 2015); redMaGiC galaxies (Rozo et al., 2016; Prat et al., 2018); and luminous red galaxies (LRGs), which are used in combination with redshifts from other sources to train photometric redshift algorithms (Sánchez et al., 2014; Bonnett et al., 2016; Gschwend et al., 2018).

Starting in 2013, OzDES ran for six observing seasons, with an initial allocation of 12 nights in Y1 (AAT semester 2013B), and allocations of 16, 20, 20, 20, and 12 nights in subsequent years. In 2012, when aspects of the DES survey strategy were being tested during a period of science verification (SV), 5 nights were allocated to an OzDES precursor survey. As for DES, these 5 nights were used to develop and test the observing strategy that would be used for the next six years.

Observations typically started in late August or early September and terminated four to five months later in December or January the following year. The observing log for Y4, Y5, and Y6 can be found in Appendix A and follows the format presented for earlier years in Yu15 and Ch17.

In total, there were 41 numbered runs, with the last night of run 41 occurring on 09 January 2019. Some of the nights during these runs were shared with other programmes, such as 2dFLens (Johnson et al., 2017) during Y2 and Y3, XXL (Pierre et al., 2016) in Y4, DEVILS (Davies et al., 2018) in Y5, and the HSC Transient Survey (Yasuda et al., 2019) in Y6. During Y2 and Y3, OzDES and 2dFLens shared a common run numbering scheme, and there are a number of runs during those years in which only 2dFLens fields were observed.

The end of Y5 marked the end of the DES programme to discover transients in the 10 DES deep fields. DES continued to target fields in the wide survey during Y6. OzDES also continued into a sixth year, but was no longer targeting transients. Instead, OzDES continued with its programme to monitor AGN (King et al., 2015; Hoormann et al., 2019) and to obtain redshifts of galaxies that hosted transients in earlier years. DES continued to target the 10 DES supernova fields during Y6, but with a reduced frequency of about once a month.

These data are used to spectrophotometrically calibrate the spectra of the AGN in the deep fields that were being monitored by OzDES.

As noted in Ch17, redshifts have been measured using MARZ (Hinton et al., 2016) since the start of Y3. Earlier than this RUNZ, which was developed by Will Sutherland, had been used. Measuring redshifts is a task that is performed by two redshifters, whose results are collated, scrutinised and then merged by the chief redshift whip. Merging is also done using MARZ. The merged results then enter the OzDES database, from which all other products are derived.

OzDES uses a redshift flag to quantify the confidence that the reported redshift is correct. As reported in Yu15 and Ch17, the flag can take on five main values.

• •

  $Q=4$, redshift based on multiple strong spectroscopic features matched, $>99$% confidence.

• •

  $Q=3$, redshift based typically on a single strong spectroscopic feature or multiple weak features, $>95$ % confidence.

• •

  $Q=2$, potential redshift associated with typically a single weak feature, low confidence, not to be used for science.

• •

  $Q=1$, no matching features, thus no constraints on redshift.

• •

  $Q=6$, securely classified star.

As the survey progressed, we monitored the growing redshift catalogue for sources that had conflicting redshifts. These are sources in which the redshift changes between observing runs but the redshift quality flag, $Q$, remains unchanged and $Q>2$. An example is a source that is assigned a $Q=3$ redshift during one run, but with the addition of more data in a later run, is assigned a different redshift ($\Delta z>0.05$) with the same quality flag.

On two occasions, once at the end of Y4 once at the end of Y6 , we collated all cases where there was conflict, defined on these occasions having a redshift difference larger than 0.05, and reprocessed and re-redshifted these sources in the same way we redshifted sources during each observing run. The revised redshifts were then added to our redshift catalog. At the end of Y4, there were a total of 191 sources that had conflicting redshifts. At the end Y6, there were a total of 49, perhaps illustrating a degree of improvement by the human redshifters.

In preparation for the second OzDES data release, we revisited cases with conflicting redshifts, this time decreasing the threshold from 0.05 to 0.01 in redshift. We also included sources that were classified with $Q=3$ quality redshift at some stage during the survey but ended up with a lower redshift quality flag once all data were combined. Combined, about 820 sources were re-examined.

This paper marks the second OzDES release, which will be referred to as OzDES-DR2. OzDES-DR2 differs from OzDES-DR1⁵⁵5The redshift catalogue for OzDES-DR1 is available from https://datacentral.org.au., the first OzDES data release, in two key aspects.

• •

  The redshift catalogue includes all objects observed during the 6 years of OzDES and the OzDES pre-survey. This includes SN hosts, SNe, and targets for which a redshift could not be obtained.

• •

  The fully processed spectra are provided, including all individual spectra, as well as the stacked spectrum of each object.

In OzDES-DR2, we release 375,000 spectra of almost 39,000 objects, of which three quarters have a redshift with a quality flag of 3 or greater. We release all spectra, including the spectra of objects that do not have a redshift. These data can used to help secure redshifts for these objects if they are observed in the future with the AAT or other spectroscopic facilities. A full description of the data release is provided in Appendix C.

Second to SN hosts, AGN were the next most frequently observed target type. As shown in Fig. 2, about a quarter of the fibre hours available to OzDES were allocated to the AGN reverberation mapping (RM) programme. The first year of the OzDES RM programme was used to define the sample that was followed in the subsequent five years. Details of the selection are provided in King et al. (2015). In brief, AGN were selected on the basis of the quality of the spectra that were taken in Y1, on the presence of emission lines, while avoiding AGN that occasionally fell in chip gaps, maximising the number of AGN with two reverberating spectral lines in our wavelength range, and ensuring that the AGN selected sampled the redshift interval $0<z<4$. Most AGN that pass these cuts have have $r_{\mathrm{A}B}<21.2$. From an initial sample of 3331 AGN, 771 AGN were followed during the next five years.

Apart from active transients, AGN in the OzDES RM sample have the highest priority, so the percentage of AGN that were allocated a fibre in any given observation was near to 100%. However, not all fields could be observed during every OzDES run, because of a combination of the finite duration of the observing runs, time lost to poor observing conditions, and the differing priority of the ten DES deep fields. Considering the latter and the bar chart shown in Fig. 3, we expect that the AGN in the two E-region fields and the X3 and C3 fields will have the greatest number of observations, both in terms of the raw number of times they were observed, and the number of runs in which they were observed.

In Fig. 15, we show the distribution of the number of runs in which AGN were observed. Two plots are shown. In the lower plot, all observations are included, regardless of the observing conditions or the quality of the data. In the upper plot, only data that had a relative zero point greater than $-2$ (see Fig. 4) and a quality control (QC) flag of ‘ok’ are included (see Appendix B for a discussion on the QC flag). The vertical solid line shows the number of runs on which OzDES observations were scheduled during the six years that OzDES ran. The two runs from the OzDES pre-survey are excluded, since the AGN RM pre-sample had not been defined at that time. This solid line represents an upper limit to the number of epochs. The median values for the four regions are shown in Table 4. The median values for the X and C regions include X3 and C3, which were the highest priority fields, so we list the median values for the C3 and X3 separately as well. For the E region fields and X3 and C3, the median is close to the 25 epochs that were used in the simulations of the OzDES RM programme in King et al. (2015).

With this number of epochs, simulations in King et al. (2015) predict that one should recover lags for 35-45% of the AGN depending on the accuracy with which one can calibrate the line flux. The simulations used a 10% accuracy for the line measurements. Hoormann et al. (2019) demonstrate that we are able to achieve this level of accuracy, so we anticipate that we will recover this fraction of lags in the two E fields and the C3 and X3 fields, and a lower fraction in the other six fields. The first two lags, using the CIV line and based on a subset of the OzDES data, were published in Hoormann et al. (2019).

As illustrated in Fig. 2, slightly more than one-third of all the fibre hours allocated to targets during the six years of OzDES were allocated to SN hosts. Unlike other targets, SN hosts were observed until a secure redshift (see Sec. 4) was obtained. SN hosts also extend to fainter magnitudes than most targets. Both of these factors meant that the integration time for some hosts are very long, as long as 2 days in a number of cases.

We use SN hosts to examine how well the measured noise changes with time, and explore the implications of deselecting SN hosts once they reach a $Q=3$ quality redshift and before they reach the more secure $Q=4$ quality redshift.

By the early 2020s, LSST will have started surveying the Southern sky and several new multi-object spectroscopic facilities fed by rapidly configurable fibre positioners covering wide fields of view will be operational. Of the planned surveys using these new facilities, the Time-Domain Extragalactic Survey (TiDES, Swann et al., 2019), using 4MOST (de Jong et al., 2019) in combination with LSST, is the one that most closely matches the OzDES concept. A comparison of the OzDES and TiDES surveys is shown in Table 5. The survey characteristics of TiDES listed in Table 5 are subject to change as the survey strategy is still evolving.

On practically every metric, TiDES will be superior to OzDES. It will observe at least 10 times more live transients, mostly in the wide survey, it will measure up to 10 times more host galaxy redshifts, and it will monitor AGN at twice the frequency that OzDES did and with shorter seasonal gaps. TiDES will also observe transients over the majority of the southern sky, something that OzDES did not do.

Although the science goals and strategy of TiDES are heavily based upon that of OzDES, practically there are some differences. The first difference between OzDES and TiDES is that TiDES will not drive the pointings of 4MOST. Wherever 4MOST is scheduled to point there will be previously discovered transient targets, a majority of these targets anticipated to be discovered by LSST. If these transients are still bright enough to obtain a spectrum for spectral typing, TiDES will attempt do so. Otherwise TiDES will target the host galaxy of the transient to obtain a spectroscopic redshift for type Ia SN cosmology.

The second difference between OzDES and TiDES is that TiDES will have two components, a wide field component and a deep field component. The wide field follow-up will cover the majority of the southern sky, where low redshift transients and host galaxies will be targeted with a maximum exposure time of two hours. The second component to TiDES will be spectroscopic follow-up in the deep fields of 4MOST, where higher redshift live transients will be observed, the AGN reverberation mapping will take place, and where one can repeatedly target SN host galaxies to build up redshifts on the faintest of targets.

The 4MOST Deep fields currently correspond to four pointings, each covering 4.2  sq. degrees and contained within the LSST deep drilling fields (Guiglion et al., 2019). LSST data will provide the high precision photometry and high cadence light curves required for sub-percent precision cosmology. As three of the DES deep fields (ECDFS, XMM-LSS, ELAIS S1) are also covered by the 4MOST and LSST DDFs, this offers the exciting possibility of continued monitoring a subsample of the AGN observed by OzDES over a 15 year baseline, if the time gap between OzDES and TiDES can be covered. The higher cadence and longer observing season will also mean that TiDES will be more sensitive to AGN with shorter lags and less sensitive to seasonal gaps. Therefore TiDES will obtain a higher fraction of lags compared to OzDES.

With more than 4 times the number of fibres, and a field of view that is only slightly larger that 2dF, TiDES on 4MOST will be able to observe SN hosts in the deep fields longer without running out of fibres. Given the higher throughput, broader spectral coverage, and higher spectral resolution of 4MOST compared to AAOmega, and the better observing conditions on Cerro Paranal compared to Siding Spring, TiDES will be able to obtain host galaxy redshifts more quickly and to a fainter magnitude limit than OzDES.

Over six consecutive observing seasons, starting in 2013, OzDES obtained over 375,000 spectra of almost 39,000 sources in the ten DES deep fields. These spectra were obtained with the 2dF fibre positioner and AAOmega spectrograph on the AAT. Over that time, each DES SN field was observed between 18 and 25 times.

In these fields, OzDES has obtained redshifts for almost 7,000 galaxies that hosted a transient, and classified almost 270 transients. The strategy of targeting each field repeatedly has enabled redshifts to be obtained for sources as faint as $r_{\mathrm{AB}}\sim 24$. We show that one is not limited by a noise floor, even after two days of exposure, and that one can reach high spectroscopic completeness at these faint limits.

The strategy of targeting each field repeatedly has allowed OzDES to monitor 771 AGN. These data will be used to measure the lag between the continuum from the accretion disk and the broad lines from the broad line emission region. We anticipate that we will recover lags and derive black hole masses for approximately 30% of the AGN we monitored.

The OzDES observing strategy allowed a number of science programmes to run in parallel. Just over 60% of the 192,000 fibre hours were allocated to monitoring AGN and obtaining redshifts for galaxies that hosted a transient. The remaining 40% was used to obtain redshifts for wide variety of targets, such as radio galaxies, cluster galaxies, and galaxies to train photometric redshift algorithms.

OzDES can be used as a template for future surveys that use the next generation of wide-field multi-object fibre-fed spectroscopic facilities to target and monitor sources in the LSST deep drilling fields. These new facilities will have up to an order of magnitude more fibres than 2dF, and they will enable one to follow a larger number of fainter sources more frequently and for longer.

Parts of this research were supported by the Australian Research Council under grants DP160100930, FL180100168, and FT140101270.

Based in part on data acquired at the Anglo-Australian Telescope, under programme A/2013B/012. We acknowledge the traditional owners of the land on which the AAT stands, the Gamilaraay people, and pay our respects to elders past, present, and emerging.

We acknowledge the support of James Tocknell and Simon O’Toole from Data Central in helping us prepare OzDES-DR2.

L.G. was funded by the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No. 839090.

F.H.P. was supported by an Australian Government Research Training Programme Scholarship.

Funding for the DES Projects has been provided by the U.S. Department of Energy, the U.S. National Science Foundation, the Ministry of Science and Education of Spain, the Science and Technology Facilities Council of the United Kingdom, the Higher Education Funding Council for England, the National Center for Supercomputing Applications at the University of Illinois at Urbana-Champaign, the Kavli Institute of Cosmological Physics at the University of Chicago, the Center for Cosmology and Astro-Particle Physics at the Ohio State University, the Mitchell Institute for Fundamental Physics and Astronomy at Texas A&M University, Financiadora de Estudos e Projetos, Fundação Carlos Chagas Filho de Amparo à Pesquisa do Estado do Rio de Janeiro, Conselho Nacional de Desenvolvimento Científico e Tecnológico and the Ministério da Ciência, Tecnologia e Inovação, the Deutsche Forschungsgemeinschaft and the Collaborating Institutions in the Dark Energy Survey.

The Collaborating Institutions are Argonne National Laboratory, the University of California at Santa Cruz, the University of Cambridge, Centro de Investigaciones Energéticas, Medioambientales y Tecnológicas-Madrid, the University of Chicago, University College London, the DES-Brazil Consortium, the University of Edinburgh, the Eidgenössische Technische Hochschule (ETH) Zürich, Fermi National Accelerator Laboratory, the University of Illinois at Urbana-Champaign, the Institut de Ciències de l’Espai (IEEC/CSIC), the Institut de Física d’Altes Energies, Lawrence Berkeley National Laboratory, the Ludwig-Maximilians Universität München and the associated Excellence Cluster Universe, the University of Michigan, the National Optical Astronomy Observatory, the University of Nottingham, The Ohio State University, the University of Pennsylvania, the University of Portsmouth, SLAC National Accelerator Laboratory, Stanford University, the University of Sussex, Texas A&M University, and the OzDES Membership Consortium.

Based in part on observations at Cerro Tololo Inter-American Observatory, National Optical Astronomy Observatory, which is operated by the Association of Universities for Research in Astronomy (AURA) under a cooperative agreement with the National Science Foundation.

The DES data management system is supported by the National Science Foundation under Grant Numbers AST-1138766 and AST-1536171. The DES participants from Spanish institutions are partially supported by MINECO under grants AYA2015-71825, ESP2015-66861, FPA2015-68048, SEV-2016-0588, SEV-2016-0597, and MDM-2015-0509, some of which include ERDF funds from the European Union. IFAE is partially funded by the CERCA programme of the Generalitat de Catalunya. Research leading to these results has received funding from the European Research Council under the European Union’s Seventh Framework Program (FP7/2007-2013) including ERC grant agreements 240672, 291329, and 306478. We acknowledge support from the Brazilian Instituto Nacional de Ciência e Tecnologia (INCT) e-Universe (CNPq grant 465376/2014-2).

This manuscript has been authored by Fermi Research Alliance, LLC under Contract No. DE-AC02-07CH11359 with the U.S. Department of Energy, Office of Science, Office of High Energy Physics. The United States Government retains and the publisher, by accepting the article for publication, acknowledges that the United States Government retains a non-exclusive, paid-up, irrevocable, world-wide license to publish or reproduce the published form of this manuscript, or allow others to do so, for United States Government purposes.

We are grateful for the extraordinary contributions of our CTIO colleagues and the DECam Construction, Commissioning and Science Verification teams in achieving the excellent instrument and telescope conditions that have made this work possible. The success of this project also relies critically on the expertise and dedication of the DES Data Management group.