1. Introduction
Type Ia supernovae (SNe Ia) were used to discover the accelerating expansion of the universe (Riess et al.1998; Perlmutter et al.1999) and remain one of the key probes for understanding the nature of the mysterious “dark energy.” Over the last two decades, there have been considerable improvements in the calibration and size of samples at low redshift (Jha et al.2006; Hicken et al.2009b,2012; Contreras et al.2010), intermediate redshift (Holtzman et al.2008), and high redshift (Astier et al.2006; Wood-Vasey et al.2007; Conley et al.2011; Betoule et al.2014; Rest et al.2014). When combined with cosmic microwave background (CMB) data, these samples have been used to demonstrate that the dark energy equation of state,w, is consistent with a cosmological constant (w = −1) with a precision ofσw = 0.04. The recent Pantheon analysis combines >1000 SNe Ia from several surveys, resulting inw = −1.026 ± 0.041 (Scolnic et al.2018).
The Dark Energy Survey Supernova program (DES-SN) is striving to find even greater numbers of SNe while reducing systematic uncertainties on the resulting cosmological parameters. A top priority of this effort has been to accurately model each component of the DES-SN search and analysis, and to accurately simulate bias corrections for the SN Ia distance measurements. DES has also made improvements in instrumentation and calibration, including: (i) detectors with higherz-band efficiency to improve measurements of rest frame supernova (SN) colors at high-redshift, and (ii) extension of the photometric calibration precision over a wide color range by correcting each charge-coupled device (CCD) and exposure for atmospheric variations and the spectral energy distribution (SED) of the source (see Section3). These improvements enable DES-SN to make a state-of-the-art measurement of dark energy properties.
This Letter reports “DES-SN3YR” cosmological constraints from the spectroscopically confirmed SNe Ia in the first three years of DES-SN in combination with a low-redshift SN Ia sample from the literature. The results presented here are the culmination of a series of companion papers that contain details of the SN search and discovery (Goldstein et al.2015; Kessler et al.2015; Morganson et al.2018), spectroscopic follow-up (D’Andrea et al.2018), photometry (Brout et al.2018a), calibration (Burke et al.2018; Lasker et al.2018), simulations (Kessler et al.2018), and technique to account for selection bias (Kessler & Scolnic2017). The cosmological analysis method and validation are detailed in Brout et al. (2018b,B18), which presents the full statistical and systematic uncertainty budget for these new results. Hinton et al. (2018) tested a new Bayesian Hierarchical Model for supernova cosmology. In this Letter, we summarize these contributions and present our measurements of the equation of state (w) and matter density (
). Data products used in this analysis are publicly available online.77 In addition, Macaulay et al. (2018) measured the Hubble constant (H0) by applying these DES-SN3YR results to the inverse-distance-ladder method anchored to the standard ruler measured by baryon acoustic oscillations (BAO; Alam et al.2017; Carter et al.2018), and related to the sound horizon measured with CMB data (Planck Collaboration et al.2016).
In Section2 we discuss the data sets used in our analysis. In Section3, we summarize the analysis pipeline. In Section4, we present the cosmology results. In Section5, we present our discussion and conclusions.
2. Data Samples
The DES-SN sample for this analysis was collected over three 5-month-long seasons, from 2013 August to 2016 February, using the Dark Energy Camera (DECam; Flaugher et al.2015) at the Cerro Tololo Inter-American Observatory. Ten 2.7 deg2 fields were observed approximately once per week in thegriz filter bands (Abbott et al.2018). The average depth per visit was 23.5 mag in the eight “shallow” fields, and 24.5 mag in the two “deep” fields. Within 24 hr of each observation, search images were processed (Morganson et al.2018), new transients were discovered using a difference-imaging pipeline (Kessler et al.2015), and most of the subtraction artifacts were rejected with a machine-learning algorithm applied to image stamps (Goldstein et al.2015).
A subset of light curves was selected for spectroscopic follow-up observations (D’Andrea et al.2018), resulting in 251 spectroscopically confirmed SNe Ia with redshifts 0.02 <z < 0.85, and
SNe Ia that satisfy analysis requirements (B18) such as signal-to-noise and light curve sampling; this sample is called the DES-SN subset. The spectroscopic program required a collaborative effort coordinated across several observatories. At low to intermediate redshifts, the primary follow-up instrument is the 4 m Anglo-Australian Telescope, which confirmed and measured redshifts for 31% of our SN Ia sample (OzDES collaboration; Yuan et al.2015; Hinton et al.2016; Childress et al.2017). A variety of spectroscopic programs (D’Andrea et al.2018) were carried out using the European Southern Observatory Very Large Telescope, Gemini, Gran Telescopio Canarias, Keck,Magellan, MMT, and South African Large Telescope.
We supplement the DES-SN sample with a low-redshift (z < 0.1) sample, which we call the low-z subset, comprising 122 SNe from the Harvard-Smithsonian Center for Astrophysics surveys (CfA3, CfA4; Hicken et al.2009a,2012) and the Carnegie Supernova Project (Contreras et al.2010; Stritzinger et al.2011). We only use samples with measured telescope+filter transmissions, and thus CfA1 and CfA2 are not included.
3. Analysis
Supernova cosmology relies on measuring the luminosity distance (dL) versus redshift for many SNe Ia and comparing this relation to the prediction of cosmological models. The distance modulus (μ) is defined as
For a flat universe with cold dark matter density
, dark energy density
, and speed of lightc, the luminosity distance to a source at redshiftz is given by
with
Observationally, the distance modulus of a supernova is given by
For each SN Ia, the set ofgriz light curves are fit (Section3.4) to determine an amplitude (x0, with
), light curve width (x1), and color (
).γ describes the dependence on host-galaxy stellar mass (
, Section3.5), where
if
, and
if
. A correction for selection biases (
) is determined from simulations (Section3.6).
All SNe Ia are assumed to be characterized byα,β,γ, andM0. The first three parameters describe how the SN Ia luminosity is correlated with the light curve width (αx1), color (
), and host-galaxy stellar mass (
).M0 accounts for both the absolute magnitude of SNe Ia and the Hubble constant. In the rest of this section we describe the main components of the analysis pipeline that are needed to determine the distances (Equation (4)) and cosmological parameters.
3.1. Calibration
The DES sample is calibrated to the AB magnitude system (Oke & Gunn1983) using measurements of theHubble Space Telescope (HST) CalSpec standard C26202 (Bohlin et al.2014). DES internally calibrated roughly 50 standard stars per CCD using a “Forward Global Calibration Method” (Burke et al.2018; Lasker et al.2018). Improvements in calibration at the 0.01 mag (1%) level are made using SED-dependent “chromatic corrections” to both the standard stars and to the DES-SN light curve photometry. The low-z sample is calibrated to the AB system by cross-calibrating to the Pan-STARRS1 (PS1) photometric catalogs (Scolnic et al.2015). We also cross-calibrate DES to PS1 and find good agreement (see Section 3.1.2, Figure 3 ofB18).
3.2. Photometry
To measure the SN Ia flux for each observation, we employ a scene modeling photometry (SMP) approach (Brout et al.2018a) based on previous efforts used in SDSS-II (Holtzman et al.2008) and SNLS (Astier et al.2013). SMP simultaneously forward models a variable SN flux on top of a temporally constant host galaxy. We test the precision by analyzing images that include artificial SNe Ia, and find that photometric biases are limited to <0.3%. Each CCD exposure is calibrated to the native photometric system of DECam, and zero points are determined from the standard star catalogs (Section3.1).
3.3. Spectroscopy: Typing and Redshifts
Spectral classification was performed using both the SuperNova IDentification (Blondin & Tonry2007) and Superfit (Howell et al.2005) software, as described in D’Andrea et al. (2018). All 207 events are spectroscopically classified as SNe Ia. Redshifts are obtained from host-galaxy spectra where available, because their sharp spectral lines give more accurate redshifts (σz ∼ 5 × 10−4; Yuan et al.2015) than the broad SN Ia spectroscopic features (σz ∼ 5 × 10−3). While 158 of the DES-SN events have host galaxy redshifts, the rest have redshifts from the SN Ia spectra. For the low-z sample, we use the published redshifts with a 250 km s−1 uncertainty from Scolnic et al. (2018). Peculiar-velocity corrections are computed from Carrick et al. (2015).
3.4. Light Curve Fitting
To measure the SN parameters (
), the light curves were fit withSNANA78 (Kessler et al.2009) using the SALT2 model (Guy et al.2010) and the training parameters from Betoule et al. (2014).
3.5. Host Galaxy Stellar Masses
For the
term in Equation (4), we first identify the host galaxy using catalogs from Science Verification DECam images (Bonnett et al.2016), and the directional light radius method (Sullivan et al.2006; Gupta et al.2016).
is derived from fitting galaxy model SEDs togriz broadband fluxes withZPEG (Le Borgne & Rocca-Volmerange2002). The SEDs are generated with Projet d’Etude des GAlaxies par Synthese Evolutive (PEGASE; Fioc & Rocca-Volmerange1997). In the DES-SN subset, 116 out of 207 hosts have
. The low-z host galaxy stellar masses are taken from Scolnic et al. (2018).
3.6. μ-bias Corrections
We use a simulation-based method (Kessler et al.2018) to correct for distance biases arising from survey and spectroscopic selection efficiencies, and also from the analysis and light curve fitting. For each SN Ia we calculate the bias correction in Equation (4),
, where
is the average in bins of measured redshift, color, and stretch. The distanceμ is determined by analyzing the simulated data in the same way as the real data (but with
), and
is the true distance modulus used to generate each simulated event. The correction increases with redshift, and for individual SNe Ia it can be as large as 0.4 mag (Section 9 of Kessler et al.2018).
The simulation accurately models DES-SN3YR selection effects. For each generated event it picks a random redshift, color, and stretch from known distributions (Perrett et al.2012; Scolnic & Kessler2016). Next, it computes true SN Ia magnitudes at all epochs using the SALT2 SED model, intrinsic scatter model (Section3.7), telescope+atmosphere transmission functions for each filter band, and cosmological effects such as dimming, redshifting, gravitational lensing, and galactic extinction. Using the survey cadence and observing conditions (point-spread function, sky noise, zero-point), instrumental noise is added. Finally, our simulation models the efficiencies ofDiffImg and spectroscopic confirmation. The quality of the simulation is illustrated by the good agreement between the predicted and observed distribution of many observables including redshift, stretch, and color (Figures 6 and 7 in Kessler et al.2018, and Figure 5 inB18).
3.7. Intrinsic Scatter Model
We simulate bias corrections with two different models of intrinsic scatter that span the range of possibilities in current data samples. First is the “G10” model, based on Guy et al. (2010), in which the scatter is primarily achromatic. Second is the “C11” model, based on Chotard et al. (2011), which has stronger scatter in color. For use in simulations, Kessler et al. (2013) converted each of these broadband scatter models into an SED-variation model.
3.8. Generating the Bias-corrected Hubble Diagram
We use the “BEAMS with Bias Corrections” (BBC) method (Kessler & Scolnic2017) to fit for {α,β,γ,M0} and to fit for a weighted-average bias-correctedμ in 18 redshift bins. In addition to propagating the uncertainty from each term in Equation (4), the BBC fit adds an empirically determinedμ-uncertainty (
) to each event so that the best-fit
. This redshift-binned Hubble diagram is used for cosmology fitting as described in Section3.9. Figure1 shows the binned Hubble diagram, and also the unbinned Hubble diagram using individual bias-corrected distances computed in the BBC fit.
Figure 1. Hubble diagram for the DES-SN3YR sample. Top: distance modulus (μ) from BBC fit (black bars, which are used for cosmology fits) and for each SN (red, orange circles). The dashed gray line shows our best-fit model, while the green and blue dotted lines show models with no dark energy and matter densities
and 1.0 respectively. Bottom: residuals to the best-fit model; 1σ error bars show 68% confidence.
Download figure:
Standard imageHigh-resolution image
3.9. Cosmology Fitting
Cosmological parameters are constrained using the log-likelihood
and minimizing the posterior with CosmoMC (Lewis & Bridle2002).
for redshift bin
,
is the BBC-fitted distance modulus in theith redshift bin, and
is given by Equation (1). The covariance matrix (
) is described in Section 3.8.2 ofB18, and incorporates systematic uncertainties from each analysis component in Section3.
and
are computed separately using the G10 and C11 scatter model in the bias-correction simulation. Each set of quantities is averaged over the G10 and C11 models, and these averages are used in Equation (5). The purpose of averaging is to mitigate the systematic uncertainty related to our understanding of intrinsic scatter (Section 4.2 ofB18).
Finally, we combine these SN Ia results with priors from CMB and BAO as described in Section4.
3.10. Blinding and Validation
The cosmological parameters were blinded until preliminary results were presented at the 231st meeting of the American Astronomical Society in 2018 January. The criteria for unblinding (Section 7 ofB18) included analyzing large simulated DES-SN3YR data sets, and requiring (i)w-bias below 0.01, and (ii) the rms ofw-values agrees with the fittedw-uncertainty, for simulations with and without systematic variations. Following this initial unblinding, several updates were performed (Section 3.8.4 ofB18), again blinded, and the final results presented here were unblinded during the DES internal review process. Compared to the initial unblinding,w increased by 0.024 and the total uncertainty increased by 3% (0.057 to 0.059).
4. Results
We present the first cosmological results using SNe Ia from DES. We begin with the BBC-fitted parameters (
) in Section4.1, then present our statistical and systematic uncertainty budget forw in Section4.2 and Table1. Finally, we present our cosmological parameters in Section4.3 and Table2. For our primary results we combine DES-SN3YR with the CMB likelihood from Planck Collaboration et al. (2016) using their temperature power spectrum and low-ℓ polarization results. We also present results without a CMB prior, and with both CMB and BAO priors. All reported uncertainties correspond to 68% confidence. To evaluate consistency between our primary result and BAO, we compute the evidence usingPolyChord (Handley et al.2015a,2015b), and compute the evidence ratio (R) defined in Equation (V.3) of DES Collaboration et al. (2018). Consistency is defined byR > 0.1.
Table 1. w-uncertainty Contributions forwCDM Modela
| Descriptionb | σw |  |
|---|---|---|
Total Stat ( ) | 0.042 | 1.00 |
| Total Systc | 0.042 | 1.00 |
| Total Stat+Syst | 0.059 | 1.40 |
| [Photometry and Calibration] | [0.021] | [0.50] |
| Low-z | 0.014 | 0.33 |
| DES | 0.010 | 0.24 |
| SALT2 Model | 0.009 | 0.21 |
| HST Calspec | 0.007 | 0.17 |
| [μ-Bias Correction: Survey] | [0.023] | [0.55] |
| †Low-z 3σ Cut | 0.016 | 0.38 |
| Low-z Volume Limited | 0.010 | 0.24 |
| Spectroscopic Efficiency | 0.007 | 0.17 |
| †Flux Err Modeling | 0.001 | 0.02 |
| [μ-Bias Correction: Astrophysical] | [0.026] | [0.62] |
| Intrinsic Scatter Model (G10 versus C11) | 0.014 | 0.33 |
| †Twoσint | 0.014 | 0.33 |
 ,x1 Parent Population | 0.014 | 0.33 |
† in sim. | 0.006 | 0.14 |
| MW Extinction | 0.005 | 0.12 |
| [Redshift] | [0.012] | [0.29] |
| Peculiar Velocity | 0.007 | 0.17 |
| †z + 0.00004 | 0.006 | 0.14 |
Notes.
aThe sample is DES-SN3YR (DES-SN + low-z sample) plus CMB prior. bItem in[bold] is a sub-group and its uncertainty. cThe quadrature sum of all systematic uncertainties does not equal
because of redshift-dependent correlations when using the full covariance matrix.† Uncertainty was not included in previous analyses.
Download table as: ASCIITypeset image
Table 2. Cosmological Resultsa
| Row | SN Sample + Prior (ΛCDM) |  |  | |
|---|---|---|---|---|
| 1 | DES-SN3YRb+flatness | 0.331 ± 0.038 | 0.669 ± 0.038 | |
| 2 | DES-SN3YR | 0.332 ± 0.122 | 0.671 ± 0.163 | |
| 3 | DES-SN3YR+CMBc | 0.335 ± 0.042 | 0.670 ± 0.032 | |
| 4 | DES-SN3YR+CMB+BAOd | 0.308 ± 0.007 | 0.690 ± 0.008 | |
| Row | SN Sample + Prior (FlatwCDM) |  | w | |
| 5 | DES-SN3YR+CMB R | 0.321 ± 0.018 | −0.978 ± 0.059 | |
| 6 | DES-SNe+CMB | 0.341 ± 0.027 | −0.911 ± 0.087 | |
| 7 | DES-SN3YR+CMB+BAO | 0.311 ± 0.009 | −0.977 ± 0.047 | |
| 8 | DES-SN+CMB+BAO | 0.315 ± 0.010 | −0.959 ± 0.054 | |
| 9 | CMB+BAO | 0.310 ± 0.013 | −0.988 ± 0.072 | |
| Row | SN Sample + Prior (Flat ) |  | w0 | wa |
| 10 | DES-SN3YR+CMB+BAOR | 0.316 ± 0.011 | −0.885 ± 0.114 | −0.387 ± 0.430 |
| 11 | CMB+BAO | 0.332 ± 0.022 | −0.714 ± 0.232 | −0.714 ± 0.692 |
Notes.
aSamples in bold font are primary results given in the abstract. bDES-SN3YR: DES-SN + low-z samples. cCMB: Planck TT + lowP likelihood (Planck Collaboration et al.2016). dBAO: SDSS DR12 (Alam et al.2017); SDSS MGS (Ross et al.2015); 6dFGS (Beutler et al.2011). eDES-SN alone (no low-z).Download table as: ASCIITypeset image
4.1. Results for Standardization Parameters
While the cosmology results are based on averaging distances using the G10 and C11 intrinsic scatter models, here we show best-fit BBC values fromB18 using the G10 intrinsic scatter model:α = 0.146 ± 0.009,β = 3.03 ± 0.11,γ = 0.025 ± 0.018, and
. Ourα,β, and
values are consistent with those found in previous analyses, whileγ is smaller compared to those in Kelly et al. (2010), Sullivan et al. (2010), Lampeitl et al. (2010), Betoule et al. (2014), and Scolnic et al. (2018). Results with the C11 model (Table 5 ofB18) show similar trends.
We also check the consistency among the DES-SN and low-z subsets. Whileα andβ are consistent, we find
for DES-SN, the lowest value of any rolling SN survey. This value differs by 3.3σ from
for the low-z subset, and the systematic uncertainty in adopting a single
value is discussed below in Section4.2 and also in Section 7.3 ofB18. Ourγ values differ by 1.5σ:γDES = 0.009 ± 0.018 (consistent with zero) andγlowz = 0.070 ± 0.038.
4.2. w-uncertainty Budget
Contributions to the systematic uncertainty budget are presented inB18 and shown here in Table1 for flatwCDM fits combined with the CMB likelihood. The statistical uncertainty onw (
) is determined without systematic contributions. Each systematic contribution is defined as
where
is the total (stat+syst) uncertainty from including a specific systematic, or a group of systematics. The uncertainty inw has nearly equal contributions from statistical and systematic uncertainties, the latter of which is broken into four groups in Table1.
The first three systematic groups have nearly equal contributions: (1) photometry and calibration (
), which includes uncertainties from the DES-SN and low-z subsets, data used to train the SALT2 light curve model, and theHST Calspec standard; (2)μ-bias corrections from the survey (
), which includes uncertainties from rejecting Hubble residual outliers in the low-z subset, magnitude versus volume limited selection for low-z, DES-SN spectroscopic selection efficiency, and determination of DES-SN flux uncertainties; and (3)μ-bias corrections from astrophysical effects (
), which includes uncertainties from intrinsic scatter modeling (G10 versus C11, and two
, parent populations of stretch and color, choice ofw and
in the simulation, and Galactic extinction. The fourth systematics group, redshift (
), includes a global shift in the redshift and peculiar velocity corrections.
Finally, the Table1 systematics marked with a dagger (†) have not been included in previous analyses, and the combined uncertainty is
. Most of this new uncertainty is related to the low-z subset, which is almost 40% of the DES-SN3YR sample. For previous analyses with a smaller fraction of low-z events (e.g., Pantheon, JLA) we do not recommend adding the full
w-uncertainty to their results.
4.3. Cosmology Results
4.3.1. ΛCDM
Using DES-SN3YR and assuming a flat ΛCDM model, we find
. Assuming a ΛCDM model with curvature (
) added as a free parameter in Equation (3) (e.g., see Section 3.1 of Davis & Parkinson2017) we find the constraints shown in Figure2 and Table2 (row 2). Solid contours show our result with both statistical and systematic uncertainties included, while dashed contours show the statistical-only uncertainties for comparison. Figure2 also shows that the CMB data provide strong flatness constraints, consistent with zero curvature; the impact of using this CMB prior is shown in row 3. The impact from adding a BAO prior is shown in row 4, where the evidence ratio
shows consistency between the SN+CMB and BAO posteriors.
Figure 2. Constraints on
for ΛCDM model (68% and 95% confidence intervals). SN contours are shown with statistical uncertainty only (white dashed), and with total uncertainty (green shaded). Constraints from CMB (brown) and DES-SN3YR+CMB combined (red) are also shown.
Download figure:
Standard imageHigh-resolution image
4.3.2. FlatwCDM
For our primary result, we use DES-SN3YR with the CMB prior and a flatwCDM model (
) and find
and
(Table2, row 5). Our constraint onw is consistent with the cosmological-constant model for dark energy. The 68% and 95% confidence intervals are given by the red contours in Figure3, which also shows the contributions from DES-SN3YR and CMB. We show two contours for DES-SN3YR, with and without systematic uncertainties, in order to demonstrate their impact. In Table2, row 6, we show the impact of the low-redshift SN sample by removing it; thew-uncertainty increases by 25% and the constraint lies approximately 1σ fromw = −1.
Figure 3. Constraints on
–w for the flatwCDM model (68% and 95% confidence intervals). SN contours are shown with only statistical uncertainty (white dashed) and with total uncertainty (green shaded). Constraints from CMB (brown) and DES-SN3YR+CMB combined (red) are also shown.
Download figure:
Standard imageHigh-resolution image
Next, we consider other combinations of data. Adding a BAO prior (Beutler et al.2011; Ross et al.2015; Alam et al.2017) in addition to the CMB prior and SN constraints, our best-fitw-value (Table2, row 7) is shifted by only 0.006, the uncertainty is reduced by
compared to our primary result, and the evidence ratio between SN+CMB and BAO is
showing consistency among the data sets. If we remove the low-z SN subset (row 8), thew-uncertainty increases by only ∼8%. Furthermore, we remove the SN sample entirely and find that thew-uncertainty increases by nearly 50% (row 9).
4.3.3. Flat w0waCDM
Our last test is forw evolution using the
CDM model, wherew = w0 + wa(1 − a) anda = (1 +z)−1. Combining probes from SNe, CMB, and BAO, we find results (Table2, row 10) that are consistent with a cosmological constant (
) and a figure of merit (Albrecht et al.2006) of 45.5. Removing the SN sample increases thew0 andwa uncertainties by a factor of 2 and 1.5, respectively (row 11).
4.4. Comparison to Other SN Ia Surveys/Analyses
The DES-SN3YR result has competitive constraining power given the sample size (
with 329 total SNe Ia), even after taking into account additional sources of systematic uncertainty. While our DES-SN3YR sample is less than one-third of the size of the Pantheon sample (PS1+SNLS+SDSS+low-z+HST,
), our low-z subset is 70% the size of Pantheon’s low-z subset, and we included five additional sources of systematic uncertainty, our improvements (Section1) result in aw-uncertainty that is only ×1.4 larger.
5. Discussion and Conclusion
We have presented the first cosmological results from the DES-SN program:
and
for a flatwCDM model after combining with CMB constraints. These results are consistent with a cosmological constant model and demonstrate the high constraining power (per SN) of the DES-SN sample. DES-SN3YR data products used in this analysis are publicly available athttps://des.ncsa.illinois.edu/releases/sn. These products include filter transmissions, redshifts, light curves, host masses, light curve fit parameters, Hubble Diagram, bias corrections, covariance matrix, MC chains, and code releases.
We have utilized the spectroscopically confirmed SN Ia sample from the first three years of DES-SN as well as a low-redshift sample. This 3-year sample contains ∼10% of the SNe Ia discovered by DES-SN over the full five-year survey. Many of the techniques established in this analysis will form the basis of upcoming analyses on the much larger five-year photometrically identified sample.
To benefit from the increased statistics in the five-year sample it will be critical to reduce systematic uncertainties. We are working to improve calibration with a large sample of DA white dwarf observations, including twoHST Calspec standards. Other improvements to systematics are discussed in Section 7.2 ofB18. We are optimistic that our systematic uncertainties can remain at the level of our statistical uncertainties for the five-year analysis. This progress in understanding systematics will be critical for making new, exciting measurements of dark energy and for paving the way toward Stage-IV dark energy experiments like the Large Synoptic Survey Telescope and the Wide Field Infrared Survey Telescope.
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 No. 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 program 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, 306478, and 615929. We acknowledge support from the Australian Research Council Centre of Excellence for All-sky Astrophysics (CAASTRO), through project No. CE110001020, and the Brazilian Instituto Nacional de Ciênciae 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.
This Letter makes use of observations taken using the Anglo-Australian Telescope under programs ATAC A/2013B/12 and NOAO 2013B-0317; the Gemini Observatory under programs NOAO 2013A-0373/GS-2013B-Q-45, NOAO 2015B-0197/GS-2015B-Q-7, and GS-2015B-Q-8; the Gran Telescopio Canarias under programs GTC77-13B, GTC70-14B, and GTC101-15B; the Keck Observatory under programs U063-2013B, U021-2014B, U048-2015B, U038-2016A; theMagellan Observatory under programs CN2015B-89; the MMT under 2014c-SAO-4, 2015a-SAO-12, 2015c-SAO-21; the South African Large Telescope under programs 2013-1-RSA_OTH-023, 2013-2-RSA_OTH-018, 2014-1-RSA_OTH-016, 2014-2-SCI-070, 2015-1-SCI-063, and 2015-2-SCI-061; and the Very Large Telescope under programs ESO 093.A-0749(A), 094.A-0310(B), 095.A-0316(A), 096.A-0536(A), 095.D-0797(A).