Prospects for future supernova surveys are discussed, focusing on the European Space Agency’s Euclid mission and the European Extremely Large Telescope (E-ELT), both expected to be in operation around the turn of the decade. Euclid is a 1.2 m space survey telescope that will operate at visible and near-infrared wavelengths, and has the potential to find and obtain multi-band lightcurves for thousands of distant supernovae. The E-ELT is a planned, general-purpose ground-based, 40-m-class optical–infrared telescope with adaptive optics built in, which will be capable of obtaining spectra of type Ia supernovae to redshifts of at least four. The contribution to supernova cosmology with these facilities will be discussed in the context of other future supernova programmes such as those proposed for DES, JWST, LSST and WFIRST.
The discovery of the accelerating expansion of the Universe [1,2] was one of the biggest breakthroughs of the last 20 years and was awarded the 2011 Nobel Prize for Physics. However, the fundamental question remains: what drives this acceleration? One possibility is that it is driven by a mysterious component of the Universe, termed ‘dark energy’ that exerts negative pressure and that constitutes about 70 per cent of the energy density of the Universe. Focus has now turned to measuring the equation-of-state parameter w (= pressure/density) of this dark energy. Current measurements of w are consistent with −1, the value expected if dark energy is equivalent to the cosmological constant (Λ). Although this would perhaps be the simplest explanation, large problems remain; for example, there is no natural explanation for such a small and yet non-zero value for the vacuum energy density. Other explanations such as ‘quintessence’ and modified gravity have been proposed, and distinguishing between these models is one of the major challenges of modern physics. A measurement of w not equal to −1 (at any redshift) would rule out the cosmological constant explanation and would have profound consequences for physics. Therefore, there is considerable effort being directed towards improving measurements of w using several techniques, including weak lensing, baryon acoustic oscillations (BAOs) and type Ia supernovae (SNe Ia).
The best constraints on dark energy to date measure w consistent with −1 at approximately 7 per cent (including statistical and systematic uncertainties), assuming a flat universe and a constant equation of state [3,4] (figure 1). Presently, systematic errors are estimated to be comparable with the statistical uncertainties in SN cosmology. The major sources of systematic error are related to photomeric calibration, particularly when comparing distant SNe with nearby SN samples that have been compiled in a different way and with different rest-frame wavelength coverage. New nearby searches are addressing this problem (for examples, see papers by B. Schmidt and J. Tonry at this meeting).
A second source of systematic error arises from our lack of understanding of the colours of SNe Ia. SNe Ia show a range of colours, and the colour correlates inversely with brightness. Such a relation could be caused by dust extinction in the SN host galaxies or could be an intrinsic property of SNe Ia themselves, or a combination of the two. To make progress, any future SN survey should measure SNe Ia in multiple bands, particularly towards redder wavelengths where the effects of dust are reduced and there is evidence for reduced intrinsic dispersion .
Finally, another potential systematic effect is evolution of the SN Ia population. We know that SN Ia spectra are similar at low and high redshift (up to z∼0.9) [8,9] although differences in the UV have recently been seen . We also know that SN Ia rates and broadband SN Ia lightcurve properties are dependent on the host galaxy type. The luminosity of SNe Ia depends on host galaxy type even after correction using the usual stretch and colour correction method . Extending the redshift range over which SNe Ia are measured will allow better control of these effects. To date, only a handful of SNe Ia have been observed at z>1 at any wavelength. Such observations are extremely difficult from the ground because (i) the SNe are faint and (ii) the peak of the SN Ia spectral energy distribution moves into the near-infrared (NIR) where the sky background is higher. The Hubble Space Telescope has therefore been the primary route for finding such SNe [4,12,13].
(a) Future facilities
In the following sections, several upcoming projects are described that will make dramatic advances in SN Ia cosmology. These projects approach the problem in two main ways. The first is improving statistics (which can also lead to improved control of systematics through construction of subsamples). The second is extending the wavelength coverage of high-redshift SN samples into the NIR.
(i) Dark Energy Survey and VISTA
The Dark Energy Survey (DES, see http://www.darkenergysurvey.org) is an international private–public partnership involving the USA, Spain, UK, Brazil and Germany with the goal of mounting and operating a 3 sq deg CCD camera on the CTIO Blanco 4 m telescope. DES will carry out a 5000 sq deg survey over 5 years starting in late 2012. It will survey the sky in g,r,i,z,Y bands.
The SN component of DES has been described in detail by Bernstein et al. . 30 sq deg will be repeat-imaged in the g,r,i,z bands and is expected to result in 4000 well measured SNe Ia in the redshift range 0.05<z<1.2. An external spectroscopy programme is being planned that will provide host galaxy spectra (which aids SN photometric classification) and spectra of a subset of the SNe themselves (less than 20%), some of which will have detailed multi-epoch spectroscopy.
The 4 m VISTA telescope, equipped with a 67 Mpix NIR camera and operated by European Southern Observatory (ESO), has been carrying out public NIR surveys in Z,Y,J,H,Ks bands since 2009. The SN component within the VIDEO survey  is expected to produce about 100 SNe Ia with z<0.5 observed in Y and J bands. These SNe will overlap with the DES sample and will therefore have both optical and NIR lightcurves.
(b) Large Synoptic Survey Telescope
The Large Synoptic Survey Telescope (LSST, see http://www.lsst.org) is a US-based public–private partnership that aims to construct and operate an 8.4 m survey telescope (6.7 m effective diameter) equipped with a very wide field camera (9.6 sq deg) with u,g,r,i,z,y filters. The main survey will cover approximately 20 000 deg2 every 3–4 days, and by the end of the survey each field will have been imaged over 1000 times. The deep drilling fields will cover a smaller area but reach a deeper limiting magnitude per visit and with faster cadence (to be determined), thereby producing SN lightcurves of higher quality and reaching higher redshift than the main survey. Twenty to 40 such fields are envisioned, of which the locations of four have already been selected. LSST is due to start operation at the end of the decade and will have an approximately 10 year survey lifetime.
The SN Ia case has been described in detail in Wood-Vasey et al. 16. For SN Ia cosmology, the key gain is massive statistics: 50 000 SNe Ia per year are expected, with redshift up to approximately 0.8 (main survey) and up to approximately 1 (deep drilling fields). Such statistics also improve systematics by allowing the sample to be split into subsets, for example depending on host galaxy type. However, because of the very large sample size, only a small fraction will have spectroscopic redshifts and classifications. The LSST SN survey will also allow tests on isotropy and homogeneity, and tests of SN Ia evolution. LSST will also find core collapse SNe and measure SN rates (for all types).
(i) James Webb Space Telescope
The James Webb Space Telescope (JWST, see http://www.jwst.nasa.gov/) is a joint project between NASA, the European Space Agency (ESA) and the Canadian Space Agency. It consists of a 6.5 m space observatory with four instruments optimized to the infrared wavelength range. Launch is currently expected in 2018. Being an observatory as opposed to a survey instrument, JWST’s science use will be determined mainly from PI programmes. JWST’s key advantage for SN cosmology is sensitivity at infrared wavelengths, allowing it to reach well above z=1. Several possible ways that JWST could advance SN cosmology have been suggested.
z>1: in the white paper ‘James Webb space telescope studies of dark energy’, Gardner et al.  demonstrate that in 1 year using 1080 h, JWST could find and follow 60 z>1 SNe, including obtaining spectra with JWST itself.
z>2: Riess & Livio  make the case in that observing at z>2, where the effects of dark energy are expected to be very modest, will provide a key test for evolution in the SN Ia population. They estimate that approximately 1.5 SNe Ia could be found within each NIRCAM field reaching 10 nJy in the K band. They suggest monitoring a few NIRCAM fields with cadence of approximately 100 days
z∼4: the JWST exposure time calculator predicts that JWST can reach two magnitudes below peak brightness of a typical SN Ia lightcurve at z=4 in 10 000 s. By monitoring 10 NIRCAM fields for 5 years, a sample of approximately 50 objects could be found. Spectroscopic confirmation could be obtained using future ground-based extremely large telescopes (see §1b(ii)).
(ii) The European Extremely Large Telescope
Three extremely large telescope projects are currently underway: the Giant Magellan Telescope (GMT, see http://www.gmto.org/), the Thirty Meter Telescope (TMT, see http://tmt.org/) and the European Extremely Large Telescope (E-ELT, see http://www.eso.org/sci/facilities/eelt/). These projects are at a comparable stage of development and all aim for first light around the turn of the decade. In this paper, I focus on the E-ELT, but the broad conclusions are similar, or scalable, for all three projects.
The E-ELT project (figure 2) aims to design and construct a 39 m diameter optical–IR telescope, which will be the largest optical/IR telescope in the world. The project is run by ESO on behalf of its member states. The design is a novel 5 mirror concept with adaptive optics built in, and the telescope will be situated on Cerro Armazones in Chile. The first three instruments have been selected: a diffraction-limited camera and an integral-field spectrograph for first light, and a mid-IR imager-spectrograph to arrive shortly afterwards. The full instrumentation suite will be built up over the first decade of operation. Since the time of writing, the ESO Council has approved the E-ELT programme subject to confirmation of funding from some member states.
The E-ELT is a general-purpose observatory, and the science case is very general . In addition to its clear power for identification of varying sources from other facilities (e.g. gamma-ray bursts), examples of time domain science with E-ELT include Solar System observations (including study of weather and volcanic activity), exo-planets (including radial velocity, direct detection and transit measurements) and study of the motions of stars and stellar flares in the vicinity of the Galactic centre. Furthermore, with specialized high-time-resolution detectors, E-ELT could study extreme physics (pulsars, neutron stars, black holes) stellar phenomena, transits and occultations .
The E-ELT will be particularly powerful for spectroscopic observations of SNe Ia at high redshift. Because they are point sources, SN observations benefit from the telescope’s adaptive optics capability. Such spectroscopic observations are needed in order to unambiguously classify the SNe and to determine their redshifts. Simulations show that with the HARMONI integral-field spectrograph  and using the telescope’s adaptive optics system, the SiII feature near a rest-frame wavelength of 4000 Å is clearly detected even at z=4 (at which redshift the feature is observed in the K band) . Because the presence of SiII and absence of hydrogen is the defining signature of SNe Ia, this means that it will be possible to confirm SN types even out to z=4. To obtain E-ELT spectroscopy of a sample of 50 objects, 1<z<4 would require of the order of 400 h spread over the 5 years of the survey.
Euclid is a 1.2 m optical–IR space telescope within ESA’s Cosmic Vision 2015–2025 programme . Its primary science goal is precision measurement of cosmological parameters via weak lensing and galaxy clustering techniques. ESA’s downselection process in October 2011 resulted in selection of Euclid for the second M-class launch slot. In June 2012, the mission passed the important milestone of adoption of the mission by ESA. Launch is currently scheduled to take place in the last quarter of 2019.
The instrumentation consists of an optical imager and a NIR imager and spectrograph, both with fields of view of 0.5 sq deg. The satellite will be launched by a Soyuz rocket and will operate at the L2 Lagrange point for a 6 year mission duration.
Euclid’s wide survey will cover greater than or equal to 15 000 sq deg, with imaging in a single broad optical band (R+I+Z) to a depth of AB=24.5 (10σ for a point source), imaging in three NIR bands (Y, J, H) to a depth of AB=24 (5σ, extended source) and NIR slitless spectroscopy to a depth of 3×10−16 cm−2 s−1 (3.5σ unresolved line flux).
Additional deep fields will cover greater than or equal to 40 sq deg, reaching two magnitudes deeper than the wide survey in both optical and NIR imaging and NIR spectroscopy. The deep fields are primarily for calibration purposes, but the approximately 40 repeat visits will enable a vast range of additional science, including detection and study of variable, moving and transient objects.
In addition, ideas for specialized dedicated surveys for SNe and microlensing of exoplanets are being considered, although these are not currently in the baseline. These will be further explored as the survey design is optimized.
For SN Ia cosmology, one example strategy has been developed that would allow exploration of a new redshift range, to z∼1.5. Simulations show that six months of Euclid survey time could be used to carry out a survey of 20 sq deg with 4 day cadence. When combined with simultaneous ground-based observations in I and z bands, this results in 1700 well-measured SNe Ia in the redshift range 0.75<z<1.5 with measurements covering a consistent rest-frame wavelength range. Such a survey would make a significant improvement in the measurement of cosmological parameters from existing surveys at the time, and would add an independent method to Euclid’s primary cosmological probes, thereby enhancing the mission’s overall impact.
(iv) Wide-Field Infrared Survey Telescope
The proposed NASA mission Wide-Field Infrared Survey Telescope (WFIRST) emerged as the top priority large space mission in the US ‘new worlds new horizons’ 2010 decadal survey. Its science drivers are measurement of expansion history of the Universe and galaxy clustering (exploring dark energy and modified gravity theories via the weak lensing, supernova and BAO techniques); exoplanets (via microlensing); deep NIR surveys; a galactic plane survey; high-z quasi-stellar objects; and a guest observer programme.
Definition of the hardware is in progress. The interim report of the Science Definition Team study  described a 1.3 m off-axis telescope operating from 0.6 to 2.0 μm, and an approximately 0.3 sq deg field of view covered by imaging in five filters and slitless spectroscopy with R∼200 from 1.1 to 2.0 μm. An additional R∼75 prism would be available for SN spectroscopy. Launch is anticipated for approximately 2020, assuming phase A starts in 2013.
The WFIRST SN Ia programme goals are stated as more than 100 SNe per Δz=0.1 redshift bin in the range 0.4<z<1.2 per dedicated six months (spread out during the mission lifetime). The goal is to achieve an error on distance modulus of less than 0.02 mag per Δz=0.1 redshift bin. They also consider an ‘optimistic’ case where the redshift range is extended to 1.5 and systematic effects are reduced.
There is a spectacular suite of new facilities on the horizon that will make dramatic advances in the quantity and quality of SN samples for cosmology. Several of these facilities are already planning SN programmes and have made detailed predictions of the achievable results in terms of numbers of objects and in some cases cosmological constraints (e.g. in terms of the dark energy task force figure of merit). Such figure of merit calculations depend critically on the assumptions made for systematic effects and priors, so comparison is not straightforward and is not attempted here. However, by comparing the N(z) distributions of existing and future samples, we can obtain a simple visual impression of the gains we can expect (figure 3).
In summary, the gains likely to emerge are
— an enormous gain in statistics (at least two orders of magnitude) of optical SN Ia measurements, from a combination of current nearby searches, DES and then LSST;
— an enormous gain (two orders of magnitude) in numbers of observed-frame NIR observations of distant SNe, from VISTA and then space-based surveys with Euclid (survey strategy permitting) and WFIRST, potentially allowing extension of the SN Ia Hubble diagram up to z∼1.5;
— better control of systematic effects, arising from (i) the ability to create sub-samples from large parent samples and (ii) the improved wavelength coverage resulting from combined optical–NIR observations of the same SNe; and
— extension of the SN Ia Hubble diagram to currently unexplored redshifts above 2 and up to about 4, from JWST and ELTs.
In summary, the prospects are very exciting. However, coordination of survey execution between facilities is required to maximize the gains; for example, survey fields must be chosen so that they are observable by both space and ground based facilities, and scheduling must permit such coordination. In particular, exploitation of the synergies between JWST and the ELTs and between Euclid/WFIRST and LSST would be particularly valuable.
I.H. is supported by grants from STFC, OPTICON (EC FP7 grant no. 226604) and the UK Space Agency for work on the European ELT and Euclid projects. The E-ELT/HARMONI simulations in §1b(ii) were carried out by Tim Goodsall and I.H. Calculations for JWST observations of z>2 SNe (§1b(i)) were carried out by W. Taylor and I.H. Section 1b(iii) is based on work undertaken within the Euclid SN and transients SWG (part of the Euclid Consortium), and makes use of calculations performed by P. Astier, S. Spiro, K. Maguire and J. Guy. Helpful input on the LSST SN search (§1b) was provided by M. Wood-Vasey.
One contribution of 15 to a Discussion Meeting Issue ‘New windows on transients across the Universe’.
- © 2013 The Author(s) Published by the Royal Society. All rights reserved.