The five key science goals for astronomy in the coming decades were identified by the committee as follows:
In assessing the kind of observations that would be most effective in addressing these questions, the committee identified the following as key ones.
1. Are we alone in the Universe?
Obviously the single most important observation would be the detection of signals of an unambiguously artificial nature that could only have originated from intelligent life. The SETI searches in the radio waveband are the best hope for this. This is outside the purview of the NGC study. Likewise, exploration of the Solar System (e.g. Mars or the moons of the outer planets) may yet turn up evidence that life started independently elsewhere. Failing these dramatic developments (and rather more plausibly), we envisage a step-by-step approach involving the following.
a. Characterization of extra-solar planetary systems using indirect search techniques (i.e. the reflex motions of the primary). At present, the emerging view of extra-solar planetary systems is driven by the selection bias in favour of massive planets near to their parent stars. Future improvements can be imagined in astrometry (e.g. with the Space Interferometer Misson) and in higher precision radial velocity measurements that will lower the planetary masses that can be detected through these indirect means. However, there is likely to be a limit imposed here by the velocity noise of the star itself so that more than just light gathering power by itself will be required.
b. Direct detection of terrestrial planets through the isolation of their light. Terrestrial planets are not outstandingly faint (typical visual magnitude 26 around nearby stars), but their direct detection requires exquisite suppression of the light from the much brighter star (e.g. 109 times brighter at 1 3#3, 0.2 arcsec away). Detection of the planet's own thermal radiation (e.g. at 10 micron for a planet in the ``habitable zone'') reduces the required contrast relative to the star to 106 at the expense of poorer resolution. Thus direct detection requires a combination of very high resolution, super-smooth optics and advanced suppression mechanisms (i.e. coronagraphic or null-interferometric techniques).
As a concrete example of what will be required for such a detection consider a terrestrial planet orbiting the nearby star 6#6 Ceti at 1 AU. The separation from the star is 0.3 arcsec and the planet's H-magnitude is about 26. Assuming that we can largely correct for the speckle noise and have a Strehl ratio of 0.8 in the image, this planet could be detected with a signal-to-noise ratio of 3 in about 12 hours of observing time with a 25m ground-based telescope. This estimate neglects any contribution from possible zodiacal light in the 6#6 Ceti system and it also assumes that such a high Strehl ratio could be obtained with a large, likely segmented mirror. This has yet to be shown and requires a real technological advance before it can be achieved.
In any case, to achieve the resolution required to adequatedly suppress the stellar light in the IR/optical and detect the planet's reflected radiation, interferometric baselines or apertures of at least 25m are needed.
c. Detection of life gases. A spectrum of a terrestrial planet orbiting a nearby star would take about a week of time with a 25m telescope and be restricted to finding O2 in the optical/near infrared region as the remaining ``life'' gases only have spectral signatures in the 3 - 20 3#3 region. This is a part of the spectrum much better accessed from space and will be the prime goals of DARWIN and TPF. However, this 25m telescope could provide candidates for further spectroscopy with NGST and other space observatories.
2. What is the Universe made of and what is its overall geometry?
The nature of the ``dark matter'', believed to be the dominant form of matter in the Universe, is unknown and it is, by nature, difficult to detect with telescopes. It is possible that direct detection by particle detectors on Earth or a breakthrough in theoretical particle physics may solve this question, but this is by no means certain. The role of astronomers is to characterize the average density of the dark matter, the power spectrum of density perturbations in the Universe, to determine the relationship between the distribution of luminous material and to determine how this changes with cosmic epoch. There are three ways to detect accumulations of dark matter - through observable scattering effects at the last scattering surface of the CMB, through gravitational lensing effects on background sources of light and through large-scale dynamical effects acting on luminous test-particles (i.e. galaxies).
The overall geometry and dynamics of the Universe is described by the classic cosmological parameters and will reflect the dark matter and other components to the mass density. At present, there is renewed speculation in the possible existence of a false vacuum term which would produce an acceleration of the Universe as opposed to the more conventional deceleration, and much work over the next decades will be focussed on pinning this down. Demonstration of the existence of this term would represent a profound contribution to physics to rival the discovery of gravity in the 1700s and of the expansion of the Universe in the 1920's.
The primary observational capabilities that will be required to tackle these problems are enumerated below.
(a) Measurements of the CMB anisotropies on all angular scales greater than 2 arcmin (e.g. with MAP and Planck). If there are no surprises, this has the potential for determining the cosmological parameters to exquisite precision (though with some degeneracies) and for determining the spectrum of fluctuations of the dark matter.
(b) Undertaking the classical cosmological tests to yield estimates of the age of the Universe and the deceleration parameter q0 can break the degeneracies present in (a) and provide a fundamental consistency check. Much of this work, for instance the study of supernovae at high redshifts, is well suited to wide field 4m class telescopes (e.g. CFHT) with follow-up from 10m class telescopes such as Gemini. NGST should extend this program to z > 1 and a very large ground-based telescope would be a powerful complement to NGST, especially in the area of spectroscopy.
(c) Mapping the large scale structure of the Universe will be best undertaken by 2 - 4m class dedicated telescopes like the SDSS which will map the distribution of 106 galaxies over a scale of 1/10 of the horizon. Dynamical studies of bulk flows and large collapsed structures such as clusters will also be carried out on 4m class telescope. At high redshifts, 10m class telescopes on the ground will pursue these programs to 7#7 and likely beyond. The highest redshifts will demand levels of sensitivity that could be attained only with NGST or a very large ground-based telescope.
(d) Mapping the dark matter distribution through gravitational lensing effects is a key goal of CFHT's wide field imaging surveys - the requirement is for the highest possible image quality sustained over the largest possible fields. Again, NGST or a very large ground-based telescope will extend these studies in a more specialized regime at z > 1.
3. How did our and other solar systems form and evolve?
Developing a better understanding of the star formation process (and at the same time how planets might form) means investigating phenomena on very different scales. The key word thus describing any facility for this work is probably versatility. Superb image quality with both adaptive optics systems and wide field imagers appear necessary as well as an integral field spectrograph (or a Fabry-Perot interferometer) and a high resolution spectrograph. Two options thus appear suitable for the next CFHT facility.
(a) A large (8#8 25m) telescope with superb image quality (adaptive optics system) equipped with an integral field spectrograph and/or a high resolution spectrograph. This facility should be optimized for the visible and the near-infrared and could study the star formation rate in galaxies at high redshift and even view individual high mass stars being formed at redshifts beyond 7#7. It would also open up the possibility of measuring oscillation frequencies of stars in the nearest open clusters, thus providing decisive tests of stellar evolution theory.
One clear advantage of such a telescope on the ground compared to NGST will be its superior spectroscopic capabilities. For example, for a protostellar system in our Galaxy, a 25m NGC could trace the gas falling down from the outer edge of the protostellar envelope in to the stellar seed (i.e. resolve protostellar disks on scales of 2 - 5 AU). A spectral resolution of 100,000 coupled with the large collecting area would allow for measurements of velocities to a few km/s of faint gas clumps and permit us to estimate the mass infall rate. If planets are formed during the accretion phase, high spectral resolution is necessary to observe the gaps in disks via the shapes of the absorption lines. The spectroscopic capabilities of the 8m-class telescopes will just barely allow this kind of observations but the integration times will be very long (2#210 hrs). By contrast a 25m NGC will not only allow investigation of very specific targets but also permit surveys of protostellar systems to be carried out searching for planetary systems in formation.
The star formation activity in galaxies of different morphologies, environment and epoch varies by many orders of magnitude. The origin of these differences is still unclear but a large aperture telescope will allow one to study star formation in faint protogalactic systems (a supporting role for NGST) and study the distribution of star forming regions in individual objects at high - z. Equipped with integral field units and a spectral resolution of a few thousand, imaging spectroscopy of faint irregular galaxies up to 9#9 will allow us to derive their star formation properties (i.e rates, distribution, abundances). Stellar populations of nearby galaxies (<25 Mpc) could be studied spectroscopically for individual objects to derive the star formation history of these systems, including their chemical evolution which is a long-standing problem in galaxy evolution studies.
(b) Since a detailed understanding of the physical processes at work inside a star is required before complex models of its formation and evolution can be developed, a future CFHT can be optimized for high spectral resolution studies including asteroseismology. For this purpose a telescope of the 4 - 8m class equipped, at a minimum, with an efficient echelle spectrograph/spectropolarimeter will be needed for most of the stellar physics programs foreseen for the next decade or more. Such a facility will indeed give access to the study of the physical processes in both individual nearby stars and members of young open clusters and pre-main sequence objects. Also, based on existing asteroseismological results on 2m class telescopes, it can be expected that solar-like oscillations will be detected down to about 7th magnitude with an 8m telescope. This will give access to a robust stellar sample.
Another important observational aspect of stellar physics in the future will be the need for continuous spectroscopic and spectropolarimetric monitoring. This need implies that if the new CFHT is in the 4 - 8m class range it must be operated in such a way that long observing runs can be attributed to specific stellar physics programs, and that CFHT observations can be part of world-wide multi-site campaigns involving other telescopes of the same class.
4. What were the first sources of light in the Universe, and how did galaxies like our own come into being?
This science goal requires extreme sensitivity and good angular resolution because the interesting objects are extremely distant as they must be observed in the early Universe. Because the light from the earliest objects is highly redshifted, observations in the near-infrared are extremely important, both for detecting older stars at high redshifts (z >> 1) and, crucially, for detecting the first stars formed at z > 6. We also know that about 50% of the light from star-formation at high redshift is obscured by dust and reradiated in the far-IR and observations in the sub-mm waveband are required to follow this component. Thus many facilities will be required to yield a comprehensive picture of how galaxies such as our own formed and evolved and how the Universe first ``lit up'', ending the darkness which followed the Big Bang.
(a) The first stars and galaxies and the assembly of mass: NGST will have exceptional sensitivity in the near-IR and will make major contributions to this area. With its spatial resolution of a fraction of 0.1 arcsec, corresponding to kpc scales at cosmological distances, it will allow the separation of different morphological components of galaxies and the determination of gross physical properties (such as chemical abundances) across the galaxies. Thus NGST should enable us to follow the assembly of material into galaxies by detecting sub-galactic fragments even to very high redshifts. A very large aperture ground-based telescope (25m+) would act as a powerful complement to NGST in the optical and near-infrared (even without full diffraction limited-performance) by having enough aperture to perform the required spectroscopy on the galaxy fragments.
(b) The hidden phases of galaxy evolution: FIRST and the future millimeter-wave arrays (LSA/MMA) will also be very important as detectors of dust emission and line studies of the gas. The latter will provide images with comparable (kpc-scale) resolution as will be obtained with NGST.
(c) Breaking open galaxies for observations: Finally, a very large aperture diffraction limited ground-based telescope would offer a whole new capability - especially at visible wavelengths where the gain relative to NGST would be maximized. First, the exquisite resolution and sensitivity would be sufficient to study individual objects (such as star-forming regions, even the brightest individual stars) within galaxies at redshifts as high as 7#7, i.e. when the Universe was about 1/3 its present age. This would transform our view of galaxy evolution - galaxies would be observed as composed of individual interacting objects rather than as overall ``systems''. Likewise the gain in visible sensitivity and resolution would enable detailed studies to be made of the stellar content of local galaxies - a 25m telescope would detect moderate mass stars in the Virgo cluster, vastly increasing the range of galaxies for which their past history could be determined.
While many questions posed above will be tackled by the NGST, this telescope will have its own limitations as it will be a modest sized instrument (with limited capabilities in the optical waveband, spectroscopy and wide field imaging). It appears necessary to complement NGST (and the future LSA/MMA) with a large aperture ground-based telescope, in much the same way HST and the 8 - 10m telescopes are complementing each other today. To access LMC-type building bricks of galaxies at z 2#2 3 - 10 will require a facility capable of reaching magnitudes in the range of 28 - 29 in spectroscopic mode. A ground-based telescope (> 25m), with full adaptive optics, multi-object and integral field spectroscopy will be required for this task.
5. Are there things in the Universe that we haven't as yet dreamed of?
By the very nature of this question, it is impossible to predict what observational facilities will be required to discover completely new types of objects or phenomena. The history of astronomy suggests that performance gains of several orders of magnitude are required to discover previously unsuspected phenomena or objects. Smaller gains in performance, while invaluable in the progress of the discipline, lead to evolutionary gains rather than revolutionary gains. A recent example of such revolutionary gains might be the BeppoSax satellite which, with its wide spectral coverage and extraordinary positional accuracy, has revolutionized the study of gamma-ray bursters. Among the range of facilties now being built or planned, we identify the following as instruments which have the potential for major surprises.
1) NGST and FIRST in the 10 - 200 micron range where they will achieve 1000 - fold gains in sensitivity.
2) Kilometer-scale interferometric arrays on the ground or in space, operating at visible to mid-IR wavelengths and offering the potential for 1000 fold gains in spatial resolution.
3) The sub-millimeter/millimeter array, offering enormous gains in resolution and sensitivity.
4) Extremely large centimeter arrays on the ground such as the SKA.
5) The MAP and Planck microwave satellites which will undertake the mapping of CMB anisotropies over all angular scales greater than 10 arcminutes, to a level of accuracy set by astrophysical limits,
6) A ground-based optical-IR telescope having a minimum diameter of 25m. Such an instrument would represent a factor of 6 increase in the collecting area of the primary mirror over planned or existing facilities. This increase is identical to that achieved by the Keck telescope when it began operation in 1993. Many of the scientific gains realized by Keck are better classified as revolutionary rather than evolutionary, particularly in the fields of high-redshift galaxies, quasar aborption line systems and gamma-ray bursters. A 25m+ telescope will certainly play a similar role and enjoy comparable gains. It is also worth noting that such an instrument would result in a factor of 5 gain in spectroscopy over NGST (at 1 3#3) -- itself a revolutionary instrument -- and potentially much larger gains in the optical, where NGST's performance is expected to be limited or non-existent.