The following plan follows naturally from the discussion in the previous sections. It is clear that the telescope which will have the strongest impact on the key science goals in the next few generations is an instrument with an aperture in the range of 25m or more. A 25m ground-based telescope would have a resolution of .005 arcseconds in V, and would reach V = 35 in 10 hours integration for point sources. A ground-based telescope with a diameter of 25m would be, at 1 3#3, equivalent or superior to NGST, with better spatial resolution, and higher collecting efficiency. For shorter wavelengths the ground-based instrument completely dominates for both imaging and spectroscopy. Above 2#2 2 3#3 the thermal background of the ground telescope starts to dominate and in this region NGST will have unsurpassed efficiency. It is therefore important to seek a complementary approach between NGST and the next generation of ground-based telescopes, as has been done successfully between HST and the 8 - 10m instruments on the ground.
Below we provide three options for a Next Generation CFHT. The committee was close to unanimous in the choice for a large instrument but each proposal suggested in what follows had its champion. The first option is an instrument that will be capable of impacting strongly the scientific goals outlined in the preceeding sections. The second option will have less of an impact but will be a more conservative project providing a telescope which is an important contributor to certain restricted areas of the science program that has been developed. The third option will produce an instrument which will serve mainly as a support instrument for the larger telescopes.
The first option of the committee, and the one which it overwhelmingly recommends, is composed of 3 parts.
(1) The existing CFHT should be operated with most of its current suite of instruments (except for the addition of a wide field IR-imager) for about another decade. With MEGACAM, OASIS, the Adaptive Optics Bonnette, a high resolution spectrograph and a wide field IR-imager CFHT will continue to be competitive in many areas of research (galaxy surveys, weak lensing, stellar population studies, detailed observations of active galactic nuclei, stellar spectroscopy).
(2) Beginning almost immediately the CFHT Board should initiate a ``Phase A'' study of a large optical/IR telescope, in the range of 25m. The CFHT staff should be directly and actively involved in this exercise. This telescope should be optimized for imaging and spectroscopy, have an adaptive optics system and a suite of instruments that includes wide field imagers in the visible/IR, and integral field, multi-object and high resolution spectrographs. Such a telescope could be built on the current CFHT site. This site, by all accounts, remains the premier locale on Mauna Kea and the community will want to maintain its use of it. This telescope will be able to accomplish the prime scientific goals outlined in the previous sections and in addition, at 25m, will have the minimum size capable of exploiting new discoveries made by the NGST currently scheduled for launch by NASA in 2007.
(3) If it is decided to build this large telescope, construction could begin in about 2008 at which time the current CFHT operations would cease.
The committee is well aware that this telescope requires some technological developments before it can be realized. In particular, for it to have its optimal impact, the telescope will have to operate with a full adaptive optics system. This point and other related issues are discussed more completely in the following section.
A second possible option for a Next Generation CFHT, and one that had only modest support from committee members, is construction of a modern, moderate-sized (8m) replacement for the existing CFHT. This telescope could likely fit in the existing CFHT dome and be much less costly than the large instrument proposed above. However, most committee members have clear concerns that such an instrument will not have the scientific impact of the larger instrument. However, building a very large (25m+) telescope on Mauna Kea is not without its risks. Some of these are detailed below in making the case for a more modest refurbishment of the CFHT.
It is almost certain that telescopes of the future that will be 10m or more in diameter will have segmented primary mirrors. This will likely limit the quality of the images that will be obtained with such instruments due to scattered light at segment boundaries and misalignment of individual segments. In a short technical note M. Espiard, former director of REOSC, claims that an inexpensive refurbishment option for CFHT would be an 8m with the best 8m mirror constructed to date. Such a mirror could have a Strehl ratio as high as 0.92 and an RMS wavefront error at 550nm of about 20nm. Effectively this translates into a mirror that is diffraction limited in the V band. A further claim is that a telescope with this mirror could be installed in the existing CFHT dome in 3 years at a cost of about 300M French francs.
Assuming that none of the above has been overstated, we can ask what the impact of such an instrument will be. Allowing for detailed ``Phase A'' studies and financial arrangements means that this telescope will likely not be available for observations before about 2006, 7 years from now. By this time 8 northerm 8m class instruments will be in operation as well as 7 in the south. But this telescope will be somewhat unique with its superb optics. Even so it will not impact strongly on our science goal #4 (First Stars and Galaxy Formation) which requires either long wavelength sensitivity (31#31) to see forming galaxies at high redshift or a very large aperture to carry out spectroscopy of extremely faint objects. Similarly, science goal #2 (What is the Universe Made of and Geometry of the Universe) which explores the dominant source of mass in the Universe demands a telescope capable of investigating the kinematics of the visible matter in distant galaxies. The dominant spectral features that will be used for such a study will all be significantly redshifted so that a modest (8m should now be considered modest!) ground-based telescope will find it difficult to compete with NGST. Although the images with this proposed new CFHT will surpass that of any other ground-based telescope currently planned, it will still not place the telescope into a regime of parameter space that entirely new phenomena are likely to be discovered. Its step forward from the existing and planned 8m instruments will be only incremental. Hence it will have little impact on science goal #5 (Things Not Yet Dreamed Of).
Where this new CFHT could have its most important influence is on goals #1 (Are We Alone?) and #3 (Formation of Solar Systems). Adaptive optics imaging with a coronograph of nearby stars searching for zodiacal light-type disks as indicators of planetary systems would be a prime observational program for such a telescope. It is not at all clear that infrared imaging (at 32#32) would be more advantageous for such a project as at least the one system that has been studied in most detail (our Solar System) does not appear to have a large infrared excess. The smooth optics of this NGC telescope, with its minimum of scattered light, would also be useful in microlensing studies of large samples of stars in the direction of the Galactic centre. This program could eventually provide statistics of the number of planetary systems as the presence of a planet (in the proper position) yields a microlensing signature which is unique and which can be produced even by terrestrial-type planets. An 8m telescope will not be able to directly image a terrestrial planet. This feat will be reserved for large ground-based instruments (33#33)m or sophisticated space interferometers.
The smooth mirror of this NGC will also be important in other studies not directly defined by the stated science goals. Wide field imaging, both in the optical and near IR, will greatly benefit from the superb images of this telescope. Such imaging studies could be important in large scale structure investigations, in searches for distant supernovae as well as in trying to locate the Galactic microlensing candidates.
It is only reasonable to evaluate the impact of this telescope on the primary science goals in the same manner that we have done for the others. If we add a column to Table 5 for this 8m ``Super Smooth'' instrument, the evaluations would be B B B B C.
This particular choice for the next generation of CFHT is likely to present many fewer technological challenges than the 25m option. The current new generation of 8 to 10m telescopes has been made possible by a technical breakthrough: active optical control. No equivalent breakthrough has been made today that can ease the construction of a much larger telescope. The use of a thin meniscus mirror for a 25m telescope is out of the question. Only segmented mirrors can be used and the likely price to pay is a loss of optimal optical quality. This is clearly demonstrated by the first VLT images, the quality of which already surpasses that of the Keck images at the same wavelength. At its current level of development, adaptive optics does not properly correct for discontinuities in a segmented mirror surface. This is an area where technological advances will be required in order to make the 25m+ feasible.
It can be argued that the mere collecting power of a 25m telescope amply justifies the investment. Unfortunately, under seeing limited conditions, instrument size grows in proportion to telescope size which makes instrumentation as problematic as the telescope to construct. Worse, detector pixel size must also grow in the same proportion to the detriment of detector performance. The problem is already severe on current 8 to 10m telescopes; it may be insurmountable on a 25m telescope. The only way out is the use of adaptive optics. The committee's primary recommendation assumes that adaptive optics can be implemented on a segmented-mirror 25m telescope; it is indeed the case that this is a large extrapolation from current knowledge.
The first VLT results have demonstrated that actively supported thin meniscus mirrors can deliver images with an optical quality which surpasses that of any other ground-based telescope. Development costs for such mirrors have all been paid for, and companies such as REOSC can now routinely produce top quality 8m mirrors at bargain prices. Moreover, the CFHT dome is so oversized that a modern 8m telescope would fit inside it. The argument that 15 other similar telecopes will already be operating must be viewed in context of the history of CFHT. When the CFHT saw first light, a number of 3.6m or larger telescopes were already in operation, including the Palomar 5m and Soviet 6m telescope. It did not prevent the CFHT from being considered, 20 years later, as one of the most productive telescopes in the world. This is because of the telescope's unique performance and highly advanced instrumentation. Today the CFHT is still unique for both wide-field imaging and adaptive optics. There is no fundamental reason why an 8m CFHT would not remain competitive.
There are a number of other issues that make the 8m proposal attractive. Canada currently has about 155 nights on the CFHT. Her share of time on a 25m telescope is unlikely to be as large (other partners will probably be required to financing the building of such an instrument). With closure of the present CFHT to build the large instrument, Canada's access to ground-based telescopes will be very minimal. France is in a somewhat better position with its connection to ESO, but even here her access to the northern sky will be curtailed. One benefit that will accrue from moving quickly to an 8m option is a sharing of instruments and instrumentation development costs with other 8m telescopes on Mauna Kea. There is also the potential for considerable savings in operational costs if appropriate arrangements with the Gemini consortium can be developed.
We offer one other option for a refurbished CFHT, and that is a simple upgrade of the existing telescope coupled with a strong investment in a next generation of instrumentation. Such a plan would be low cost relative to the other options mentioned above and could keep CFHT in a competitive position for the foreseeable future in the limited areas of research in which it has decided to specialize.
Such a refurbishment might involve overhauling the telescope mechanically and electronically, improving the dome environment and possibly replacing the mirror. Simultaneously, CFHT must develop a strong and unique instrumentation plan that will continue to exploit the superb CFHT site.
If this option is adopted, CFHT should start to define and implement, as soon as possible, a new operating mode for the telescope. The operating mode should be dedicated to vast key programs: wide-field imaging with Megacam (it is unlikely we will need a next generation of this instrument): place the wide field IR camera high on the priority list of instrumentation and plan eventually for an even wider field IR camera, possibly to be located at the prime focus, and which will also have spectroscopic capability: high resolution cross-dispersed spectroscopy/polarimetry for asteroseismology and other long term spectroscopic monitoring programs (a second generation instrument after Espadons could be planned with even higher resolution).
This new mode of operation of CFHT could start to be phased in immediately and be fully operational around 2003-2004 at which time some of the next generation instruments (larger IR camera, next generation spectrographs) could undergo serious development. A CFHT operated in such a mode will, to a great extent, be a support telescope for the larger instruments. It would be useful for the construction of databases of multicolour photometry that would be helpful in identifying suitable candidates for spectroscopy or higher resolution imaging with the bigger telescopes.