Document: ICAP-AB-002 Revision no.: 1 Last updated: 03.10.2001 First created: 02.10.2001 Author: Anthony Brown, brown@strw.leidenuniv.nl Description: Questions concerning the building of a database of real spectra ----------------------------------------------------------------------- GAIA Classification working group: Identification, Classification and Astrophysical Parametrization (ICAP) ----------------------------------------------------------------------- Questions concerning the building of a database of real spectra --------------------------------------------------------------- I was asked to present at the RVS workshop (Meudon, October 1-2 2001) the plans that exist within ICAP for building up a database of observed spectra in support of the development of the GAIA photometric system. This is a summary of that presentation. I think that at this moment we are not ready to devise a concrete plan for obtaining spectra because we lack a precise enough set of requirements for these spectra. At the moment the requirements on the observed spectra coming from the PWG/ICAP are much too vague. They amount to asking for flux-calibrated spectra over the full GAIA wavelength range at certain resolutions. Taken at face value, this a requirement which is very difficult to meet and will certainly take a very large amount of time and effort. Therefore below I list a number of relevant questions that need to be answered first in order to get to a much more focused plan of what kind of spectra and how many we need to obtain. --- Questions to be addressed --- The following broad areas need to be addressed and each on of those will be covered in more detail below. + What characteristics should the spectra have? + Which objects, and how many, are to be observed? + What libraries of observed stellar spectra are presently available and what are their deficiencies? + On what timescale are they needed? + Which observational facilities can we count on? Of these questions the first and second need to be answered most urgently in order to be able to start planning observations. --- What characteristics should the spectra have? --- + What wavelength range needs to be covered? Is it necessary to have continuous coverage from 250 to 1000 nm or can we live with gaps? This will depend on whether or not we can exclude a priori certain regions of the spectrum because they will not be useful for photometry. This issue is important because there are not many spectrographs available that offer (in an easy way) full coverage of this wavelength range. + What resolution is needed? One single resolution would be preferable in view of the required amount of observing time. + Flux-calibrated spectra have been requested. Does this mean absolute flux calibration or relative? + How accurate should the flux scale be? It makes a very large difference whether one is talking about, say, 2 per cent or 10 per cent accuracy. --- Target list --- We need to come up with a list of objects to be observed which covers as much of stellar parameter space as possible. Below a preliminary list as suggested by various people is given. Filling in the details of this list is a top priority. Preliminary target list: - metal poor stars - peculiar stars (WR, Be, Ap, Carbon) - young/active stars - open clusters with a range of extinctions - globular clusters over a range in [M/H] - Bulge fields for studies of the effects of extinction - 47 Tuc for alpha-element variations - omega Cen for CNO anomalies ==> Keep in mind that we also need to define set of standard stars to observe! An important point here in view of the timescales involved is that if cosmic variance is to be taken into account this list may comprise thousands of objects to observe! --- Time Scales --- These are the timescales as I understand them at present: + The present three proposals for the photometric system are to be merged into one by July 2002. + Major design details of the photometric system (Location MPB and BBP; approximate number of filters) are to be fixed by the end of 2001. + Fix other parameters during 2003--2004 (number of filters, response curve, manufacturing feasibility). + Final details can be left open to mid-2005 + Do not count on changes to the filter system beyond 2005 This is a tight schedule if many observed spectra are needed. How much room do we have to relax this schedule? --- Observational Facilities --- Going through normal time allocation procedures at large observatories will probably not be efficient enough. Other possibilities: + Make use of existing programmes and telescopes to which someone has "private" access. (e.g., Asiago, Hamburg survey). Here one can also think of taking advantage of planned spectroscopic programmes which will provide data useful to the GAIA project. + Propose a large programme (at ESO for example) with results to be made publicly available. Building a homogeneous database of observed spectra for stars across the HRD Would be an important contribution in its own right. + Consider getting funding for acquiring a large amount of observing time on an existing facility or for buying/building our own instrument. This would be advantageous in terms of the homogeneity of the data and data reduction procedures.