----------------------------------------------------------------------- Proposed AVO SRM: Galaxy Formation and Evolution from z = 10 to z = 0.1 Andrea Cimatti & Dave de Young - 18 October 2004 Scientific Justification The recent detection of objects at redshifts near 6.5 (and perhaps even as high as z ~ 7 - 10), together with the the impressive success of the Cold Dark Matter (CDM) scenario to account for the CMB power spectrum, have led to statements that Cosmology is now a "solved" problem. However, one of the most crucial and profound astrophysical problems of our time remains unresolved. This is the issue of the formation and evolution of galaxies from the earliest epochs of the Universe to the present day. An understanding of this phenomenon will provide insight into the formation of the first stars, the construction of galaxies as a function of cosmic epoch and the origins of metals in the IGM. It will also provide knowledge of how galaxy morphology changes with epoch and how star formation is enhanced or reduced as a function of dynamical, chemical, and environmental factors as they change within an evolving universe. It is the objective of this proposal to outline a means by which this problem can be addressed, using techniques, capabilities, and resources that are only now becoming available through Virtual Observatory efforts worldwide. More specifically, the current hierarchical Lambda-CDM model predicts that the most massive galaxies are formed at late times via a long process of the merging of smaller galaxies. The early work of Madau and subsequent efforts that suggests a peak in star formation and the occurrence of active galaxies around z ~ 1-3 with little subsequent evolution has provided impetus to this view. However, the data, though sparse, do not provide strong support for this simple model. In particular the high redshift passive elliptical galaxies and the large number of lower mass systems required at large redshift have not yet been found. In addition there is no evidence for the very high redshift "first light" population of objects that are needed in very large numbers to support this model. It seems clear that much more data at rather small redshift intervals from a few tenths to z ~ 8 is needed to examine these questions. Unraveling this mystery of the earliest formation of stars and galaxies and their subsequent evolution is a problem of the highest priority, and it is clear that very large datasets, organized in a manner that make them accessible to a wide segment of the worldwide astronomical community, is essential to developing a solution to this problem. In particular, an organized effort in the form of a Science Reference Mission for the AVO could address the following key scientific questions: - When did the first objects form ? - What are the progenitors of the present-day massive ellipticals ? - What type of galaxies populate the Universe at z >1, 2, and 4 ? - How many massive galaxies are already in place at z >1, 2 and 4 ? - How do the star formation and galaxy stellar mass densities evolve ? - What is the history of metallicity of the IGM at z > 1, 2, and 4? ----------------------------------------------------------------------- THE REQUIRED DATA The ideal dataset for this kind of studies is represented by a deep multi-wavelength coverage of a significant sky area such as the GOODS and the COSMOS fields. The existing database of the GOODS-South (CDFS) field can be taken again as an ideal laboratory for training and to optimize the use of AVO. The GOODS-South (150 arcmin^2) public database includes: - HST+ACS images in bviz bands (deeper in the Ultra Deep Field) - VLT and 2.2m images in UBVRIZJHK bands - optical spectroscopy taken in a variety of surveys - VLA and AT radio data (See note below) - Chandra and XMM-Newton X-ray data - Spitzer Mid-IR imaging (3.6-24mum) - Future submm imaging (APEX ? SCUBA ?) - Future GALEX UV imaging Another valuable dataset that could be used is the SDSS, where DR3 has just been released. Most images and catalogs are already available at STScI and ESO web sites. NB: There is an urgent need for radio imaging datasets with good resolution. The current data "archives" available at NRAO and other radio facilities is in the form of visbility data that are virtually unusable to the vast majority of astronomers. VO organizations need to urge radioastronomy facilities to rectify this situation. ----------------------------------------------------------------------- THE UNIQUE CAPABILITIES PROVIDED BY THE VO Until very recently the joint use of such multi-wavelength datasets required very specialized expertise in both the knowledge of data characteristics from x-ray to radio and in the reduction of that data. Thus combining such datasets required the coordinated effort of many different persons and consumed a great deal of time and funds. The net result was that such efforts have been very rare, and the astrophysical insights made possible via multiwavelength studies have not been realized. The Virtual Observatory concept completely revolutionizes this situation because the VO framework will enable easy access to datasets at all wavelengths for all astronomers. The tools and capabilities to do this are already in place or under development at VO sites around the world. This unique software capability, when coupled with the extremely large datasets now becoming available, provides an astrophysical problem solving ability that has never been possible in the history of astronomy. Only now, and only through the VO, are the datasets large enough and the software and network tools mature enough that the Formation and Evolution of Galaxies can be examined in a meaningful way. The VO now makes it possible to determine The History of Light. SOME POSSIBILITIES FOR AVO The key aim of AVO should be to provide output results **scientifically usable and reliable**, and not only to act as a simple "cross-correlator" of datasets. Obviously, some human intervention will be needed to develop the AVO tools and to test their results. The main steps required by the AVO tools can be (from a purely scientific point of view) summarized as follows: - Select the band where the sample is to be extracted - Extract the sample from the whole image or from a sub-section of it, and do photometry with appropriate tools (e.g. the usual SExtractor for optical/near-IR, or simply selecting the objects from an already existing catalog). This step will require inputs from the users to optimize the extraction and photometry parameters. - Cross-correlate the extracted sample with the existing images, catalogs and spectra. This is one of the crucial steps. In order to build reliable and scientifically usable multi-band catalogs and SEDs, this step should include: matching of the PSFs of the used images to the PSF of the first image where the sample was originally extracted, use of consistent photometric apertures, correct treatment of the noise and estimate of upper limits for undetected objects, sanity checks such as comparing the observed and predicted colors of stars. These steps are critical for optical/near-IR/mid-IR images, but less for data characterized by a larger PSF (e.g. radio, submm, X-ray data). Correction for Galactic extinction should also be included here. -- The output of the last step should be in the form of: (i) a multi-band catalog with positions, fluxes, flux errors, upper limits, and available spectra (if any) (ii) color-color diagrams defined by the user. I strongly suggest to make the color-color diagrams *CLICKABLE* on individual plotted objects or to have the possibility to mouse-select regions of the plots. -- Photometric redshifts: if the SEDs are built accurately, the user should have the possibility to run photo-z code(s) (e.g. Hyper-z) on all sample objects or only on specific objects selected from the color-color diagrams. This step will require some inputs and iterations from the users to optimize the code parameters, to choose the SED template library, the preferred dust extinction curve, etc etc... -- Physical parameters: photo-z codes can also provide as outputs some physical quantities which characterize the properties of the selected galaxies such as Luminosity, E(B-V), SFR, M/L ratio and hence stellar mass. -- It would be absolutely great to have on-line libraries of synthetic spectra (e.g. Bruzual & Charlot, Jimenez, PEGASE, ...) with a wide range of ages, metallicities and star formation histories and IMFs in order to allow the user also to perform a comparison/fitting between the available spectra for the objects in the sample and the synthetic spectra. -- It would be great also to have the chance to ask for morphological information when HST imaging is available and to have software which computes the CAS parameters (Abraham et al., Conselice et al.) for faint galaxies, and public software (e.g. GALFIT, GIM2D, GASPHOT) which fits the surface brightness profiles for objects bright enough. -- Another extremely useful performance would be the possibility to produce multi-band thumbnail images to visualize and check how the selected objects look like in all selected images. -- The possibility to "stack" images at the same or at different wavelengths, or spectra at different redshifts would be extremely valuable in order to search for ultra-faint objects (e.g. visible only in J+H+K bands), to confirm their absence of flux in some bands (e.g. to confirm no flux in U+B+V+R+I bands for I-band dropouts), or even to build "average" spectra for specific classes of objects. -- It shouldn't be impossible also to implement some routines which perform an analysis of the angular clustering of the selected sources in the selected field (e.g. the w(theta) estimator of Landy & Szalay 1993). -- The availability of theoretical simulations from models of galaxy formations, such as "mock catalogs", would make this game even more promising as the user will have a chance to quickly compare the observations with the model predictions. Most theory groups are already used to make the simulations publicly available on the web (VIRGO, GALICS, ...). Thus it's a question to make them readable and usable in the AVO context. -----------------------------------------------------------------------