EUROVO FC

AVO
SWG
Management
Science
Standards
Technology
Reports&Minutes
Forum
Events
Contacts
Help

EGG

Main

Know

Test

TWiki

Demo Science

A draft of the science for HDF(N) Demo

GOODS - see ESO Messenger article for excellent intro, at non-specialist astronomer level.

Here is an attempt at an overview of how the demo might appeal to the audiences, firstly the SWG and other astronomers, secondly the media. It is very incomplete and needs input on:

1. Suitable images
2. Technical aspects (faster bigger ....)
3. Diversity/international nature of data
4. Science drivers from anyone who knows anything about galaxies, cosmology or wavelengths < 1cm.

Astronomers' view

• The HST observed the HDF(N) and HFF using WFPC2 in 1995. Since then observations have been made by
  • CFHT and other ground-based telescopes (wider fields and/or better spectral coverage)
  • VLA+MERLIN (and other radiotelescopes - EVN, WSRT)
  • CHANDRA
  • SCUBA
  • ISO
  • and more focussed observations of individual objects
  • etc. what have i missed, - elaborate above.
• Southern field (CDF(S))....
  • ...
  • ...
  • VLA data

In 6 yrs since the first data were released, although ?over 100 papers have been published, there are still data to be extracted from the northern field images, and scientific analysis of the results is far from complete. The full picture - understanding changes in the galaxy population with age, locating rare or very distant objects etc., can only be achieved by using data at all frequencies, but the sheer volume of data and the many different formats and conventions make this a very slow process.

There are ~3000 objects in the 4.7 arcmin² of the central WFPC2 triple HDF(N) field, and thousands more in the 8x3 HFF fields. There are nearly 200 ISO sources, and nearly 400 CHANDRA detections. There are ~100 secure detections of radio galaxies within a 10' square, some of which have no obvious optical counterparts, or counterparts too faint to have an obvious optical morphology or redshift.

The AVO will greatly speed up the interpretation of the images, comparison with catalogues and construction of new or extended data sets. This requires harmonising units, resolutions etc. as required as well as astrometric and photometric alignment, and implies that each catalogue and image is supplied with the necessary metadata.

This will achieve new scientific results in several ways including:

• Providing access to existing catalogues of sources and properties via a single interface - as by Vizier, but allowing more operations such as evaluating colours or giving errors on position cross-matches
• Plotting properties such as SEDs from catalogues, and allowing the original image data to be selected interactively
• Providing an interface to image analyise tools in order to
  • re-extract source list according to user criteria and meaure sizes, flux densities etc.
  • measure the flux density in other images at the positions and in apaertures corresponding to the sizes of sources from the first image (thus obtaining homogenous samples and upper limits, or allowing the statistical properties of faint objects to be deduced from stacking)
  • etc.

Example science projects which may be suitable for the demo:

• North: Does the nature of radio emission change with distance?
  • The VLA observations of a region 10' square centred on the HDF(N) are complete to 40 microJy at 2" resolution; the MERLIN+VLA data are complete to 27 microJy (7sigma) at 0".2 - 0".5 resolution. The latter have only been fully analysed to data within the inner 3'x4', but no new radio sources were found. Most of the existing sources were resolved. A few retained compact bright cores and at least one, also detected at 0".025 resolution by the EVN, is at z=4.4. This almost certainly contains an AGN. The local population of bright radio galaxies is dominated by AGN, as are the only three galaxies in the 10' field brighter than 1 mJy at 1.4 GHz. However about 2/3 of the radio galaxies in the HDF(N)/HFF are starbursts. A comparable trend is observed in CHANDRA data.

    Taking the ISO source list and measuring the MERLIN+VLA flux densities in the central 3' field in a suitable aperture at each position reveals a statistically significant excess of 3-7sigma sources. Comparison with optical data shows a similar phenomena at I<24. Downes et al. found that the properties of the SCUBA detections compared with other HDF(N) observations suggests that there is a population of dust-enshrouded starbursts at z= a few. Can we find these in the radio data?

  • The MERLIN+VLA data have been imaged in a region 8'.5 square for the first time for the demo (5x the area previously mapped). It is unlikely there will be many, if any, new >7sigma detections (which would be very compact objects). However we can use catalogues derived from other images to investigate the distribution of faint radio flux (binned by query catalogue properties):
    • At the position of ISO sources;
    • At the position of CHANDRA sources;
    • At the position of optical sources, or subsets chosen for colours indicative of starburst or other nature or by z.
    • At the position of sources with a reasonably secure classification, e.g. by optical type, so that data can be stacked (as done e.g. for CHANDRA data ...), to obtain the radio end of the SED for faint sources of various types.
    • Compare flux density at a single wavelength on different scales (by using different apertures and/or comparing different telescopes e.g. VLA, MERLIN, EVN) is it more compact for objects of a particular type or at high redshift?
    NB
    • In each case it is desirable to select a resolution comparable to the query catalogue, to perform a control measurement with positions offset by ~10".
    • At present we only have the HST and MERLIN+VLA data 'SExtractor ready' so source positions and flux densities from other telescopes e.g. CHANDRA, ISO will have to be taken from catalogues.
    This will reveal whether there is a statistically significant excess of radio flux associated with other emission processes and/or with particular types of source (Starburst etc.), and if so, out to what distance. This can be compared with e.g. X-ray/optical comparisons to see if radio flux is associated with candidate dust-enshrouded starbursts.

  We need to clarify what can be wrapped as part of the demo and what might be done by SWG members etc. exporting demo results and using other packages for further processing.

• South: What are the properties of faint high-z galaxies? (based on a conversation with Richard McMahon who can expand/correct misconceptions....)
  • To date, most high-z galaxies (?>4) detected are the brightest or in some other way special (e.g. lensed).
  • The Southern field data can be used to find and classify a homogenous, less flux-limited sample.
    • Use existing SED catalogues to select potential high-z sources (from optical data?) - outliers
    • Use Aladin-based tool to inspect outliers and refine list
    • Re-extract SED from all available data (SExtractor) (at several resolutions, e.g. at full resolution for optical data, repeat at slightly lower resolution to also include CHANDRA, lowest to also include ATCA)
    • Interface with hyperZ or export SEDS to feed to hyperZ to classify galaxies and measure photometric redshift (possibly use low-res, widest wavelength coverage for classification, and then higher-res optical only for z?)
    • Compare results with catalogues for any known classifications/spectroscopic z as checks.
    • Also compare flux at given wavelength at different scales as in northern project
    NB - some considerations, probably off-line from the demo itself:
    • Correction for Malmquist bias needed (drop out of fainter objects at greater distances)? (for both projects?) And K-correction (e.g. the colour ratio indicating a starburst is itself redshifted at high z).
    • Although the limiting flux density is similar in each particular image set (in some cases some loss of sensitivity at large angular separations from pointing centre, or between fields), it is different at different wavebands, so compensation should be attempted based on model SEDs or something.

    We need to clarify what can be wrapped as part of the demo and what might be done by SWG members etc. exporting demo results and using other packages for further processing.