AgDemoDec2004 Extra Galactic

Overview

A general technique used in determinibg the distances to faint objects at high redshift from optical image surveys, is to use a technique known as 'photometric redshift determination' to estimate object redshifts based on a match of their observed spectral energy distribution with assumed galaxy energy distributions. The knowledge of only a small number of optical broadband fluxes can provide usable redshift estimates, in a statistical sense. A number of statistical techniques have been developed, and are in use today.

This demo will provide a powerful demonstration of the Virtual Observatory system through means of automating the discovery and manipulation of image data, through source extraction and catalogue cross federation, with the input data being fed to a range of redshift determinators. The output to the scientist will be a range of statistical redshifts for the object set in the input target field data.

Methodology

This case will demonstrate the discovery of optical image data from a deep public access survey. The data over a user specified area will be retrieved using standard protocols, and delivered to the user Myspace. The data files will be run through a source extraction application (SExtractor). The outputs from the multicolor images will be cross correlated to form an input catalogue. This will be fed to a number of redshift determinators.

Applications

Data Sets

Preliminary IDHA trees for AVO-Aladin access to ELAIS N1 cutouts idha-tree_ELAIS_Spitzer.xml, idha-tree_ELAIS_INT-WFS_1-2.xml, idha-tree_ELAIS_INT-WFS_3-4.xml.

Spitzer/matching INT info added by AnitaRichards - 01 Dec 2004

Work Steps

1. RA, Dec query of the WFS data image files - use the SIAP interface
   • e.g. for the ELAIS N1 field, take RA=245.0985 , Dec=53.5 , search radius = 0.05 deg
   • so from http://www.ast.cam.ac.uk/~wfcsur/status/wfsj1610.html can select wfsj1610+5430_020 at 16 08 12.4 +53 30 00
   • or try this SIAP call is http://archive.ast.cam.ac.uk/cgi-bin/wfs-siap/queryImage?POS=244.759158,53.833325&SIZE=0.0001 and this returns this votable file siap.out.vot
     • BUT - a gotcha - this returns images from pre and post 1999 - check the equinox: if 2000 it is a file with image size of 4100 pixels in y, if J2000.0 it is 4096 pixels. therefore some of these images have a different size than the others which foxes sextractor.
   • solution for the demo - choose a field where all images are from post 1999 - see http://www.ast.cam.ac.uk/~eglez/avo/workflow_sex_simple.html for details of such a field - http://archive.ast.cam.ac.uk/cgi-bin/wfs-siap/queryImage?POS=240.94,55.10&SIZE=0.1
2. This will return ugriz image files for each patch (one file is 11x22 arcins^2)
   • these files need to be returned to the user MySpace - how do we know which is the correct filter image?
     • to get round this we'll choose one pointing which has one set of UgriZ data, all taken after 1999, so of the same image size. We then have 20 files, one image for each CCD, times the five filters. The list of these files is wfslist.vot
3. For each colour image
   • run SExtractor - with the r band image as the reference image for detection

1. This returns for each field - 5 source lists
2. Cross federate each source list into the cross match tool
3. Output of this is one file: RA, Dec, 5 colours
4. Input this file to:
   • hyperz
   • bpz
5. outputs give redshifts for each object
6. cross federate the two output files to generate a master list of object, fluxes, redshifts (from the two techniques)
7. As an example the input WFS image can be loaded into Aladin-AVO (version 1.106 handles the WFS ZPN WCS system correctly). The output file from hyperz or bpz can be overlaid on top of this to show sources with catalogue overlays.