Generating galaxy colour libraries for multi-colour classification

The multi-colour classification performed in the COMBO-17 and MANOS surveys relies on a comparison of measured colour data with a-priori known object libraries of colour templates. A common way of generating such colour libraries involves the preparation of spectral libraries and the subsequent conversion of spectra into colours. Here, we explain how libraries for galaxy colours are prepared. If we work with spectra from synthetic population synthesis codes, this procedure involves three phases:

A.) The preparation of a base of synthetic spectra, e.g. from the PEGASE code

B.) The assembly of a suitably structured library from these base spectra

C.) The calculation of colour templates from the spectral templates.

In the following, we explain the individual steps taken with our current software setup. You need the PEGASE package compiled on your machine, in particular the executables SSPs.exe, scenarios.exe, spectra.exe, and the compiled C program sortieren.exe, as well as the MIDAS scripts crea_PEGlib.prg, inter_550_3250.prg and apply_ex_SMC.prg for section A. For section B, you need to write your own scripts following your specific intentions and needs. Section C relies on the MIDAS context classy.ctx and specifically a script defining the filter set such as farben_all.prg available on the /photo/user/exe/MCC/-area, just like the other relevant files including the PEGASE package.

A.) The preparation of a base of synthetic spectra:

Run SSPs.exe only once to generate relevant input files for PEGASE. This step needs to be done only once and forever if you don't erase the files which are produced now and subsequently required. See the PEGASE manual for details.
Run scenarios.exe to setup a star formation history (SFH) and further parameters for your first set of spectra. You can define several different SFH setups within one call of this program, until you exit the query loop. The setups will all be stored in the same scenario file. Alternatively, you can rerun scenarios.exe any time to generate further scenario files with further specifications of alternative SFHs. In COMBO-17 we do not use PEGASE-internal dust extinction with chemically consistent curves but we apply extinction laws like external screens following an independent recipe.
Run spectra.exe to produce continuum spectra and emission-line fluxes for all SFHs defined in one scenario file. PEGASE outputs the spectral data for a given grid of time steps running from 0 Myr to 20000 Myr after the onset of the first SF.
Start a MIDAS session, and run @@ crea_PEGlibs {filename.dat} to turn the spectral data from one SFH, i.e. one PEGASE spectra output file, into a MIDAS table with spectral templates. The first column of the resulting table holds the wavelength axis :lambda (in nm) and the subsequent columns hold the f_nu spectra of all time steps calculated by PEGASE. This step also generates artificial emission lines following their calculated fluxes which are inserted and added on top of the continuum spectra.
Run @@ inter_550_3250 {Spec-table} to add further time steps in the grid via interpolation. Certain areas with too low grid density are thus filled to ensure that the colour library obtained later will not contain any holes that reduce classification performance. You have now a spectral template library on the desired grid of time steps for the desired SFH assuming no dust extinction.
If you want a library with a certain amount of dust extinction applied, you need to run @@ apply_ex_SMC {Intable} {Outtable} {exA_B} and specify the amount of extinction in terms of B-band absorption A_B in magnitudes. For the SMC law defined in this script, you can assume A_B = 4.2 * E_(B-V) such that a grid in E_(B-V)=[0.0,0.1,...0.5] is realized with a set of absorption values A_B=[0.00,0.42,...2.10]. You can easily modify the extinction law by adjusting the parameters describing the three optical-UV components of the extinction curve. Pei (1992) tabulates appropriate parameters for the SMC, LMC and Milky Way, but a 3-paramter family of extinction laws can be defined by varying the parameters as you wish.

B.) The assembly of a suitably structured library from these base spectra:

You now want to define a spectral library with precisely the SFHs, extinctions, composite populations and grid structure desired for the multi-colour classification. If you want to allow for the estimation of library parameters besides redshift, then you need to give your library a regular grid structure in these parameters (see MCCLASS/LIBRARY). You can also generate composite populations from your base spectra by just adding spectra of different age in some weighted form, although a consistent treatment of metallicity would require to define a complex star formation history and rerun PEGASE. However, the first option may be simpler.

C.) The calculation of colour templates from the spectral templates:

You can now turn your spectral template library into a colour template library using MCCLASS/LIBRARY

Back to COMBO-17 index page

Last update July 20, 2005, CW