Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.adass.org/adass/proceedings/adass00/reprints/P2-48.pdf Дата изменения: Wed May 30 01:32:27 2001 Дата индексирования: Tue Oct 2 11:24:26 2012 Кодировка: Поисковые слова: trees

Astronomical Data Analysis Software and Systems X ASP Conference Series, Vol. 238, 2001 F. R. Harnden Jr., F. A. Primini, and H. E. Payne, eds.

An Extragalactic Point Source Simulator for SIRTF Pip eline Testing
Sonali Kolhatkar, Fan Fang, Mehrdad Moshir SIRTF Science Center, California Institute of Technology, Pasadena, CA 91125 Abstract. We designed and developed a package which simulates extragalactic point source data within a given field of view and SIRTF band. Our method uses input empirical relations, such as known luminosity functions and galaxy spectral energy distribution templates. It also uses instrument filter information to produce a list of sources, their redshifts, positions, and fluxes. The output fluxes in the specified band are based on the user-specified luminosity and redshift evolution modes. We designed the simulator so that the empirical input and filter information can be easily modified. The simulator is modular and extensible to enable multi-wavelength simulations, has been optimized by utilizing parallel processing capabilities, and will be used to test pipeline software.

1.

Design

The extragalactic simulator has a modular, linear design and consists of six stand-alone C modules. These modules are represented in Figure 1 as (red) boxes labeled: in (1), "GENERATE...", "EVOLVE", and "DISTRIBUTION...", in (2), "GENERATE FLUXES", and in (3), "POSITION..." . There are several additional input tables of data and parameters which are bound to the modules via a single C-shell script. The design of the simulator is well-suited to future extensions. Each module can be run separately at the command line for testing and debugging purposes and also produces its own output and log files. All user-specified parameters reside in a separate file that can easily be up-to-dated with new inputs (e.g., SED templates, filter responses, etc.). The simulator has been streamlined for high-speed processing and is capable of generating tens of thousands of extragalactic sources in minutes. This is enabled by splitting the most time consuming process over several processors (currently up to 20) and running them in parallel. Users can simply list available machines and run processes in the background of each machine. 2. Algorithm

In the first stage of the simulator, a set of extragalactic sources is generated for a given set of input parameters. The simulated sources fall within a user-specified area on the sky and have three basic attributes--see part (1) of Figure 1:

329 c Copyright 2001 Astronomical Society of the Pacific. All rights reserved.


330

Kolhatkar, Fang, and Moshir

FIELD OF VIEW
The user specifies the center of the simulation FOV in RA and Dec, and the x and y dimensions of the FOV.

GENERATE NUMBER OF SOURCES
This module calculates the number of sources that fall within the FOV.

(1) generating sources

INTRINSIC LUMINOSITIES

GENERATE LUMINOSITIES
This module uses observed luminosity functions of galaxies to generate ngal luminosities for the simulation, at lambda microns

NUMBER OF SOURCES
ngal

WAVELENGTH
A user-specified wavelength, lambda, (microns) in which the simulation is carried out.

EVOLUTION INDEX
User-specified cosmological parameters for the simulation to abide by.

EVOLVE
This module generates ngal redshifts and adjusts the intrinsic luminosities according to redshift

REDSHIFTS AND LUMINOSITIES

GALAXY TYPE DISTRIBUTIONS
A database of observed galaxy type distributions, per wavelength interval. Eg. From 7.0-7.5 microns, 25% of galaxies are dominated by the starburst galaxy type.

DISTRIBUTIONS OF GALAXY TYPES
This module generates the relative proportions of galaxy types associated with our sample of redshifts and luminosities. Eg. A starburst-dominant galaxy could consist of 30% photodissociation region (PDR), 55% starburst (SB) and 15% Active Galactic Nucleus (AGN)

SED TEMPLATES
Spectral Energy Distribution (SED) template files for the three galaxy types: PDR, SB, AGN.

SIMULATED GALAXY TYPE DISTRIBUTIONS

(2) "observing" fluxes

CPU LIST
A user-specified list of available machines for parallel processing capabilities.

GENERATE FLUXES
This module creates composite SEDs for the simulated galaxies, and uses them in conjunction with the appropriate filter, to deduce "observed" fluxes.

FILTER RESPONSIVITIES
Tables of responsivity functions for all filters through which simulations will be performed.

GALAXY FLUXES

(3) positioning sources
POSITION GENERATOR
This module generates correlated positions for the simulated sources, according to a "dumb-bell" model.

FINAL OUTPUT TABLE SOURCE POSITIONS
This contains the 2dimensional pixel positions of the sources, as well as the RA and Dec, the fluxes, and the redshift of the simulated galaxies.

Figure 1. Schematic diagram of the extragalactic simulator illustrating its three main stages; see text. (In the on-line version of this article, blue boxes represent data input modules; red, data processing modules; and green, output modules.)


Point Source Simulator for SIRTF Pipeline Testing

331

-2 log(# galaxies per bin)

-4

-6

-8 7

8

9 10 log(L/Lsolar)

11

Figure 2. Distribution of simulated 8 µm luminosities. Overlaid smooth line is the parent 12 µm function from which luminosities are drawn. · Intrinsic luminosities are chosen on the basis of an appropriate luminosity function (see Figure 2). · Redshifts are derived independently for each galaxy using a (1 + z )- distribution, appropriate for a co-moving density evolution. · Galaxy types are determined from intrinsic luminosities and the band of interest and emission is modeled as a linear combination of the average emission from "Photo-dissociation Regions (PDR)", "Star Bursts (SB)" and "Active Galactic Nuclei (AGN)", determined from empirical data. In the second stage of the simulator, available Spectral Energy Distribution (SED) templates for PDR, SB, and AGN galaxy types are used--see part (2) of Figure 1--to construct individual composite SEDs based upon the proportions of types calculated in the previous stage. With the filter response function of the specified SIRTF wavelength, the "integrated luminosities" of the galaxies are calculated by multiplying by the normalized SEDs and integrating to determine flux densities from the following relation: Fl uxBand = Lint /4D2 where Lint is the integrated luminosity, and D is the luminosity-distance of the galaxy. In the last stage of the extragalactic simulator, galaxies are placed on the 2dimensional sky--see part (3) of Figure 1--with correlated positions calculated according to a dumb-bell model of fractal structure (Soneira & Peebles 1978). The final output of the simulator contains luminosities, redshifts, fluxes, and positions in Right Ascension and Declination of the simulated galaxies. 3. Discussion

The extragalactic simulator has been tested for four filters on the SIRTF IRAC instrument: 3.6 µm, 4.5 µm, 5.8 µm, and 8.0 µm. The table look-up method used to generate the luminosities of the simulated galaxies works well at all four wavelengths (see Figure 2).


332

Kolhatkar, Fang, and Moshir

Figure 3. vs. flux.

Log plot of number counts of 34,557 extragalactic sources

The extragalactic simulator is capable of generating large numbers of galaxies within a given patch of sky for SIRTF pipeline testing. Tens of thousands of galaxies can be simulated within minutes and further techniques to speed up the process are under consideration. This includes "table look-up" methods in the flux calculation module that could replace calculation of individual fluxes. When available, luminosity functions can be replaced by newer ones recorded at wavelengths close to SIRTF instrument wavelengths. SED templates can also be replaced by newer, more accurate ones when available. In order to simulate an observation of the real sky, this simulator will soon work in conjunction with the SKY model galactic simulator (Cohen 1993) to produce data sets that contain external galaxies as well as ob jects within our own. A common test of cosmological models is a log N - log S plot of galaxies. An effective test requires a complete sample of galaxies out to relatively high redshifts, and Figure 3 shows a plot of log N versus log S for 34,557 simulated extra-galactic sources with redshifts in the range z =0 to 5, "observed" using the IRAC 8.0 µm filter. The extragalactic simulator produces galaxies with fluxes that are at the limit of IRAC sensitivity ( 1 µJ y ). References Cohen, M. 1993, AJ, 105, 1860 Soneira, R. M. & Peebles, P. J. E. 1978, AJ, 83, 845