HELP project - a dreamed-of multiwavelength dataset for SED fitting: the influence of used models for the main physical properties of galaxies

TL;DR

The HELP project provides a comprehensive multiwavelength galaxy dataset to test and improve spectral energy distribution fitting models, focusing on the influence of different modules on physical property estimation.

Contribution

This work introduces a large multiwavelength galaxy catalog and evaluates the impact of various modeling modules on derived galaxy properties.

Findings

Identified how different modules affect stellar mass estimates.

Developed a new procedure for selecting peculiar galaxies.

Provided a valuable dataset for future SED fitting improvements.

Abstract

The Herschel Extragalactic Legacy Project (HELP) focuses to publish an astronomical multiwavelength catalogue of millions of objects over 1300~deg$^2$ of the Herschel Space Observatory survey fields. Millions of galaxies with ultraviolet--far infrared photometry {make} HELP a perfect sample for testing spectral energy distribution fitting models, and to prepare tools for next-generation data. In the frame of HELP collaboration we estimated the main physical properties of all galaxies from the HELP database and we checked a new procedure to select peculiar galaxies from large galaxy sample and we investigated the influence of used modules for stellar mass estimation.

Tables

Table 1. Table 1: Overview of 23 fields used for HELP project.

HELP field name	number of objects	area [deg2]
AKARI-NEP	531 746	9.2
AKARI-SEP	844 172	8.7
Bootes	3 367 490	11
CDFS-SWIRE	2 171 051	13
COSMOS	2 599 374	5.1
EGS	1 412 613	3.6
ELAIS-N1	4 026 292	14
ELAIS-N2	1 783 240	9.2
ELAIS-S1	1 655 564	9.0
GAMA-09	12 937 982	62
GAMA-12	12 369 415	63
GAMA-15	14 232 880	62
HDF-N	130 679	0.67
Herschel-Stripe-82	50 196 455	363
Lockman-SWIRE	4 366 298	22
HATLAS-NGP	6 759 591	178
SA13	9 799	0.27
HATLAS-SGP	29 790 690	295
SPIRE-NEP	2 674	0.13
SSDF	12 661 903	111
xFLS	977 148	7.4
XMM-13hr	38 629	0.76
XMM-LSS	8 704 751	22
Total:	171 570 436	1270

Table 2. Table 2: Main modules and input parameters used in CIGALE for the analysis of the high-z sample. The first column lists the CIGALE model, the second provides a brief description of the main parameters, and the third one shows the range of the selected values.

CIGALE module	main parameter	description
SFH delayed + additional burst	$\tau$ of the main stellar population model [Myr]	3 000
	$\tau$ of the late starburst population model [Myr]	10 000
	mass fraction of the late burst population	0.001–0.300
SSP: [Bruzual & Charlot (2003)]	initial mass function	[Chabrier (2003)]
dust attenuation: [Charlot & Fall (2000)]	$A_V$ in the BCs	0.3–3.8
	power law slopes (BC and ISM)	-0.7
dust emission [Draine & Li (2007)]	minimum radiation field ($U_{\min }$)	5.0, 10.0, 25.0
	mass fraction of PAH	1.12, 2.5, 3.19
	power law slope dU/dM ($U^{\alpha }$)	2.0, 2.8
AGN emission: [Fritz et al. (2006)]	fractional contribution of AGN	0.0, 0.15, 0.25, 0.8

Full text

††volume: 341††journal: Title of your IAU Symposium††editors: A.C. Editor, B.D. Editor & C.E. Editor, eds.

1 Introduction

The primary objective of the Herschel Extragalactic Legacy Project (HELP project, Oliver et al., in preparation, [Vacdustatt_modified_CF00cari:2016, Vaccari 2018]) founded by FP7 European Union is to provide homogeneously calibrated multiwavelength catalogues covering roughly 1300 deg2 of the extragalactic Herschel Space Observatory surveys (HSO, [Pilbratt et al. (2010), Pilbratt et al. 2010]) at wide redshift range. Millions of galaxies with good coverage of ultraviolet–far infrared spectral range make HELP a perfect sample to prepare tools for next-generation data. The detailed description of a final master list creation of 170 million objects, selected at 0.36—4.5 $\mu$m from HSO, depth maps etc. can be found in Shirley et al., 2019 MNRAS (under review). The catalogues supported by spectroscopic (if possible) or photometric redshift ([Duncan et al. (2018), Duncan et al. 2018]) will allow for colour-colour/colour-flux analysis, multi-wavelength spectral energy distribution (SED) fitting and many more statistical studies of the low-to-intermediate redshift galaxy population formation and evolution over cosmic time.

Tab. 1 shows the list of the HSO fields used for HELP project. It demonstrates that HELP not only created a huge multiwavelength, homogenized database, but also focuses both on wide and deep fields, with different area on the sky. This careful selection and the final data product can remove the barriers to multiwavelength data studies on the statistical level.

2 Data and short overview of the method

The European Large Area ISO Survey North 1 (ELAIS N1, 13.51 deg2 area centred at 16h10m01s +54o30${}^{{}^{\prime}}$36${}^{{}^{\prime\prime}}$, [Oliver et al. (2000), Oliver et al. 2000]) was a pilot field for HELP. The HELP homogenized catalogue of ELAIS N1 includes 50 135 galaxies with good ultraviolet (UV)–far infrared (IR) measurements (quality criterion requires at least two optical – near IR measurements and at least two two of five Herschel measurements with signal to noise ratio $\geqq$2). We used the sample of 50 135 galaxies and we estimated the key physical parameters (i.e. stellar mass, star formation rate, dust luminosity) by fitting SED to all of them using Code Investigating GALaxy Emission (CIGALE, [Burgarella et al. (2005), Burgarella et al. 2005], [Noll et al. (2009), Noll et al. 2009], and [Boquien et al. (2018), Boquien et al. 2018]).

CIGALE is designed to estimate the physical parameters by comparing modelled galaxy SEDs to observed ones. CIGALE conserves the energy balance between the dust-absorbed stellar emission and its re-emission in the IR. A more detailed description of the code can be found in [Boquien et al. (2018)].

All adopted parameters used for modules are presented in Table. 2. More detailed discussion of used parameters and description of addition quality tests for SED fitting procedure for ELAIS N1 field can be found in [Małek et al. (2018)]. An exemplary fit of SED, showing typical photometric coverage of the spectra is shown in Fig. 1.

3 Impact of the dust attenuation law on the stellar mass

Based on the statistically significant sample of $\sim$50 000 galaxies we check the influence of different dust attenuation recipes on the main physical parameters calculated for all HELP galaxies; stellar mass, star formation rate and dust luminosity. We perform the SED fitting of ELAIS N1 galaxies by assuming three different dust attenuation laws separately: [Charlot & Fall (2000)], widely used in the literature [Calzetti et al. (2000)], and [Lo Faro et al. (2017)] – dust attenuation recipe created in the framework of the HELP project for Ultra Luminous Infrared Galaxies at redshift $>$ 2. This test allows us to analyze the impact of the assumed law on estimated physical parameters. We find that the attenuation law has an important impact on the stellar mass estimation (on average leading to disparities of a factor of 2), and we derived the relation between stellar mass estimates obtained by those three different attenuation laws. Found recipes (published in [Małek et al. (2018), Małel et al. 2018]) can help to homogenize estimated stellar masses from different attenuation laws, and allow to make more precise comparisons, sample selection or study of so called main sequence (stellar mass versus star formation rate relation) between different SED fitting procedures.

We check that the differences in obtained stellar masses are closely related to the shape of each attenuation law at near IR wavelengths. Fig. 2 shows relation between attenuation in near IR band and far UV band for all three attenuation laws used in our analysis. This figure presents that the range and distribution of attenuation in ultraviolet band is similar for [Charlot & Fall (2000)], [Calzetti et al. (2000)], and [Lo Faro et al. (2017)], however the attenuation obtained in near infrared band is meaningly different. Similar result, showing that Calzetti recipe leads to steeper slopes, not consistent with radiation transfer models results, was found by [Buat et al. (2018)] based on the infrared complete sample of galaxies in the COSMOS 3D-HST CANDELS field at 0.6$<$z$<$1.6. Similar impact of the attenuation law on the stellar mass was found by [Mitchell et al. (2013)] based on the semi-analytic galaxy formation model GALFORM ([Cole et al. (2000), Cole et al. 2000]).