2017A&A...602A..96S

This paper introducesEGG, the Empirical Galaxy Generator, a tool designed within the ASTRODEEP collaboration to generate mock galaxy catalogs for deep fields with realistic fluxes and simple morphologies. The simulation procedure is based exclusively on empirical prescriptions - rather than first principles - to provide the most accurate match with current observations at 0<z<7. We considered that galaxies can be either quiescent or star-forming, and used their stellar mass (M_*) and redshift (z) as the fundamental properties from which all the other observables can be statistically derived. Drawing z and M_* from the observed galaxy stellar mass functions, a star-formation rate (SFR) is attributed to each galaxy from the tight SFR-M_* main sequence, while dust attenuation, optical colors and simple disk plus bulge morphologies are obtained from empirical relations that we established from the high quality Hubble and Herschel observations from the CANDELS fields. Random scatter was introduced in each step to reproduce the observed distributions of each parameter. Based on these observables, an adequate panchromatic spectral energy distribution (SED) is selected for each galaxy and synthetic photometry is produced by integrating the redshifted SED in common broad-band filters. Finally, the mock galaxies are placed on the sky at random positions with a fixed angular two-point correlation function to implement basic clustering. The resulting flux catalogs reproduce accurately the observed number counts in all broad bands from the ultraviolet up to the sub-millimeter, and can be directly fed to image simulators such as SkyMaker. The images can then be used to test source extraction softwares and image-based techniques such as stacking.EGG is open-source, and is made available to the community on behalf of the ASTRODEEP collaboration, together with a set of pre-generated catalogs and images.