Astron. Astrophys. 326, 950-962 (1997)

1. Introduction

The recent development of NIR observational astronomy, as well as the progress of the theory of stellar evolution (overshooting, nuclear rates, winds, opacities, see Maeder & Meynet (1989); Chiosi (1986)), notably of the evolved phases (Groenewegen & de Jong 1993; Vassiliadis & Wood 1993) up to the PAGB terminal phases (Blöcker 1995), put new constraints on the evolution of galaxies. The UV-optical range of currently star-forming galaxies is dominated by young stars, and results of models limited to this domain are unable to determine the past star formation history. To break this degeneracy, the NIR emission of the bulk of giants is worthy to be examined as a meaningful indicator of mass and age. Moreover, the NIR light is less obscured by dust than at shorter wavelengths and has proved very useful in the spectral analysis of dusty starbursts (Lançon et Rocca-Volmerange 1996). Following star formation history over a long timescale thus needs a continuity of the wavelength range. That continuity is also necessary to analyze distant galaxies at large redshift ranges in the deepest galaxy counts. From the long series of models compared by Arimoto (1996) and Charlot (1996) at the Crete meeting, most current models agree in the visible. However, none of them is published in the NIR by Leitherer et al. (1996a), possibly because they show a flux deficit from J to K in the spectral energy distribution (SED) of early-type galaxies (Arimoto 1996). The poor knowledge of the atmospheric parameters of cold stars dominating the NIR and the rapid evolution of the latest phases are evident difficulties for the modeling of NIR emission. Another difficulty is the connection of the NIR stellar emission to the visible at about 1 µm, where cold stars of 2000 to 3000 K peak. Moreover, in that domain, only a few data are available because receivers in the optical as in the NIR have minimal sensitivity.

Our main goal is to build a new atlas of evolving standard synthetic spectra for the types of the Hubble sequence. We present here UV to NIR energy distributions for 8 spectral galaxy types. Evolutionary parameters are constrained by fitting synthetic spectra at on integrated spectra and colors of nearby galaxies. The choice of templates requires a peculiar attention to aperture effects and a significant identification of Hubble types. These criteria have led us to use statistical color samples, because of the lack of significant spectral templates in the NIR wavelength range. The samples of optical and NIR data show however a high dispersion, which is partly observational, but is also due to uneven star formation histories leading to an intrinsic scatter of colors for a given morphological type. This makes it difficult to identify observational templates.

A first application of PEGASE solves the puzzling question of the bright galaxy counts. The slope at bright magnitudes is shown to be in agreement with recent multispectral observations of Gardner et al. (1996), excluding the strong evolution of giant galaxies at low redshift advocated by Maddox et al. (1990) from their steep counts. Moreover, the normalization of the luminosity function is in accordance with the high value of Marzke et al. (1994).

The structure of this paper is as follows. In section 2, we present our model, the algorithm, the stellar library and other related data, the evolutionary tracks, the nebular emission and finally the extinction model. In section  3, we examine the star formation history for starbursts and evolved galaxies and propose evolutionary scenarios fitted on UV to NIR observations of galaxies of the Hubble sequence. In section  4, we check that the evolution of our standard scenario is realistic by comparison with bright galaxy counts, and we finally discuss in the conclusion the advantages and limitations of the model.

© European Southern Observatory (ESO) 1997

Online publication: April 8, 1998
helpdesk.link@springer.de