Astron. Astrophys. 363, 517-525 (2000)

4. The calibration sample

To further constrain the evolutionary parameters in these models, in particular n and , a calibration sample of HDF galaxies with available spectroscopic redshifts is compiled. The calibration sample contains galaxies bright enough to allow spectroscopic redshift measurements, consisting of both objects with UBVI detections and the UV `drop outs' (i.e. high redshift objects). The advantage of using the HDF galaxies as the calibration sample here is two-fold: (a) HDF galaxies cover the spectral range required for photometric redshift measurement, containing galaxies in the range ; (b) they are selected using a uniform criteria, for both nearby and distant galaxies, making this an unbiased calibration sample.

The calibrating galaxies are individually inspected and for galaxies with close neighbours, a smaller aperture is adopted to avoid contamination of their light by nearby objects. For the objects with no UV detection (i.e. UV `drop outs'), a U-band magnitude of 28.01 mag. was assumed for the photometric redshift measurement. Although this is likely to introduce a bias due to a colour-magnitude relation (i.e. fainter galaxies are bluer), we do not expect it to be significant. This is examined by exploring the range for each of the galaxies in the UV- drop out sample and estimating their respective photometric redshift. Using models with the same parameters, we find, on average, only a small sensitivity of the estimated photometric redshifts () to changes in the UV magnitudes in the above range.

The calibrating sample, consisting of 73 galaxies, is listed in Table 1 together with their UBVI photometry and spectroscopic redshifts. The reliability of the spectroscopic redshifts and the photometric accuracy of individual galaxies are discussed in the footnote to this table. The calibration sample in Table 1 was selected to be in the HDF area, to allow accurate UBVI photometry and to have unambigious spectroscopic redshifts. Moreover, objects with non-stellar sources of energy (i.e. gravitationally lensed candidates; Zepf et al. 1996) are not included. The magnitudes are in the AB system and are measured over an aperture of 3 arcsec diameter (unless stated otherwise in the footnotes to Table 1).

The EPS models for different types of galaxies are therefore further constrained by minimizing the rms scatter between the photometric redshifts, predicted by our models, and their spectroscopic counterpart, using the calibration sample in Table 1. Thus, the final EPS models for different types of galaxies, which we define as templates (see below), have local () SEDs which match the local, observed, SEDs of the corresponding galaxy types and evolutionary properties constrained, using the calibration sample here.

© European Southern Observatory (ESO) 2000

Online publication: December 11, 2000
helpdesk.link@springer.de