2025A&A...693A.102D

We present photometric redshifts for 1 341 559 galaxies from the Physics of the Accelerating Universe Survey (PAUS) over 50.38 deg² of sky to i_AB = 23. Redshift estimation was performed using DEEPz, a deep learning photometric redshift code. We analysed the photometric redshift precision when varying the photometric and spectroscopic samples. Furthermore, we examined observational and instrumental effects on the precision of the photometric redshifts, and we compared photometric redshift measurements with those obtained using a template method-fitting BCNz2. Finally, we examined the use of photometric redshifts in the identification of close galaxy pairs. We find that the combination of samples from the W1 and W3 fields in the training of DEEPz significantly enhances the precision of photometric redshifts. This also occurs when we recover narrow-band fluxes using BB measurements. We show that DEEPz determines the redshifts of galaxies in the prevailing spectroscopic catalogue used in the training of DEEPz with greater precision. For the faintest galaxies (i_AB = 21 - 23), we find that DEEPz improves over BCNz2 both in terms of the precision (20-50% smaller scatter) and in returning a smaller outlier fraction in two of the wide fields. The catalogues were tested for the identification of close galaxy pairs, showing that DEEPz is effective for the identification of close galaxy pairs for samples with i_AB < 22.5 and redshift 0.2 < z < 0.6. In addition, identifying close galaxy pairs that are common between DEEPz and BCNz2 is a promising approach for improving the accuracy of the catalogues of these systems.