I. Apples to apples $A^2$: realistic galaxy simulated catalogs and photometric redshift predictions for next-generation surveys

Abstract: We present new mock catalogues for two of the largest stage-IV next-generation surveys in the optical and infrared: LSST and Euclid, based on an N-body simulation+semi-analytical cone with a posterior modification with \texttt{PhotReal}. This technique modifies the original photometry by using an empirical library of spectral templates to make it more realistic. The reliability of the catalogues is confirmed by comparing the obtained color-magnitude relation, the luminosity and mass function and the angular correlation function with those of real data. Consistent comparisons between the expected photometric redshifts for different surveys are also provided. Very deep near infrared surveys such as Euclid will provide very good performance ($\Delta z/(1+z) \sim 0.025-0.053$) down to $H\sim24$ AB mag and up to $z\sim3$ depending on the optical observations available from the ground whereas extremely deep optical surveys such as LSST will obtain an overall lower photometric redshift resolution ($\Delta z/(1+z) \sim 0.045$) down to $i\sim27.5$ AB mag, being considerably improved ($\Delta z/(1+z) \sim 0.035$) if we restrict the sample down to i$\sim$24 AB mag. Those numbers can be substantially upgraded by selecting a subsample of galaxies with the best quality photometric redshifts. We finally discuss the impact that these surveys will have for the community in terms of photometric redshift legacy. This is the first of a series of papers where we set a framework for comparability between mock catalogues and observations with a particular focus on cluster surveys. The Euclid and LSST mocks are made publicly available in the following link: http://photmocks.obspm.fr/.