Accuracy of photometric redshifts for future weak lensing surveys from space

Photometric redshifts are a key tool to extract as much information as possible from planned cosmic shear experiments. In this work we aim to test the performances that can be achieved with observations in the near-infrared from space and in the optical from the ground. This is done by performing realistic simulations of multiband observations of a patch of the sky, and submitting these mock images to software usually applied to real images to extract the photometry and then a redshift estimate for each galaxy. In this way we mimic the most relevant sources of uncertainty present in real data analysis, including blending and light pollution between galaxies. As an example, we adopt the infrared setup of the European Space Agency proposed Euclid mission, while we simulate different observations in the optical, modifying filters, exposure times and seeing values. Finally, we consider directly some future ground-based experiments, such as Large Synoptic Survey Telescope (LSST), Pan-STARRS and Dark Energy Survey (DES). The results highlight the importance of u-band observations, especially to discriminate between low (z≲ 0.5) and high (z˜ 3) redshifts, and the need for good observing sites, with seeing FWHM < 1 arcsec. The former of these indications clearly favours the LSST experiment as a counterpart for space observations, while for the other experiments we need to exclude at least 15 per cent of the galaxies to reach a precision in the photo-z values equal to ?.