作者: Christopher Bonnett , MA Troxel , William Hartley , Adam Amara , Boris Leistedt
DOI: 10.1103/PHYSREVD.94.042005
关键词:
摘要: We present photometric redshift estimates for galaxies used in the weak lensing analysis of Dark Energy Survey Science Verification (DES SV) data. Four model-or machine learning-based methods-ANNZ2, BPZ calibrated against BCC-Ufig simulations, SKYNET, and TPZ-are analyzed. For training, calibration, testing these methods, we construct a catalogue spectroscopically confirmed matched DES SV The performance methods is evaluated spectroscopic catalogue, focusing on metrics relevant analyses, with additional validation COSMOS photo-z's. From shear which have mean 0.72 +/- 0.01 over range 0.3 < z 1.3, three tomographic bins means = {0.45;0.67;1.00}. These each systematic uncertainties delta <= 0.05 fiducial SKYNET photo-z (dz). propagate errors distributions through to their impact cosmological parameters estimated cosmic shear, find that they cause shifts value sigma(8) approximately 3%. This shift within one sigma statistical catalogue. further study potential differences critical surface density, Sigma(crit), finding levels bias safely less than power recommend final Gaussian prior n(z) width bins, show this sufficient model corresponding cosmology analysis.