Photometric Redshift Calibration with Self Organising Maps

Accurate photometric redshift calibration is central to the robustness of all cosmology constraints from cosmic shear surveys. Analyses of the Kilo-Degree Survey, KiDS, re-weighted training samples from all overlapping spectroscopic surveys to provide a direct redshift calibration. Using self-organising maps (SOMs) we demonstrate that this spectroscopic compilation is sufficiently complete for KiDS, representing $99\%$ of the effective 2D cosmic shear sample. We use the SOM to define a $100\%$ represented `gold' cosmic shear sample, per tomographic bin. Using mock simulations of KiDS and the spectroscopic training set, we demonstrate that the mean redshift of the `gold' sample can be recovered by the SOM with an accuracy better than $| \Delta \langle z \rangle | < 0.004$, with the exception of the $0.7 < z_B < 0.9$ tomographic bin with $ |\Delta \langle z \rangle | = 0.011$. Photometric noise, sample variance, and spectroscopic selection effects induce a combined maximal scatter of $\sigma_{\Delta \langle z \rangle} < 0.007$ in all tomographic bins. We demonstrate that the previous direct redshift calibration method applied to the full cosmic shear sample is accurate to $| \Delta \langle z \rangle | < 0.025$. We find that photometric noise dominates the calibration dispersion, and that neither sampling variance nor a realistic fraction of spectroscopic misidentifications in the training set introduce significant bias.