Calibrating photometric redshift distributions with cross-correlations

TL;DR

This paper explores a cross-correlation method to calibrate photometric redshift distributions in large galaxy surveys, aiming to reduce the need for extensive spectroscopic follow-up by analyzing mock data and identifying key limitations.

Contribution

It develops and tests a pipeline for redshift distribution calibration using cross-correlations, highlighting its effectiveness and current limitations.

Findings

The method is effective but limited by noise in correlation function measurements.

Disentangling redshift distribution from bias evolution remains challenging.

Further refinement could reduce spectroscopic follow-up requirements.

Abstract

The next generation of proposed galaxy surveys will increase the number of galaxies with photometric redshifts by two orders of magnitude, drastically expanding both redshift range and detection threshold from the current state of the art. Obtaining spectra for a fair sub-sample of this new data could be cumbersome and expensive. However, adequate calibration of the true redshift distribution of galaxies is vital to tapping the potential of these surveys. We examine a promising alternative to direct spectroscopic follow up: calibration of the redshift distribution of photometric galaxies via cross-correlation with an overlapping spectroscopic survey whose members trace the same density field. We review the theory, develop a pipeline, apply it to mock data from N-body simulations, and examine the properties of this redshift distribution estimator. We demonstrate that the method is effective, but the estimator is weakened by two factors. 1) The correlation function of the spectroscopic sample must be measured in many bins along the line of sight, rendering it noisy and interfering with high quality reconstruction of the photometric redshift distribution. 2) The method is not able to disentangle the photometric redshift distribution from evolution in the bias of the photometric sample. We establish the impact of these factors using our mock catalogs. Although it may still be necessary to spectroscopically follow up a fair subsample of the photometric survey data, further refinement may appreciably decrease the number of spectra that will be needed to calibrate future surveys.