Mitigating the impact of fiber assignment on clustering measurements from deep galaxy redshift surveys

Abstract

We examine the impact of fiber assignment on clustering measurements from fiber-fed spectroscopic galaxy surveys. We identify new effects which were absent in previous, relatively shallow galaxy surveys such as Baryon Oscillation Spectroscopic Survey (BOSS). Specifically, we consider deep surveys covering a wide redshift range such as 0.6≤ z≤ 2.4, as in the Subaru Prime Focus Spectrograph (PFS) survey. Such surveys will have more target galaxies than we can place fibers on. This leads to two effects. First, it eliminates fluctuations with wavelengths longer than the size of the field of view, as the number of observed galaxies per field is nearly fixed to the number of available fibers. We find that we can recover the long-wavelength fluctuation by weighting galaxies in each field by the number of target galaxies. Second, it makes the preferential selection of galaxies in under-dense regions. We mitigate this effect by weighting galaxies using the so-called "individual inverse probability". Correcting these two effects, we recover the underlying correlation function at better than 1% accuracy on scales greater than 10 h-1 Mpc.