Estimating the large-scale angular power spectrum in the presence of systematics: a case study of Sloan Digital Sky Survey quasars

The angular power spectrum is a powerful statistic for analysing cosmological signals imprinted in the clustering of matter. However, current galaxy and quasar surveys cover limited portions of the sky, and are contaminated by systematics that can mimic cosmological signatures and jeopardise the interpretation of the measured power spectra. We provide a framework for obtaining unbiased estimates of the angular power spectra of large-scale structure surveys at the largest scales using quadratic estimators. The method is tested by analysing the 600 CMASS mock catalogues constructed by Manera et al. (2013) for the Baryon Oscillation Spectroscopic Survey (BOSS). We then consider the Richards et al. (2009) catalogue of photometric quasars from the Sixth Data Release (DR6) of the Sloan Digital Sky Survey (SDSS), which is known to include significant stellar contamination and systematic uncertainties. Focusing on the sample of ultraviolet-excess (UVX) sources, we show that the excess clustering power present on the largest-scales can be largely mitigated by making use of improved sky masks and projecting out the modes corresponding to the principal systematics. In particular, we find that the sample of objects with photometric redshift $1.3 < z_p < 2.2$ exhibits no evidence of contamination when using our most conservative mask and mode projection. This indicates that any residual systematics are well within the statistical uncertainties. We conclude that, using our approach, this sample can be used for cosmological studies.