Paper detail

Impact of our local environment on cosmological statistics

We conduct a thorough investigation into the possibility that residing in an overdense region of the Universe may induce bias in measurements of the large-scale structure. We compute the conditional correlation function and angular power spectrum of density and lensing fluctuations while holding the local spherically averaged density fixed and show that for Gaussian fields this has no effect on the angular power at $l>0$. We identify a range of scales where a perturbative approach allows analytic progress to be made, and we compute leading-order conditional power spectra using an Edgeworth expansion and second-order perturbation theory. We find no evidence for any significant bias to cosmological power spectra from our local density contrast. We show that when smoothed over a large region around the observer, conditioning on the local density typically affects density power spectra by less than a percent at cosmological distances, below cosmic variance. We find that while typical corrections to the lensing angular power spectrum can be at the 10% level on the largest angular scales and for source redshifts $z_s \lesssim 0.1$, for the typical redshifts targeted by upcoming wide imaging surveys the corrections are sub-percent and negligible, in contrast to previous claims in the literature. Using an estimate of the local spherically averaged density from a composite galaxy redshift catalogue we find that the corrections from conditioning on our own local density are below cosmic variance and subdominant to other non-linear effects. We discuss the potential implications of our results for cosmology and point out that a measurement of the local density contrast may be used as a consistency test of cosmological models.