A NEW STATISTIC FOR ANALYZING BARYON ACOUSTIC OSCILLATIONS

We introduce a new statistic w(l)(r(s)) for measuring and analyzing large-scale structure and particularly the baryon acoustic oscillations. w(l)(r(s)) is a band-filtered, configuration space statistic that is easily implemented and has advantages over the traditional power spectrum and correlation function estimators. Unlike these estimators, w(l)(r(s)) can localize most of the acoustic information into a single dip at the acoustic scale while avoiding sensitivity to the poorly constrained large-scale power (i.e., the integral constraint) through the use of a localized and compensated filter. It is also sensitive to anisotropic clustering through pair counting and does not require any binning of data. We measure the shift in the acoustic peak due to nonlinear effects using the monopole w(0)(r(s)) derived from subsampled dark matter (DM) catalogs as well as from mock galaxy catalogs created via halo occupation distribution modeling. All of these are drawn from 44 realizations of 1024(3) particle DM simulations in a 1 h(-1) Gpc box at z = 1. We compare these shifts with those obtained from the power spectrum and conclude that the results agree. We therefore expect that distance measurements obtained from w(0)(r(s)) and P(k) will be consistent with each other. We also show that it is possible to extract the same amount of acoustic information by fitting over a finite range using either w(0)(r(s)) or P(k) derived from equal volume surveys.