Estimating beta from redshift-space distortions in the 2dF galaxy survey

Given the failure of existing models for redshift-space distortions to provide a highly accurate measure of the beta-parameter, and the ability of forthcoming surveys to obtain data with very low random errors, it becomes necessary to develop better models for these distortions. Here we review the failures of the commonly-used velocity dispersion models and present an empirical method for extracting beta from the quadrupole statistic that has little systematic offset over a wide range of beta and cosmologies. This empirical model is then applied to an ensemble of mock 2dF southern strip surveys, to illustrate the technique and see how accurately we can recover the true value of beta. We compare this treatment with the error we expect to find due only to the finite volume of the survey. We find that non-linear effects reduce the range of scales over which beta can be fitted, and introduce covariances between nearby modes in addition to those introduced by the convolution with the survey window function. The result is that we are only able to constrain beta to a 1-sigma accuracy of 25% (beta=0.55 +/- 0.14 for the cosmological model considered). We explore one possible means of reducing this error, that of cluster collapse, and show that accurate application of this method can greatly reduce the effect of non-linearities, improving the determination of beta. We conclude by demonstrating that, when the contaminating effects of clusters are dealt with, this simple analysis of the full 2dF survey yields beta=0.55 +/- 0.04. For this model this represents a determination of beta to an accuracy of 8% and hence an important constraint on the cosmological density parameter Omega_0.