Have baryonic acoustic oscillations in the galaxy distribution really been measured?


  • Anna Cabré,

    Corresponding author
    1. Department of Physics and Astronomy, University of Pennsylvania, 209 South 33rd Street, Philadelphia, PA 19104-6396, USA
      E-mail: annanusca@gmail.com
    Search for more papers by this author
  • Enrique Gaztañaga

    1. Instituto de Ciencias del Espacio (IEEC/CSIC), F. de Ciencias, Torre C5- Par-2a, Bellaterra, 08193 Barcelona, Spain
    Search for more papers by this author

E-mail: annanusca@gmail.com


Recent publications claim that there is no convincing evidence for measurements of the baryonic acoustic oscillation (BAO) feature in galaxy samples using either monopole or radial information. Different claims seem contradictory: data are either not consistent with the BAO model or data are consistent with both the BAO model and featureless models without BAO. We investigate this point with a set of 216 realistic mock galaxy catalogues extracted from MICE 7680, one of the largest volume dark matter simulation run to date, with a volume of 1300 cubical gigaparsecs. Our mocks cover similar volume, densities and bias as the real galaxies and provide 216 realizations of the lambda or ω=−1 cold dark matter (ωCDM) BAO model. We find that only 20 per cent of the mocks show a statistically significant (3σ) preference for the true (input) ωCDM BAO model as compared to a featureless (non-physical) model without BAO. Thus the volume of current galaxy samples is not yet large enough to claim that the BAO feature has been detected. Does this mean that we cannot locate the BAO position? Using a simple (non-optimal) algorithm we show that in 50 per cent (100 per cent) of the mocks, we can find the BAO position within 5 per cent (20 per cent) of the true value. These two findings are not in contradiction: the former is about model selection and the later is about parameter fitting within a model. We conclude that current monopole and radial BAO measurements can be used as standard rulers if we assume ωCDM type of models.