Measuring the baryon fraction using galaxy clustering

The amplitude of the baryon signature in galaxy clustering depends on the cosmological baryon fraction. We consider two ways to isolate this signal in galaxy redshift surveys. First, we extend standard template-based Baryon Acoustic Oscillation (BAO) models to include the amplitude of the baryonic signature by splitting the transfer function into baryon and cold dark matter components with freely varying proportions. Second, we include the amplitude of the split as an extra parameter in Effective Field Theory (EFT) models of the full galaxy clustering signal. We find similar results from both approaches. For the Baryon Oscillation Spectroscopic Survey (BOSS) data we find f_b≡Ω_b/Ω_m=0.173±0.027 for template fits post-reconstruction, f_b=0.153±0.029 for template fits pre-reconstruction, and f_b=0.154±0.022 for EFT fits, with an estimated systematic error of 0.013 for all three methods. Using reconstruction only produces a marginal improvement for these measurements. Although significantly weaker than constraints on f_b from the Cosmic Microwave Background, these measurements rely on very simple physics and, in particular, are independent of the sound horizon. In a companion paper we show how they can be used, together with Big Bang Nucleosynthesis measurements of the physical baryon density and geometrical measurements of the matter density from the Alcock-Paczynski effect, to constrain the Hubble parameter. While the constraints on H_0 based on density measurements from BOSS are relatively weak, measurements from DESI and Euclid will lead to errors on H_0 that are competitive with those from local distance ladder measurements.