Baryon acoustic oscillations with the cross-correlation of spectroscopic and photometric samples

The baryon acoustic oscillation (BAO) experiment requires a sufficiently dense sampling of large-scale structure tracers with spectroscopic redshift, which is observationally expensive especially at high redshifts $z∼gt 1$. Here we present an alternative route of the BAO analysis that uses the cross-correlation of sparse spectroscopic tracers with a much denser photometric sample, where the spectroscopic tracers can be quasars or bright, rare galaxies that are easier to access spectroscopically. We show that measurements of the cross-correlation as a function of the transverse comoving separation rather than the angular separation avoid a smearing of the BAO feature without mixing the different scales at different redshifts in the projection, even for a wide redshift slice $Δ z∼eq 1$. The bias, scatter, and catastrophic redshift errors of the photometric sample affect only the overall normalization of the cross-correlation which can be marginalized over when constraining the angular diameter distance. As a specific example, we forecast an expected accuracy of the BAO geometrical test via the cross-correlation of the SDSS and BOSS spectroscopic quasar sample with a dense photometric galaxy sample that is assumed to have a full overlap with the SDSS/BOSS survey region. We show that this cross-correlation BAO analysis allows us to measure the angular diameter distances to a fractional accuracy of about 10% at each redshift bin over $1∼lt z∼lt 3$, if the photometric redshift errors of the galaxies, $σ_z/(1+z)$, are better than 10-20% level.