The jackknife method gives an internal covariance estimate for large-scale structure surveys and allows model-independent errors on cosmological parameters. Using the SDSS-III BOSS CMASS sample, we study how the jackknife size and number of resamplings impact the precision of the covariance estimate on the correlation function multipoles and the error on the inferred baryon acoustic scale. We compare the measurement with log-normal mock galaxy catalogues with the same survey geometry. We build several jackknife configurations that vary in size and number of resamplings. We find that it is useful to apply the tapering scheme to estimate the precision matrix from a limited number of jackknife resamplings. The results from CMASS and mock catalogues show that the error estimate of the baryon acoustic scale does not depend on the jackknife scale. For the shift parameter $\alpha$, we find an average error of 1.6% and 1.2%, respectively from CMASS and mock jackknife covariances, consistent with pre-reconstruction analyses. However, when relatively few resamplings are used, the jackknife error estimate becomes unreliable. Future large-scale structure surveys will map even greater volumes allowing percent-level estimation of the covariance matrix using a jackknife approach.

The Baryon Acoustic Oscillation (BAO) scale in the clustering of galaxies is a powerful standard ruler to measure cosmological distances and determine the geometry of the Universe. Past surveys have detected the BAO feature in the clustering of different galaxy samples, most of them composed of redder, quiescent galaxies and bluer, star-forming ones out to redshift $z\sim1$. Besides these targets, new upcoming surveys will observe high-redshift galaxies with bright nebular emission lines out to $z\sim2$, quasars and Lyman-$\alpha$ quasars at $z\gt 2$. All these different galaxy targets will be used as multi-tracers of the same underlying dark matter field. By combining them over wide cosmological volumes, we will be able to beat cosmic variance and measure the growth of structure with unprecedented accuracy. In this work, we measure the BAO scale in the two-point auto- and cross-correlation functions of three independent populations of multi-tracers extracted from the SDSS DR7 Main galaxy sample at redshift $0.02\lt z\lt 0.22$. Combining their covariances, we find accurate constraints on the shift parameter $\alpha=1.00\pm 0.04$ and $D_{\rm{V}}(z=0.1)/r_{\rm{s}}=2.92\pm0.12$.

In this work we perform an analysis on the recently proposed conjoined cosmic growth and cosmic expansion diagram [1] to compare several dark energy models using the Figure of Merit showed in [2], which consists in the inverse of the $1\sigma$ confidence region in the $f\sigma_8(z)-H(z)$ plot. Our analysis also consists of comparing the models by performing different statistical criteria: Bayes factor [3], the Bayesian Information Criteria [4] and the Akaike Information Criterion [5]. We also developed a 3-dimensional Figure of Merit to account simultaneously for the errors on the growth rate and the Hubble parameter. The main idea is to consider several cosmological models and compare them with the different statistical criteria in order to highlight the differences and the accuracies of each single criterion.