1: \begin{abstract}

2: 

3: \textcolor{blue}{\noindent Data re-sampling methods such as delete-one jackknife, bootstrap or the sub-sample covariance are a common tool for estimating the covariance of large scale structure probes. We investigate different implementations of these methods in the context of cosmic shear two-point statistics. Using log-normal simulations of the convergence field and the corresponding shear field we carry out realistic tests of internal covariance estimators. For a survey of $\sim 5000 \deg^2$ we find that jackknife, if implemented in what we call the \emph{galaxy-scheme}, provides the most reliable covariance estimates. Bootstrap, in the common implementation of duplicating sub-regions of galaxies, strongly overestimates the statistical uncertainties. }

4: 

5: In a forecast for the complete 5-year DES survey we show that internally estimated covariance matrices can provide a large fraction of the true uncertainties on cosmological parameters in a 2D cosmic shear analysis. The volume inside contours of constant likelihood in the $\Omega_m$-$\sigma_8$ plane as measured with internally estimated covariance matrices is on average $\gtrsim 85\%$ of the volume derived from the true covariance matrix. The uncertainty on the parameter combination $\Sigma_8 \sim \sigma_8 \Omega_m^{0.5}$ derived from internally estimated covariances is $\sim 90\%$ of the true uncertainty.

6: \end{abstract}

7: