\begin{table}%t1
%\centering
\par
\caption{\label{model_prop}Properties of the simulations and time averaged value of~$\alpha$.}
\begin{tabular}{@{}cccc}
\hline\hline \noalign{\smallskip}
Model & Resolution & $Re$ & $\alpha$ \\
\hline \noalign{\smallskip}
$Re3125$ & $(128,192,128)$ & 3125 & $7.9 \times 10^{-3} \pm 4.5 \times 10^{-4}$ \\
$Re6250$ & $(256,384,256)$ & 6250 & $5.9 \times 10^{-3} \pm 1.8 \times 10^{-4}$ \\
$Re12500$ & $(512,768,512)$ & 12~500 & $8.4 \times 10^{-3} \pm 3.8 \times 10^{-4}$ \\
\hline 
\end{tabular}
\tablefoot{The errors on $\alpha$ are computed following \citet{longaretti&lesur10}: the time history of $\alpha$ is divided in  $N$ bins of size $\tau$. $\tau$ is varied between $0.1$ and 8~orbits. For each $N$, the standard deviation $\sigma_N$ is computed according to $\sigma_N=[\Sigma (\alpha_i-\alpha)/N]^{1/2}$, where $\alpha_i$ is the mean value in bin~i. For large~$N$, $\sigma_N$ scales like $N^{-1/2}$. The errors reported on $\alpha$ use that scaling to estimate $\sigma_N$ when $\tau = 40$~orbits, the time duration over which the mean values of $\alpha$ are calculated.}
\end{table}