1: \begin{abstract}

2: 

3:   We study the statistics of the \Lyaf in a flat $\Lambda$CDM cosmology with the

4:   $N$-body + Eulerian hydrodynamics code Nyx. We produce a suite of simulations,

5:   covering the observationally relevant redshift range $2 \leq z \leq 4$. We

6:   find that a grid resolution of 20 $\hinv$kpc is required to produce one

7:   percent convergence of \Lyaf flux statistics, up to $k = 10$ $\hinv$Mpc. In

8:   addition to establishing resolution requirements, we study the effects of

9:   missing modes in these simulations, and find that box sizes of $L > 40$

10:   $\hinv$Mpc are needed to suppress numerical errors to a sub-percent level. Our

11:   optically-thin simulations with the ionizing background prescription of

12:   \citet{haardt_and_madau_2012} reproduce an IGM density-temperature relation with $T_0

13:   \approx 10^4$ K and $\gamma \approx 1.55$ at $z = 2$, with a mean transmitted

14:   flux close to the observed values. When using the ionizing background

15:   prescription of \citet{faucher_giguere_et_al_2009}, the mean flux is 10-15 per

16:   cent below observed values at $z = 2$, and a factor of 2 too small at $z = 4$.

17:   We show the effects of the common practice of rescaling optical depths to the

18:   observed mean flux and how it affects convergence rates. We also investigate

19:   the practice of `splicing' results from a number of different

20:   simulations to estimate the 1D flux power spectrum and show it is accurate at

21:   the 10 per cent level. Finally, we find that collisional heating of the gas

22:   from dark matter particles is negligible in modern cosmological simulations.

23: 

24: \end{abstract}

25: