Comment on 'Radiative transfer in CO2-rich atmospheres: 1. Collisional line mixing implies a colder early Mars'.

Ozak et al. [2016] claimed that explicitly including the effect of CO$_2$ collisional line mixing (LM) in their radiative transfer calculations yield CO$_2$ atmospheres that are more transparent to infrared radiation than when spectra calculations are made using sub-Lorentzian line shapes. This would in particular imply significantly colder surface temperatures (up to 15K) for early Mars than estimated in previous studies. Here we show that the relative cooling that Ozak et al. [2016] associated to the effect of collisional line mixing is in fact due to a wrong choice of broadening species (air instead of CO$_2$). We then calculated Line-by-Line spectra of pure CO$_2$ atmospheres using a line-mixing model developed for self-broadened CO$_2$. Using the LMD Generic model (in 1D radiative-convective mode), we find that calculations made with the proper collisional line mixing model and with sub-Lorentzian line shapes lead to differences between early Mars surface temperatures smaller than 2 Kelvins only.