Dear Hélène,

Thanks for this new version of your paper. I have now read it as well as your reply to the 2 reviewers and the 2 comments. I have few remarks, mainly technical, before your paper can be published in TC (see below).

Best regards,
Olivier

- Eq. 1 and below. I guess i already made this remark in one of the ISSM paper, but why are you not using the classical fluidity parameter A instead of B, which by the way is more a viscosity than a hardness (which is related to elasticity...). Especially in this paper where you are referencing to the MISMIP experiments which give the value of A. (If you change this, think to change the notation for the element area, which is also A).

- page 10, lines 9-11: I am not really convince that starting with h=10m or h=500m will really change your steady state solution, as both will be obtained for an advancing GL. Moreover, you are here discussing the spread in GL position, not the position itself. Did you test this with your model, i.e. starting with h=500m decreases the spread of the GL position?

- page 10, line 24: in Table 1...

- page 14, line 18-19: repetition: We tested this solution on the 5 km and 1 km meshes

- page 14, line 19: Results in Fig. 7

- page 17, below A3: where $C_g$, Eq. (7), is the ...

- Legend Fig 3: delete : results of this experiment: the

- Figs. 3 and 4, and Figs 5 and 6, should be only 2 figures.

- Table 2: $\Delta GL$ has not the same definition in Table 2 and Table 1. Consider using different notations.