Dear Hélène,

Due to the substantial revisions made, your revised manuscript has been gone through a second review. Fortunately, both original referees accepted to re-review the paper, which facilitated the procedure. Both are positive towards this version, as their queries were clearly addressed to. They find the manuscript acceptable for publication, if some minor issues are corrected.

Referee 1 has some issues with the incompleteness of the figures as well as some wording in the text. Referee 2 would have like to see the BP model actually verified with the MIMSIP3d setup. Therefore you can not state explicitly that the model has been tested. However, I agree that showing the effect of of mesh resolution on the results (which is the main limiting factor in the MISMIP3d experimental results) shows the robustness of the model to cope with grounding line migration.

The referee reports are pasted below for your convenience. I invite you to carefully address the different points raised by both referees before submitting your final manuscript. At this stage i would also like to have a rebuttal that details your responses to the (remaining) remarks.

Thank you very much for your contribution to The Cryosphere.

Kind regards,

-Frank



Second review: reviewer 1

It is good to see the additional figures and explanations. The authors have clearly made a good effort to respond to the reviewers comments. I think the reader can now have more confidence that these results are robust. However, I still must ask for a few minor changes.

The new figures are not completely convincing. In particular, fig 2 appears to be on a different scale to the other figures, and shows a rather strange shaped domain (is this really the model domain? Why the long thin blocky white region on the right extending up from the lower edge of the domain?). Most importantly, the reader needs a way to compare figs 2 and 6. I think a box in fig 2 indicating the sub-region shown in fig 6 would be the simplest way to do this.

The figure text is in several instances too small, especially fig 5.
What is the green line in fig 6b?

Is the fig 10 caption correct? Should beta = 1 be the black line, and Favier melt the red line?

Page 9 line 18 “retreats” -> “retreat”

Page 11, lines 10-21. I simply don’t agree with this. You’ve shown that imposing a melt rate of 60m/a (high but not ridiculously high) in newly ungrounded regions has comparable effect in terms of contribution to sea level rise to doubling melt under the current ice shelf region. And yet you claim that imposing such a melt rate does not affect your results. You then say that Favier show the same thing, which is also not true. I guess you are talking about their m2 and m5 experiments. I agree, these are very similar for Elmer/Ice. But they are quite different for BISICLES (blue and red lines in their SLR plot). So I would argue that neither your study, nor Favier, present an unambiguous case for melting in newly ungrounded regions being unimportant.

Lines 25-26. I agree with reviewer 2. Could you change “conservative” to “likely to be conservative”, thereby acknowledging the less likely but still plausible possibility of a reduction in melt?

Page 12, line 14 “this parameter controls most of the dynamics” – please modify wording. You are saying that melt is the most important of the climatic forcing factors you have considered. But you haven’t considered rheology or basal processes. With this in mind, the current wording is a bit too general.

Lines 18-19 repetition of “qualitative”; I suggest removing the first one.

Line 21 “not exactly the same time period” – can you be more specific and say that glacier response lags increased melting?



Second review: reviewer 2

Seroussi et al. explore the sensitivity of the Pine Island Glacier (PIG) to various climate forcings (change in surface mass balance, calving front position and sub ice shelf melting) using the Ice Sheet System Model (ISSM). They demonstrate the importance of sub ice shelf melting and that a better description of this boundary condition is required to improve the reliability of sea-level rise projections. This work is well described, with relevant results and authors fairly answered to previous reviewers’ comments. This clearly deserves publication. I however have one regret and one suggestion.


ISSM, grounding line dynamics and MISMIP3d
Authors answer to one comment from the second reviewer that ISSM has been validated against MISMIP3d. Then, one could question why it is not mentioned in the paper? Actually, only the Shallow Shelf Approximation (SSA) implemented within ISSM has been evaluated so far against the MISMIP3d setup (and results were in conflict with other SSA models owing to parametrization of grounding line position, see Seroussi et al. 2014). The present study is made using the Blatter-Pattyn (BP) approximation. To that respect, the model used in this work has not been evaluated against MISMIP3d. This therefore justifies to not refer to the MISMIP3d benchmark in the present work.
Authors stressed in the past that solving higher-order approximations of the Stokes equations is important for proper modeling of the PIG flow (Morlighem et al., 2010). Consequently, they use PB in the present work. I regret to not see their BP model evaluated against MISMIP3d. I personally believe that any marine ice sheet model producing mass balance projections of actual glaciers should be first evaluated against MISMIP3d. This is not always the case (e.g. Joughin et al. 2014) and not here as well. It has to be noted that the authors made significant efforts to show that their results are not affected by mesh resolution (annex). I let at the appreciation of the editor whether further validations are required or not. To my opinion, validation against MISMIP3d would be a valuable annex to this paper or maybe a part of the Seroussi et al. 2014 (currently under revision in TCD). This would give more confidence in the result obtained in the present study and any further similar work the authors could produce. But I would understand that it is out of the scope of that paper.

Modeled grounding line retreat in disagreement with current observations?
Authors shows the evolution of their modeled grounding line in Figure 6. Modeled grounding line retreat seems in contradiction with current observations (Rignot et al., 2014): the model is not retreating in the main trunk (observations shows that it does) but exhibits large retreat in the eastern part of the ice shelf (not observed). I believe this should be clearly stated in the present work and some related caveats should be formulated when discussing the coming evolution of the PIG deduced from their simulations.
May this contraction with observations come from an inappropriate description of the grounding line dynamics (and evaluation against MISMIP3d may help to exclude this hypothesis)? I am more inclined to believe that it is mainly due to the melt rate distribution applied, here imposed from MIT GCM outputs. Then, concluding « coupling of the ice sheet with ocean circulation models is therefore desired for future studies to conduct more accurate simulations » seems a bit overstated to me. Authors rather show that imposed melt rate is a very sensitive parameter for a marine ice sheet model and that the melt distribution currently computed from the MIT GCM does not allow to properly reproduce some observed critical changes using this particular ice sheet model. I would suggest to reformulate related sentences.


Minor comment
I would suggest to use the same vertical axes in panel a, c, e and b, d, f of Figure 5. Comparison between panels would be easier.



References




Joughin, I.; Smith, B. E. & Medley, B. Marine Ice Sheet Collapse Potentially Underway for the Thwaites Glacier Basin, West Antarctica Science, 2014.
Morlighem, M.; Rignot, E.; Seroussi, H.; Larour, E.; Ben Dhia, H. & Aubry, D. Spatial patterns of basal drag inferred using control methods from a full-Stokes and simpler models for Pine Island Glacier, West Antarctica Geophysical Research Letters, 2010, 37.
Rignot, E.; Mouginot, J.; Morlighem, M.; Seroussi, H. & Scheuchl, B. Widespread, rapid grounding line retreat of Pine Island, Thwaites, Smith and Kohler glaciers, West Antarctica from 1992 to 2011. Geophysical Research Letters, 2014.
Seroussi, H.; Morlighem, M.; Larour, E.; Rignot, E. & Khazendar, A. Hydrostatic grounding line parametrization in ice sheet models. The Cryosphere Discussions, 2014, 8(3), 3335-3365.