This is a well written and timely paper that uses numerical modelling constrained by palaeo-glaciological and geochronological data to produce a set of ice sheet histories for the Icelandic Ice Sheet through the last glacial cycle. I think the subject of the paper is suitable for Climate of the Past and could be published subject to some minor-moderate revision that address the following points:

1. The interpretation that hydrofracturing drove ice shelf disintegration and thus collapse, clearly relies upon the presence of ice shelves fringing the ice sheet during deglaciation. But what is the direct geological evidence that ice shelves did indeed fringe the IIS during retreat? Rather than an ice shelf could the retreating ice sheet margin have been in the form of a grounded tidewater margin? Can the authors rule out the latter scenario and, if not, would this have an impact on their interpretation of the role of hydrofracturing? There needs to be a more explicit justification for the existence of ice shelves based on marine geological data. I am not saying I disagree with the authors but rather that this is important to their study and they need to provide a more convincing justification for ice shelves presumably from published geological data. This could be included within #2 below.
2. I think it would be helpful if the authors included a summary of the key points from the geological data of the ice sheet history – perhaps summarised by different sector and noting in particular the geochronology on retreat timing and rate. This could be an additional section or included in the Introduction. It does not have to be very long but it would be helpful. At the moment there is only a section on ‘Palaeo-constraints which appears as section 2.2 in Methods.
3. At the very end of the paper (lines 424-428) sea level rise is mentioned as a potential driver of marine ice sheet collapse. The authors state that to address the impact of sea level they carried out a sensitivity experiment with sea level held constant at its 15 ka value. This had little impact. How realistic is holding sea level constant given that the sea level jump associated with Meltwater-Pulse 1A occurred shortly after this time and was broadly coincident with the timing of the rapid deglaciation and collapse of marine-based ice at 14.6-14.0 ka?
4. Line 27 – presumably reconstructing IIS evolution is also challenging due to the absence or sparsity of geochronological control?

Overall comments

An overall well-written and interesting manuscript on an important aspect of ice-shelf hydrofracturing and its impact on ice decay, but several aspects listed beloww would benefit from clarification and expansion to strengthen the interpretation of the results presented in the current version. Presentation of results (i.e., figures) could also be enhanced to support the key points.



Ice dynamics, deglaciation drivers, and isostacy

The statement “We then disentangle the drivers and controls of its subsequent deglaciation” would benefit from a broader discussion of glacial isostatic adjustment. In particular, consideration of its sensitivity to heterogeneous mantle viscosity and the resultant impact on regional sea level would strengthen the interpretation of deglaciation processes, as well as their rates. Then , the  discussion of ice stream behaviour could be expanded. In the statement “Most of these ice streams activate and deactivate independently…”, basal velocities periodically drop to zero, implying complete shutdowns. How is subglacial meltwater involved in these shutdowns? A more detailed discussion of basal melting, geothermal heat flux, and meltwater production would be useful here.



Representation of hydrofracturing and ice shelves

The paper would benefit from a clearer discussion of the limitations associated with representing hydrofracturing at a 5° (~7 × 6 km) grid resolution. How well can hydrofracturing processes be captured at this scale? Furthermore, ice shelves appear to occupy relatively limited areas in the northern sector (Fig. 8), whereas southern part lacks floating ice. Quantification of ice-shelf area relative to the total ice-sheet area would be helpful as this is central part of the study and heterogeneous variability would add nuance to the results. In addition, further discussion is needed on how much hydrofracturing occurs in the southern ice-sheet sector and how this compares with the north.



Geological constraints and model–data comparison

I think that perhaps a more explicit figure showing empirically reconstructed ice margins, including dating uncertainties, is needed to allow a clearer visual comparison between geological reconstructions and the model runs. Finally, heat-flux reconstructions and borehole locations from Flóvenz and Sæmundsson (1993) and Hjartarson (2015) should be shown on one of the maps to better contextualise the geothermal forcing used in the model.



Minor comments

Line 20: Robel et al. (2019) is cited twice.

Line 270: Section heading “3.2.1 pre-LGM” should be capitalised.

Figure 7: Ensure that minimum and maximum extent labels are clearly identified and, if possible, differentiated by colour.