Clayton et al. derived analytical equations and conducted LEFM analysis to study the influence of firn-layer material properties (depth-varying density and modulus) on surface crevasse propagation in glacier and ice shelves. They found that the firn layer has a stabilizing effect on grounded glaciers (free slip boundary condition), whereas a destabilizing effect on ice shelves, with regard to fracturing and calving. The study has important implications for assessing the stability of ice sheets or ice shelves. However, there are two major limitations in the assumptions of the models: i) Poisson ratio is assumed to be depth-invariant; ii) firn is assumed to be impermeable when evaluating the depth of meltwater-driven hydrofracture, neglecting the fact that meltwater will penetrate the porous firn layer instead of fracturing it.  I suggest the authors reconsider the model assumption, or at least highlight the limitations, before it can receive further consideration.

1. Why do the authors neglect depth-variations in Poisson ratio? Shouldn’t Poisson ratio and Young’s modulus both strongly depend on the density? Will a depth-varying Poisson ratio (which is more realistic) significantly affect the results? Below attach some references on the depth variations in Poisson ratio [1, 2, 3]. One possible way to represent the depth varying mechanical properties could be developing empirical relationships between Poisson ratio/Young’s modulus, and the firn density.
2. The longitudinal stress was derived for compressible linear elasticity (Eqn.1 in the manuscript), why not viscous model? I think that a common approach, when looking at calving for example (e.g. Benn et al 2007, Ann [4]), is to calculate the background stresses from a viscous model (associated with long- term creep of the ice, and estimated perhaps from satellite-derived estimates of strain rate) instead of using an elastic model to calculate that background state. The authors might need some explanation justifying why they use linear elasticity to calculate the longitudinal stress.
3. Once the authors start to consider meltwater within the crevasse, it confuses me that the porous nature of firn is completely ignored. LEFM no longer holds for porous material and poromechanics [5] should be considered. Could the authors at least highlight the limitations of current results (Figure 5&7 in the main text)?

[1] Schlegel, R., Diez, A., Löwe, H., Mayer, C., Lambrecht, A., Freitag, J., ... & Eisen, O. (2019). Comparison of elastic moduli from seismic diving-wave and ice-core microstructure analysis in Antarctic polar firn. Annals of Glaciology, 60(79), 220-230.

[2] Smith, J. L. (1965). The elastic constants, strength and density of Greenland snow as determined from measurements of sonic wave velocity (Vol. 167). US Army Cold Regions Research & Engineering Laboratory.

[3] King, E. C., & Jarvis, E. P. (2007). Use of shear waves to measure Poisson's ratio in polar firn. Journal of Environmental and Engineering Geophysics, 12(1), 15-21.

[4] Benn, D. I., Cowton, T., Todd, J., & Luckman, A. (2017). Glacier calving in Greenland. Current Climate Change Reports, 3, 282-290.

[5] Coussy, O. (2004). Poromechanics. John Wiley & Sons.

    The idea that it is worth considering how the properties of firn layers could affect the stresses that control surface crevasse opening is very compelling.  However, this analysis makes the radical assumption that the stresses in an ice sheet or shelf are controlled solely by the elastic deformation of compressible ice.  This is fine if one simply wants to go through a mathematical exercise, but the title, abstract and body of the paper imply that the results of this analysis applies to real ice sheets and shelves.  I was particularly disturbed by the fact that the abstract does not make clear that this is an exercise based on ignoring viscous flow of ice.  The fact that the Maxwell time of ice is on the order of days means that an ice sheet or shelf would have to have formed in less than a day for this analysis to be applicable. 

   The major conclusion of the paper is that inclusion of low-density firn produces opposite effects for idealized ice sheets versus floating ice shelves.  The abstract and a cursory reading of the paper makes this seem like a general conclusion.  Upon closer reading it is clear that the ice shelf result is only for a particular region close to the edge of the shelf.  The authors correctly note that assumption of perfect elasticity results in compression everywhere far from the shelf edge so that no surface crevasses should result for any assumed firn densities or Young’s Moduli!  This is confusing because the paper only discusses analytic solutions for the stress field far from a shelf edge.  To get surface crevassing on a compressible ice shelf with infinite viscosity requires bending stresses close to the edge of the shelf.  The authors then use a finite element model to compute those stresses at a fixed position (250 m) from the shelf edge.  At that position the predicted crevasse depth is increased by a decreasing firn density and Young’s Modulus.  I assume that this is a robust result but it is hard to evaluate given the information in this paper.  More importantly, the paper makes it seem that this is general result based on the analytical results derived in the paper, as is clear from the opening of the “Conclusions” section:

 “In this paper, we derived analytical equations for the far field longitudinal stress including the effects of surface firn layers, described by depth-varying density and Young’s modulus profiles based on field data. These analytic expressions were used to perform fracture propagation studies on isolated air/water-filled surface crevasses in grounded glaciers and ice shelves …”

This certainly gave me the wrong idea when I first read the paper.

   The other major result of the paper is that low-density firn results in smaller crevasse depths for a grounded glacier compared to a uniform ice case. The authors note that this result contradicts the previous Linear Elastic Fracture Mechanics analysis of van der Veen (1998).  I suspect that the difference with the previous study is caused by the assumption of purely elastic horizontal stresses which are less extensional at the ice sheet surface than the stresses assumed by van der Veen (1998).  Thus, again I am not convinced that the results of the new analysis apply to the real world.

It is incumbent on these authors to make a case that the assumption of perfect elasticity gives insight into the opening of surface crevasses on real ice sheets and shelves.