|I appreciate the work the authors have invested and found most of the raised questions answered satisfactorily. The new figures are very interesting and helped to evaluate the model, however I have some general concerns which I will try to outline below:|
The new Fig. 8 exhibits a poor comparison to the SMB from the regional model MAR. Even though I am aware that model evaluation on the basis of other models is problematic, I would in this context consider MAR to be some kind of gold standard. The disagreement is substantial and I have doubts that the model is reliable, especially as FAMOUS was using the same ocean forcing as used in MAR. I am still not sure if here the climate model or the surface module is the problem, and this question should be answered. If the climate model is the problem (it seems there is too much cloud cover), it would be helpful to also show the near surface temperatures. But neither in this manuscript nor in Gregory et al 2020 I have found an evaluation of the 2m air temperature distribution. (I take it that the model is generally slightly too warm over Greenland). Furthermore I wonder if the disagreement in shortwave radiation is related to the coarse resolution rather than to a "wrong" background climate (in that case downscaling of the atmospheric transmissivity might be considered). Is the shortwave radiation also biased outside of Greenland? How would Figure 2 look if the native FAMOUS resolution was used and MAR/RACMO was also remapped conservatively to the same coarse resolution? How does figure 2 look for the 2080-2099 period? In any case it is worrying that the albedo does not exhibit the typical spatial pattern. After tuning it is likely that albedo compensates biases in the climate, and in consequence the sensitivity of the surface scheme would be biased, too (good results for the wrong reasons) which could also explain the poor agreement in Fig. 8.
Finally I would like to point out that Kapsch et al, 2021 also present Greenland SMBs based on T31 resolution simulations, which seem to be fine; and from my own experience, PDD schemes are likewise able to produce a reasonable SMB response based on lapse rate corrected T31-temperatures (if insolation is similar to today's Greenland and temperatures are reasonable).
Kapsch, M.-L., Mikolajewicz, U., Ziemen, F. A., Rodehacke, C. B., and Schannwell, C.: Analysis of the surface mass balance for deglacial climate simulations, The Cryosphere, 15, 1131–1156, https://doi.org/10.5194/tc-15-1131-2021, 2021.