In the paper entered “the impair of surface melt rate and catchment characteristics on greenland ice sheet moulin inputs” the authors apply the SaDS surface meltwater routing model to a group of catchments located in west Greenland. The authors compare low and high intensity melt seasons to determine the relative importance of surface melting on meltwater inputs to moulins and the impact of supraglacial drainage system evolution. The authors find that supraglacial drainage system develop has a more pronounced influence on meltwater delivery to moulins in years with lower melt rates, while going on to provide recommendations for when to apply the computationally expensive SaDS model over other less expensive models. The authors have addressed many of the concerns raised in the last round of review and as a result the manuscript is much improved. I have included a few major concerns that should be addressed before publication, however once addressed I think he manuscript will make a significant contribution to the literature. A robust model such as the SaDS model will be increasingly significant as our in situ glacial hydrology observational record continues to expand.
Major Comments
In this paper lag times are presented as between solar noon (~15:22) and peak moulin inputs, rather than the lag time between peak melting and peak moulin inputs. Using the later definition would be inline with previous studies (). As is written now, it is unclear whether the timing of solar noon is allowed to change (presumably due to the approximate time stated in the manuscript), nor is it clear if there is a lag between the timing of peak melting and solar noon. This may become problematic for example if peak melting during transient melt events did not coincided with local solar noon (ref to the results presented on L113-115). If so, the longer lag time in moulin input would be an artifact of the timing of peak melt rather than caused by the supraglacial drainage system.
Statistical Analysis:
A few things regarding the statistical analysis presented on lines 126-136 and in Table 2 are unclear. First, the R^2 typically represents the coefficient of determination whereas r^2 would be the square of the Pearson correlation coefficient. I assume the text is referring to the former as it is stated that R^2 is equal to the proportion of variance explained by the independent variable. I know this is very in nit-picking but it is important to be precise here. The Pearson correlation coefficient (r) should also be included in this analysis. In the same regards, I am confused by the stated maximum and minimum values for R^2 as there should be a single value given for each of the correlations. Table S1 does not give statistics or p-values as stated in L134-L136. Additionally, even though p-values are small <10^-6 they should not be represented as 0 it is in violation of the definition of a p-value which is a probability. And finally, there should be a figure added to the supplement showing these relationships as graphs are essential to correctly interpret regression analysis results.
Regarding the interpretation of the statistical tests it is unclear how the p-values alone are being used to determine there are good correlations between variables while R^2 values range from 0.09—0.9, this issue here is not a low R^2 value but the variance between variables, years, and smoothing choices (e.g., diurnal vs daily). Moreover, I wonder if the lower R^2 values for the diurnal variables are a result of the lag time between variables. This is a common problem that is either solved by imposing a lag-time adjustment (e.g., Smith et al., 2021), or by instead analyzing forcing-response plots (e.g., Extended Data Figure 4 in Andrews et al., 2014). Due to the significant amount of text in the Results and incorporation within the Discussion (e.g., L165-170), I recommend explaining this analysis in more detail.
In section 4.1 the manuscript states internal variability is most important on timescales shorter than one day, as evidenced by the statistical analysis. Is this conclusion supported by model results? Specifically, how are model parameters (e.g., channel water depth, incision depth, density, flow, etc.) changing on daily vs. sub daily timescales? In high melt vs low melt years (e.g., 2012 vs other years)? Figure C1 shows that there is diurnal variability in channel length, so how does this fit in? What is the breakdown between catchments (e.g., is this only important for large or small catchments? What controls this variability and how does this vary between years?
In the discussion comparing model outputs to other works from Rio Behar catchment lag times are compared to work by Smith et al., 2017. It is important to note here that the lag time reported in that paper are the time between peak melting and peak moulin inputs (this is different from how lag times are described in the present manuscript), and are accordingly not directly comparable.
The discussion also describes the models limitations on refining the potential influence of supraglacial lakes on moulin inputs, is there a way to look at the outlet channels that drain the lakes and compare changes within those to other parts of the supraglacial drainage system to see if there are localized effects there on the draining lake? Alternatively, how do the lag times for catchments with lakes compare to a simple parameterization such as that used in Smith et al., 2017 (already cited within the manuscript)? I think understanding the influence of lakes on lags and meltwater inputs to moulins is very interesting and would be a significant contribution to the field of supraglacial hydrology. While I understand that lakes cannot be disentangled from your model domain, I wonder if comparison with a synthetic unit hydrograph could help parse out the lake’s influence on lag times. (As stated previously, I would suggest redefining the lag determination used in the text to be consistent with other models (e.g., SUH/UH models).
Channel density is spoken of throughout the entire manuscript yet there are no figures showing channel density evolution (only Figure 1)
Minor Comments
L2-6: Runon sentence
L4: change to “and the timing of surface meltwater inputs to moulins”
L13: Add citations to Smith et al., 2021, and Mejia et al., 2022. (Full citations at the end of this document)
L16: Add citations to Yang et al., 2020
L20: Define “efficient” here
L21: Define what you mean here by “evolution of drainage density”, the processes you describe typically control the evolution of a single channel (e.g., hydraulic capacity of that single channel), from the text it is not clear how these processes modify drainage density.
L28-29: It is not clear what “supraglacial drainage characteristically acts to reduce the diurnal amplitude of moulin inputs” means, would the concentration of flow by supraglacial drainage systems not increase the amplitude of diurnal meltwater inputs to moulins?
L105: Do you mean diminished diurnal amplitude for smaller catchments? Over time? Be specific.
L105-111: consider combining paragraphs.
L108-111:
L119: change & to “and”
L126: Figures 2—5
L128: R^2 is the coefficient of determination
L173-176: runon sentence
L172-177: Hard to understand paragraph
L295 (and elsewhere): I understand the use of the normalized or relative moulin input amplitude but this line is misleading, because the actual amplitude of moulin input variability is larger for your large catchments (it is only smaller/lower when you normalize by the overall larger discharge rates)
Supplement
Figure S1: Please add a legend corresponding to the colors used in the plots as to not make the reader flip back and forth between the main text and the supplement. It is also not clear what you mean by bold colors vs. light colors, do you mean the black line here? Please make this more clear in either the legend or in the figure’s caption. Further, it appears the colors used in the main text are different from those in the supplement.
Citations
Andrews, L. C., Catania, G. A., Hoffman, M. J., Gulley, J. D., Lüthi, M. P., Ryser, C., Hawley, R. L., & Neumann, T. A. (2014). Direct observations of evolving subglacial drainage beneath the Greenland Ice Sheet. Nature, 514(7520), 80–83. https://doi.org/10.1038/nature13796
Smith, L. C., Andrews, L. C., Pitcher, L. H., Overstreet, B. T., Rennermalm, Å. K., Cooper, M. G., Cooley, S. W., Ryan, J. C., Miège, C., Kershner, C., & Simpson, C. E. (2021). Supraglacial River Forcing of Subglacial Water Storage and Diurnal Ice Sheet Motion. Geophysical Research Letters, 48(7). https://doi.org/10.1029/2020gl091418
Mejia, J. Z., Gulley, J., Trunz, C., Covington, M. D., Bartholomaus, T. C., Breithaupt, C. I., Xie, S., & Dixon, T. H. (2022). Moulin density controls the timing of peak pressurization within the Greenland Ice Sheet’s subglacial drainage system. Geophysical Research Letters, 49, 1–13. https://doi.org/https://doi.org/10.1002/essoar.10511864.1
Yang, K., Sommers, A., Andrews, L. C., Smith, L. C., Lu, X., Fettweis, X., and Li, M. (2020)Intercomparison of surface meltwater routing models for the Greenland ice sheet and influence on subglacial effective pressures, The Cryosphere, 14, 3349–3365, https://doi.org/10.5194/tc-14-3349-2020.
Yang, K., & Smith, L. C. (2016). Internally drained catchments dominate supraglacial hydrology of the southwest Greenland Ice Sheet. Journal of Geophysical Research : Earth Surface, 121, 1891–1910. https://doi.org/doi:10.1002/ 2016JF003927 |
