Dear A. Wirbel,
your manuscript has received 2 in depth reviews that were very positive and highlighted in particularly the thoroughness and detail as well as the significant advance and relevance in model development (with high potential for addressing future real-world science questions).
The comments and issues raised by the reviewers have in general been addressed well and carefully in the revised manuscript and related author response (for which I thank the authors).
The only major critical point raised by both reviewers was that the focus of the manuscript is rather technical and more on model development and testing of benchmark experiments and thus they suggest that a more geophysical method based journal may be more applicable, but leave the decision to the editor. Somewhat surprising, the authors have not commented on/justified in their response why they originally chose to submit it to TCD (rather than a more geophysical methods journal).

I agree with the reviewers that the focus of the manuscript is rather on the technical side but I note that TC does not ‘per se’ exclude methodological modelling papers and the manuscript presents besides the more theoretical and technical numerical tests relevant glaciological applications (this was also why I accepted this paper for the discussion stage).
The focus and relevance of the modelling tests on glaciers could however easily be strengthened by (i) clearer present the performance regarding issues of numerical diffusion/smearing for advection the very narrow and small debris features (a point that has been indicated by referee 2 as well: demonstrate strength of model e.g. concentration tracking) and (ii) a slightly improved discussion of the glaciological applications and related implications. Such strengthening of the glaciological application (details see list below) would better justify publication in TC (rather than in a more methodological paper).

Thus before publication in TC I suggest to address:
• these (relatively minor) improvements linked to the glaciological application indicated above (and detailed in list below)
• to correct some other minor points listed below
• and to comment/justify briefly their motivation for the choice of TCD as the journal (rather than a more geophysical methods journal)

Strengthening/improving glaciological application:
This relates somewhat back to some comments of referee 2 of not visualizing concentrations in Fig. 9 well enough and hence ‘underselling’ the capability of tracking concentrations instead of just features or particles. I completely agree with the authors that the original concentration (which is for these debris bands chosen as 100%) should not change during transport, hence if the model is perfect we would only see red in Fig 9 after transport. However, I also agree with referee 2 that transport for such an example (with 100% concentration features) could also be produced by ‘simple particle tracking’ methods, even for time dependent cases (with updating velocity field with changing surface). Thus the change of shape of the advected features is not a unique feature of your method and does in my view not really give enough credit to the strengths or your methods/Modelling framework (tracking concentrations, little numerical diffusion).
What I would really want to see here (as a glaciologist/TC-reader), to trust and see the advantages of the proposed advection method/modelling, is some clear demonstration that your method (advection scheme) for debris transport works well and accurately enough FOR SUCH A GLACIOLOGICAL APPLICATION (an issue mentioned in conclusions (p. 19 last line/p.20 first line), but unfortunately not really shown anywhere properly for the GLACIER case). More specifically, I want to know how much/little for example tracking these narrow debris bands are affect by numerical diffusion/smearing, or how well concentration gradients are preserved (and thus how valuable your method is for glaciological applications).

Currently one has to gain this ‘quality assessment’ from the earlier non-glaciological applications. These earlier theoretical tests are very valuable, but to make it more relevant for glaciers (and glaciologists, TC audience), I think a better demonstration of the method regarding numerical diffusion/smearing FOR THE GLACIER APPLICATION is crucial, and would also better justify the choice of TC.

I think such a glacier case assessment is easy to do by:
1) Adding (to Fig 9) some line-plot of profiles of concentrations for example across 2-3 debris features (bands, crevasse fills) before and after (and potentially during). You could plot these profile lines all in same figure with the centre of the feature in the middle of the cross-direction axis. This would further illustrate/quantify better how for example the shape (thickness) of such features has changed during transport.
2) Similar for Fig. 10, currently I really struggle to see in a 3D plot how well concentrations are preserved (remember tracking concentrations is the strength of the modelling method). So maybe add a plot similar to above, that shows a 1-d profile (along or across flow) of concentration through the debris feature (again with a relative spatial-coordinate, with the centre of the feature in the middle of the along or across-flow coord.). This would allow much better to demonstrate and see the behaviour and performance regarding variable debris concentration.

This would also allow or require a better/more detailed discussion of the performance and usefulness (regarding concentration, numerical diff…) of the proposed methodology FOR GLACIOLOICAL APPLICATIONS.
Currently, the glaciological application and discussion is almost entirely focussed on ‘tracking shapes’, which is a shame and really underselling the part and capability of the modelling framework relevant for glaciers (and TC).

Andreas Vieli (editor) 21 Oct 2017


Further rather minor points:

P. 6 line 13; I assume the numbers in the brackets refer to min and max, why not say this more explicitly e.g. (min 18.2, max. 1.4).


p. 11, section 4.2.2.: I think this 3D surface-geometry requires some more explanation where it is from, or how it has been constructed. I ask this because you want an as realistic as possible geometry and related flowfield. Is it a result of some build up (spinning up) experiment, or even related to a real world case?

p. 17: here the layer shaped features are mentioned, BUT NOT SHOWN, which apparently deform to arcuate concave features. It is a bit tricky for the reader to digest, and just not clear why they are not shown. I would either show it (maybe in appendix) or leave this sentence away as the reader has no way of judging this without evidence.

p. 17, line 15: i am a bit critical about rockfall events producing thick debris deposits (sheprical). Maybe this applies for very small rockfalls, but the ones I am aware off or have seen in field are rather relatively thin extended sheets (layers). Maybe rephrase this a bit.

P. 17 line 23-25: this is more a note than an urgent requirement to change anything. Yes I agree that this results in a non-uniform debris thickness, but if one considers that the debris once melted out at the surface and just advected on the surface downstream one probably gets a surprisingly smooth thickness distribution at least in an along flow sense and after some time. So ‘uniform’ debris thickness may not be so wrong. Maybe this could be formulated a bit more nuanced.

p.19 last line/p20 first line: ‘..but are still suitable for glaciological application’. Yes this may be the case but this applicability (accuracy, perforambce in concentration tracking) should be better shown/visualized/demonstrated (see main comment above)).

Fig. 1: the ‘grey’ line is very hard to see, if it is important make it clearer (maybe black and thicker).

Fig. 4 caption line 2: ‘…the location and extent of the additional debris layers deposited later at the surface at the indicated times.‘

Fig. 5: (d) I personally find the 3D-profile plot along the flow line (d) not very helpful as it is ‘3d’-again, its is very hard to quantify anything.
Why not just plot a 2-dim profile along the flowline in along-flow and z dimensions and then indicate the flowline in 5(a)
(e) at least in my printed copy it is very hard to see any bed structure in this figure, as it appears just as a blue blob. Can this be improved, maybe plot thicker ‘wires’ or add colours as elevation?

Fig. 9: add perhaps some additional figure here that shows zoomed in profile plots visualizing concentration (changes, diffusion) across some features (see main comments)

Fig. 10: again add perhaps some additional figure here that shows zoomed in profile plots visualizing concentration (shape change, concentration) across some features (see main comments)

Dear authors,
After carefully analysing the revisions following 2 very positive reviews, I identified one more substantial and some minor points that still needed addressing. The more substantial point on strengthening and clarifying of the debris concentration tracking as a glaciological application has been well addressed by adding a further figure (although it may need slight clarification in caption that you mean ‘vertical’ profiles). Further, o this point a detailed and very useful discussion regarding capability and uncertainties of tracking debris in glaciers has been added.
I also thank the authors for their justification of the journal choice, it confirms my view as the editor (and initial decision at the access review stage).
The authors also addressed all the minor points (or justified why nothing is changed).
Thus, the paper is ready to be accepted for final publication after the very minor point on the caption of the new figure 11 has been addressed. At this stage I would also like to thank the authors for their excellent handling of revisions process.
Best regards

Andreas Vieli, the editor, 20th Nov 2017

Minor point:
Fig11 caption: I somewhat struggled initially how these profiles are oriented (and had to consult fig A8) and how they are meant. I would help the reader here by say that they are ‘vertical’: ‘Vertical cross-profiles at different models times through the debris features D1,…. The vertical profiles ….’.
You may also clarify that the absolute position on the x-axis is somewhat arbitrary but that for better comparison, the same features have for different times be plotted with the centre at the same x-position