Dear authors,

thank you for responding to the reviews in detail. The reviewers have raised some important points, and the revised manuscript has certainly improved. There is no doubt that you have compiled an important dataset, that deserves publication as it is outstanding particularly regarding the temporal coverage. However, I am not yet convinced by structure and presentation of the paper. Below you will find a long list of editorial and structural changes that I suggest. I am not a native english speaker myself, and I don't require that all editorial remarks are put into practice.

However, particularly section 4 and section 6 require revision in terms of ordering, clarity, and succinctness. Details of the GNSS measurements are not yet clear to me, and also the derivation of errors (and their temporal variability) is not rigorously enough described for a paper in ESSD.

Note that none of my comments require further data analysis, and I echo the reviewer's judgement that the dataset is scientifically meaningful. Another round of revisions with a focus on clear structure and presentation style will give this dataset the framework it deserves.

Kind regards, Reinhard Drews

General Comment: Please provide a version including track-changes next time.

P1, l3: Suggestion of rephrasing: “Here we present a unique dataset of yearly-averaged ice-flow velocity measurements at stakes and stone lines covering more than 100 years on Hintereisferner and more than 50 years on Kesselwandferner. Moreover, the dataset contains subseasonal variations of ice flow from Gepatschferner and Taschachferner covering almost 10 years.”

P1, l 13 - Awkward opening sentence with “established”, “confirmed”, “later confirmed” and “explained”. Suggestion: “…after Aggassiz (1847) hypothesized the theory of ice ages, which was then confirmed by Penck and Brückner (1909) and further substantiated with isotope analysis on polar ice cores (REF?) and theoretical work by Milankowitch (REF?)”

P1, l15: The monitoring efforts in the 17th and 19th century require a reference, or at least some details on who and where.

P1, l 20: Split than sentence up into two. It is unnecessarily convoluted.

P2, l 8: Replace Jacob et al., 2012 with more recent Zemp et al, 2019?

P2, l 8: Again a compound-complex sentence structure containing two add-ons with “which” and one inset with “, and estimates … ,”. This needs to be disentangled because it is unnecessarily hard to read.

P2, l 10: “For example ELA..” -> “For example, ELA…”

P2, l 14: I think “ALS” is more commonly referred to as “Airborne Laser Scanning” (not Scan)

P2, l16: “at annual resolution” -> “annually”

P2, l16: Make two sentences. Second one “The stones advect with ice flow ( i.e. recording ice-flow in a Lagrangian system). The stakes, on the other hand, are relocated every year to their original position (i.e. recording ice-flow in an Eulerian system).

P2, l17 “net” -> “network” and then later “from THE Little Ice Age onwards” .

P2, l20 Remove “Without further discussion….” and start with “This paper presents ice-flow velocity records on well ….”

Close the introduction with one sentence on how this data can be applied (maybe as a baseline dataset for using glacier length/velocity as proxy for climate change over centennial timescales ?)

P2, l 24: “Ötztal Alps (Austria)” later, “that differ” -> “differing”

P2, l 30 “behind” > “upstream”

P3, l 8: What was the main outcome of this comparison?

P3, l 19: several measurements during summer do not allow discussion of seasonality (but sub-seasonal variations during summer or variability between different summers).

P3, l26 “. After than DGPS was used.”

P4, l10 Ok, forget my previous comment about Lagrange and Euler, but I hope you see how this was written confusingly.

P4, l11 : What is the approximate lateral variation in a stone line, and why do you “only” report the average value? Also “depends on the number of stones” is correct but does not give credit to the reviewer comment (referring to a statistical bias) nor to the lateral variability (simply speaking, if all stones are located at the same place the mean does not depend on thenumber of stones). Rephrase and add details.

P4 l15: It is unclear to me where this ratio comes from, why it was used for comparison and what the take away message is.

P4. L 24: Whether velocities are positive or negative depends on your choice of coordinate system (some people to z vertically downwards). Be precise here and don’t outsource such conventions to an external reference. Also add “ relative to the surface” for emergence/submergence velocities.

P4, l 30 “correlate” -> “correspond”

P5 , l 1 “Aim at high accuracies” ? Rephrase: Measurements done with theodolites and tachymeters are accurate to the lower cm-level. This is possible, because of statistical adjustment …. . Also add: Further details on the derived errors for this specific dataset can be found in Schneider, 1979.

P5 , l 1 I admit that I don’t know what “resection” means in this context and my online translation tool only finds medical terms. Please rephrase or explain for the wider ESSD audience.

P5 l 7 what is “staking-out” ?

P5 l 1: First: Collecting all uncertainties in one section is a good idea. However, then this section should only contain this and details of the individual methods (e.g. what stakes are used, the use of wood wool, what rigid connections were used, where the base station was located and so on) should be transferred to the sections 3.1- 3.2. Second: This section needs more structure. What I expect here (given the section heading) would be something like:

Measurements based on theodolits and tachymeters are with XX cm. This depends on a,b,c and details can be found here. Differences between the observed glaciers are a,b,c :

Measurements based on DPGs are within XX cm. This depends on a,b,c. Differences between the observed glaciers are a,b,c.

So for KWF you used your own base station and for GPF & TSF a base station in Italy. Why?

P5, l19 State-of-the art GNSS base stations should receive not only GPS and GLONASS, but also Galileo, Beidou and so on. Is there a reason why you specifically mention GPS and GLONASS ?

P5 l20 “linear distance” – “baseline”

P5, l25 what is the order of magnitude for observe sub-seasonal velocities in summer, and how do these relate to the errors of up to 1 m ?

P5, 25 The use of RTK – DGPS is unclear to me. RTK to me means “real-time kinematic” and requires a radio-link between base station and rover (I doubt this was the case when using the base station in Malles). Also, why would you use RTK for measuring the stake positions? Wouldn’t it be better to determine the position in the post-processing ? Why is the stake position required in real-time?

Following up on this: I am not convinced by the details given regarding the DGPS data. Additionally to the confusion whether the position was determined during post-processing or in real-time, it is also unclear to me what processing scheme has been applied. For example, how did you get the position of your own local base station? In my experience those details are not unimportant when aiming for an absolute horizontal precision of less than 5 cm. Moreover, there is no distinction between errors in horizontal and in vertical direction. Shouldn’t it be more difficult to get the vertical displacement? If all of these points result in smaller errors compared to other sources (e.g. the tilt of the stakes, or slipping of stones) this is ok but should be explicitly stated and quantified.

p.5 l 30: GPS measurements should be characterized with an absolute error. 5 % is additional useful information but requires the local flow velocity for context.

Section 4 is important, but it is currently too unstructured and partially also unclear. This needs to be revised.

P6, l 10 Suggestion add: This is XX m/a (or XX %) above the mean flow velocity measured at this location.

P6 l 20 “During the first period, “ (comma) and then later “advanced” (instead of “was in an advancing state”)

P6 l 21 “presented positive length changes” sounds complicated. Why not “advanced”?

P6 l 32 would it be better to say “Submergence transitioned to emergence around 1990…”

P8 l 11 “and highest ant the lowermost stakes”

Section 6 has a quite narrative character and I am not a friend of merging Discussions and Conclusions. Many people only read abstract and conclusion of a paper (and they tend not read the conclusions if they are intertwined with discussion elements). Consider taking these two sections apart, and provide a succinct conclusion emphasizing the uniqueness of your dataset and possible applications.

P8 l 26: That the accuracy of measurements depends on topographic shading has already been mentioned above. Also it is not quantified. I wonder if this statement is important enough to be repeated in the conclusion.

It is unclear to me, why the “peak in 2014” for GPF has the highest uncertainty. What event in time makes this measurement less certain than the others? Section 4 estimates errors at GPF to be within pm 20 cm “at its best”. I don’t see how this explains the variability in errors shown in Figure 7.

P 8 l 28: The term “jerky movement” requires a reference (Goldthwait ?). This is another somewhat speculative sentence which is ok in a discussion but misplaced in a conclusion.

P 8 l 31: “As conventional…” is this sentence complete? After reading it multiple times I think I understand that “ELA” and “specific mass balance” are other glaciological parameters to characterize a glaciers, but you find them less well suited compared to the ice-flow velocity.

P 8 l 30 “extra effort” compared to what?

I apologize, but the meaning of the last paragraph is opaque to me. Please rephrase. I suggest ending the paper with your main finding, uniqueness and further applications of your dataset. Keep it succinct.

Dear Authors,

thank you for responding to my final list of comments. I feel a number of misunderstandings have now been clarified, and the paper reads more fluently. The reviewers have already evaluated the paper positively, I and I follow their recommendations of publishing the paper in ESSD. Congratulations for compiling (amd maintaining) a long-term dataset that is unique it is temporal coverage, and can surely be useful for other researchers.

Kind regards, Reinhard Drews