Decision after re-review of revised version
The authors addressed in their first revisions in dertaiul the major points made by the 2 initial reviews that were thorough and critical but overall positive and highlighted the value of the undertaken research and presented data (comprehensive long-term dataset of new areas (North), valuable insights into controls on dynamics). The major improvements encompassed above all a much better focussing of the analysis and discussion on terminus types and geometric factors and removal of the previously somewhat vague discussion of the influence from climatic and oceanic forcing. The categorization has also been clarified and simplified and the writing and structure improved.
Due to the major revisions undertaken, the revised version has been sent out to review again, to a third 'new' reviewer as the two previous were not available for a re-review. This review is again positive (highlighting again value of long-term dataset on dynamic changes justifying publication) but also being thorough and critical and flagging some more major points with some constructive suggestions for improvemnet. Some of these points further draw on points made in earlier reviews but in general also flagging the need for a more concise and clear writing.
I suggest to the authors to address the list of points made by the re-review which I summaries and moderate below briefly in order to take into account the earlier revisions/reviews (which the 3rd reviewer does not comment on).

Recommended revisions moderated by editor (following numbering of major points by re-review):

#1 This point on considering more detailed temporal trends/resolution of velocities and elevation changes has already been risen in the first round of reviews and explained but the authors why they did not change this. I agree with the re-reviewer that for the explanation of some of the dynamic changes the detailed timing of acceleration or thinning is crucial (see major comment #3) but I understand that calculating continuous rates is with the given dataset very challenging.
I suggest at least in the discussion to include a more detailed/temporally resolved consideration of temporal variations (even though in the presented spatial variations of rates (in figures and tables) this is not explicitly done).

#2 I would also welcome a clearer introduction of 'terminus types' in the introduction which would allow a more focussed and clearer presentation in the results.

#3 improve and clarify the argumentation and line of thought for the discussion of the causes of enhanced dynamic changes (retreat). Relating to point #1 the more 'continous' time series of thinning and velocities could be better used in discussion.

#4 This is a tricky point, as the authors removed the extensive presentation and discussion related to climatic and oceanic forcing in order to address the first round of reviews. I support the much clearer focus on the terminus types and geometry factors improved the paper substantially but of course the observed changes are not fully independent of climatic/oceanic forcing I somewhat agree with the re-reviewer that presenting no evidence of climatic/oceanic forcing and which makes it somewhat difficult to relate the observed changes to potential triggers (timing…).
I leave it open to the authors whether they address this point but perhaps just showing for context some simple well established time series of atmospheric and oceanic (the latter being already tricky though) temperatures maybe useful.

#5 regarding the force balance I assume no explicit calculation has been done, if so rephrase this sentence somewhat to avoid confusion (if it has been done then explain it clearer).

#6 I guess the fact that some surge-type glacier occur should be included but perhaps this can be shortened a bit to avoid loss of focus on the main important points.

#7 please try to further make the writing in general more concise and clearer (simpler). Partly this will be solved by addressing the listed minor points by the editor.

And further address list of minor points (which to some degree will make the manuscript more focussed and concise, see point #7

In general the above suggestions should not need very major revisions and already be partly solved by some careful editing.

The editor, Andreas Vieli, 27th July 2018

Dear Emily Hill and co-authors,

This final decision has been taken on the basis of a 2nd round of revisions of the manuscript and a detailed author-response to a re-review and related editor comments (after a first round of revisions in response to two positive but critical reviews).

Thank you for the revised manuscript and the detailed author- response to the referee-comments. All major points raised by the referee 3 have in my view been addressed satisfactorily in the revised manuscript or justified why not changed.
This includes all the major points earlier summarised by the editor which include in brief:
-clarification on long-term focus and interpretation of data
-improvement of structure of results/discussion of grounded versus floating
-clarification and streamlining of argumentation in interpretation of dynamic changes (initial thinning…)
-further reduced interpretation away from climatic forcing and better focussed towards terminus type an dynamic behaviour
-shorted and streamlined discussion of surge-type glaciers
-addressed minor points and made manuscript more concise

Regarding the last point, the manuscript is still very long but for a manuscript that presents such extensive interlinked observational datasets this is justified.

When reading through the manuscript I spotted only a few very minor points that needs addressing before the final publication of the manuscript and which are listed below.
At this stage I would also like to thank the authors for their excellent handling of review and revision process.

Best regards
Andreas Vieli, editor, 7 Sept 2018

Minor points to be considered:
p. 2 line 33 and p. 12 line 4: at both instances the term 'dynamic ice discharge' is used which I find awkward (and I have not seen before) as 'discharge' already includes 'dynamic', maybe 'dynamic mass loss' is more appropriate

p.18 line 12: is 'correlated' here really the appropriate term. Correlation mostly refers to a statistical relationship but no such statistical correlation has been undertaken. Maybe 'associated' would be more appropriate.
A similar issue occurs on p. 22 line 12.

p. 22 line 27: there is an 'in' missing: '… The switch to terminus retreat IN the 1990s…'

p. 23 line 6: '…the initial forcing that accelerated…', I would not use the term 'forcing' here as the thinning is rather the initial PROCESS that causes the acceleratin. Forcing wold rather be climate or ocean…

p. 23 lines 14-15: maybe I am picky here but strictly speaking your data and analysis does not say anything directly about calving type, so saying '… we OBSERVE two dominant calving behaviours…' is not correct. You can only indirectly INFER anything on calving so I would replace 'observe' by 'infer'.

p. 24line 8: there is an 'of' missing: '…due to the calving OF large tabular…'

p. 24 line 16: '…substantial PARTS OF floating ice'

p. 27 line 19: awkward formulation '..too..also…' one of the two is probably enough.