Please see the attached pdf file.

Laurence Gray

The authors use Sentinel-1 ice velocity products, DinSAR-derived grounding lines, and a variety of surface elevation data sources, to understand build up to the next surge of two important glaciers in Greenland, and document evidence of subglacial drainage events and the dynamic response to them.

This paper brings up to date the work by Mouginot et al., (2018) and provides a small advance in understanding the quiescent phase and predicting the likely year of the next surge. It is well written and very well illustrated and, although it could be considered incremental, I believe it is worthy of publication as a Brief Communication in the Cryosphere subject to some revisions:

GENERAL

1) The paper focusses on SAR and InSAR methods and observations, demonstrating very well developed data analysis skills and figure-making. In contrast, the glaciological discussion (page 6) is rather brief and pays no attention to the literature (surge-related, glacial hydrology-related, or otherwise), which is a potential missed opportunity for influencing the topic and picking up citations. There could be lots to discuss here about where the water goes (ground-water? - this is a growing topic), whether the glaciers are frozen to their beds (see lots of papers about surge initiation, and subglacial water outbursts), and what actually triggers a surge. If the authors do not have the appetite for a literature review, could they co-opt someone (e.g. a well-known surge specialist), to add this extra bit of informed (and referenced) discussion? If not, it may be better to couch this section purely in terms of observations and leave out the under-developed glaciology.

MINOR

line 50: "Contrary to all other"→ "In contrast to non-surge-type"

line 51: "decrease with decreasing distance to the ice front"→ "decrease up-glacier"

section 2.1: I got very confused over all of the time periods, so I recommend revising this section. Where do the 24-day averaged velocities come from and why are these used to create the quarterly mosaics? Why not go from the natural periodicity (which is unacknowledged as 6-days for part of the time-period) direct to 3-months?

Line 58: “displacement anomalies”. This doesn’t seem like a suitable term. Anomalies are normally related to long timeseries. I think the term differences or variations is more suitable here.

Line 71: Remove “roughly”. I’m sure you were as careful as you could be.

Line 83 (and later): The term “accumulation” here is used to describe vertical uplift, which I guess is the net result of ice inflow, surface snow input and minimal surface melt. So it is not exactly wrong, but might be misinterpreted as simply the surface snow input. I suggest you find another way of expressing this.

Line 100: Here I think you are referring to the highly noisy velocity maps for each jun-aug period in figure 2. These appear to be too noisy to make sense, and the apparent (but clearly wrong) signals dwarf those that you are drawing attention to (the non-summer speed-ups). I suggest you either properly filter these data (I am surprised that the PROMISE processing chain has allowed these through), or just remove the jun-aug panels and say that the summer data is not reliable because of surface melt.

Lines 140-148: Have a rethink of the order of explanation here. It is confusing that you talk about the 2027 date, then the 2040 date, then mention them both again later. This could all be made much clearer with reference to each date, its source, and implication only once.

Line 149: “decrease in back pressure”. This would be better expressed as “increase in driving stress”