Overall assessment

Perovich et al provide an elegant analysis of legacy SHEBA data to resolve the fate of snow and ice melt water during Arctic summer. This paper is concise and clearly written, and provides new insights into the fate if summer snow/ice meltwater in a multi-year ice pack. This paper is acceptable after some minor revisions, which are relatively simple to address.

Major points

Be consistent in the use of terms related to the melt water quantified. If salinity is not taken into account it is more accurate to call it "melt water" and not "freshwater". Also make sure you use the terms consistently in the text, figure labels and legends and captions. "melt water" and "fresh water " are not interchangeable terms IMHO.

Add some relevant citations from work on melt ponds and support some statements with citations. Indicated in the attached pdf. Already Nansen did observe such meltwater layers. 

It would be also useful to early in the paper describe the composition and properties of the ice pack where the SHEBA study took place. Typical information on snow depth, ice thickness etc. of the ice pack would be useful to have at least briefly described. 

Detailed comments

The above and some more minor suggestions are included in the annotated supplement pdf

Review of tc-2021-114

Summary

The authors calculate the freshwater budget for an area of sea ice during summer, evaluating the relative contributions of the sources and sinks, and volume storage in melt ponds, in a time series. Aerial photography provides the basis for scaling of the data according to ice concentration, and floe perimeter to area, the latter important for understanding the lateral melt contribution especially late in the season (around break-up). The rich SHEBA dataset affords them the ability to do this. The need for understanding the fate of melt water is well justified, as is the lack of knowledge on the seasonal cycle. The paper is well written and certainly fits the remit of The Cryosphere pending some minor revisions.

Major comments: 

More information about the snow and sea ice at the SHEBA measurement area should be given at the outset of the paper. SHEBA is relatively well known, and of course there are references to look up, but the authors should still provide the reader some specific information about the ice condition in the measurement area (snow and ice thickness, age/type, topography). It is not until Page 14 that we learn that it is a multiyear ice floe.

Measurements from a single lead are used to estimate the relative contributions of vertical and horizontal drainage over a 10-day span. When scaling by IC, it is assumed that measurements at the lead site are representative of the broader area. How safe is this assumption, i.e. are the ponds completely cut-off from the floe edge or could there be some linkages even near the sampled lead?  

It would be helpful if the authors could place their results in the context of emerging sea ice conditions, as a point of discussion. Recognizing the lack of supporting data, some comments about expected sources and sinks could be made about a smoother, and thinner, first-year sea ice dominated Arctic with changing snowfall, floe size distributions, and increased bottom melting. In other words how the freshwater seasonal cycle would be expected to be different or similar to SHEBA. This would fit well with the points about the need for more studies and insights on data collection strategies at the end of the paper.   

 Specific comments:

(Page = P, Line = L)

P2L38: Please provide the reference(s) for the freshwater layer impact on gas exchange and aerosol particle production.

P2L41: Suggest “What are the relative contributions … and how do they change with time?”

P4L69: The term meltwater is used here. Elsewhere the term freshwater is used. Meltwater is more appropriate in cases where the water isn’t fresh such as the ice melt. If the authors choose to stick with freshwater then provide a definition early on such as all relatively freshwater is termed freshwater (or make the distinction).   

P6L108-110: Can get rid of ‘1998’ since the year isn’t specified elsewhere in the time series related analyses.   

P7L30: It may not carry much significance in relation to the calculated budget, and observed steady influx of freshwater to the ocean, but the authors should make a short note about whether or not there were any freezing events over the study period.

P13L193:  Include mention of the role of wind forcing.

P14L228: It’s probably implied that aerial observations would be included in any further studies. But it still would be worth highlighting here their importance for scaling as was demonstrated with the SHEBA data.