Overall this is a well-done paper. It is clearly written with great effort taken to actively measure those terms often used in numerical models, but to do so through great effort by going into the field and directly measuring as many inputs as possible and then compute the actual results using a box model approach which is then compared to a current and well respected community model. These types of papers are significant because they are the ground truth which reveals where the physics in the models need to focus so as to advance closer to reality. This paper clearly involved a ton of team work - which has been carefully done and so I strongly recommend that this paper be accepted for publication pending one notable recommendation to improve the paper further. 

One highly recommended extra step to significantly improve the paper is the following. When one first steps onto snow-covered sea ice, the first thing that becomes crystal clear is the variability of absolutely everything around you. Take one step and you are able to measure a whole new set of values. A simple way to bring all that rich variability into a paper like this is to create a table of each type of measurement and then explore the impact of propagated variability. Specifically, which terms in the equations will yield the highest variability in the results? In other words, which of the many collected measurements are the most influential in widening the outcome of the results? As an exercise, I recommend that the authors create a table of each measured variable and then estimate either in absolute or percent-relative terms the degree/range of variability observed within the red perimeter study area.  As example, the lateral melt range is discussed as being large but negligible to freshwater estimates. That's a good start, but that discusses away a source as negligible. What are some of the variables that have a large range but are not negligible and which of these may vary so as to impact the overall conclusion of the paper? 

One does not need to do an exhaustive study, rather pick judiciously those variables which have the greatest weighting factor on the outcome.  

Take for example, Figure 3 - cumulative freshwater (m). The Surface melt is a bigger producer than the bottom and the bottom a bigger producer than the snow. But how much do equations (4), (5), and (6) change if the ice salinity is not fixed at 3 g kg^-1 (line 90) and on line 254 the salinity of the surface meltwater is not constant at 1.1 g kg^-1. I don't find this constant in the Lange et al 2022 paper - rather I find a range of values in the Lange et al 2022 paper as an excellent source from which to gather variability statistics. What happens to the results if the ice salinity ranges 2-4 g kg^-1 and the surface meltwater is not constant but  ranges 0.5-1.5 g kg^-1 ? These are small perturbations and easily possible within a few horizontal meters. Do small perturbations like this impact the outcome ? 

Same goes for the 1 W m^-2 ocean heat flux. What happens if the heat flux is in a range of 0.5 to 3 W m^-2 and the snow variability is 10-30 cm instead of a constant 22 cm? 

Same with ridge macroporosity.  In early June before solstice, it may still be 4-6% but the freshwater equivalent time series goes into August. Is 4-6% still valid for ridge macroporosity when melt ponds are experiencing bottom drainage? Ridges are polar reefs of biodiversity in the Arctic in Summer. There is surely a wide range of salinities and porosity in a single ridge between start of June and start of August. 

In particular, Equation (3)  computes the internal meltwater volume as a residual difference between top and bottom. In that case, which of the two inputs is more responsible for the higher variability in V_internal - is it the top or the bottom and why? These are really effective sensitivity questions that can be answered by inputting a range of values instead of a constant number. 

In particular, Figure 7 would benefit from some vertical whiskers especially just prior to the surface drainage event at the end of June.

Could variability in bottom melt from 15 to 22 June have an impact on the big surface ice freshwater drainage event  between June 22 and July 1? If so, then which input(s) is/are responsible? I suspect that bottom and surface melt process are highly intertwined as the porosity in the microstructure increases toward 5% as the season progresses from June to August. 

In short, this paper will elevate from a point measurement sample study to a representative regional study if the authors run some sensitivity tests to get a sense of ranges of outcomes in their freshwater volume calculations. And especially, if they identify which variables most impact the sensitivity of these calculations. Those are the variables that need to be prioritized in models to ensure that the most critical meltwater physics are nominally incorporated. 

I suspect that the calculations are already in a code so it is the matter of making a loop around that code and running some realizations with a range of values instead of constant numbers. 

My biggest concern with constants is that the constant will stick and it will be very hard to undo once done. It is unwise - especially now - to assume something constant when in fact variability itself is on the rise. 

Melting of snow and sea ice produce relatively fresh water, adding into the upper ocean, and further impact the interactions among air, ice and ocean. This physical process is complicated, and some important quantitative fractions within which are still unknown. Based on the field investigations during MOSAiC, the manuscript presents a detailed anysis on the sea ice freshwater budget, and the fate of meltwater over time. The manuscript is well organized and easy to follow. I have only some specific comments as below.

1. In the first paragraph of introduction section, I think the importance of the sea ice meltwater should be clarified more clearly rather than only saying “The magnitude and fate of freshwater associated with sea ice melt are important for the surface energy budget, ice mass balance, ocean structure and primary productivity” on lines 24-25. The relevant description can be more specific.
2. There are some parameters I do not quite understand. For example, why do you chose 3 g kg^-1 in line 90? The salinity of sea ice surface is generally low, and the same to pond bottom ice.
3. The curves in the figure can be more differentiated. For example, figure 7. You can use both dash and dot lines to highlight the difference.
4. When compared figure 7 with figure 13, more quantitative descriptions should be added. On lines 377-378, “the modeled total freshwater input from sea ice and snow melt (Fig. 13) is moderately low compared to that observed (Fig. 7)”. How much and why?
5. On lines 391-392, “hence the underestimation from the model even more dramatic than presented here”. Can a clearer explanation be given? The comparison between obervations and model results shows a big difference. Is there any possiblity to improve the parameterization of freshwater budget in the numerical modling accroding to these observations?
6. What is the perpose of section 4.3? As you say “The net impact of the freshwater budget on the heat budget cannot be determined” on line 417.
7. Lines 396-406, I think these are not discussions.