In this paper, the authors use a combination of satellite and in-situ data to document the evolution of water masses in the Laptev and East Siberian Sea of the summer of 2018. The focus is on the surface layer, and in particular, water masses deriving from the Lena River and low salinity waters from the Kara Sea.
The authors make use of CTD transects undertaken in the ARKTIKA-2018 expedition to evaluate the accuracy of satellite SST and SSS products. For the latter, the authors themselves present a novel SSS product that they deem to be most appropriate for the geographical region under question. The authors argue that the match between the satellite products and the in situ data is sufficiently strong that the satellite products themselves can be used for much of the remaining analysis in the paper, in particular, defining separate water masses and tracing their evolution over the late summer 2018 period.
Using a TS framework to separate the water masses, the authors resolve spatial changes in the water masses over time. They also focus attention onto particular sections in order to make links between water mass changes and other processes: wind speed and the passing of cyclones.
The authors present geochemical analysis from the cruise transects that helps to verify the origins of the surface water masses between the tree end members: river water, sea ice melt, and marine water.
This paper is a useful documentation of the surface ocean in the Laptev and East Siberian Sea in the summer of 2018, and to that extent that the authors successfully make process-based findings, it has implications beyond the time frame during which the measurements were made.
The methods are mostly appropriate for the chosen questions, though I have some specific suggestions in terms of robustness of the evaluation of satellite data, and the means of comparison of certain physical data.
Novel oceanographic data and novel SSS satellite retrievals are presented, but it is not quite true that the presentation of water mass properties at a synoptic scale in this region of the Arctic is totally novel (see Osadchiev et al., 2020, Scientific Reports).
The significance of the work is largely in its implications for the usefulness of satellite SST and SSS in the Arctic, though in my view the authors have more work to do on that point.
The communication of the results and methods needs some improvement, both in text and figures.
Evaluation of satellite data.
The authors present co-located surface water (CTD and TSG) measurements and satellite data in scatter plots. The authors also provide a useful error frequency distribution, from which they show that the observed errors are more substantial than that provided by the satellite product (for both SST and SSS). After discussing the comparison to in situ data, the authors simply resolve that the two products agree well with the in situ data, and use this as support for using the products for the following analyses. A more robust justification is required, especially given that the differences to the in situ data are not insignificant.
Can the authors use existing conventions in the literature for what constitutes a “good agreement”? Another approach would be to think about how sensitive the key findings are to random error and biases in the satellite data. Can the authors comment on what threshold of random error or bias the data must be below in order for the results to be robust? If it can be demonstrated that the results are not sensitive to the present level of error in the measurements, that would be more convincing evidence to the reader that the data are appropriate to be used.
Surface water-mass TS analysis
It is inevitable that the separation of water masses into distinct groups will be based on somewhat arbitrary salinity and temperature values. However, the authors need to do a bit more to explain to the reader why they have chose the definitions they have, and how sensitive their results are to these choices.
It is somewhat concerning that the range of values displayed by the satellite measurements vastly exceeds that of the in situ data. The authors need to explain to what extent this is due to the in situ data being fewer and spatially limited, and to what extent it is due to error in the satellite data. It is a concern that the classification might be based on erroneous values; it appears that according to the in situ measurements, water-masses 1 and 3 were not sampled. Comparison of the cruise track and Figure 9 indicates that the water mass 1 might never have been properly sampled by the CTD measurements, while it is a surprise that watermass 3 is never sampled in the MIZ on sections 7 and 8.
Can the authors put the cruise sections in context of the identified water masses?
P1L20 - make it clear exactly what region was affected by this freshwater decrease
P2L13 - describe what these studies contributed. This would also give the reader some background into what is understood already about the region.
P2L14 - Statement should be revised in light of Osadchiev et al. 2020
Fig. 1 - Mark rivers clearly. Dotted lines are used for ice edge, bathymetry and grid markings; this is a bit hard on the eyes. Consider changing some (or all) of these to light, solid lines. Consider choosing a different colour for text that refers to different types of features (eg. rivers in blue, straits in red…)
P5L7 - what exactly is the standard error in this context?
P6L1 - units can be given for the standard deviations
P7L6 - what depth does the satellite sample? This is clearly relevant in the discussion here and would be worth stating explicitly.
P7L8 - given that the median profile portrays a picture of the water column that is not representative of any given profile, I am not sure it is worth describing it here as if it is physical. For instance, the smoothness of the thermo-/halocline is an artefact of averaging many profiles with sharper halo-/thermoclines at different depths.
P7L14 - use ‘is composed of’ instead of ‘composite’ (composite suggests another meaning)
P8L12 - meaning of statement in parentheses is unclear
Fig. 4 - Contour plot looks a bit messy - consider pixel plot without contours. Fig. 4c does clearly show coherent groupings with distinct biases in the satellite data (e.g. measurements from days 45-50 are consistently too cold in satellite data, while in days 25-30 the satellite data are too warm)
P9L6 - is this all data or just those from CTD casts where the mixed layer depth is below 7 m?
P10L21 - ‘potential cloudiness’ is a bit confusing for a specific date that has already passed. Was it partially cloudy?
P10L33 - justification required, as suggested above
P11L7 - provide a reference for this statement? It is not obvious to the reader that this would be the case, as density of measurements and degraded sensitivity are not obviously compensating.
P11L20 - explicitly state the implications of oversampling for interpretation of the results
P11L26 - is this an innovation or a convention? Provide a reference if a convention, provide justification if an innovation
P13 L2 - DMI SST / SST DMI - be consistent with the naming
P13L6 - quantitative assessment of what makes this a very good agreement would be valuable (as suggested above)
P13L24 - what exactly are ‘the ice charts’ from AARI? Are they available online?
P15L9 - can thermal fronts be seen in the daily satellite data?
P16L1 - provide uncertainty
P16L11 - how might one distinguish between these two sources of variability? Can any attempt to do so be made here? Can a comparison be made with the variability detected in the observational survey?
P16L23 - explain origin of the two separate branches
P16L29 - make it clear to the reader that these diagrams are not analogous to Fig. 7, as they are only for surface values
P16L1 - describe how water mass boundaries were determined (as suggested above)
P18L9 - worth explicitly stating that this water mass is not considered to be melt, but trapped river water, based on the geochemical analysis in 3.3.4
P18L4 - clarification required to make origin of CMS as referred to in L4 and L14 compatible. Could be simply saying that it is comprised of both transformed CF and transformed WF (if this is the case), or explicitly saying that there is no transformation route to CMS directly from WF, but it is produced only from CF.
Fig. 9 - use acronyms for the water masses as per main text
Fig. 10 - calendar time as y axis would assist linking to text. Wind speed colormap is not intuitive in my opinion (try another one?)
P23L6 - it is not surprising that there is a weak correlation between these variables; salinity changes will effectively integrate the variability in the wind (see, for instance Osadchiev et al. 2020, Fig. 4). Could the authors try correlating the time derivative of salinity and temperature with the wind speed? This might be more instructive.
P24L5 - is it that the “warm” river water signal is not observed anymore?
Fig. 13 - show dashed blue line over entire sections for those done entirely in the MIZ
P26L13 - comment on the implications of the findings regarding the B-V frequency?
P27L8 - comment on timing, which is naturally important here
P29L13 - sentence is ambiguous
P30L14 - it would be better to provide evidence that sea ice melting cannot create this freshwater horizon, than to say that there is no evidence that it can.
All figures showing maps should include the traces of the major rivers, and in my view it would be helpful to keep the names of the straits on at least one window in each figure.
The written text needs some improvement in grammar. A few pointers:
Use of ‘the’ and ‘a’ needs attention
Use of passive voice and tenses is sometimes confusing/ambiguous. I’d recommend using second person (‘we’), active voice, present tense to described things done by you, the authors, in this study.