Persistence of a Subsurface Water Mass in a Deep Mid-Latitude Fjord

By: L Bianucci

The authors use a high-resolution, unstructured model to investigate the persistence of a cold, oxygen-rich sub-surface layer formed during the preceding winter. They show that the presence of the layer – and the stratification changes that it brings about – changes the background circulation from a three-layer to a four-layer system, and suggest that increased mixing at the head of the fjord reduces the estuarine circulation.

The paper is clear and well written – but the scientific argument is relatively weak, and the results could be better quantified and presented. Rather than exploring how the cold anomaly can persist, which is what they set out to do according to the abstract – the paper is a comparison of the circulation within the fjord during a short period in June for experiments with and without the cold layer present. We are shown that the circulation changes – but the authors do not explain why.  Mixing is stated to be weaker due to the circulation changes, but this (or the effect on the cold layer) is not shown/quantified. Is the difference in mixing between the two scenarios larger than the difference between the model and the observations? (which is mentioned in the text and, seen in the excessive “smoothing” of the modelled T-profiles in Fig. 3). What is the “normal”  residence time for water at one level in the fjord – and how does that change with the “perturbed” stratification?

The model was run for one month – but there is no mention of how the boundary conditions change throughout the period (or the year) and how this would affect the circulation.

In addition, I find that the choice of figures illustrating the points could be improved (see detailed suggestions below for a few suggestions).

I can recommend publication only after major revision.

Specific comments:

L 47: Explain here how this changed the stratification/layering of the fjords (e.g. using text from line 220)  - and refer to Fig. A1 (which ought to be included in the main paper). Consider including also a profile from a non-Arctic outflow year in Fig A1.

L 97: The description of the river-forcing is very detailed – consider moving it to the appendix.

L127: What do you mean by “mostly limited to”

1. 145 and on: How useful are these metrics - as used here - when the larger part of the water column does not change during the short simulation, i.e. it's all about the initial conditions? In addition, there are now observations of velocity, on which the paper's main results are based.
2. 158 and on: Nor sure I agree - if you zoom in on the cold, subsurface layer and use a scale that's adapted, there are quite some differences in the depth, “sharpness” and the vertical extent of the cold layer already a fortnight after model initialization. I'd suggest plotting the profiles of temperatures from the three occasions on top of each other (modeled in one panel and observed in one) to show the time development in the model vs. obs. If one assumes advection to be negligible, one can (I think?) use the differences in the profiles to infer an estimate of observed/modeled diffusivity.
3. 160. Well this is not very surprising, you initialized the model two weeks earlier with the cold layer present. How long is your run compared to the normal residence time of water in the fjord?

L 170 It is not easy to see the structural difference between Fig. 4 a and d that you describe in the text. To me there’re four layers in both of the figures: red, blue, red, blue – but I understand that what you refer to as four layers are red, blue, blue red, where the two blue layers are separated by white?

1. 173. How can there be a net along fjord circulation below sill depth?
2. 185 I presume that this is because the effect of the salinity changes on the density is greater than the effect of temp. changes (Please quantify) – they would also have different signs, right? If you lower the temperature you make the water at that level denser, while if you make it fresher you make it lighter.
3. 193 (give depth intervals)
4. 217 Why/how did the temperature minimum create a separation of the return flow?
5. 291 “velocities were weaker” – please quantify (and or make figures where the reader can directly compare the velocity profiles)
6. 222 “salinity and density were impacted” – please describe how (and quantify) – e.g. referring to a new version of Fig A1 where a profile from a “normal year” is included. As of now, the reader has no means to judge whether the salinity profile used in the sensitivity profile (which determines the density and hence the stratification) is realistic.
7. 225 Rather than referring to Fig. 6c, refer to a new version of Fig A1 which also includes a panel comparing the initial N2-profile
8. 226 Give depth range
9. 226: Please include figures/numbers that show the reduced density difference/estuarine circulation.

L228: “decreased mixing”. What is this statement based on? Fig 6b shows that the density decreased all the way to the bottom for the baseline exp? I presume / guess / hope that the difference below e.g. 250 m between the simulations is zero at the start of the simulations.

L 240 responds? I do not understand this sentence.

L242 give depth range

L 245 Is it relevant to mention deep water renewals here?

L 249 Do you think/mean that the findings from Bute inlet is universal? Would it not depend on the stratification outside of the fjord?  What are the “mechanisms” that you refer to?

Fig 1

Lon/lat are exchanged

Consider using color to show the resolution and move panel (a) to the appendix

Why do the two maps appear different – is the aspect ratio not the same?

Consider including a length scale in (b) to help the reader.

What about showing bathymetry rather than resolution?

Fig 2.

See comment above about the usefulness of these metrics – move to appendix

Fig 3.

Plot only three profiles – and let us know where each one is from. Especially for modelled temperature they are different. Are these differences there initially, or are they “produced” within the model.

Consider “cutting” the profiles, so that you show the upper layer with a different x-axis than the deeper waters. Using the large scale needed for the upper layer, means that changes in the lower layer are not shown. One alternative could be to include a row showing also initial conditions in the same way. The observed structure in salinity/density above about 100m but below the surface layer appears to be missing in the model. Is this feature not there initially, or do they disappear during the run.

Fig 4

What happens at about 70 km – and why is this not commented in the ms?

How do you explain the velocities below sill depth?

Consider helping your readers see the four layers.

Figure 4 and 5 basically shows the same thing, right? Maybe you only need one of them?

Fig 5

For clarity, use a velocity scale suitable for the lower layers – and only give the upper layer outflow velocity as numbers?

Fig 6

a)Use smaller dots. Not sure this figure is necessary?

b-c) I think you need to include panels showing delta ro/delta N2 from the initial conditions for this figure to be meaningful. And would it not be better to (instead or in addition) compare the changes in density/N2 between the start and the end of the run (In the “end of the run” you’d likely have to average over some sensible period, but I think one could use a number less than 29 days? )

Table 1

Move to appendix

The MS presents the study of dynamics of intermediate layers in about 730-m deep fjord - Bute Inlet, a mainland fjord in British Columbia, using short-period calculations with a numerical model that was validated by observational data. The used FVCOM finite-element model has variable mesh size from 13 to about 1000 m. The baseline 1-month model run in summer, simulating subsurface temperature minimum due to adopted initial conditions, is complemented by another experiment where the subsurface cold layer was removed from the initial conditions for temperature and salinity. Comparison of the two numerical experiments revealed that layered circulation depends on the initial stratification – usual three-layer flow is replaced to a four-layer one, when cold subsurface layer of Arctic origin is found in the region. The results are this way interesting and worth of publishing.

Reading further, I was not always able to find justifications for the interesting statements.

A. “Persistence” is an interesting interpretation that could be discussed, but two one-month model studies do not allow its quantitative evaluation; therefore, this term should be avoided in the title. Two times of this term in the abstract is also not justified. I recommend reformulation of the MS title. Also, the abstract could be rewritten, since about half of it is general introduction not directly connected to the conducted studies.

B. Instead of sufficiently long maximally realistic simulation study with variable forcing and boundary conditions, that could describe formation, evolution and decay of intermittent cold subsurface layer, the authors have adopted a simplified approach where open boundary conditions were kept unchanged for a one-month run with modified initial conditions. The authors should carefully justify: (a) Why one-month simulation is appropriate for a process of seasonal duration // from February 2019 to fall, L45-46 //. (b) Why the “sensibility” experiment with altered initial conditions but unchanged forcing is physically feasible. Perhaps it is useful to make an alternative full simulation for the period of missing cold sub-surface layer, when three-layer flow is evident.

C. The paper could reproduce and/or elaborate the observational background of Arctic outflow, the main headline of the MS, and its response in the Bute Inlet, in order to support the present modelling study. There are general papers by Jackson et al. (2022) and (2023) referenced, but meteorological and oceanographic anatomy of the modelled period would be nice to be read from this paper.

Minor remarks

1. The terms “baseline” and “sensitivity” are commonly used in other meanings than here, please consider reformulation.
2. 1 does not reflect the location of study area in wider geographical context, it was not easy to find it e.g., from Google Maps.
3. L191: “strong winter mixing event” is introduced, but it remains uncovered (see also conclusions L243 and L247).
4. L247: “Our study highlights how a fjord’s circulation can be changed for the whole year by an extreme wind mixing event in winter.” Where this statement comes from?

I propose revision of the MS.