Please ignore the *** marks, these are for our own documentation.

Dear reviewer,

Thank you for your review on our manuscript. From reading your review, we identify four major concerns: 1. Presenting presumed published data. 2. Speculation on benthic respiration. 3. Not elaborating on internal waves and tidal constituents. 4. Missing ADCP data. We will address these points one by one, after which we reply to your specific comments.

1. Using published data. The reviewer states that we use data that was already published as a deliverable from the EU-Horizon project that this study was (partly) funded by. Although we get the point of the reviewer, we do not think we present published data, as the project deliverables are not scientifically peer reviewed. In addition, the deliverable that the reviewer refers to is inaccessible for the general public and scientists. Therefore we strongly suggest to keep this data (sediment traps, isotopes) in this paper, as it is the first time it is presented in a peer-reviewed publication.
2. Benthic Respiration. We agree with the reviewer that this paragraph contained too much speculation, therefore we will remove this text.
3. Internal waves. We will perform an extra analysis on critical slope at the study sites. In addition, we will compare tidal ellipses from the lander with the barotropic tidal ellipses from a model, and elaborate if barotropic tide or baroclinic tide is more dominant at our sites.
4. Missing ADCP data. We mistakenly did not report that the ADCP devices used in this study were single point measurement devices. We will address this.

Below I will reply in detail to your specific comments:

L2-3 The title should be reconsidered. The strong bottom currents are not the only important factor for the sponge ground in this manuscript but also the nutrient and food availability. That all is excluded with this title. All other than the benthic measurements are also excluded with this title (CTD and ARGO float are not benthic).

Reply: thank you for this suggestion. Indeed the title does not cover all the aspects that the MS covers, and we will assess if the title is still covering the story after updating the MS.

L38 Please remove “In Canadian waters”. The Labrador Shelf is already a unique area, and “Canadian waters” makes it sound political. Is your study region still in Canadas Exclusive Economic Zone?

Thank you, will be removed.

L40 The sensor most likely measures fluorescence and not chlorophyll-a. The lander is equipped with one sediment trap, not multiple.

Thank you, will make sediment traps singular. We would suggest to keep chlorophyll-a in, as that is the variable that we are reporting in the results.

L41 Check the spaces, looks like there is too much space between some words (whole paragraph).

Thank you, I’ve checked the text for double spaces and it seems that it looks like this due to justification of the text (formatting issue).

L42 I am confused about the ARGO floats. Were they deployed for this study? Where is the explanation? If not, why is there such a big focus on these floats? Presenting the data so extensively is not necessary and gives no additional information since the general flow pattern are known already.

Thank you for this comment. We see that we did not explain the ARGO float data well enough to prevent confusion in the MS. Argo float data were retrieved from existing datasets, and not deployed for this study. Data from Argo floats were used to compare the seasonal dynamics in temperature at the two benthic landers with the surrounding areas. We think that by showing that the seasonal trend in bottom temperature measured by the lander compares well with Argo float data, we verify that our measurements are valid and reliable. The Argo float data also serves as material to explain differences in temperature between the two benthic landers.

We will assess where we can improve the explanation for the use of ARGO float in the manuscript and explain why it is useful to take into account.***

L44 It seems like the study was not designed to identify potential food sources, since no food source except SPM was considered. Bacteria, dissolved OM and zooplankton are missing for example.

Thank you for this comment, we see that our choice of words is confusing here. We propose to rephrase this to***:

Benthic fauna stable isotopes were analysed to investigate food web structure at the sponge grounds.

 L47ff It was never really shown that tidal currents cause these things. Where are the correlations?

In figure 9 A and E, you can see that temperature increases when current direction is directed southwestward at HSB and LSB. However, this was not the case for the whole period the benthic landers were deployed. We will analyse if current direction was correlated with temperature for periods as shown in figure 9 (1-9 september), and report this in the manuscript. ***

L51 How could this lead to downwelling?

Downwelling can occur in areas where multiple currents converge and an obstacle, like the Labrador continental shelf obstructs these current from flowing any further. This could result in water being pushed down in the area of convergence.

We leave this statement out of the abstract, as it is not reported as a main finding.***

L53 It's not clear for the reader of the abstract why silicate is important.

To clarify, we will add  “ , which could benefit growth in deep-sea sponges.” behind that sentence.

L58…compared to the low-sponge-biomass site.

Adjusted accordingly.

L65-77 The line of thoughts is a bit confusing.

We see what you mean, we will switch the final sentences of the paragraph around, so that we finish with the statement of the VMEs. The last two sentences will now look like this:

Sponge grounds form complex habitats that provide breeding grounds and shelter for (commercially important) fish species, increasing demersal fish biomass and diversity (Kenchington et al., 2013; Kutti et al., 2015; Meyer et al., 2019; Brodnicke et al., 2023). Finally, they are often classified as Vulnerable Marine Ecosystems (VMEs) as defined by the Food and Agriculture Organization of the United Nations (FAO, 2009).

L75 Not limited to commercially important fish (also see Brodnicke et al. 2023)

Thank you, we will put commercially important fish between brackets, and add the reference to the citations at the end of the sentence.

L79 Maybe also mention deep-sea mining (Wurz et al. 2021).

We will add deep-sea mining in the first sentence of the paragraph, and add the following sentence in the middle of the paragraph (line 89-90):

“In addition, prolonged exposure to elevated concentrations of suspended sediments, i.e. due to deep-sea mining, could adversely affect deep-sea sponges (Wurz et al., 2021).”

 L84 Maybe also mention reduced variation due to trawling (Morrison et al. 2020).

Thank you, we have added the reference to “ (Colaço et al., 2022),” after the first part of the sentences. As Morrison also describes reduced density and diversity at deep-sea sponge grounds due to benthic trawling.

L102 Food availability was explained in the previous sentence. Gamete dispersal is also connected to currents (maybe connect the sentences).

Thank you for your suggestion. We will rewrite these sentences to accommodate this. ***

L123 remove “any”

Done.

L126 I would be a bit careful with extending this too much. The Canadians have very good environmental datasets of their areas.

Thank you for the suggestion. Indeed DFO and Canadian universities and research institutes have good environmental datasets, but year-long recordings of environmental conditions at the seafloor are relatively scarce. Therefore we think this statement remains relevant. Changed the word can for “could” in the last part of the sentence.

L136ff Please mark these areas on the map (with their name) otherwise it is impossible to follow.

Thank you for the suggestion, will adapt this accordingly in figure 1.

Figure 1 There are no general circulation patterns marked in the map. Make the map bigger and add currents and names.

Thank you for pointing this out. By mistake we uploaded an old version of the figure to the MS. We will add currents and names accordingly.

L154 remove “:”

Removed.

L157 Rock boulders are no animal. It doesn’t make sense to list them in the same sentence with the fauna. The sediment plays a critical role for the occurrence for sponges and should be explained further. Please check the description of the sediment. It is not coherent within the manuscript.

Thank you for pointing this out. We see our writing was unclear on the substrate composition at both benthic lander sites. Because we qualitatively assessed the substrate composition in two ways: video images and the rock dredge, the observations got mixed up in the writing.

The videos show more pebbles and coarse sediment at LSB, but with the rock dredge we collected 5 totes of 64L of soft sediment at LSB. We will elaborate on this, and distinguish the methods, in this paragraph. In addition we will first characterize substrate at each site, and afterwards the biology. ***

 L158 It is mentioned later (L613f) that at the LSB site consists of soft, muddy sediment. Figure 2 C&D also shows the opposite.

See comment at L157.

L163 You do not see the slope on the pictures.

Thank you for your comment. In this line we refer to Figure S2 in the supplements. To further clarify, we will adapt this sentence by writing the following:

The west-to-east slope angle was directed downhill (Figure S2 D & H), and north-to-south slope angle was directed uphill at both lander sites (Figure S2 B & F).

Figure 2 Here you define HSB again as “high-sponge-biomass” without the “the” or “site”. Do not abbreviate words you don’t use further (DFO or CSSF). Pictures are difficult to compare because one is vertically downward with a drop camera and the other one more horizontally. Picture C and D could be the same area when you would take a vertical downward picture for example in the right corner of picture A. Maybe try to find other pictures from the same camera system. What distance are the lasers?

Thank you for your suggestions. The abbreviations refer to the locations where the two benthic landers were deployed. One lander was deployed at a high-sponge-biomass site (HSB), and one lander was deployed at a low-sponge-biomass site (LSB). See also our comments on use of abbreviations above. The lasers are 6 cm apart. We will adapt the caption in the following way:

Figure 2: Images of benthic lander deployment sites, at the high-sponge-biomass lander site (HSB; A,B) and low-sponge-biomass lander site (LSB; C, D). ROV image credits: ArcticNet/Canadian Scientific Submersible Facility (CSSF)/Department of Fisheries and Oceans (DFO). Laser points in panel C & D are 6 cm apart.

L184 “MHz” (make a capital M). Maybe state which particle size classes are observed with a 2 MHz ADCP. These settings will only detect really small particles.

MHz is adjusted accordingly. We added the following sentence on particle size class:

The 2 MHz ADCP have a lower particle size detection limit for particles 12 μm in diameter, and a maximum sensitivity for particles of 242 μm diameter (Haalboom et al., 2021, 2023). 

L185 FLNTU abbreviation is not explained. Sensor measures fluorescence not directly Chlorophyll-a.

Thank you for your comment. We couldn’t find a definition of the FLNTU on the website of the manufacturer nor found a definition of FLNTU in the literature. We therefore believe it is probably a brand name of the product (likely meanings something like Fluorescence and Nephelometric Turbidity Units). However, as we are not sure about the precise abbreviation we just mention the name of the product, and adjusted the sentence into the following:

The landers were each equipped with a 2 MHz ADCP (upward-looking, Nortek Aquadopp), a sediment trap, and a combined sensor for turbidity and fluorescence (Wetlabs ECO-FLNTU; Table S1).

L186 I do not understand. The ADCP does not measure a 3D velocity field.

Thank you for this comment, we see that our description has lead to some confusion. In this study we used two single point measurement ADCPs (see reply in L190). We will adapt this text to make this clear.***

L188 Are you sure about altitude?

See comment above.

L190 Where is the complete data? There is no plot of the ADCP data. Are you sure the first bin is okay? Normally you have to discard (at least) the first two bins (~2-3m distance) since the lander is an obstacle in the flow field and it is highly influencing the (close) currents around it which influences the measurements.

Thank you, we caused confusion by not mentioning that the ADCPs used in this study were single point measurement devices. The ADCPs were set to measure at single bin at 1.14 m distance, this single bin had a cell size of 0.75m. As our current velocity (u,v,w) measurements were consistent throughout the deployment, with, as figure 8 shows, a clear tidal component, we think that the benthic lander did not interfere with the velocity measurements.

L192 Later you state that data was transformed using the program MATLAB. Please clarify.

Thank you, we will clarify this was done in MATLAB.***

L192 Is the FLNTU not one sensor?

Indeed, but we deployed a FLNTU in each benthic lander. We will clarify this. ***

L193 The sediment trap is one device.

Thank you, will be adapted. ***

L195 Remove second dot.

Thank you, will be adapted. ***

L195 What’s the end of the collection? Did the last bottle last for the complete cycle? Was the last bottle closed before the lander was retrieved? If it is pulled through the water column while being open, data is not useable.

Thank you for pointing this out. Lander collection happened 1-2 july 2019, and the bottle scheme showed that last sediment trap bottle remained open until 15 july 2019. Therefore, the last bottle closed after retrieval of the lander, therefore the data is not usable and the last datapoint will bediscarded from the analysis. This has no impact on conclusions of the manuscript.

In addition, date of retrieval mentioned at L182-183 will be adapted accordingly to 1-2 july 2019.***

L204 Where are the plots of the CTD transects? Was there no sensor for turbidity? It would be interesting if turbidity around the high sponge biomass area is increased.

The CTD transects are plotted in figure 1B. This will be emphasized more clearly in the text in the revised version***. During the Amundsen 2019 research cruise leg 1b the CTD frame was not equipped with a sensor for turbidity.

L214 You probably measured SPM and not sPOM. Where is the data?

We indeed sampled particulate matter, which was later analysed for POC and PN (POM). Therefore we refer to sPOM. We will adapt text by referring to SPM here and at L240. SPM data will be added to figure S6. ***

L217 Cote et al. 2019 is not accessible (internet page with pdf is not existing anymore). Please give another reference (or a way to access it -DOI?)

Will be adapted.

L219 Explain why two different methods.

The CCGS Amundsen cruise in 2019 leg 1B was the first time that a rock dredge was used on the Amundsen. Therefore the crew and scientists were testing out different modes of deployment of the Rock dredge. At the LSB lander station, the rock dredge collected a lot of material: 5 totes (volume 64L) of very soft sediment were sieved. At both sites (HSB and LSB) substantial amounts of animals were collected. The difference in fauna between the two sites is not caused by the different modes in using the rock dredge.

To accommodate this comment, we will add a sentence on why we used the tow mode at LSB***

L240 POM or SPM?

See comment L214.

L244 Please be more specific about the isotopic measurements. How reliable were the measurements? This data was already published before. Consider removing this part.

We checked with the analytical lab what the margin of error is of the measurement devices for the isotopic measurements, which was ± 0.15‰. The isotopic measurements of the same species were close to eachother (i.e., Geodia), therefore we think measurements were accurate. We will add information on precision and reliability in the text.***

See our reply in major comments section on presumed published data.

L245 Were these particles and swimmer also analyzed?

Thank you for your comment, we counted the number of swimmers from the sediment traps, and identified swimmers into broad taxonomic groups. In L473 – 476 and L676-680 we mention and discuss the results of this analysis. We will add a figure on number of swimmers per bottle in the supplements. 

L270 Where is the data?

Data on the lipid flux can be found in figure 12F, and additional data on unsaturated alcohol/PUFAs/Sterols are represented in Figure S11.

L271 Were specific parts of the fauna sampled or only whole animals?

The fauna was mostly subsampled on board of the CCGS Amundsen, as other labs/researchers used the samples for identification purposes. Therefore only parts of the bodies of the fauna were analysed in isotopic composition. We will clarify this in the text ***.

L276 Was this the same as for the water column samples?

Yes this was the same for water column samples, we will clarify this in the text. ***

L277 What is this standard?

We mean the Vienna Pee Dee Belemnite (δ13C) and air (δ15N) standards, we will adapt this in the text ***.

L280 see L192

See L192 comment for answer.

L281 Why different programs? And which packages?

MATLAB and R were both used because of the collaboration between two authors of which one does not know how to do MATLAB and the other does not know how to do R.  

The R packages will be added to the text. As well as the MATLAB toolboxes used.***

L286 What is the reason for a temporary shift? Did the lander move back again? That doesn’t make a lot of sense in my opinion. Where is the pitch and roll data?

We believe the temporary disturbance of the pitch/roll/heading data was due to moving of the lander, and the lander likely moved back again in place as pitch/roll/heading data was identical before and after this disturbance. We will add a time series figure of the pitch/roll/heading data, which shows this disturbance in the supplements. ***

L289 Explain NTU. Turbidity in NTU is not a concentration.

Thank you, we will adapt this ***.

L293ff Which time-lag and how was the time leg assessed? Where is the outcome of these tests?

Thank you for your comment. Correlation analysis with time lag was done to see at which time lag two variables were correlated with each other. Results are stated in L411 – 419, but we will adapt this to make it more clearly. ***

L297f First it was stated that unfiltered data was used and then it was stated that data was filtered before. What is right?

Thank you for your comment, time series data were indeed smoothed prior to spectral and coherence analyses, and we will adapt the text accordingly.***

L300 Before it was stated that all analyses were performed in R. Please show the results of the harmonic analysis.

Analyses were done in R and Matlab, we will remove statement that all analyses were done in R. A table with results of harmonic analysis will be given***. In addition, Figure 8C shows also the results of the harmonic analysis.

L303 Ice data not shown. Please add to supplements.

Sea ice data are shown in Figure S10 D, but we see we made a mistake with the cross-reference. Will be adapted.***

L312 remove “yet”. If significant show statistics.

We will remove yet and rephrase “significant” to “substantial”.

L308ff It would be much nicer to show the transects.

Thank you for your comment. We think the reviewer means a coloured chart in which transect length is on the X-axis, depth is on the y-axis, and the variables are given with a color code (oxygen, temperature, etc). However, as we only have 5 CTD measurements per transect, and the transects are ~200 km long, a large area would need to be interpolated. Therefore we chose to combine the transects into one figure.

L316 Sentence is more discussion.

Thank you will be moved to discussion. ***

L324 It makes no sense to give a range for a mixture of nutrients. The thermocline is not in the Figure.

Indication of thermocline will be given in figure.*** Range of mixture of nutrients will be removed.

L328 Remove “to a lesser degree” or give a more specific indication.

Thank you, we will provide a more quantitative description of the relative increase in nutrient concentration. ***

L334- 338 This seems weird here. It is not your results.

We understand the confusion. These paragraphs describe the results from the ARGO float data. To accommodate this, we will make a separate subheading that indicates we are talking about Argo float data.

Figure 3 In A: Why pressure and not depth? All other graphs have depth as an y axis. In general, use the same units (Temperature or potential temperature/ salinity (PSU) or practical salinity). Both graphs have the same legend, combine and make uniform. Explain grey lines in figure C.

Thank you, pressure will be replaced with depth. Figure will be adapted that the same units are used. Legends will be combined. Grey lines will be explained in caption. ***

Figure 4 X-axis description is cut off. Make uniform capital or not (e.g. “Depth (m)” and not “depth(m)”). Maybe combine the legend for region. Its rather confusing. In general: Why not show the data as a transect from west to east?

Figure will be adapted, x-axis cut-off will be removed, we will use same y-axis label. The legend for region is not combined because figure 4A is a line (from oxygen sensor), and figures 4B and C are point measurements (from Niskin bottles). For comment on transect see reply at L308.

Figure 5 A: Is this really important to show? I think the general currents could be shown in Figure 1. This figure is overloaded with data that is not further used and has some explanation issues. From which depths are the currents? The yellow arrows are not visible. And it is not clear what the difference with the blue arrows is (StE not explained). The dots are above the arrows, which makes it hard to see, which currents are at these positions.

Thank you for this comment. Indeed the general currents are roughly known, but we think that providing a description of the currents in the area that is backed up by data shows a more powerful message. We agree that we haven’t explained this figure well enough in the text, and will provide an extra text that explains the figure.

This is drifter data, which shows mean surface flow over the period of 1995 – 2020. We will emphasize this in the text. We will explain StE, move the abbreviatons HSB/LSB out of the figure to make it more clear what kind of currents are on the locations of the HSB and LSB landers. ***

B: I don’t really understand why this data is important to show. On the ARGO float locations from Figure S3 it seems that the ARGO float “HSB” is covering both stations and “LSB” is south of both stations. Make the y axis the same, so it is better comparable.

See major comment section on why we used ARGO float data in our study. We think that showing that the seasonal cycle in temperature over the year matches between the ARGO float data and our lander data, shows that our lander data is valid.

L361 Northward velocity was directed southward sounds wrong; I would describe it in a different way.

Thank you for this comment, we see it might be confusing, but Northward velocity (v) is an international renowned parameter unit (Parameter Database (ecmwf.int)). When northward velocity has a negative value, then the direction of the current is southward.

L363 is upward possible? Where is the water coming from?

The vertical velocity was slightly upward, due to the slope angle at which the lander was positioned. The north-to-south slope angle was slightly upwards at both lander locations. Since Nortward Velocity (v) was mostly directed southward, vertical velocity was directed slightly upwards.

L365 What is meant with bottom currents? Horizontal currents or the separate velocities? If not, horizontal currents please calculate them.

Yes horizontal currents, will be emphasized in the text.***

L367 Remove “signal” and talk about pressure (or water depth).

Will be changed accordingly.***

L371 Give table with tidal analysis.

Thank you, we will provide this.***

L395 Why not recalculate the ABS to decibel sound pressure level to make it comparable to other data? Please show the complete data set of all bins (at least in supplements).

We here only aim to make a qualitative comparison of ABS between the landers. Recalculating ABS to decibel sound pressure level is not straightforward with the set-up for the ADCPs that we used. Therefore, we suggest to keep ABS in echo intensity (as is also used in many other benthic lander studies).

See reply in L190 for reply on the bins.

L400 Which turbidity? From ADCP or FLNTU?

Thank you, here we refer to turbidity measured by the FLNTU. We will adapt the text to clarify which turbidity is meant here.***

L405 Reference to Figure S7 is wrong.

Thank you for pointing this out, we will review the cross-referencing. ***

Figure F8: cut off the period when the lander was lifted, then you would very likely remove the peak in fluorescence and turbidity.

We think that this comment refers to figure 10. The data in this figure was cut-off at 00:00 1^st of July 2019, while the LSB lander was retrieved at 08:35  1^st of July 2019, we therefore presume peak in fluorescence and turbidity was not caused by lifting of the lander.

L412 Please give the complete correlation, not only r² and show the correlation plot.

Thank you for this comment, we are not sure what is meant with complete correlation. We will add the p-value to the text, and show correlation plot in the supplements. ***

L413 ABS was explained before à ABS is also a measure for turbidity. It's confusing if ABS and turbidity are used but describing data from different devices. In general, I would not talk about “signal” but rather what the signal stands for. Every sensor gives just a “signal”. In L420 another “signal” (in the data) is mentioned. Very confusing.

Thank you for this comment. We will emphasize which sensor is used for turbidity, and review the use of signal in our text.***

L417 Show and give complete correlations.

We are not completely sure what is meant with complete correlations, but we will add the p-value to the text, and if a correlation was negative or positive. ***

L421 Figure S9 does not show ice cover. If S10 is meant I still don’t see how the ice cover is influencing Chl-a. Why is the data for this paragraph only showed in the supplements?

Our apologies, we indeed mean S10. This data was put in the supplements, as it should only be seen as a qualitative estimate of ice cover in the area directly above the landers. In the text we mean that during the spring peak of chl-a at both landers (start April), ice cover was higher at HSB, which could reduce the chl-a quantity in the surface waters and therefore reduce chl-a signal also in bottom waters. However, we see that this is more a point for discussion. We will clarify this in the text, and move part of it to the discussion. We will colour the period of peak-chla in the Ice cover data chart to emphasize differences between sites.

Table 1: What statistics are shown? It seems like these are just mean values and their SD. HSB and LSB are again defined differently than before. Bottom current speed= horizontal current speed? ABS is also a measure for turbidity. How is along and across slope velocity calculated?

We indeed only report mean values and SD. We will review definition of HSB/LSB here. Bottom current speed = indeed horizontal current speed, we will adjust text accordingly. We see we didn’t explain cross and along slope velocity, we will explain this in the methods, thank you for pointing this out. ***

Figure 7 I do not think that this is the best way of showing the data. First of all, progressive vector plots might be useful to show transport for short time periods, but after more than some days these plots are invalid. At every new position a parcel of water experiences different forcing, that are not comparable to the initial position. Especially factors like temperature are very different at each position because of the general water mass distributions. I do not believe that you can extract any useful information from this plot. Additionally, it was shown before that tidal currents are influencing temperature which are not even included in this figure. Figure 7C&D are on the other hand a nice way of showing the current direction distribution, whereas it is not clear why not the same way of visualization like for example current speed was used (maybe not useful).

We see your point. We will remove the Progressive vector plot, and replace it with a scatter plot, that also shows the residual current direction.***

Figure 8 C is difficult to see the difference between temperature, ABS and Chl-a. I would suggest different colors and to move the legend to the right side.

We will make the figure more clear and move the legend. We will keep the same colours as these are used throughout the whole MS.***

Figure 9 This is in principle a more acceptable that Figure 7. Turbidity (counts) for example is depending on the current speed which is not at all visible in this graph. Turbidity is not differing on small scale spatial patterns as it seems in this plot. The same for temperature. Temperature depends on the current direction of the water mass with a certain temperature. This graph suggests that temperature varies on a small spatial scale. Data needs to be shown differently.

See comment Figure 11.

Figure 10 Nice way of showing the data. ABS should also be called turbidity or any other coherent way (acoustic vs. optical backscatter). Chl-a needs to be with a capital C. The two different turbidity signals should be used to describe the particle sizes of the SPM (peaks in spring= bigger particles because not seen in the ADCP data). The data of the last days before the retrieval of the lander can probably be cut off, since it is likely showing the time period when the lander was lifted, or other bottom touching activities were nearby.

We will adjust definition of ABS, capitalize chl-a, describe particle sizes in the text. The time period in when the lander was lifted was already cut-off from the data.

Figure 11 This makes Figure 9 obsolete. Much nicer way of showing the data. Same comments for the y axis description like before. Again, comparing the turbidity sensors here would help with the particle description à Fine sediments are dependent on tidal currents and are peaking with a high horizontal velocity when currents are in a certain direction (I assume when currents are coming from the shelf, current direction is missing here). A correlation matrix would help here as well.

We will add current direction here to this figure, and move figure 9 to the supplements. This reduces the number of panels/figures in the MS, which was also a wish of the other reviewer.

L457-495 Remove everything that is redundant from the already published data (Wolff et al. 2020). Or mark it better as a repetition if needed to be included.

See our reply in major comments section.

L486 “indicated a lower trophic level..” is for the discussion.

Thank you, we will remove this. ***

Figure 12 Most of this figure was published already.

See our reply in major comments section on presumed published data.

Figure 13 This is not a bi-plot (includes sample data and variable data). The plot was already published before and is therefore redundant. Other food sources are missing (zooplankton, dissolved food).

See our reply in major comments section on presumed published data.

In this study we did not take samples on zooplankton or dissolved food, we rephrase our papers in a way that we do not refer to the data as food sources, but more to isotopic signatures or food web structure.

L499 remove “more specifically.”

Will adjust this.***

L521 “Our ARGO float profiles”. The ARGO data is not from this study, this is misleading.

We will adapt this in the text by referring to Argo float profiels/data, and remove the word “our”. ***

L529 How was this calculated?

We divided distance between the landers (~130km) by the lag at which correlation was most significant (5 days). This gives the speed at which the water parcel presumably moves, and this aligns well with the residual bottom current speed at HSB.

L532 Why is the water temperature explaining salinity?

In our study we only measured temperature on the benthic landers, and we do not have info on the salinity. In our study we therefore used temperature to investigate connectivity between the two sites, and in Sutcliff et al, and Myers et al, they looked at salinity to look at connectivity between Hudson strait outflow and Labrador Shelf.

We will review this paragraph to prevent confusion.

L540 What is in-situ remineralization supposed to be? Remineralization in the Baffin Bay? Why is there more organic matter in the deep-water that is demineralized? Is it not primary production or terrestrial OM that is demineralized? Normally deep water has higher nutrients but not higher OM.

Thank you, we see that there are some words missing in this sentence. It is indeed OM which is remineralized, we will adjust this sentence.

L567 The tidal amplitude? Or tidally driven horizontal currents?

Here we refer to the tidal amplitude. We will clarify the sentence.

L568 This is highly dependent on the sponge species. Are these the same sponges?

White et al, 2003 investigated glass sponges, and here we looked at sponge ground what was a mix of glass sponges and demosponges. So partly the same sponges. We will add information on this in the text.

L578 Why is the ARGO float data important here?

See comment L42.

L584 I would expect current direction to have an effect but not tidal currents.

We indeed mean current direction, we will clarify this.

L587ff Currents at the shelf break are likely very different than on the shelf and this is likely not a valid argument even though the time frame would maybe allow assumptions like this.

Thank you for your comment. We see it might be a stretch to assume transport of 5 km. We will look into this in combination with the tidal ellipses analysis (see reply major comments).

L606 Logically not completely right. Higher resuspension is clogging the sponge due to more particles in the water column. The particles are also retained within the sponge and are not removed from currents.

An important point is also that resuspension is delivering food since bacteria and organic substances are binding to particles, which will be eaten. Higher turbidity is associated to higher amounts of bacteria.

Thank you, we agree higher turbidity could enhance food availability, but could also cause clogging of the sponges. We will adjust the text.***

L614 contrary to what was mentioned before and also pictures show something else.

See comment L157.

L616 Check your data. Is there a time lag between high currents and high turbidity? Then SPM is likely coming from somewhere else. Otherwise, OM might be collected between the sponge (and other fauna), like mentioned in the introduction.

Thank you, we will check this in our data.

L618 I don’t agree that it prevents smothering.

We here mean smothering from a layer of sediment, or sedimentation. We will adjust this to make it more clear what we mean.

L619 It is known that OM/bacteria are binding to particles and can act as a food source. Third? What is first and second?

Thank you, we will add this.

Third will be removed.

L635 repetition.

Thank you, we will rephrase this sentence that will avoid this repetition.

L638 Show prim. production.

We did not measure primary production in our study, but we will add the chl-a profiles to figure 4. ***

L650 Which export? I assume from the Hudson Strait. Why is the export tidal?

We mean export from surface waters to the seafloor. We mean that chl-a showed a tidal periodicity during this period, and we see that back in Figure S9B.

L677 Why would that have an influence? They produce offspring when they have enough energy but that can happen to any time of the year.

Thank you, we are not sure what the reviewer means with this comment. Does the reviewer review refer to zooplankton offspring here?

L689 Was not Chlorophyll-a measured here and not fluorescence?

Yes, thank you, adapted.

L706 It seems that a more important factor is the horizontal distance to the food source from the Hudson Bay or Baffin Bay.

Thank you for this comment. We think that, as primary production is likely comparable between the two sites (Frajka-Williams et al, 2010), food sources are comparable between  the two sites and the water column dynamics play a more important role. In addition, surface waters of the Northern Labrador Sea or Baffin Bay are quite distant from the study area and therefore unlikely serves as a source of organic matter.

L708-730 This paragraph is not okay in this way. There is no explanation about the estimate of the sponge biomass by the ROV or trawl method, which is a number that is not easy to obtain. Sponge biomass of a trawl cannot be referenced with personal communication. The image analysis was already published in an earlier report and also discussed there. The paragraph is based on a comparison of these results and the data from personal communication and there are no results concerning this paragraph in this manuscript. Therefore, this paragraph should be removed. The respiration potential can be discussed in one sentence when needed but it should be clear, how estimates are made or from where numbers were taken. I would recommend concentrating on the long-term measurements of environmental factors, which is the strength of this manuscript.

Thank you, we tried to give an estimate of OM utilization at the two sites, but agree it is quite speculative. We will remove this paragraph.

L731- 767 The isotope data was already discussed in Wolff et al. There are some more points in the discussion of this manuscript, but they are not connected to any new results. The author has to consider if the isotope part should stay in the manuscript. Again, I think there is no substantial gain in presenting the same results again and they are not relevant for the time-series of the environmental data.

See our reply in major comments section.

L776 Please rephrase the sentence. Why should primary production alone be a good predictor, this was never mentioned before? Maybe better state your positive result: Primary production in an area that is connected by water currents is important for the delivery of food to the sponge area. Please reflect these results also in your title.

Following comments from the other reviewr as well, we will carefully review our conclusion section.

Figure S1 and S2 Is this really necessary? Nothing was really done with this data.

Figure S1 and S2 serve as background information on the slope angles, which explains vertical velocity for instance. We would prefer to keep them in place.

Figure S3 Seems like the “HSB- ARGO” was above both areas. And the “LSB-ARGO” was mostly south of the study area.

Thank you, we chose the cut-off point of the HSB/LSB regions at the latitude of the LSB lander, as the general current direction is southward and therefore HSB-Argo locations cover the area where the three currents converge, and the LSB – Argo locations cover the area south of it, and where less sponges are present.

Figure S4F The buoyancy frequency shows a very interesting peak in ~230 m above the high sponge area. This is a bit higher than the sponge ground itself but might be very important for the food delivery. This should be mentioned.

Thank you for this observation. The buoyancy frequency shows a peak in the profiles of the CTD casts taken on the shelf, likely related to the boundary layer. CTD transect above the HSB lander shows also a peak in BVF, but more close to the Cold-Intermediate-Layer. We will mention this in the results***

Figure S5 and S6 remove à is already published data.

See our reply in major comments section.

Figure S7 y-axes are different, which makes it hard to compare. This figure is not referenced in the text. Refer to it or remove it.

Thank you for your comment, we want to compare the temperature and salinity between landers in each respective water layer (350 – 450 m depth at HSB; 550 – 650 m depth at LSB). If we would adapt the y-axes, the LSB water layer (550 – 650 m depth) would shrink considerably in size and difference between HSB and LSB would not be visible anymore. We therefore suggest to keep the scale of the y-axes different between the subpanels.

We will reference to the figure in the text. ***

Dear reviewer,

Thank you for your comments and suggestions on our manuscript. As the reviewer suggests, we will carefully review our discussion section, use of figures, and write a new conclusion section. Below we reply to the reviewers' comments more extensively:

SECTION 2: METHOD

- L. 139: "This region is known for intense mixing and water mass transformation (Dunbar, 1951; Kollmeyer et al., 1967; Griffiths et al., 1981; Drinkwater and Jones, 1987; Yashayaev, 2007)"



-> The Yashayaev paper is not about mixing in Hudson Strait and should be removed.

Thank you for this suggestion, reference will be removed.

- L. 140: "four distinct flow components can be identified (Figure 1A...)"



-> There is no flow component in Figure 1...

Thank you for pointing this out. By mistake we uploaded an old version of the figure to the MS. We will add currents and names accordingly.

- Figure 1: There is a big void at 61N. Is it because there is no sponge or because it was not sampled? (e.g. it is a closure?). If the latter, it should be clearly stated that the sampling is not representative. Overall, it looks like there is sponge almost everywhere that was sample (very little number of red dots).

- Figure 1 (related to point above): It looks like the LSB site was just not sampled. How can you tell it has low biomass?

The sponge biomass data was retrieved from Kenchington et al., 2010. It seems that the data presented in Figure 1 does not contain all the sponge biomass data, and that an error occurred. We will review this.

- L. 186: What is the vertical resolution of ADCP bins?

Thank you for this comment, we see that our description has lead to some confusion. In this study we used two single point measurement ADCPs. We will adapt this text to make this clear.***

- L.286: "ADCP sensor at HSB were shifted for a small period of the deployment, implying the lander was occasionally moving a bit"



-> Like sea floor displacement? Or oscillation and coming back? I have trouble to picture what you mean and what could have caused such a displacement.

The sensor showed small pertubations in the middle of the time series. We will elaborate on this, and show the heading/pitch/roll data in the supplements.

- L. 304: "extracted from weekly ice charts (Canadian Government, 2022)."



-> Please verify is this is the correct way to cite this document.

Thank you, we will verify this.

RESULTS:

- L. 316: "The temperature changes from cooling to warming with depth signify the Cold Intermediate Layer (CIL)."



-> This is a weird sentence. Can you expand?

Thank you, we mean that a cold intermediate layer is visible, we will adapt the text. ***

- L. 322: "The bottom oxygen concentrations at the lander stations were, for both transects, relatively depleted compared to the deep water CTD transects at similar depths."



-> 250 uM/L is still pretty high!

Thank you, we will add this to the text.

- Figure 3 caption:



-> is it the "surface circulation"?

We think the reviewer means figure 5. Yes it is surface circulation we will add this.



-> What are the dots and the lines in panel B?

Dots are data from Argo float profiles (averaged over the water layer), the lines are smoothed fit, that shows the seasonal pattern. We will add this to the captions and how this was calculated to the methods.

- Figure 12: Would it be possible to have error bars?

We can add error bars (SD) the dotted lines (the mean), not to individual measurements.

DISCUSSION:



This is that should be revised in depth.

We will review the correct use of cross-referencing extensively.

- L. 519: "Our findings confirm previous work which showed that Irminger Water is gradually cooled while moving southward by mixing with the Baffin Island Current (Cuny et al., 2002)"



-> How your study confirms this?

Thank you for your comment. Cuny et al, 2002 show that Irminger water is gradually cooled moving southward by misign with the Baffin Island Current. Our study does not confirm this, but we do see the same pattern. Namely warmer waters at HSB (containing IW), which cools moving southward in the Labrador current. We will rephrase this sentence and choose more appropriate wording.

-L. 521: "our Argo float profiles..."



-> Argo profiles are public. I would rephrase, these are not "yours".

Thank you, we will adjust this.

- L. 523: "For example, the 350-450 m depth layer in the HSB area regularly showed presence of Irminger Water (>4.5 ºC), while Irminger Water was only sporadically measured at LSB (Figure 5B). Irminger Water might therefore be cooled and freshened in the area around HSB due to convergence and consequently mixing with the Hudson Outflow and Baffin Island Current."



-> I don't understand the reasoning. How do your discard different advection patterns? Or the fact that the moorings are at different depths?

We infer this from the argo float data, in which we compare temperatures from the same water layers. The advection patterns are retrieved from the drifter data (Figure 5). We will try to clarify our reasoning with this information.

1. 528: " This time lag corresponds to an along slope velocity of 0.3 m s-1, which is close to the mean bottom current speeds measured at HSB (0.25 m s -1) and on the Labrador Slope (0.11 – 0.23 m s-1 ; Lazier and Wright, 1993). This supports earlier findings on the Labrador Shelf that found a connection between the Hudson Strait outflow strength and the southern Labrador Shelf water salinity (Sutcliffe et al., 1983; Myers et al., 1990)"
   
   
   
   -> Again, I don't see how the former support the latter.

Thank you, we will review if this statement is still relevant in addition to L 523.

1. 536: "These observations are thought to be related to the sources of the bottom water and circulation. Thus, intermediate water flows from Baffin Bay via the Davis Strait southward along  the  continental  slope  (Curry  et  al.,  2014).  This  water  mass,  referred  to  as  Baffin  Bay  Water (BBW), contains higher nutrient concentrations (e.g., 41.6 ± 25.5 μM Si(OH)4, 18.5 ±2.6μM NO3-; Sherwood et al., 2021)  due to  in situ  remineralization of deep water circulating in the Baffin Bay basin (Jones et al., 1984; Tremblay et al., 2002; Lehmann et al., 2019). BBW mixes  with  water  masses  on  the  Labrador  Shelf  and  Slope  and  Hudson  Strait  outflow  water  while flowing southward along the Labrador Slope, resulting in lower nutrient concentrations at the LSB compared to the HSB (Figure S4). The absence of high nutrient concentrations at the  shelf/deep  CTD  station  at both sites  supports  this  interpretation."
   
   
   
   -> I am not sure if your study shows this...

See reply L 557.  We will provide a hypothesis more carefully.



L.557: "Namely, such an efflux from the sediments would be quickly advected away by the high bottom tidal currents, while nutrient concentrations were elevated up to 100 meters above the bottom (Figure 4 B & C)"



-> Can you rule out advection from upstream? sediment efflux at the site would be advected, but efflux from upstream can be advected at the site of the measurements... I don't think that your explanation is satisfactory."

We agree that advection upstream is likely the source of these nutrients. But sediment efflux is still unlikely as this efflux would have to be very high to increase nutrient concentration of a water layer of 100 meter high. We will state our idea more carefully, that the source is uncertain, but will mention that sediment/sponges is an unlikely source. As we can estimate the retention time of a water parcel in the area, we will give an estimate on how high the efflux needs to be to increase nutrients in a water layer of 100 meter high.



L. 563: "This study provides the first concurrent long-term measurements of hydrodynamic and environmental conditions at a high-and low



sponge-biomass site."



-> You mean specific to these sites or worldwide?

Thank you, we believe this is the first worldwide assessment of two concurrent measurements of environmental conditions at a high- and low sponge biomass site. We will emphasize this.

1. 
   
   
   
   578:  "Temperature increased gradually from summer until December, which is measured previously on the Labrador upper slope and attributed to Irminger Water (Cuny et al., 2002)."
   
   
   
   -> The temperature increase is attributed to the IW? Can you clarify? A larger proportion of warmer IW throughout the year?

Indeed, this is shown by Cuny et al. 2002, We will clarify this statement.



L. 589: "As the lander was placed ~500 m from the shelf break (Figure S2C&D), and bottom water could be transported ~5 km in the north-easterly direction in one semidiurnal tidal cycle (Figure 9A), this means that colder bottom water is transported on to the Labrador Shelf



from beyond the shelf break to the HSB lander site."



-> I am not sure I follow this reasoning... The temperature is colder above (on the shelf) compared to the slope. You say here that colder temperature is transported on the warm shelf...

Thank you for this comment, we will remove this statement because it seems to speculative.



L. 593: "Therefore, although higher variability in bottom water temperature has been attributed to the presence.



of internal waves at other sponge grounds (Roberts et al., 2018; Davison et al., 2019), we attribute the variability in our study area to tidal-induced cross-slope transport of bottom water."



-> The main generation mechanisms for internal waves are the tides. So in this case do you mean the barotropic tide? If so, you would need to demonstrate that there is not internal waves generated (or propagating) at these sites.

Thank you for this comment, we will add an analysis on comparing tidal ellipses from the benthic landers with the barotropic model data, and from that elucidate if baroclinic or barotropic tides dominate in this area. We will adjust the text accordingly to the outcome of this analysis.



L. 612-628: This paragraph is convincing (strong current = better feeding opportunities). All the rest that precedes seems highly speculative.

Thank you, as stated above, we will assess those paragraphs on how to reduce speculation.

1. 
   
   637: "Our CTD profiles show elevated chl-a concentrations in the CIL (~150 m depth)"
   
   
   
   -> It would be nice to have the chl-a profile in figure 3A.

Thank you, we will add this to the figure.

1. 638: The fact that primary production rates are comparable above the two lander station site s (Frajka-Williams and Rhines, 2010), suggests that differences in primary production alone are insufficient to explain the differences sponge biomass between regions."
   
   
   
   -> This statement is problematic. Frajka-Williams paper is not about these 2 lander sites but rather compared the Northern vs Central Labrador Sea. 

Thank you for this comment. Indeed the Frajka-Williams paper focusses on the Northern vs Central Labrador Sea. However, in figure 2a of that paper, it shows increased mean chlorophyll concentrations of phytoplankton bloom in surface waters at both the HSB and LSB landers. Therefore, we argue that primary production alone is not sufficient to explain differences in sponge biomass between the two lander sites. We will adapt the text as follows:

The fact that primary production rates are comparable above the two lander station sites (see Figure 2 A in Frajka-Williams and Rhines, 2010), suggests that differences in primary production alone are insufficient to explain the differences sponge biomass between regions.

L.645-669: This whole paragraph is highly speculative. The authors talk about seasonal sea ice and spring bloom dynamics as if there was no interannual changes (while they are actually quite large!). The whole paragraph needs to be better backed up by the literature review, or by new data.

In this paragraph we tried to describe and explain arrival of chl-a in bottom waters in early March, which is surprisingly early for this region. Indeed there is yearly variability in sea ice and spring bloom dynamics, but the start of spring bloom in this region and the Labrador shelf has been studied. We think we can still make a statement that chl-a arrived earlier than the timing of the onset of the spring bloom. A recent paper on spring bloom initiation on the Labrador Shelf (Cyr et al., 2023) shows that standard deviation of spring bloom initiation is around 21 days. Figure 2B in Frajka-williams & Rhines 2010 shows bloom onset is around June. This gives us an indication that spring bloom starts later than chl-a arrives at the seafloor, which has likely to do with stratification.

We will review this paragraph and phrase our ideas more carefully, and add information from Cyr et al. 2023 and, if possible, other papers to it.

CONCLUSION



-> So what? how this study contributes to science? Can you highlight the main findings? The results that are currently recalled in the Conclusion looks to me things that were already known...

We see that our statements in the conclusion are to generic to pin point what our study has contributed to science. We will make this section more specific to the outcomes of this study.

MINOR COMMENTS:



- L. 184: mHz -> MHz

Thank you, adjusted.

- Figure 4: xlabel is cut.

Thank you, will be adjusted**.



- L. 505: "is known" -> "are known"

Thank you, adjusted.