Review of  The Scotland-Canada overturning array (SCOTIA): twenty years of meridional overturning in the subpolar North Atlantic

By Alan D. Fox, Neil J. Fraser, Kristin Burmeister, Sam C. Jones, Stuart A. Cunningham, Lewis A. Drysdale, Ahmad Fehmi Dilmahamod, and Johannes Karstensen

General comments

The manuscript introduces a new transatlantic section, the Scotland-Canada overturning array (SCOTIA), to compute subpolar overturning transports combining existing moorings from the OSNAP array, CTD profiles, Argo data, satellite ADT, and GLORYS reanalysis. One of the main advantages of this new estimate is that the monthly gridded merged product to analyze AMOC variability on SCOTIA extends the subpolar AMOC record back in time before the OSNAP measurements started providing the possibility of analyzing long-term AMOC variability and trends at subpolar latitudes.

The extended record at SCOTIA is compared with the twenty year (2004-2024) AMOC record at 26N from the RAPID array and with the ten year record (2014-2024) from OSNAP in the subtropical and subpolar North Atlantic, respectively. Five key metrics (maximum of overturning streamfunction, density at which this maximum occurs, northward heat and freshwater transports, and density flux) are analyzed to validate SCOTIA against OSNAP on seasonal and longer time scales. SCOTIA provides an overturning structure and variability consistent with that from the full OSNAP array, offering an alternative configuration for observing subpolar AMOC and the associated heat and freshwater fluxes, allowing near-real-time updates to be generated based on the latest Argo and satellite data following mooring recovery. Given the logistical and funding difficulties often associated with maintaining sustained ocean monitoring systems, these results offer and alternative way of estimating AMOC transports in the subpolar North Atlantic with a focus on decadal and longer variability, and to extend the AMOC record back in time before OSNAP started. One of the main results from SCOTIA is that subpolar AMOC varies on pentadal to decadal timescales with an amplitude comparable to that observed in the subtropical North Atlantic.

These analyses are very valuable and useful for the AMOC community. The manuscript is well written and well organized. My impression is that the manuscript could benefit from more discussions or clarifications on to what extent SCOTIA relies on the actual mooring design at OSNAP, and if possibly SCOTIA could serve to inform alternate, or minimal, OSNAP array design. Please find some specific comments below.

Specific comments:

Line 50: it may be useful to mention here that the resulting blended gridded product has monthly temporal resolution (compared to the daily resolution from OSNAP). Line 5 and other places: I am hesitant if SCOTIA should be considered as a "new subpolar observing array" or as an alternate blended product based on combinations of in situ hydrographic profiles, moorings, satellite altimetry, and reanalyses products.

Lines 125-127: Are there are attempts to regionally validate GLORYS temperature, salinity and velocity data near the shelf?

Lines 233-243, and other places: Given that all OSNAP mooring data are included on the SCOTIA estimates, I suggest adding some discussion on how independent are OSNAP and SCOTIA time series, and how this may impact the interpretation of the correlation coefficients.

Could tests for minimal array design be done?

Line 235: please clarify, SCOTIA MOC shows higher variability compared to OSNAP MOC on which time scales?

Figure 4a, 4b (MOC or maximum overturning and density of maximum overturning, respectively), Line 245, and other places: please mention which correlation coefficients are statistically significant. SCOTIA and OSNAP (MOC) records agree fairly well with a correlation of ~0.66. The correlation considering the density of maximum overturning diminishes to nearly 0.3, implying that a small portion of the variance is shared between the two compared records. Visual inspection of the time series shown in 4b indicates that sometimes the records are out of phase.

I suggest adding more interpretation for the comparisons in Figure 4a and Figure 4b.

Figure 3b: There is little discussion in the text about the comparisons of the MOC seasonal cycles from SCOTIA and OSNAP. Is it possible to separate the seasonality from the interior transport and from the Ekman component?

I suggest adding a comparison of the seasonal cycles from the heat and freshwater transport to better understand the overturning variability from both estimates at seasonal timescales.

Line 252: correlations for the density of the maximum streamfunctions are slightly lower that this value.

Lines 264-274: How dependent is SCOTIA estimate on the maintenance of the full mooring array in the future? Could SCOTIA be used to inform a minimal or reduced array design for maintenance in the future?

Figures:

Figure 1: it may be useful to add schematics of the main ocean currents and pathways to better illustrate the dynamical connections between the arrays in the North Atlantic.

Figure 2: panel b: I suggest using different colour for the CTD (black dots) and mooring (vertical black lines) profiles respectively, to better appreciate the spatial and temporal distribution of the different data sources.

panel d: please indicate in the figure caption what shading represents.

Figure 3: please add in the Figure caption that these comparisons are made for the OSNAP period 2014-2022. How does the SCOTIA MOC seasonal cycle for the full period since 2004 compare to the estimate between 2014-2022?

Figure 4: Please indicate in the figure caption what blue shading represents.

Technical corrections:

Line 4: remove “the” between “amplitude” and “interannual”

Line 13: I suggest adding “in the upper ocean” for the transport of warm water northward or an indication that this text refers to the upper limb of the AMOC.

Line 21: I suggest highlighting that these are continuous daily observations

Line 29: revise referencing --> (McCarthy et al., 2025).

Line 51: To generate the corresponding velocity field…

Line 397: Please remove one “either”

General comments:

This paper introduces a new subpolar overturning array, SCOTIA. The array uses part of the OSNAP mooring array, but runs south of Greenland and additionally makes use of CTD, ARGO and reanalysis data. One of the added values of this array is that it allows to reconstruct subpolar overturning for the last 20 years, and therefore to compare it over this entire time period with subtropical overturning at RAPID. Additionally, it could provide a lightweight option for continued subpolar overturning monitoring.

This is a high quality paper, and the 20 year timeseries of subpolar overturning at the OSTIA section will likely be of interest for everyone working on overturning in the subpolar north Atlantic and meridional connectivity of the AMOC. I found the methods section very comprehensive, and the graphics very good. I would recommend the paper to be published with minor revisions. My comments can be found below, divided in minor comments and typos.

Minor comments

L220 : What are “these processes”? Does it refer to processes mentioned a few sentences before? Maybe repeat in the sentence to clarify.

L244: I do not understand what is meant with “increased power of the two sections”. Do you mean that there is added value in using the two sections rather than one? Maybe reformulate / clarify what is meant.

L251: Here, the authors write that heat fluxes show the least correlation of all metrics studied between scotia and osnap. However, the correlation given here (r=0.352) is higher than the correlation found for the density of maximum overturning (r=0.320, L245), which was described as “showing many similarities” between the two time series

L373-379: I am a little bit confused here about the discussion on the definition of LSW with a fixed density range. The authors first say that according to fig 9, the most prominent feature of low frequency overturning at SCOTIA is increased southwards transport in the denser LSW class between 2014-2022. They link this to results from Yashayaev et al 2024 showing LSW formation over that same time period. They then move on to ask why results in fig 8 suggest that the most variability takes place instead in the deep overflow waters, and attribute this discrepancy to the definition of LSW with a fixed density range. I’m a bit confused because this fixed density range is also applied in fig 9, where we clearly see the increase in transport in the denser LSW class. Why is the fixed density range an issue for fig 8 but not fig 9? I must be missing or misunderstanding something, and it would be nice if the authors could clarify this paragraph.

L 399-403 The authors identify an anomalously high overturning at SCOTIA in 2016-2020, which they associated with increased Labrador Sea Water southwards transport. Is that a result that necessitated the use of SCOTIA, or could it also have been inferred from the existing OSNAP observations (given that the time period is covered by OSNAP)? If so, it would be nice to more clearly state what the role of SCOTIA (the longer timeseries allow to put this higher overturning in a broader context?) is in this result.

L409 Similarly, I would appreciate a little more information on the role of specifically the use of SCOTIA in this result. Is this a result that is made possible specifically by the use of SCOTIA, because it provides a longer timeseries for instance, or is it a side result, that is demonstrated here with SCOTIA but could have been deduced from other available data? I think more clearly connecting these results with the use of SCOTIA would strengthen the case for the usefulness of this new array.

L415 Here the authors say that the SCOTIA array, because of its position, cannot discriminate between convection in the Irminger and Labrador basins. Could this impossibility to discriminate between different basins create issues in the previous discussion, which focuses on the Labrador Sea only?

Typos / reference issues

L4 : Either an “of” is missing or there’s a “the” too much

L8: “subtropics” should likely be “subtropical”

L25: should “be” be  “been”?

L50: Missing “the”

L51: “generate corresponding the velocity field”, the “the” is misplaced

L54: “extending back 2004”, is a “to” missing?

L127: The reference uses the first name, not family name of the author. The issue is also present L505 in the reference list. For that item, the first names of authors and the first letter of their family name are listed rather than the contrary.

L222: An “are” seems to be missing

L228: “Figs. 3” should be “Fig. 3”

L275: Should there be a 4.1 here? This part has heading 4. , then several paragraphs, then heading 4.1(L319), but no 4.2 . Adding a 4.1 heading here might also help the reader in identifying the arguments made in each part.

L349 “Susan Lozier et al, 2022” should be “Lozier et al 2022”