Remineralisation changes dominate oxygen  variability in the North Atlantic

Sanders, Rachael N. C.; McDonagh, Elaine L.; Lauvset, Siv K.; Turner, Charles E.; Haine, Thomas W. N.; Goris, Nadine; Sanders, Richard

doi:10.5194/os-22-225-2026

Articles | Volume 22, issue 1

https://doi.org/10.5194/os-22-225-2026

© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/os-22-225-2026

© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 22, issue 1

Research article

| Highlight paper

|

20 Jan 2026

Research article | Highlight paper |

| 20 Jan 2026

Remineralisation changes dominate oxygen variability in the North Atlantic

Rachael N. C. Sanders, Elaine L. McDonagh, Siv K. Lauvset, Charles E. Turner, Thomas W. N. Haine, Nadine Goris, and Richard Sanders

Download

Final revised paper (published on 20 Jan 2026)
Preprint (discussion started on 18 Aug 2025)

Interactive discussion

Status: closed

RC1:
'Comment on egusphere-2025-3729', Jannes Koelling, 04 Oct 2025
Review of Sanders et al, 2025, submitted to Ocean Sciences

The manuscript “Remineralisation changes dominate oxygen variability in the North Atlantic” uses data from repeat hydrographic occupations of the A05 line in the to investigate long-term oxygen changes North Atlantic. They partition the change into four different terms: the excess, redistributed, remineralization, and disequilibrium oxygen change, reflecting different underlying processes, and find that the remineralization term is responsible for most of the change from 1992 to 2015.
I think overall this paper is well written and the methods are sound, but I think some improvements could be made. In particular, I think the role of ocean circulation in the observed changes should be highlighted more, and there are some inconsistencies throughout the text, as well as some parts of the methods section that could be explained better

General comments

Reading the manuscript, my first impression from the title and much of the text was that the authors found that the rate of respiration in the North Atlantic increased since 1992 (“Remineralisation changes dominate”). But as stated in line 269-272 and in the abstract in line 15-17, it actually is likely due to circulation changes, i.e. changes in “accumulated” upstream remineralization, rather than local ones. I think some more explicit mention of the role of circulation would make this clearer.
Below are some ideas for changes that could help make this point; I don’t think all of these need to be included, but implementing at least some of them would help highlight this point:
Change the title and/or text to use something more reflective of the role of circulation in the remineralization signal; e.g. “upstream remineralization” or “accumulated remineralization”

Move the sentence in line 10-12 (“This remineralization-driven change…”) up to earlier in the abstract, e.g. to line 5 after the sentence “Mechanisms responsible for…” (might need to rewrite a bit)

In lines 35-40, explicitly mention ocean circulation. I believe there are several studies linking the “indirect” (not solubility related) oxygen loss to circulation/ventilation changes, such as Oschlies et al 2018 or Schmidtko et al 2017, and this explains part of the model-obs discrepancy (Buchanan & Tagliabue 2021). It could also be useful to move parts of the following paragraph (lines 41-49) up. Currently, you discuss drivers of ocean deoxygenation in line 32-36, then talk about the model-observation disparity in line 36-40, then go back to discussing drivers of deoxygenation more generally in line 41-49, so rearranging this could improve the flow

In section 2.2, explicitly mention circulation changes. Line 83-85 talks about “export of organic matter” and “spatial patterns of remineralization”; I think maybe the latter is meant to refer to circulation, but it would be good to spell this out more clearly. And in the same section, circulation is mentioned in line 76-77 as part of the redistribution term. Does this refer to local circulation versus global/upstream circulation for the remineralization term? If so it might be useful to state that explicitly

The beginning of the discussion (line 255-256) is another opportunity to more explicitly state that the “alterations in remineralization” could be either local or upstream

While I agree that “discerning the physical processes responsible is complicated” (line 273-274) I still think the discussion could mention some studies on physical transport for context. E.g. could part of the oxygen trend be explained by the decreased flow of AAIW mentioned in line 60 (Hernandez-Guerra et al 2014) or results from the RAPID array (Smeed et al 2018; Johns et al 2023) which showed a decrease in AMOC strength around 2008-2010?

Specific comments

Line 4: rephrase “solubility-driven deoxygenation driven by…”

Line 29: “oxygen at the surface is close to saturation” – this isn’t true in many deep/mode water formation regions like the North Atlantic (e.g. Clarke & Coote 1988, Wolf et al 2018) and Southern Ocean (Bushinsky & Cerovecki, 2022), and also seems inconsistent with having a “disequilibrium” term

Line 71: I assume the change is actually calculated as 1998 minus 1992 (etc), so the way it’s phrased currently as “1992-1998, 1992-2004, and so forth” could be confusing

Line 76: Does “globally” here refer to the actual globe, or just globally within the dataset, i.e. along the section?

Line 85: There’s something missing in the phrase “The final, disequilibrium change..” – should it be “The final term”?

Line 100: Should there be a Delta before O_d?

Line 112/eq. 4: It would be good to already briefly describe the coefficients here so the reader doesn’t have to search for them through all the sections. The equation could also be moved down to be between section 2.6 and 2.7, so that the individual equations are described first, and the system of equations is then formulated once all the terms have been defined

Line 122: Does the “gradient of the temperature-oxygen solubility curve” mean that you take a T-O plot with contours of solubility at 100% saturation, then use that to determine the solubility gradient as a function of T and O? That could be good to show graphically in addition to (or instead of) Fig. 3b showing the values of the coefficient

Line 140: Is the redistribution term just vertical redistribution for each profile? Since the equations only include vertical gradients I assume that’s the case, but if so it should be noted explicitly here and/or earlier on (e..g line 75)

Line 148-149: Does the shift from negative to positive coefficients correlate with any particular water masses?

Line 153/eq.10: Add line break between C and O equations to be consistent with other eq.s

Line 163-164: Since weights are used in eq. 4 it is no longer just defined as Ax = b as stated in line 109, correct? It would be useful to restate the form the equation takes with the weight matrix, but could do so inline instead of as a numbered equation

Line 191-192: Why was the region from 30-70W chosen? Also it would be good to show the box used for the averaging in at least one panel of fig. 4, and/or highlight it in fig 1.a

Line 200-202: Can the changes over different depth ranges be related to any particular water masses? E.g. subtropical mode waters or Antarctic Intermediate Water. Looking at fig. 1 it sems that there is a clear oxygen minimum/DIC maximum around 800m, which could be indicative of AAIW, and this is near where the changes are most pronounced

Line 207-208: The last clause seems a bit redundant; “with consumption of oxygen via remineralisation generally increasing over time in the upper 1000 m, indicating an increase in the total amount of remineralisation.”

Line 215: Rephrase “Excess increase peaks” - actually a decrease

Line 219: “around 1000m” – change to “around 800m” to be consistent with line 149?

Line 235/line 140: How is temperature redistribution defined? Is it just the redistributed oxygen change times the coefficient (eq. 10)?

Fig. 4: To provide more context for the changes, it would be helpful to show mean sections of T, O2, and DIC either here or in Fig. 1.
For panels f)-j), I think it would be better to have the same color bar range as the other rows. Alternatively, you could keep the same range but a slightly different color map, e.g. purple-orange instead of blue-red.
In the caption, I’m not sure what “the surface 150m is omitted from a)” means, since the panels all seem to go to 0 on the y axis, and I’m guessing it also wouldn’t be just for one of the 25 panels

Line 239: “Due to the relationship between excess temperature and oxygen change” – this makes it sound like the oxygen changes are driving temperature changes; this may be true for the way it’s calculated here, but mechanistically it would be more correct the opposite way, i.e. excess oxygen changes are caused by the excess temperature change. I think you could just mention this in the discussion of excess oxygen change (line 213) instead

Fig. 5: The colors aren’t quite consistent with Fig. 1 since the color scheme is the same but with 5 values instead of 6. As a result, e.g. the 1992-2004 line in Fig. 5 is essentially the same color as the 1998 line in Fig. 1. I suggest using the colors corresponding to the “end year” in F1 in F5 so that for example the 1992-1998 line in this figure and 1998 in fig. 1 are the same color

Line 256: It would be good to be more consistent with the depth ranges discussed in the text. In section 3.2, changes are discussed for 150-500m, 500-1000m and 1000-200m in the first paragraph (line ~200, table 2), but later separated into 150-1000m and 1000-2000m (Fig. 6/line 228-233). Then this paragraph goes on to talk about “the upper 2000m”, before the next one talks about “the upper 1000m”.

Line 287: “an excess temperature change of 0.73C would be required” – you could state here how that relates to the observed change in fig. 5, i.e. about 3-4x (?) the warming observed so far. Again this would be easier compare if table 2 would use the same depth range as used in the text

Line 298: Repetitive sentence – “are not correctly accounted for separately, they will be incorrectly accounted for in the remineralisation component.”

Line 307: “matching” -> “match”

Line 308: Some extra words? “this increase in remineralization oxygen consumption”
Citation: https://doi.org/10.5194/egusphere-2025-3729-RC1
- AC1: 'Reply on RC1', Rachael Sanders, 05 Dec 2025
  
  The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-3729/egusphere-2025-3729-AC1-supplement.pdf
  
  Citation: https://doi.org/10.5194/egusphere-2025-3729-AC1
RC2:
'Comment on egusphere-2025-3729', Anonymous Referee #2, 16 Oct 2025
In this study, the authors use ship-based observations along a repeat section in the North Atlantic to disentangle the mechanisms driving ocean deoxygenation, i.e., solubility and remineralization changes, and changes in ocean circulation. By using a matrix of equations, the authors have decomposed the changes in oxygen, DIC and temperature over the past nearly three decades. This study shows that the deoxygenation in upper 2000m is dominated by an increase in remineralisation via circulation changes, and the increasing temperature is becoming a more and more important mechanism to drive the upper ocean oxygen loss. The main conclusion is straight forward and the manuscript is generally well written.

We have some general comments here:

1. The writing structure of the results part is a bit fragmentary, especially in section 3.2. I encourage the authors to regroup the current paragraphs (sentences).

2. It is a good study to show how the interior ocean deoxygenized, through multiple GOSHIP transects, but we have a general impression that this study is a bit incomplete. For example, (1) the authors didn’t explain well how they deal with the potential bias between the several ship tracks, which might result from the different measurement methods, different seasons of the ship track undertaken; (2) there is a clear different pattern of the oxygen changes between the west and east parts of the transect. It’s not clear why the authors didn’t discuss this signal, but instead, they took the average of the whole section. Please refer to the more specific comments below.

3. The title suggests that remineralization is the primary cause of oxygen change, but the abstract and the main text indicate that it is actually circulation changes that lead to a change in the spatial pattern of remineralization. Please reword the title to make this clearer

Overall, agreements between treatments of independent ship transects and justification of different signal between west and east seem to be robust. The manuscript is also readable and well organized. However, before publication, several methodological problems need to be fixed (declared), discussions need to be refocused, and some typos and visualization issues need to be fixed.

Method.2.1 We think the important information about how these GO-SHIP transects deal with the samples is missing. Were there any changes in metholdogy over the three decade time period? Although the authors listed the expocode in Table A1, we suggest adding more information in the method section, such as a simple, summarized protocol (consistent part for all the transects) and a transparent reminder of any potential biases and differences from transect to transect.

Line 83: ‘However, the water mass properties we are interested in change more slowly, and so are resolved by the∼5-year frequency.’ Do the authors have any reference for this?

Figure 1.

I would put the profiles from the west above (i.e., b-d) the profiles from the east (i.e., e-g). Since we read from left to right, this is how I would expect the outline of the figure.

And we would suggest the authors use another way to plot the ship-transects in the map (1a). I understand that the authors have stated they interpolated the six transects to a uniform shape, but a black line with a part stack on the continent is not very informative. Perhaps showing the location of the casts would be helpful as a piece of information to aid the readers in understanding the transects.

What’s the purpose of separating into two sections (west/east)? It seems like the whole study discussed the transect as a whole (mainly focused on 30-70W).

The authors switch between using the 30 – 70W average and the whole transect in Figures 1 – 5. If there are important differences between the west and east section, they should be presented and discussed separately and if there are no differences, then this should be stated up front and thereafter only the 30-70W averages shown and discussed.

Figure 2. Since this figure is discussed before the acronyms are explained in the main text, I think it would help the reader to explain what e, d, rd, and rem stand for in the figure caption. Or maybe include the acronyms at the beginning of section 2.2.

Lines 100 – 102: the text preceding the equations goes through them in order of temperature, DIC, and O2. I would suggest listing the equations in that order for ease of reading. Equation 10 has two equations; please numerate these separately.

Section 2.6: Would a different Refiled Ratio change the result? I see this is discussed later, so I suggest bringing this up at this point.

Line 211 & 252-253: could the authors please also report p values?

Line 248: Suggest rephrasing “lower depths” to deeper depths, as I assume this is what was meant? Lower is ambiguous in this context.

Line 236. Steadily increasing trend? Over time?

Line 240. Could the authors please refer to a plot or table?

Figure 6. Sometimes we found it hard to see the error bars since they overlap, and the straight line between each dot makes it look like a linear trend, although it’s not. Thus, we suggest using a better way to visualize this figure, for example a bar chart with stacked components.

Line 255. We have noticed that the ship transects occurred at different seasons (see table A1). Do the authors think that the different temperatures, primary production, etc, of the different seasons could impact the results? In the abstract, the authors mentioned ‘This remineralization-driven change may be caused by a change in the supply of biological material to depth…’, We assume the large organic particles would have a different export rate in different seasons since it’s a function of primary productivity and mixed layer depth. We would expect the authors to discuss a bit more about how they treat such variability.

Line 270. The export ratio determines what fraction of productivity is exported, so a change in productivity does not necessarily have to lead to an equal change in export amount. It would be good to acknowledge this here and the consequences for the authors’ conclusions.

Line 285. ‘The excess oxygen change has already doubled in magnitude between 1992 and 2015.’ Can you report the value of this magnitude?

Line 305. The conclusion part is too simple to summarize the key results of this study, including the methods used and the trend (magnitude, proportion, caveats, etc.) found from these GOSHIP transects. This is an important section; thus suggest the authors add more detail to this part.

Line 310: Where is Appendix A?

Co-review statement:

I co-reviewed this manuscript with a PhD student who provided the listed reports together with me. This is part of the EGU Co-review scheme to facilitate training in peer review and to provide appropriate recognition for Early Career Researchers who co-review manuscripts.
Citation: https://doi.org/10.5194/egusphere-2025-3729-RC2
- AC2: 'Reply on RC2', Rachael Sanders, 05 Dec 2025
  
  The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-3729/egusphere-2025-3729-AC2-supplement.pdf
  
  Citation: https://doi.org/10.5194/egusphere-2025-3729-AC2
EC1: 'Comment on egusphere-2025-3729', Bernadette Sloyan, 16 Oct 2025

The referees have provided valuable comments. I encourage the authors to provide a brief response with how they can address the comments here, and then prepare a revised manuscript and point-by-point reply to referee comments for submission to OS.

Citation: https://doi.org/10.5194/egusphere-2025-3729-EC1

Peer review completion

AR – Author's response | RR – Referee report | ED – Editor decision | EF – Editorial file upload

AR by Rachael Sanders on behalf of the Authors (05 Dec 2025) Author's response Author's tracked changes Manuscript

ED: Publish subject to minor revisions (review by editor) (18 Dec 2025) by Bernadette Sloyan

AR by Rachael Sanders on behalf of the Authors (18 Dec 2025) Author's response Author's tracked changes Manuscript

ED: Publish as is (21 Dec 2025) by Bernadette Sloyan

AR by Rachael Sanders on behalf of the Authors (23 Dec 2025)

Co-editor-in-chief

Oxygen is fundamental to ocean biogeochemical processes, with deoxygenation potentially reducing biodiversity, and disrupting biogeochemical cycles. This paper develops and implements a new inverse method to decompose oxygen change into its constituent parts by linking each process to concomitant changes in temperature and dissolved inorganic carbon (DIC). They apply this method to a GO-SHIP repeated section of the North Atlantic Ocean. They find that alterations in remineralisation, either local or upstream, are responsible for up to half of the total oxygen decrease seen in the upper 2000 m between 1992 and 2015. This method could be repeated on any dataset, observations or model, that includes temperature, DIC, and oxygen concentration data.