Acidification in coastal waters of Ad&eacute;lie Land, Antarctica (1985&ndash;2025)

Metzl, Nicolas; Tilbrook, Bronte; Akl, John; Neill, Craig; Aymard, Alexandra; Lo Monaco, Claire; Reverdin, Gilles; Sallée, Jean-Baptiste; Barton, Aude; Chevallier, Frédéric; Gehlen, Marion

doi:10.5194/egusphere-2026-1664

Preprints

https://doi.org/10.5194/egusphere-2026-1664

Preprints

16 Apr 2026

| 16 Apr 2026

Acidification in coastal waters of Adélie Land, Antarctica (1985–2025)

Nicolas Metzl, Bronte Tilbrook, John Akl, Craig Neill, Alexandra Aymard, Claire Lo Monaco, Gilles Reverdin, Jean-Baptiste Sallée, Aude Barton, Frédéric Chevallier, and Marion Gehlen

Abstract. Ocean acidification is expected to be particularly severe in Antarctic continental shelves due to enhanced anthropogenic carbon uptake in cold waters in response to rising atmospheric CO₂, sea-ice retreat, freshening and climate-change feedbacks. Models suggest that undersaturated conditions with respect to aragonite (Ωar), a major form of calcium carbonate formed by marine species, could be reached as soon as 2052 for austral winter. Here we present new ocean carbonate system observations from cruises conducted since 2010 in the Adélie Land coastal region in East Antarctica, along with data from a BCG-Argo float and results from a neural network model for the period 1985–2025. The region is a permanent CO₂ sink and was most pronounced since 2006. The CO₂ sink leads to a positive increase of surface water total CO₂ concentrations (C_T) (+0.44 ± 0.01 µmol.kg^-1.yr^-1) and to a progressive decrease of pH (-0.013 per decade) and Ωar (-0.035 per decade) for the winter season. The lowest surface Ωar of 1.2 was observed in winter 2024 from the float data, a critical limit for some marine species such as pteropod. A projection of the C_T concentrations in the future, based on observed anthropogenic CO₂ concentrations and emissions scenarios, suggests that aragonite saturation state (Ωar = 1) will occur in surface waters as soon as 2055 in the Adélie Land region, which is part of a larger area of East Antarctica proposed as a Marine Protected Area by the Commission for the Conservation of Antarctic Marine Living Resources since the early 2010s.

Received: 24 Mar 2026 – Discussion started: 16 Apr 2026

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.

Download & links

Preprint (PDF, 2135 KB)

Supplement (1490 KB)

Download & links

Nicolas Metzl, Bronte Tilbrook, John Akl, Craig Neill, Alexandra Aymard, Claire Lo Monaco, Gilles Reverdin, Jean-Baptiste Sallée, Aude Barton, Frédéric Chevallier, and Marion Gehlen

Status: final response (author comments only)

RC1:
'Comment on egusphere-2026-1664', Anonymous Referee #1, 26 May 2026
Review
Title: Acidification in coastal waters of Adélie Land, Antarctica (1985–2025)

Author(s): Nicolas Metzl et al.

MS No.: egusphere-2026-1664

MS type: Research article

The paper entitled “Acidification in coastal waters of Adélie Land, Antarctica (1985-2025)” uses cruise observations and neural network approaches to estimate the carbonate chemistry in the upper levels of Adélie Land over the last 40 years. It highlights the temporal variability of carbon dioxide fugacity, total dissolved inorganic carbon, total alkalinity, pH on a total scale, anthropogenic carbon, and the saturation state. Furthermore, the authors estimate the variability trends for most of these parameters and extend these trends into the future, correlating the expected saturation states with respect to aragonite with the critical conditions for precipitation associated with the estimated anthropogenic carbon levels over the course of the century.

The paper’s approach is innovative for the region and includes new data from summer 2024-2025. Another innovative perspective is the application of neural-network outputs to the carbonate system in high latitudes.

There are a few points in the paper that need to be addressed before it can be properly published:
The reproducibility of the neural-network model in the dataset. The steps for validation and testing are lacking, which means that reproducibility is also lacking. This may be justified by the relatively good representation of fCO₂ and CT, but not so much for pH and Omega aragonite. Although the provided link guided us to the L4 time series of monthly global reconstructed surface ocean pCO₂, air-sea fluxes of CO₂, pH, total alkalinity, dissolved inorganic carbon, and the saturation state with respect to calcite and aragonite, I could not identify any validation that would (1) corroborate the potential representation of real carbonate chemistry in the target region, or (2) justify annual variability based solely on data from summer cruises (there is a few years with winter data). I suggest performing comparisons to indicate the representation of the in situ data and the neural network model before expanding it back in time, especially for winter scenarios using the BGC-Argo data. The robustness of the model can be improved by testing more equations for AT reconstructions beyond the global ones mentioned in the paper. There are a few other papers that develop such approaches in the coastal zones surrounding Antarctica, which may or may not be applicable to the Adélie region, but this has not been tested. The point here is that equations derived from global datasets may miss regional features; therefore, comparisons with equations derived in coastal areas around Antarctica are likely to be more representative. I was hoping to see at least one Taylor diagram (https://doi.org/10.1029/2000JD900719) when I read the abstract.

The hidden mechanisms behind the trends. Most of the trends observed in the paper are posed on two points (e.g. year 1 and year 2), divided by the interval between them (i.e. the inclination from year 1 to year 2 / [year 2 – year 1]). The study area is close to Antarctica, where the Southern Annular Mode (and the El Niño Southern Oscillation, which has a minor impact) and all the impacts stated in the introduction (lines 39 - 41) affect the CO₂-carbonate chemistry in the surface layer. Interannual variability is important here, but trends based on extreme values (i.e. the first and last concentrations within the time series) could obscure this important aspect of the carbon cycle. I suggest two options: (i) in situ data supporting deeper reconstructions of long time series (1985 – 2025), which might highlight the role of anthropogenic carbon over the other drivers, especially for more detailed interannual variability; or (ii) combining time scales to deconvolve the trends in more detail, paying special attention to other drivers over time. My criticism here is based on the absence of parameters other than fCO₂ from R2R3R4/L’Astrolabe fCO₂ showing a more consistent time dependence, whereas the other data does not clearly demonstrate this (see Fig 6a).

I believe that combining both points could provide the authors with more information and improve the reproducibility of the data.

####################
General comments:
# line 18: standardise the subscript for CO₂ and omega aragonite throughout the paper.
# line 21: remove the space before “Here we present …”.
# line 24: add a space between ± and 0.01, and remove “. “ in the units, both throughout the paper.
# line 38: suggest replacing “offshore ocean” with “open ocean”, as cited in line 43
# lines 44-45: several other papers in the coastal zone of the Northern Antarctic Peninsula that are not cited, but should be. These could be useful for the discussion (especially comparisons between Adélie and other sensitive areas to climate change) and cover aspects of CO₂ fluxes (Monteiro et al 2020, https://doi.org/10.1038/s41598-020-71814-0) and anthropogenic carbon (Kerr et al 2018, https://doi.org/10.1016/j.dsr2.2017.07.007). These are just a few examples.
# line 47: add a space between ± and std in the first trend.
# line 123: add space between ± and 2
# line 123: I did not identify the description of the flags used. Either describe it or reference it in the bridging papers and method.
# lines 131-133: To standardise quality control flag system, mentioned the one used for AT and CT here (correct me if I am wrong, but I believe you are referring to the one by Boyer, T. P., Baranova, O. K., Coleman, C., Garcia, H. E., Grodsky, A., Locarnini, R. A., et al. (2018), World Ocean Database 2018).
# line 134: remove “.” between µmol and kg.
# line 139: O₂ with subscript 2.
# line 144: remove “.” Between µmol and kg. Please correct it throughout the manuscript; there are many of these typos.
# line 151: Why did the authors use the dissociation constants of Lueker et al (2000) instead of those of Goyet and Poisson (1989), the latter of which are known to perform well in high latitudes? If based on Dickson (2007) and Sulpis et al (2020), then cite them, please.
# lines 163-165: beyond the open-ocean AT-salinity equations, there are a few others for coastal regions in Antarctica. Have the authors compared these, too? It would be useful to test the open-ocean and coastal region equations and plot the results as a Taylor diagram (https://doi.org/10.1029/2000JD900719).
# line 167: I am sceptical about how representative this neural network might be of the region. I would suggest adding a short paragraph comparing the in situ data from summer 2024-2025 with neural networks from summer 2024-2025, similar to the comparison between lines 211 and 220. Previous cruises could be included in this comparison (those in Table S1). The paragraph between lines 177 – 192 could be modified to be more method-like than result-like. Additionally, the paper lacks a discussion on modes of climate variability in terms of their regional impact on circulation and biological activity. This is important because it can influence CO₂ emission and uptake, as well as long-term trends. Incidentally, sessions “2.2” and “2.4” could be reorganised for easier reading. Example: I think BCG-Argo could have its own session.
Correct me if I am wrong. CO2Sys and LIAR were used to estimate AT data, which was then used to estimate pH and omega (lines 149 - 150). If so, how did you propagate the errors of the different estimates into the pH and omega outputs?
# lines 227 – 228: From this line onwards, I have seen the trend estimates as the slope/difference between two points divided by the time interval. Based on the high interannual variability in Antarctica, I am a bit sceptical of this approach. In winter, with only a few in situ measurements, assumptions are required to do so. However, in the summer, several cruises could provide the interannual variability required to determine the long-term trend. Modes of climate variability and signal decomposition could support interannual controls, improving the interpretation and impact of the paper. The same points are valid for lines 242 – 247.
# lines 238 – 241: This should be method.
# lines 249 – 251: data for Chl-a in 2002 support this interpretation (S6), but it is not valid for 2015. Data for Chl-a 2015 in Fig.S7, please reference this.
# lines 255 – 257 & 331 - 335: How does this trend compare with that in other regions surrounding Antarctica?
# lines 277 – 278: Do the mixing and Redfield ratios in this region support the application of this method in the region? What is the error associated with such an estimate?
# lines 373 – 374: What is the relative contribution of anthropogenic carbon uptake versus CDW upwelled during winter to the impact on CT in the study area? Other regions (e.g., Monteiro et al 2020) have suggested more intense CDW intrusions in the Northern Antarctic Peninsula. Could this be experienced in Adélie Land? Could the sea ice block sea-atmosphere interactions during winter and freshening (+ biological activity) during summer, spiking carbon uptake? I think this deeper discussion would be essential to improve impact.
# Fig 1. Suggest adding isobaths to mark the bathymetry (200 m and 1000 m, for instance). This would improve visualisation.
Citation: https://doi.org/10.5194/egusphere-2026-1664-RC1
RC2: 'Comment on egusphere-2026-1664', Anonymous Referee #2, 15 Jun 2026

This is an interesting study on ocean acidification of a region of the ocean that is important but undersampled. It shows the level of OA and extrapolates it to the future, where important results emerge. Not surprisingly, biological activity plays a major role in the observed changes/variability. The combination of different sources of data makes the study more robust. The contents of the manuscript is fine, but I think it needs more explanation on the methods used and how the results were obtained.
It would have been helpful to the reviewer to have the figures embedded in the text. I think the journal also prescribes this format.
The numbering of the sub-sections of the results is erroneous. I think it should be 3.1, 3.2. 3.3 and 3.4.
At several place in the text: xCO₂ not XCO₂.

Below a list with mostly minor comments, but also a few major ones:
L24 … a permanent CO₂ sink WHICH was most pronounced …
L25 … to AN increase (delete positive)
L26 delete progressive
L28 pteropods
L36 delete enhanced
Line 59 Shadwick
L39-40 … multiple effects SUCH AS anthropogenic CO₂ …
L44-46 “Apart from the Drake Passage to the north of the Antarctic Peninsula (Takahashi et al, 2014; Hauri et al 2015; Munro et al, 2015) no long-term time-series have been used to evaluate OA in coastal Antarctic zones.” I suggest to refrain from the negative position (no long-term time series) and write this in a positive sense: There are two long-term time series …
L50 per decade (no exponents use with words, only with symbols) (also at other places in the text)
L52 delete comma
L53 TO the deep ocean, implying …
L61-64 This sentence is grammatically incorrect. Please reformulate
L68 Please define C_T
L70 What kind of CO₂ is meant here? Dissolved CO₂ or DIC? Please be more specific
L80 I think you mean: in the region off Adélie Land
L81 again which CO₂?
L85-88 Please reformulate this sentence
L115 “metadata of SOCAT” needs a reference
L118 cruises
L123 is, not are
L130 Prof. not Pr.
L131 Which one is the R2 cruise? Please explain. It was not mentioned before
L135-136 and earlier: The authors used CRMs and write: “These data were used for the validation of calculated carbonate system parameters using fCO₂ data with an A_T/salinity relationship” How were the CRMs used? Were the measured data corrected with a factor that was determined from the CRM value and the measured value? Please provide details in the text.
L139 nitrate (no capital)
L161 This relationship was determined with data south of 60°S (as seen in Fig. S4). Why was this region different from the region which was taken for other analyses (south of 63°S)?
L169 and further: The info as to the FFNN, how it works and which parameters, variables and data are used is minimal. For the reader to comprehend and reconstruct the data produced this is not sufficient; the results presented are as if coming from a black box.
L176 Why were only results used for this small region? This region is again different from the two other regions mentioned before.
L177-192 This paragraph does not belong to the methods section, but rather to the results section.
L185 Shadwick
L187, 189 What is the origin of the SAM indices? Please provide reference

Figure 2: The orange square is said to be for year 2010, but it covers at least 3-4 years around 2010. This is confusing. Please choose another way of showing the 2010 value.
Figure 3: The unit of fCO₂ is uatm, but the unit of the atmospheric CO₂ concentration is ppm. This is not correct in the figure. Why don’t they have the same unit? How can the units be reconciled? Same for Fig. S1 and other
Figure 4: How was normalization of fCO₂ and C_T done? This could be added to the Methods section
Figure 6: In panel (b), which shows CT, fCO₂ is shown in the inset, but there is no scale. I presume that this is not fCO₂ but something else, possibly C_T calculated from fCO₂? Please correct

L204 The ship’s name is not mentioned in the Methods section. There, only L’Astrolabe is mentioned as ship for the three cruises.
L242-247 Did you only calculate the trend using the first and last values? Why not using all data in the time period. This would increase the reliability of the trend calculation.
L280 “Here we selected the data below 100m depth for 1994-2018” Why again a different time period? The argument should be given in the text. And why 100 m? At least a reference should be given with respect to the depth of the mixed layer in this region.
L285-287 “that d13C observations in January 2025 were lower than in January 2003 suggesting that surface waters captured the anthropogenic Suess effect signal (Figure S13) as previously observed in this region (McNeil et al, 2001)” This asks for a better and more extensive explanation in the text.
L323 delete version
L324-325 “The estimated C_T using the C_ant trend (Eq. 5) is close to the FFNN model and observations (Figure 10a)” This is not completely surprising, as the results were based on the same data.
L325 “The mean difference of C_T” Which difference is meant here? Difference with what?
L326-327 “but other process could explain the full trend in surface C_T.” I think you mean: … but additional processes need to be included to explain the full trend in surface C_T.
L328, 392 C_T-rich, not rich-C_T
L331-332 I suggest: … that the anthropogenic CO₂ uptake is responsible for 70% of the CT increase over 1985-202. But what about the remnant 30%; what factors are responsible for that? Please explain in the text.
L340 ... remnant winter WATER layer …
L342 Suggestion: An important question is when the aragonite undersaturation …
L344 … show SIMILAR results …
L344 Suggestion: Differences between the two future scenarios …
L353 will reach, not reached
L362 Place shipboard before fCO₂ and BGC-Argo before pH_T
L365 measurements
L394 surface is mentioned twice; one should be deleted
L404 Australian
L405, 422 Earlier the ship was called L’Astrolabe

Citation: https://doi.org/10.5194/egusphere-2026-1664-RC2

Nicolas Metzl, Bronte Tilbrook, John Akl, Craig Neill, Alexandra Aymard, Claire Lo Monaco, Gilles Reverdin, Jean-Baptiste Sallée, Aude Barton, Frédéric Chevallier, and Marion Gehlen

Supplement

https://doi.org/10.5194/egusphere-2026-1664-supplement

Data sets

An updated synthesis of ocean total alkalinity and dissolved inorganic carbon measurements from 1993 to 2023: the SNAPO-CO2-v2 dataset. Nicolas Metzl et al. https://doi.org/10.17882/102337

Nicolas Metzl, Bronte Tilbrook, John Akl, Craig Neill, Alexandra Aymard, Claire Lo Monaco, Gilles Reverdin, Jean-Baptiste Sallée, Aude Barton, Frédéric Chevallier, and Marion Gehlen

Viewed

Total article views: 417 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
312	81	24	417	75	25	21

HTML: 312
PDF: 81
XML: 24
Total: 417
Supplement: 75
BibTeX: 25
EndNote: 21

Views and downloads (calculated since 16 Apr 2026)

Month	HTML	PDF	XML	Total
Apr 2026	170	54	16	240
May 2026	132	20	3	155
Jun 2026	10	7	5	22

Cumulative views and downloads (calculated since 16 Apr 2026)

Month	HTML	PDF	XML	Total
Apr 2026	170	54	16	240
May 2026	132	20	3	155
Jun 2026	10	7	5	22

Viewed (geographical distribution)

Total article views: 416 (including HTML, PDF, and XML) Thereof 416 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 22 Jun 2026

Short summary

Ocean acidification is explored in the Adélie Land coastal region in East Antarctica. The region is a permanent CO₂ sink leading to a positive increase of total CO₂ concentrations and to a progressive decrease of pH and aragonite Ωar. The lowest surface Ωar of 1.2 was observed in winter 2024, a critical limit for some marine species such as pteropod. A projection in the future suggests that aragonite saturation state will occur as soon as 2055 in the Adélie Land region.


Total:	0
HTML:	0
PDF:	0
XML:	0