the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 18 Aug 2026
A comparison of metrics for CO2 uptake efficiency of regional ocean alkalinity enhancement in an Earth system model
Abstract. Ocean Alkalinity Enhancement (OAE) is a method proposed for ocean-based Carbon-Dioxide Removal and describes the sequestration of CO2 from the atmosphere through a deliberate increase in alkalinity in the ocean. In this study, we compare multiple approaches for the calculation of the CO2 uptake efficiency for regional OAE using an Earth system model. We find that for regional alkalinity deployment at the European Atlantic coast the change in the global ocean inventories for dissolved inorganic carbon and alkalinity is not a feasible approach. The OAE-induced removal of CO2 from the atmosphere leads to phase shifts in the large-scale climate pattern El Niño, but it does not influence its frequency or intensity. This alters the timing of interannual variations in global carbon fluxes, which obscures the detection of the immediate signal of alkalinity addition on ocean carbon uptake on annual to decadal timescales. Approaches that track the regional distribution of the added alkalinity or the OAE-induced regional strengthening of the air-sea CO2 flux provide more robust estimates of CO2 uptake efficiency, when the following criteria are taken into account: (i) ocean conditions in and close to the deployment region (e.g., circulation, upwelling and subduction) (ii) the natural variability of surface alkalinity and air-sea CO2 flux, and (iii) the possible impacts on global climate modes such as El Niño.
- Preprint
(1989 KB) - Metadata XML
-
Supplement
(1146 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2026-4781', Anonymous Referee #1, 02 Sep 2026
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-4781/egusphere-2026-4781-RC1-supplement.pdfCitation: https://doi.org/
10.5194/egusphere-2026-4781-RC1 -
RC2: 'Comment on egusphere-2026-4781', Anonymous Referee #2, 28 Sep 2026
The authors compare serveral ways of calculating ocean alkalinity enhancement (OAE) efficiency in a regional setting along the European Atlantic / Nordic Sea coast in an Earth system model. They find that defining efficiency based on changes in global ocean Alk and DIC inventories is not feasible, as these signals can not be detected from the difference in the simulatinos with and without OAE as they are superimposed by natural variability in both simulations.
They highlight that tracking the regional distribution of added alkalinity or additional air-sea CO2 flux and then calculating efficiency only over this region is suited better for estimating OAE efficiency in fully coupled Earth system models and relatively small OAE signals.The fully coupled Earth system model setup with the nested high resolution setup over the North Atlantic, combined with the comparably realistic alkaliniy addition scheme, is very interesting. However, reading the manuscript, I found a couple of aspects unclear in the current version. Most importantly, it is not clear to me yet for which applications the truncated efficiency metrics introduced here are going to be useful. To my understanding, these metrics only partially represent the ocean carbon cycle feedback to the OAE-induced ocean carbon uptake, as the OAE-induced outgassing signal occurs throughout the ocean. And if one was not interested in the ocean carbon-cycle feedback, wouldn't an ocean-only model be better suited to isolate the negative emissions arising due to OAE? Please find below my main comments and a couple of smaller comments.
Main comments:
Reiterating my point from above: When calculating the efficiency metrics in the emission-driven simulations only over a certain ocean region, as defined in sections 2.1 and 2.2, don't you miss the negative ocean carbon cycle feedback from all other regions? As such, it is not clear to me why you take the challange of simulating regional OAE in an emission-driven ESM simulation when not fully characterizing the feedback. Clarifying this aspect would be important from my perspective. My cricism here would not apply to BASE and OAE simulations in an Earth system model with prescribed atmospheric CO2, where you may also have issues with dephased internal variability between the two simulations, but not carbon cycle feedbacks, such that the truncated efficiency metrics would likely represent unbiased estimates of the global efficiency.Somewhat related - it is not entirely clear to me whether you regard the four different approaches for calculating OAE efficiency as four different estimates for the same ground truth or whether they should represent different underlying concepts. For example in L261, you write "Accordingly, the calculated uptake efficiency with values between 0.16 and 0.53 underestimated the actual efficiency." What would be the correct representation of that actual efficiency, if it was not obscured by internal variability? I think it would be one based on the global net air-sea CO2 flux or ocean carbon budget.
Do you expect the coastal efficiency approach (constrained to the deployment area) to be useful in practice? I think it is interesting to analyze how much carbon is already taken up or stored in the deployment region, but I would propose to clearly state that this is not likely a practical approach for calculating efficiency.
At some places in the manuscript I understand that you interpret the changes in ENSO phasing to be mechanistically related to the OAE interventation, for example in L247, L252, L338, L343-345. I am not sure this perspective is correct - I would rather think you are dealing with a signal-to-noise problem, between the OAE signal on e.g. ocean carbon uptake, and the noise (internal variability as manifested in modes such as ENSO that dephases betweeen the OAE and BASE simulations through chaos)
Small comments:
L48: "Alkalinity varies through..." you discuss here factors of variability in the global ocean alkalinity budget, however in the same paragraph above you propose to use added alkalinity as denominator in the efficiency, which would not be subject to the variability you list here. You use both concepts in the manuscript, however it would be good to separate them more cleanly here I think.L85: What do you mean by 'ideal case' here? one, where life-cycle emissions are minimal? I think it is good that you explicitly discuss the plausibility of the assumed OAE scheme.
L99: For the threshold-based budgets: you integrate DIC and alkalinity here only over regions where Delta surface alkalinity exceeds a certain threshold. But what if subducted DIC and alkalinity move somewhere that is not visible from the surface alkalinity? You pick up this aspect in L279f, but it should be more visibly highlighted as a caveat of the definition from my perspective.
L126f: This is a very interesting aspect, you find emerging signals against background variability for the air-sea CO2 flux, but not for alkalinity? Do you know why the signal-to-noise ratio is better for air-sea CO2 flux than for alkalinity, which is the primary variable altered by the alkalinity addition?
L136: Do you know why alkalinity is slightly decreasing over the 25year period in BASE? In some models, an increase in alkalinity is simulated as a result of ocean acidification - what is driving the decrease in your case? this question becomes also relevant towards the end of the paragraph, where you describe regional patterns of alkalinity changes, and also when you show the alkalinity increases in the ocean to be mostly smaller than the added amount (Fig. 2a) - if this is actually a forced response rather than internal variability. It was somewhat surpising to see on the one hand alkalinity reducing under ocean acidification, and on the other hand added alklanity also to be decreasing.
Fig. S1: I do not understand the units of this figure, is it Tmol m-2 (this would be too much I guess), or normalized to another unit of area?
L138f: the high storage of additional carbon in the North Atlantic is not necessarily equivalent a carbon uptake in the same region - I'd rephrase it.
L141f: how is loss of alkalinity in the Atlantic caused by increasing CO2 concentrations in the atmosphere? relates to my comment in L136
L144: I would highlight here that you now turn to the OAE simulations and budget based efficiency estimation, away from characterising BASE changes over the period.
Fig 2c: Where can I see the ratio for Eu in the figure? is it simply not visible because of the large y axis scale? would be worth mentioning in this case.
L154f: The analysis in the paragraph is super interesting and the correlations are striking, but I am a bit confused by the signs of change. El Nino events cause ocean outgassing. The cumulative El Nino index change relative to BASE is negative (Fig 3b), i.e. a relative tendency towards La Nina in the OAE simulations, wouldn't that imply enhanced carbon uptake relative to BASE?
L163ff: is the model formulated in a way where calcification and redissolution of CaCO3 are not dependent on the saturation state of the respective CaCO3 mineral? Otherwise, this could be a relevant factor for the response of natural alkalinity to alkalinity addition.
L196ff: you analyze here the probability that annual mean alkalinity exceeds the 25-year temporal mean by 10-100 mumol kg-1 for at least one year?
L224f: "When the CO2 uptake efficiency is calculated based on ∆DIC and ∆ALK, temporal variability is in most cases of similar size or smaller than when only the threshold criterion is applied to constrain the area." I am a bit unsure I understand. You here state that the temporal variability of the combined plume efficiency is similar or smaller than that of the inventory efficiency using alkalinity thresholds only?
Table 2: The caption lacks some information about what the data in the table displays. You show the efficiency over the combined area of the two plume tracking methods, and then either quantify the ratio of dic and alk inventory changes or the ratio of cumulative air-sea CO2 flux over added alkalinity, right? Section 3.3 is generally a bit hard to read for me in the current form.
Caption Fig 5: "Probability density function (PDF) in BASE for surface alkalinity concentration to exceed the thresholds of (a) 10 µmol/kg,.." shouldn't it be simply the "Probability for suface alkalinity concentrtion in BASE to exceed thresholds of..."?
L232f: Connect to my main comment above: I don't think they should be expected to be fully consistent, as they quantify different parts of the global response to the OAE intervention in the coupled system.
L238: " internal variability the baseline simulation" there is a 'in' missing
L241f: "Some of that CO2 can be emitted in other places when the water mass re-equilibrates with the atmosphere due to, e.g., changes in water temperature (Müller et al., 2025)." I am not sure what is meant here - the requilibration of the ocean with the atmosphere with reduced atm. CO2 content leading to CO2 emissions into the atmosphere (i.e. the ocean carbon cycle feedback) would also act without a marine heatwave or a similar perturbation, right?
L244: 'vgl' should that be 'compare'?
Fig S6: Panel a) Why do you get this consistent land-outgassing across the simulations, given the small effect of the OAE intervention on atmospheric CO2? Is it relatied to the general tendency of your OAE simulations to be in a certain ENSO state compared to BASE? Please provide an interpretation for this clear signal. Panel b): is the line for EU missing there?
L247f: " our study reveals that these feedbacks can lead to phase shifts in the internal variability of the climate system." I feel this is perspective is not entirely correct - I have interpreted that divergence in the phasing of climate modes between OAE simulations and BASE simulation to be simply associated with chaos at play, as would be also obtained by just perturbing the system somewhere slightly. You seem to suggest instead that OAE itself has a mechanistic effect on the phase of climate modes. In line 252f you also write "feedback effects within the climate system are larger than the OAE signal in the carbon sequestration through DIC" -> isn't it instead simply internal variability in the fully coupled climate system that exceeds the OAE signal? Your perspective also comes back in L345: " (iii) to possible impacts on global climate modes such as El Niño." I doubt there is a forced link between the OAE intervention and the phasing of the ENSO but I may miss something here. Please clarify.
Figure 7: please provide more information in the caption, most importantly: what is represented by the error bars? why are the flux plume error bars so small compared to the inventory error bars?
L260: shouldn't that be "alkalinity-enriched"? DIC-enriched would point to an efficient carbon uptake
L300f: "To further reduce uncertainties from variability of the surface alkalinity and biological activity, a combination of the two plume tracking mechanisms has been applied." How is the compination of the two plume tracking mechanisms helping for reducing uncertainties connected to biological activity responding to the added alkalinity? Or do you rather mean uncertainties from variability in DIC in general?
L322f: " Similarly to the efficiency metric based on global-ocean ∆DIC, this metric can be applied to simulations without carbon feedback that avoid changes in the climate system" I don't see how this could be done, given that without carbon feedbacks there is no effect on atmospheric carbon. Please explain. I think you could add Grosselindemann et al in L324, as they also apply strong OAE on centennial timescales.
L338f: " Alternatively, very large quantities of alkalinity are added to the ocean that have a larger impact on the carbon system than any feedback effects (Jeltsch-Thömmes et al., 2025)." Also here: from my perspective the issue with the detectability is rather one of internal variability than feedabck effects. The feedback effects on the other hand should scale with the OAE intervention. Please clarify.
L344: how is i) and ii) different from each other? is i) referring to the physical abd biogeochemical mean state of the region?
L350-352: Do you specifically refer to the research question how carbon cycle feedbacks react in a specifc small-scale OAE application? otherwise I think a simulation without interactive carbon cycle, such as in ocean-only simulations, provides a clearer signal of the intrinsic / gross capture efficency.
L353-357: I think this highlights a strength of the setup in this study: you nicely see how the signal of small-scale OAE stands up against internal variability, a challenge that is just worsened in the real world where there is no BASE simulation, highlighting that measurement-based MRV is near impossible. Would it make sense to expand on this aspect?
L366-368: "In contrast to ocean DIC, the integrated carbon flux into the ocean only represents the additional uptake of carbon at a given time while the change in DIC represents the net change, including carbon gain and loss through the surface and biological processes in the ocean’s interior." This is interesting, I would argue the opposite, stating that the efficiency defined as air-sea CO2 flux over added alkalinity is actually most important, as it is directly linked to negative emissions and atmospheric carbon drawdown. Furthermore, I think this flux-based efficiency is sensitive to carbon and alkalinity changes in response to OAE through the oceans biological and calcium carbonate cycling, as these effects influence how much carbon is taken up from the atmosphere in response to an addition of alkalinity. I am looking forward to your thoughts here.
Citation: https://doi.org/10.5194/egusphere-2026-4781-RC2 -
RC3: 'Comment on egusphere-2026-4781', Anonymous Referee #3, 28 Sep 2026
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-4781/egusphere-2026-4781-RC3-supplement.pdf
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 234 | 102 | 81 | 417 | 55 | 65 | 63 |
- HTML: 234
- PDF: 102
- XML: 81
- Total: 417
- Supplement: 55
- BibTeX: 65
- EndNote: 63
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
