the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 10 Nov 2025
Organic Alkalinity modulates pH from the Sea-Surface Microlayer during a mesocosm study
Abstract. The ocean plays a central role in climate regulation by exchanging carbon dioxide (CO2) with the atmosphere. This exchange depends on the transfer efficiency across the air-sea boundary layer, the sea-surface microlayer (SML) known to be an organic-rich boundary with a thickness of less than 1 mm. The parameters dissolved inorganic carbon (DIC) and total alkalinity (TA) describe the state of the marine carbon system (MCS). However, organic alkalinity (OA), which arises from weak acid-base functional groups in dissolved organic matter, remains poorly constrained. It is known to modulate pH in organic-rich environments. Yet, to our knowledge, it has not been quantified directly in SML before. Here, we show that the enrichment of OA in the SML modulates pH and that its effect propagates further down into the underlying water (ULW). We track the evolution of the MCS during a 35-day mesocosm study where we induced a phytoplankton bloom. Three distinct bloom phases were identifyed by different biological processes dominating within the system. Dissolution dominated during the pre-bloom phase; photosynthesis and calcification prevailed during the bloom; and CO2 invasion, together with respiration, was most pronounced in the SML during the transition to the post-bloom phase. These processes provided the context for the observed variability in OA. We measured OA directly by differential potentiometric back-titration as a second titration on the same titrated TA samples. OA in the SML was persistently enriched (Enrichment Factor (EF) > 1) and reached concentrations up to 264 µmol kg⁻¹. On average, it contributed 8.4 % of TA, compared to 3.1 % in the ULW. Concurrently, the vertical pH differences between SML and ULW decreased towards zero as the bloom began and occasionally became negative. Over the study period, OA EF and ΔpH were negatively correlated (Spearman ρ = -0.75, p = 0.024), indicating that stronger OA EF dampens the pH rise associated with the bloom onset and its effect propagates further down to the ULW. Recognising that OA enrichment modulates pH in both the SML and the ULW, routine inclusion of OA in near-surface measurements and a three layer air-SML-ULW framework should guide future evaluations of air-sea CO2 exchange.
- Preprint
(1047 KB) - Metadata XML
-
Supplement
(521 KB) - BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2025-5265', Anonymous Referee #1, 21 Mar 2026
-
RC2: 'Comment on egusphere-2025-5265', Anonymous Referee #2, 30 Jun 2026
Referee Comments
Title:
This title reflects what the authors would like to have achieved. As will become apparent in this review, this is judged as unattained, and so a more modest title is probably needed. Something like: On the co-variation of Organic Alkalinity and pH in the surface layer and bulk solution of a mesocosm experiment during an induced phytoplankton bloom.
Abstract:
Much of what is covered in the paper is identified in the Abstract, but it is rather haphazard.
A statement about OA made here is: ‘Yet, to our knowledge, it has not been quantified directly in SML before’. I mention this because it could be the one thing upon which the entire paper could be hung. This would be its unique selling point.
A second point, in the last sentence is: Recognising that OA enrichment modulates pH in both the SML and the ULW, routine inclusion of OA in near-surface measurements and a three-layer air-SML-ULW framework should guide future evaluations of air-sea 𝐶𝑂2 exchange. This suggestion would go well with the above-mentioned sparsity of measurement of OA. Together, they would give a better paper, and one more provable.
Intro.
I find that the Introduction wanders about rather too much. The issues are dumped on the reader’s doorstep, for them to sort out – however, that is the author’s job! It needs to follow an argument. I suspect the authors are trying to ‘cover all popularity bases’, to get acceptance. In the Intro, better to concentrate on those which are important, and even consider neglecting those that are peripheral. Sometimes, stuff that appears to be appropriate for the Intro, is better placed in the Discussion. For example, is this paper aimed at Climate Change amelioration, or CO2-exchange at the surface of highly active biological sites? Of course, globally these two are eventually going to be linked, but maybe only tangentially in this paper! My hunch is that you are primarily in the CO2-exchange zone!! Moreover, you aren’t going to go into the depths of physically and chemically modelling the system; instead, you are essentially wanting to report experimental findings. In that sense the paper ought to be much simpler, as it is merely a description of a set of results.
Methods
The methods section is written better than is the Intro, but even so, it is not ‘stark’ enough. At each stage, it needs to be a very pointed statement of what methods were used, and if something is complicated, an ‘in principle’ description given, as to how it works. Each of the individual methods used, should be described. You see, a reader, finding some unexplained point much farther down in the ms will often say to themselves, ‘Now, how exactly was that done?’ That’s the level of delivery needed in the methods section.
In line 208 you say: By applying both methods (statistics), we aimed to determine whether OA enrichment in the SML could systematically explain variations in ΔpH, thereby revealing potential mechanistic links between OA accumulation at SML and (210) the observed vertical pH anomalies between both layers.
I discern here a confusion over mechanisms and experimental results such that you think statistics will give you a mechanism – it won’t! Inspiration about mechanisms comes from patterns in data – the shapes of plots, and a grasp of the processes involved! Statistics tests the reliability of the patterns. So, I suggest that the statement in 208: ‘… we aimed to determine whether OA enrichment in the SML could systematically explain variations in ΔpH’ is really about the relationship between ΔOA and ΔpH, the former being the difference in OA in surface layer and the bulk (ΔOA is the same as your, OA EF).
Now we come to modelling. The rule is, ‘find the model that explains everything, with the tiniest of assumption(s)’. Consequently, we start with direct proportion (y = mX + C) and work upwards! (Of course, direct proportion means that variable Y follows variable X, with a bit added on!) The nearest you get to that is in your fig. 4 which, given the seriousness of the relationship, arrives very late on. I think you need to ask yourself, ‘what am I trying to prove, on the basis of what can I rationally attempt to prove’ and start from there. Then ‘drive your writing’ at that!
But the phrase: ‘what can I rationally attempt to prove’, is another complication which involves an enormous amount of theory about the CO2-CO32- system. Basically, you are hypothesising from the beginning, that if the pH changes in the surface layer, it will be because buffers, created and released by the algae, have migrated disproportionately into the surface layer. Before doing anything else then, I think you are obliged to imagine how the ΔOA and ΔpH might be linked, physically. There is an entire mathematical theory called, ‘the Carbonate Thermodynamics of Seawater’ that will take you some way with this. Zeebe and Wolf-Gladrow (2001) is one exposition of it and is cited in your ref list. If this is beyond your current competence, I think you must ‘drop-back’ to a more modest objective in which you describe your experiments as a prototype. (See how Kerr et al. (2023) in your list, manged this.) Whether this is practicable, I cannot say. To some extent it will depend upon what else is in the collection of papers in the Special Issue. I think it would be reasonable to argue that there are many, many marine biologists who could run a mesocosm experiment, but relatively few who could design their way through the carbonate thermodynamics. The situation is probably even more confusing if the experimenter has CO2-SYS as a programme with which to handle the data. Such a device gives the impression of being able to control the CO2-system competently, but most-likely, it fails to do so!
Results
There should be a statement about the variability of the system in all its different ways, and in its prime variables, i.e., before you mess around with the numbers by various means! (E.g., line 212) If a reader finds problems later, they will want to come back to the basic data. If there are no features to report, then report just that. However, I think the authors did get salinity variation, and I suspect that that is why they went into normalising to a standard salinity. The question arises then though, as to whether it is rational to normalise some of the data. For example, it won’t be for pH. I suspect, too, that it won’t be for OA, either. One can rationalise data when the variable is a conservative tracer. OA will depend upon complex relationships between equilibrium constants, and they won’t be conservative. Given all of this, I advise you to describe what you observe before you fiddle around with the numbers. Your comment about first recognising the overall effect through the normalised data is interesting; was that trend not evident in the raw data?
If you follow a ‘dispassionate’ (factual) approach to your Results you will find that you will not generally need to explain why you did this, or that, manoeuvre to demonstrate what is happening. For example, someone showing timelines for variables A and Salinity, might obviously benefit from plotting a rationalised variable versus time curve.
There needs to be clearer acknowledgement, and later discussion, of the fact that the pH in the SML drops by ~0.5 pH units during the pre-bloom phase despite the alkalinity and DIC remaining relatively stable. Furthermore, if you are arguing that the algal production of organic alkalinity in the SML alters the pH, what is the rationale for including the pre-bloom data in figure 4?
The numbering of the sub-sections in section 3, is incorrect. The only reason to start writing prose immediately after the Heading, 3., is that you needed a short Intro to section 3 that applied to all sections within it. Otherwise, start immediately with 3.1, whence your 3.1 becomes, 3.2.
Figures
Fig. 1. The figures are nicely drawn but are possibly badly interpreted. Care is needed over this. Essentially, the ordinates of 1a and 1b are extremely expanded. It would be sobering to re-draw them on their full scale, e.g., 0 – 2200 and 0 – 2400. There is a chance that your measurements are NOT accurate and precise enough to justify the analysis you are imposing. The general rule in science is that one needs to choose systems that give big differences in signal change, so that statistical and systematic errors don’t matter, that much. I fear that this is the direct opposite of that circumstance.
While on this topic of error estimation, with the numbers mentioned directly above, it is sanguine to review the fact that OA is calculated as the difference between two very large numbers. This is a numerical trap that everyone needs to avoid because the errors in the two sets of measurements combine in the difference.
Finally, in interpretating the plots I do not see any mention being made of the chemistry of the material in which the mesocosm sits. I suspect it is concrete which, if it is, will probably be alkaline, and could therefore swamp the measurements being made. The upshot of that is that there ought to have been a blank run conducted alongside the experiment described. Of course, one can appreciate how costly that might have been, but that still does not excuse one not being included.
Final general point
This study adds an extra dimension to the original published mesocosm paper (Bibi et al.,) in Biogeosciences. This tags-on the OA variation. However, the paper has been written as if it were a new initiative.
Many of the problems in the writing would disappear if the title were something like: On the Alkalinity variations in a recent Mesocosm experiment. Oriented thus, the paper would be much shorter, with the Intro relying on the ‘mother paper’. However, the entire general justification could be much shorter. The methods section will still be required but it too could be a lot shorter, explaining the techniques only, 'in principle'. We would expect, too, that some of the figs in the main paper might re-appear in the ms, to build-up the context. In this way, the entire paper would be reduced to a ‘Note’.
See marked-up pdf for further comments and technical corrections
-
EC1: 'Comment on egusphere-2025-5265', Peter S. Liss, 30 Jun 2026
This manuscript needs to go back to the authors for them to deal with the critical comments of the referees, particular Referee(s) No.2 which rates the paper quite lowly.
Citation: https://doi.org/10.5194/egusphere-2025-5265-EC1
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 1,824 | 723 | 202 | 2,749 | 286 | 155 | 166 |
- HTML: 1,824
- PDF: 723
- XML: 202
- Total: 2,749
- Supplement: 286
- BibTeX: 155
- EndNote: 166
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Review of “Organic Alkalinity modulates pH from the Sea-Surface Microlayer during a mesocosm study” -egusphere-2025-5265
General comments: This manuscript presents a well-designed mesocosm study investigating organic alkalinity (OA) dynamics in the sea-surface microlayer (SML) and its impact on pH modulation across phytoplankton bloom phases. The work addresses an important gap in marine carbon system research by providing direct OA measurements in the SML. The methodology is generally sound and well-described, and the conclusions are largely supported by the data. However, several points require clarification and revision before it can be accepted for publication.
Specific Comments:
Page 1, Line 29: “Spearman ρ = -0.75, p = 0.024” – The correlation is mentioned but the sample size (n=10 ?) should be stated for transparency.
Page 2, Lines 56-57: The sentence “Emiliania huxleyi consume HCO3- to form coccolith and produces acidic polysaccharides associated with dissolved oranic matter, whereas diatoms. From diatoms, Cylindrotheca closterium, release organic material relaterd to carboxylate-rich extracellular polymeric substances” contains multiple errors. “oranic” → “organic” ,“relaterd” → “related”. The sentence structure is fragmented and grammatically incorrect.
Page 3, Line 77: “under ssurfactant-enriched” → under surfactant-enriched
Page 4, Lines 100-110:The mesocosm setup description is adequate, but several details are missing: What was the light:dark cycle? How was temperature controlled/monitored? What was the actual nutrient concentration added (μM)?
Page 5, Lines 125-130: The glass plate technique sampling speed (~5 cm s-1) is reported, but the film thickness achieved is not stated. This is important for comparing with other studies.
Page 7, Lines 215-225: The bloom phase definition based on chlorophyll-a is mentioned, but actual Chl-a values are not provided in the main text, including representative values for each phase.
Page 8, Line 238: “CaCO3 dissolution appears to have been the dominant process in the SML” – This interpretation is not strongly supported by the vectors in Figure 1c, which show mixed signals. Please revise or provide additional evidence.
Page 10, Figure 2: The pie charts showing OA/TA fractions are informative, but the exact values (8.1% vs 3.1%) differ slightly from the abstract (8.4% vs 3.1%). Please clarify.
Page 12, Figure 4: The confidence interval for the slope should be reported in the figure caption.
Page 13-15, Lines 349-415: The discussion section is largely descriptive, restating the results rather than exploring mechanistic explanations for OA enrichment and pH modulation. Key mechanistic questions are unaddressed: (1) Why is OA enriched in the SML during bloom phase transitions (e.g., phytoplankton exudation vs. microbial processing vs. physical accumulation)?; (2) What specific functional groups of dissolved organic matter (DOM) drive OA in the SML?
Page 14, Lines 385-395: The comparison with literature values is useful, but the authors should address why their ULW OA values (40-100 μmol kg-1) exceed typical coastal values – could this be an artifact of the mesocosm system or the measurement method?