the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 21 Nov 2024
Pelagic coccolithophore production and dissolution and their impacts on particulate inorganic carbon cycling in the western North Pacific
Abstract. Coccolithophores, a type of single-celled phytoplankton that is abundant in global oceans, are closely associated with the carbonate pump and thus play a crucial role in the marine carbon cycle. Here we investigated coccolithophore abundances, species compositions, coccolithophore calcium carbonate (CaCO3 as calcite) and particulate inorganic carbon (PIC) concentrations in the upper water column of the western North Pacific Ocean, along a meridional transect spanning the oligotrophic subtropical gyre and the nutrient-richer Kuroshio-Oyashio transition region. Our results revealed that Umbellosphaera tenuis was the dominant coccolithophore species in the former, while Emiliania huxleyi and Syracosphaera spp. dominated in the latter. Coccolithophore calcite contributed a major fraction of the PIC standing stocks above a depth of 150 m, among which E. huxleyi was the most important producer while less abundant and larger species also played a role. The coccolithophore CaCO3 production rate in the subtropical gyre (0.62 mol m−2 yr−1) was ~5-fold higher than that in the Kuroshio-Oyashio transition region (0.14 mol m−2 yr−1), indicating that inorganic carbon metabolism driven by coccolithophores is relatively strong in oligotrophic ocean waters. Using a box model including coccolithophore CaCO3 production and metabolic calcite saturation state, we demonstrated that CaCO3 dissolution associated with organic carbon metabolism can generate excess alkalinity in the oversaturated upper water column of the western North Pacific Ocean. Results of our study highlight the critical role of coccolithophores in CaCO3 production and dissolution; knowledge of these processes is important to assess PIC cycling and carbonate pump efficiency in the pelagic ocean.
- Preprint
(2579 KB) - Metadata XML
-
Supplement
(798 KB) - BibTeX
- EndNote
Status: open (until 06 Jan 2025)
-
RC1: 'Comment on egusphere-2024-3492', Alex Poulton, 20 Dec 2024
reply
GENERAL COMMENTS
The article by Han et al. represents a considerable amount of work and presents a wide-ranging number of details in terms of coccolithophore species composition, species-specific PIC concentrations, size-fractionated PIC concentrations, PIC standing stocks, estimates of CaCO3 production related to coccolithophores, statistical exploration of environmental drives of coccolithophore abundance, PIC and species composition, and an exploration of whether metabolically driven dissolution in microenvironments in sinking aggregates can explain shallow dissolution of CaCO3. This extensive list highlights that there is just too much in this single paper and its considerable breadth results in significant missing sections on methods (i.e. all the environmental data) and a full exploration of the limitations of the scaling, statistics and estimates made. A recommended step forward would be to vastly simplify the paper (e.g., focus on coccolithophore species, PIC and CaCO3 production) to allow room to fully discuss the results and insights presented.
That no methodology for the environmental data (nutrients, carbonate chemistry) is included leads to significant issues and concerns over the quality of this data. These are parameters for which there are important international norms to ensure data quality (e.g., use of internal standards, CRMS, measures of precision and replication) and so details need to be included. Further, the data availability statement only leads to a general landing page for the data rather than more direct links to the individual datasets included (e.g., coccolithophore count data, environmental data, hydrographic data). This is not to say there is anything amiss of the data and is clearly a result of the shear breadth of other material included.
Moreover, there are some confusing trends in the article, one of which is the swopping of units (from mol to g and back again), and no clear statement of what elements the units are in fact presenting (C or CaCO3). It took several readings of the paper and associated references to realise that all the mass units (mg, g) are in mg/g of CaCO3 rather than C. It would have been good to have this pointed out early in the manuscript – and sticking to either grams or moles throughout would also have eased the reading and interpretation of the paper.
As well as the missing methods there are also some examples of limited discussion where there is not room in the article to fully explore the scientific and methodological issues raised in a suitable way to allow the reader to understand the results. Several examples of this exist, most importantly: (1) the conversion of cell counts to cell CaCO3 is not a trivial conversion and a more thorough exploration of the possible errors and limitations of these methods should be included; (2) there is no explanation of whether the authors consider 150 m to be a consistent euphotic zone depth across the transect and how integrations to this depth impacts on the research questions asked and conclusions of the paper; (3) the conversion of cell CaCO3 into CaCO3 production involves numerous assumptions and room should be made to present the resulting growth rates from these estimates, and comparison of these with in situ estimates of growth rates and other more direct measurements of CaCO3 production (e.g., Daniels et al., 2018); (4) the three different statistical analysis of coccolithophore trends with environmental data (RDA, correlation, random forest) are presented and summarised in 2 paragraphs of the discussion, which barely allows any setting of the results in the context of coccolithophore ecology; (5) the modelling approach to examine whether metabolic respiration drives CaCO3 dissolution in sinking particles has numerous inherent assumptions and completely ignores the potential role of zooplankton ingestion and digestion.
This last example includes some of the strongest examples of limited discussion, where some of the literature cited makes much more nuanced (or even opposite) conclusions about the relative roles of micro-environment and zooplankton led dissolution. This is an active research field and so the authors must fully discuss the different mechanisms proposed, especially if they are going to conclude that ‘our results indicate that shallow-water CaCO3 dissolution indeed occurs in the western North Pacific Ocean mainly as a result of metabolic acidification in the particulate microenvironment’ (Lns 416-417). Are the authors confident that their model results could not also be interpreted as dissolution in zooplankton guts (i.e. another microenvironment) and that their approach accurately compares both proposed mechanisms of dissolution? This section of the discussion also brings in several totally new concepts (Lns 411-414: Alk*, TA*, CFC age-based RTA, ALK*-transit time distribution, 14C-age based RTA values) to the paper with no background to their meaning, interpretation or limitations which is completely confusing at this (late) stage of the article.
There is considerable merit to the work done by Han et al., however a single paper does not do justice to the data collected or its interpretation, or the serious limitations and assumptions included in the methods employed. Given the missing methods on the environmental data and the brevity with which the results are discussed, it is difficult to fairly assess the work presented.
SPECIFIC COMMENTS
Ln 1 (Title), Coccolithophores are only pelagic (not benthic) so suggest removing pelagic from the title. Also see general comments on content of the paper and consider what has been directly measured (cell counts, PIC) and what has only been estimated (production, dissolution).
Ln 15, When describing a species as dominant, especially in light of the various metrics used in the paper, does this refer to numerical dominance?
Ln 17, Is E huxleyi the most important producer or contributor to the PIC standing stocks? High standing stocks along the sample transect does not necessary equate to CaCO3 production rates as growth rates between species vary and it is not clear if this statement includes detached coccoliths or not.
Ln 18-19, It is strange to present the annual integrals in the abstract rather than the daily rates estimated to be associated with the observations collected. There are a lot of limitations and caveats to go from PIC concentration to growth rate, let alone to scale these to annual rates.
Ln 19-20, What does the line ‘inorganic carbon metabolism driven by coccolithophores is relatively strong in oligotrophic ocean waters’ mean? Its meaning is vague and relatively strong (relative to where?) is not a quantitative term.
Ln 27, Surely CaCO3 production and dissolution are THE two processes associated with CaCO3 cycling; the phrasing here implies that there are other processes involved. Also, it is not the ‘so-called’ carbonate pump, this is what it is called (Ln 28).
Ln 33, This line does not follow (‘this acidification feedback mechanism’) the previous line or make sense. This might also be a good point to introduce carbonate chemistry and would help the reader.
Ln 40, Contrast this line (‘a major fraction’) with the values given in the preceding lines. Please give a value.
Ln 41, What is a ‘data assessment’? Surely the Ziveri et al. (2023) paper is based on observations?
Ln 43, This line (about large uncertainties) should link to the preceding line as there are large uncertainties in the composition of CaCO3 production between the different pelagic groups mentioned in line 42.
Ln 45, Daniels et al. (2018) is not listed in the references. Also, why are the values on Ln 46 in mg m-2 d-1 rather than mol (as in the abstract)? It should also be made very clear that the values given throughout the paper are in CaCO3 (i.e. mol/g = 100). Are the other values in the paper (PIC, CaCO3 production) given in CaCO3 or C?
Lns 48-49, Unnecessary repetition (‘along with the current scarcity of studies’).
Ln 50, This is a bold statement (‘CaCO3 dissolution is generally assumed to mainly occur below the saturation horizon’), especially based on the multiple references included in the article and that shallow dissolution has been recognised since the 1990s (e.g., Milliman e al., 1999 Deep-Sea Research) – this statement should have a reference or be modified.
Ln 57-58, See new paper by Oehlert et al., 2024 (GBC, doi: 10.1029/2024GB008176) arguing that high-Mg fish calcites have a very low PIC:POC and the POC associated with fish faecal pellets provides considerable protection as they rapidly sink.
Ln 59-60, Please expand on this issue in light of: Mayers et al. (2020, doi: 10.3389/fmars.2020.569896) and Dean et al. (2024, doi: 10.1126/sciadv.adr5453) which imply that (micro)zooplankton grazing can be an important loss process for coccolithophore CaCO3.
Ln 62-65, Please consider shortening and simplifying this sentence.
Ln 65-66, Do these differences highlight deep CaCO3 production, or that the CaCO3 production is happening over a deeper euphotic zone?
Ln 66-68, This line needs the relevant citations.
Ln 69, coccolithophore should be coccolithophores.
Lns 74-76, There appears to be 4 research questions here, with (3) split into whether there is shallow dissolution, and a fourth on whether this is driven by metabolic activity. Further, while research questions (1) and (2) are clearly addressed by the data presented in the article, (3) and (4) are based on the addition of modelling to the discussion (see later comments) and it is unclear if these are addressed in the depth that they deserve.
Ln 89, Figure 1 – The colour bar used means that there are two bands of yellow (one around 0.75 and one at the maximum value), which is confusing. Also, the precise dates for the composites are not included (are these 1-30th June)?
Ln 99, Based on the use of ‘filters’ in the next line of the method (‘filters were oven-dried’) suggest adding membrane filters to the preceding line to avoid confusion – unless the SEM filters weren’t dried and stored in petri-slides?
Ln 103, Does large and small fractions of PIC need a unique abbreviation, or could they just be referred to as small and large? There use (and non-use at times) in the article is confusing. Also, does PICtotal need an abbreviation, or could it just be referred to as total PIC?
Ln 110, Do the authors mean the analytical precision was <10%? The use of ‘better than’ is confusing.
Ln 112, Phrasing confusing – sounds like only some of the filters collected were mounted, do the authors mean a portion of of the filters was mounted?
Ln 120-121, Have the authors seen the article by Sheward et al. (2024, doi: 10.1038/s41597-024-03544-1)? A comparison between the cell PIC values in these articles would be a valuable addition to the literature. As the scaling from cells to cell PIC is key to this article, some comments around potential sources of error and limitations (e.g., accurately assessing the number of coccoliths) should be included. Also, ‘all the’ is unnecessary in the second line.
Ln 124, Do the authors mean shallower than 150 m or deeper (i.e. above) 150 m? Also, why 150 m? This is unlikely to be the euphotic depth across the entire transect.
Ln 125-126, Why are the turnover times and division rates given here different than from Ziveri et al. (2024)? What were the growth rates that resulted from these estimates? What are the potential sources of error and uncertainty with this method of estimating CaCO3 production?
Ln 134, No methods are included for the environmental conditions. Much more details are needed for these parameters, such as detailed methods including internal standards, precision estimates, and CRMs. These are all included as standard with the environmental data and as presented as such there is little confidence in the quality of the environmental data as this cannot be assessed as part of the review. As these are key to the following sections of the manuscript and the calculations made, they must be presented in detail here.
Ln 142, Why is only the effect of microenvironment undersaturation considered? Would the same configuration describe dissolution driven by metabolic activity inside zooplankton guts? Surely to test the dominant driver of CaCO3 dissolution, all potential drivers need to be considered? On Ln 143, the assumption that aerobic metabolic activity consumes all ambient oxygen is a significant assumption – do the sub-surface waters go anoxic or are the authors suggesting that internal environments in sinking marine snow are anoxic? If the authors remove this assumption, can they still get these patterns of dissolution?
Ln 166, Do the authors mean nutricline, as related to multiple nutrients, or the nitracline, as related to nitrate. As the authors describe this in terms of nitrate, this seems to be the nitracline. Following on from this (Ln 175), why was the depth where DIN reached 0.1 umol/L used rather than (as more commonly used) 1 umol/L? Was this something to do with the detection limits of the measurements, in that 0.1 umol/L is a common threshold for nitrate+nitrite detection and so this effectively describes the first depth where nitrate was detectable? Without the detailed methods on the environmental parameters, it is difficult to assess this criterion. Finally, why is this depth not plotted on Figure 2?
Ln 171, What is the justification for using 0.03 mg m-3 to define the transition in density at the base of the mixed layer? The authors should give a reference as there are multiple criteria and methods used to define the mixed layer. Also, did the mixed layer depth vary by 11-25 m around a value, or do they mean that the mixed layer varied from 11 to 25 m in depth? How does this compare with their assumption of a 150 m euphotic zone depth?
Ln 181, Figure 2 – The authors should be wary of using ODV with sparse data as it tends to expand data beyond the sample depths with a horizontal bias (kms) rather than vertical (ms) one and creates (sometimes fictional) patterns. These can be seen clearly in this Figure; for example, the DCM (e) is at ~70 m at 35oN, but there are no samples to confirm that, PIC is stretched around 40oN both N and S at a depth of around 30 m but there are no sample points to support this pattern. Similar comments can be made with panels (g) and (f). As these are key parameters for the paper, the authors should consider different ways of plotting the data that more accurately represent the patterns and spatial limitations of the data.
Ln 187, How much of the apparent pattern of low PIC in the surface and increasing with depth is driven by the contouring in Figure 2? In Figure 3 why are no error bars included to give an idea of the relative patterns? Also, in terms of Figure 3, why are the three sets of measurements at different depths - they do not line up? How does this impact the interpretation of the data?
Ln 192, In the methods section (Ln 118) the detection limits for coccolithophore cell counts from SEM is given as 1.87 cells mL-1, but line 192 reports cell concentrations of 0.97 cells mL-1 (=970 L-1)? Please explain this discrepancy.
Ln 196, Here the authors talk about averages, but the preceding lines gave specific concentrations, and it is not clear what this value for coccoliths is for – all depths, all stations? As the difference between subtropical and subpolar waters is a strong theme of the paper wouldn’t it make sense to compare the averages for these?
Ln 211, What does predominantly mean when used here – do these refer to coccosphere abundances?
Ln 213, Do all species listed in lines 211 to 213 compromise >1% of total coccosphere abundance or just F profunda?
Ln 216, Did E huxleyi contribute the largest fraction in every sample?
Ln 219, What is the lower euphotic zone (LPZ) and how are coccolithophore species assigned to this depth? Is there a missing reference here?
Ln 225, Figure 4 – Beyond E huxleyi there is very little detail visible in this plot, could the authors present this as percentage contributions rather than cell abundances? Do the authors want to emphasise abundance or species diversity in this figure?
Ln 229, Is the coccolithophore CaCO3 really 0.00 umol L-1 or <0.01 umol L-1?
Ln 234, Why are exact numbers or ‘ca.’ given elsewhere in the article but “~” is used here to indicate approximately?
Ln 238, Figure 5 – What does this show? Are these averages for the all stations, all depths? How does this add to the theme of the paper in terms of latitudinal or vertical differences? Being Pie Charts this automatically presents percentage contributions, so if these are averages then there is no option to show error bars and compare the different components. Is this the best way to present the data or compare sampling depths/regions?
Ln 234, Does this mean the measurements were integrated from the surface depth to 150 m? What was the argument for 150 m (in view of changes in the euphotic zone depth)? How shallow were the shallow depths and were they consistent? Would it not be more consistent to assume that the unsampled surface depth was the same as the shallowest sampled depth?
Lns 244 onwards, Are these concentrations in mg CaCO3 m-2 or mg C m-2? Are they averages for the different areas, and if so, what are the standard deviations?
Ln 243, Figure 6 – This is a very difficult figure to understand. (a) Why are area averages given now whereas elsewhere in the paper the latitudinal pattern is emphasised? Where are the error bars for the averages? Is the data normally distributed? Why is the coccolithophore contributions surrounded by orange? Are the average bottle and sum of small and large (note no abbreviation of SSF and LSF are used) significant the same when compared? (b) The data here is obviously not normally distributed and generally the estimates of CaCO3 production tend towards lower values – however (!) the blue diamonds, which are not defined in the legend and so are we to assume they are averages, are offset higher than the data is distributed. This looks to be an example of where a geometric mean should be used, and the values should be lower than the blue diamonds – indicating that the authors are overestimating CaCO3 production at the different stations. Please explain.
Ln 258, This is the first mention of in situ pump data since the methods and so is slightly confusing in terms of SSF and LSF (small and large PIC particles). Also, are the values here mg CaCO3 m-2?
Lns 262-263, Not clear at all as to how the CaCO3 production rates presented here follow from the statement about turnover times. Would it not make sense to present the growth rates that derive from the estimates of CaCO3 production and compare these with measured growth rates in the open ocean? Are the growth rates estimated similar between the two different regions and thus the differences (or similarities) driven purely by the standing stocks of PIC? What would clarify this would be to present in Figure 6 the integrated PIC for all the stations, the growth rates estimated for each station and the CaCO3 production – this way, the reader can follow how the three interact.
Ln 267, On lines 265 to 266 the link between high CaCO3 production and high cell numbers is established, whereas the comparative information is not given in Ln 267; does the lowest CaCO3 production correspond to the lowest cell counts?
Lns 271-278, All this material appears to be introductory material and is a repeat of what has been said in the introduction.
Ln 282, The authors should consider citing Poulton et al. (2017, doi: 10.1016/j.pocean.2017.01.003) alongside Balch et al. (2019) in terms of defining depth flora in the subtropics.
Lns 282-285, Statements around coccolithophore nutrient stress and cell quotas are best backed up by physiological references, not modelling references.
Ln 288, O’brien et al. should have a capital; O’Brien et al., 2016.
Lns 292-296, It is impossible to assess the RDA without details of the methods for the environmental data and some statements around whether the authors checked for autocorrelations between the different variables. Also, what is the justification for putting in some variables, for example latitude or depth? Latitude should correlate with temperature and is depth supposed to represent a pressure effect or light availability. It would be good to see some mechanistic physiological exploration of how the variables included in the RDA impact on coccolithophore growth (as this underpins the CaCO3 production estimated). Are all the variables included necessary? This also goes for the Spearman comparison. In general, there is so little discussion of this in the discussion that this analysis comes across as very superficial (13 lines) despite including three different statistical tests of the data that are not found in the results.
Lns 325-327, The authors need to expand on their reasoning here – how do relationships between coccolith and coccosphere concentrations indicate that the detached coccoliths came from living cells and where not (e.g.) advected into the area, from viral lysis of cells or disintegration of the coccospheres after grazing or from faecal pellets or programmed cell death?
Ln 329, Figure 8 – These plots miss the sample number for the different species and it is surprising that despite the variability in presented sample number and data distribution, the p values for all the plots are p<0.01. Are these plots all forced through zero?
Ln 333, This is the first mention of aggregates in the text. The authors need to expand on this and explain how they were measured/identified/classified etc.
Ln 336, How does the statement of the role of large species link to the results presented in the article? Do the authors see these species, and do they make large contributions? Same comment on lns 337-339.
Ln 341, Is not the point being made before that large, rare species can play important roles on CaCO3 production and export? Based on the preceding references (e.g., Daniels et al., 2016) and comments this seems to better follow.
Ln 347-348, Why do the authors consider that ‘CaCO3 is largely produced in the lower layer of the euphotic zone’ when calcification is a light-dependent process? Is this what is seen in other studies of CaCO3 production (e.g., Daniels et al., 2018). References are needed here, and this should be expanded as it is a (potentially) important point. Or, do the authors mean that CaCO3 production occurs over a significant portion of the euphotic zone and not just the surface? As the authors did no measure CaCO3 production, how does this relate to their results?
Ln 353, Is ‘flux’ the correct term in this line? Do the authors mean that subsurface coccolithophore CaCO3 contributed significantly to the total water column PIC? Where does the flux relate to this?
Ln 354, Is this what the authors observed in their data – i.e. large and PIC heavy species in the subtropical gyre?
Lns 356-358, The final statement is confusing – why would low surface PIC values imply that coccolithophore CaCO3 production over the whole water column is not important?
Ln 360, Are the values in this line averages? What are the standard deviations and are these values significantly different?
Ln 361, Is the LSF PIC mentioned here from this study or Ziveri et al. (2023)?
Ln 363-366, These lines need to be better related to the results presented, here they just appear as statements without the context of the present study.
Lns 368-371, This is a large leap here from discussions of pelagic calcifiers and different groups, to how ecosystem structure impacts the relative importance of different calcifiers – for example there are no data or statements in terms of other phytoplankton groups and the nutrients they are dependent on, the availability of prey for heterotrophic calcifiers, etc. This statement needs far more background and justification. No abundances of non-calcareous phytoplankton are given or attempts to estimate the amount of the phytoplankton community that coccolithophores represent, or how these changes between sites.
Ln 375, Figure 9 – It is not possible to see all the Pie Charts of the data from the present study so represents a poor figure trying to compare studies. Also, how do the authors know that the <51 um fraction is only coccolithophores and (e.g.) not shell fragments of foraminifera?
Ln 381, As the title of the paper is around CaCO3 production it is surprising to see that only two paragraphs of the discussion cover this subject – it seems rather a light touch for a key parameter of the paper. Have the authors considered comparing their measurements with other sources of rate data (e.g., Daniels et al., 2018) or estimates of coccolithophore growth rates in the ocean?
Ln 383-384, Can the authors check the values given from Balch et al. (2007) and Hopkins and Balch (2018) as we cannot find these area specific rates in either publication, both deal with global PIC production using slightly different (but related) physiological-based models.
Lns 390-392, Does this statement relate to the data presented in this article? If so then it is a concern and considerably more should be said as to possible sources of these errors and inconsistencies (e.g., scaling from cell numbers to PIC concentrations).
Ln 399, Do the authors consider 150 m to be the euphotic zone along the entire transect? This seems highly unlikely due to the changes in vertical distribution of the phytoplankton community (i.e. Chl-a). Some statements are needed in the article about what 150 m represents and what the authors consider in terms of the difference (or not) between the euphotic zone depth and 150 m.
Ln 404-406, How did the authors do a one-way ANOVA on PIC distribution patterns? More needs to be said or shown in terms of what data has been compared and how the authors consider different water parcels or hydrographic regions. This seems a rather throw away comment that needs far more background; it would not be normally expected that an ANOVA on hydrographic data distributions is a valid statistical approach.
Ln 409, This is a significant assumption (that 100% of coccolithophore production is exported out of the euphotic zone) and needs far more discussion, including references and exploration of how if this was wrong it would impact on the article’s conclusions. Recognising that coccolithophores could be heavily grazed by microzooplankton (Mayers et al., 2019, Dean et al., 2024) and are prone to significant levels of viral infection and lysis (e.g., Vincent et al., 2023, doi: 10.1038/s41467-023-36049-3) severely questions how valid this assumption is.
Lns 412 and 414, Introducing totally new concepts (TA*-CFC age-based RTA, Alk*-TTD 14C age based-RTA) at this (late) stage of the article (i.e., concepts not mentioned in the introduction) makes this section of the discussion very confusing and difficult to follow. That these important points are only covered in 4 lines also leaves little room to fully explore them.
Lns 416-417, Is this self-validation? If the model did not assume that all production is exported (c.f. POC production), that all oxygen is consumed and that marine snow or faecal pellet sinking speeds are unlikely to be as slow as 10 m d-1, what conclusions do this lead to? Do the authors have data to support the assumptions (e.g., amount of CaCO3 production exported, characteristics of sinking particles in terms of marine snow or faecal pellets, sinking speeds, oxygen gradients in porous marine snow) in the model that validates them? Could their result be interpreted in a different way if CaCO3 dissolution occurs via the metabolic activity of zooplankton? Does their model rule out grazing as a contributor to CaCO3 dissolution or does it just assume that it doesn’t occur?
Ln 420, Figure 10 – Panel (a) is relatively easy to see, though it would be good to see an exploration of sinking speeds between 10 and 100 m d-1 (e.g., 50, 75 m) and some comparison with the literature on particulate sinking speeds of marine snow and fecal pellets (e.g., see Jansen et al., 2002 who conclude ‘respiration-driven dissolution of calcite in the water column seems to be unlikely .. as the size, settling velocity and porosity of marine snow aggregates are unfavourable for creating a microenvironment with gradients sufficient to convert an oversaturated bulk environment into a locally undersaturated state.’). Please include more discussion of the factors involved than citing only one side of the debate (e.g., Jansen and Wolf-Gladrow, 2001). However, note that Jansen and Wolf-Gladrow (2001) conclude that ‘up to 70% of the ingested carbonate maybe be dissolved in the guts’ in non-bloom situations – could the authors explain where the 25% in line 437 is from? Panel (b) is extremely difficult to understand, for example why are the depths now in density and how does this relate to panel (a) are the densities for the other points the same, how do the lines where no data points included relate in depth to the discrete measurements?
Ln 430, Subhas et al (2022) do not fully agree with the present study; instead finding that ‘a combination of dissolution due to zooplankton grazing and microbial aerobic respiration within degrading particle aggregates’ lead to shallow metabolically driven dissolution. Subhas et al. (2022) also point out that the dissolution of high-Mg carbonates is not necessary to drive shallow dissolution. This is not reflected in lines 434-436. Further, the single comment from Jansen and Wolf-Gladrow (2001) does little justice to the existence of multiple dissolution pathways (especially noting as above that this reference concludes that up to 70% of ingested carbonate may be dissolved through zooplankton ingestion).
Ln 452, Data Availability - The authors need to provide a more direct link to the relevant data from the article than the general landing page for the Science Data Bank.
Citation: https://doi.org/10.5194/egusphere-2024-3492-RC1 -
RC2: 'Comment on egusphere-2024-3492', Chloe Dean, 20 Dec 2024
reply
Review of “Pelagic coccolithophore production and dissolution and their impacts on particulate inorganic carbon cycling in the western North Pacific” by Han et al.
The authors clearly demonstrate the substantial contribution of coccolithophores to the production of calcium carbonate standing stocks in the western North Pacific through their presentation of both Niskin bottle and size fractionated PIC measurements, as well as coccolithophore cell abundances, species compositions and diversity. The importance of metabolically driven calcite dissolution is then illustrated through the use of a simple box model, which builds upon previous studies by incorporating published data for regenerated total alkalinity and PIC flux. The results presented in this paper build upon similar studies (Ziveri et al., 2023, Subhas et al., 2022), and also integrate emerging methodologies and ideas for investigating the calcium carbonate cycle. In particular, biologically driven dissolution within the supersaturated “shallow” ocean is a quickly emerging mechanism which is helping to constrain the long-standing discrepancy between calcium carbonate production and export. Broadly, I am impressed by how the author’s weaved together multiple lines of evidence to understand coccolithophores' role in the calcium carbonate cycle, especially through the lens of coccolith calcite production and dissolution. To that end, I have a few comments which I hope will help to strengthen the manuscript and improve the scope of the content.
The authors do a fairly thorough job of describing their assumptions, and for the most part, clarify how these translate into uncertainty within their analysis. There are a few assumptions in particular which I believe could be more thoroughly justified by the authors, 1) the assumption that satellite data underestimates PIC standing stocks, 2) the use of a range of reported PIC turnover times, and 3) the assumptions around PIC content of coccoliths, and the discussions around free liths vs coccospheres.
For the satellite imagery statement (lines 65-66) I am suggesting that you add a citation from Neukermans et al., 2023, which provides direct evidence for the discrepancy you are pointing out here. Additionally, given this stated uncertainty in satellite estimates of [PIC], I found it confusing to follow the author's justification of using satellite derived [PIC] (both from July 2022 and annual mean) to correct the PIC production rates for seasonal variability (lines 131-133). I would appreciate some commentary from the authors regarding their justification for this, for example, similar to the Supplementary Figure 2 from Ziveri et al., 2023, which illustrates the correction factor between satellite estimates and direct measurements of PIC for a given region.
With regards to the use of the reported range of turnover times, I think it would be worthwhile for the authors to comment on the large uncertainty that this method results in (i.e. the high production tails in Fig. 6b that stem from the order of magnitude range in turnover time estimates [0.7-10 days; Lines 124-126]), and how it may hinder interpretation of this field data. There are currently a few methods which can get at in situ turnover time, primarily the use of carbon isotopes spikes in incubations, with the carbon-14 method described in Graziano et al., 2000. While I completely understand that it is not possible to do this for this study, I do think it would be helpful for the authors to acknowledge that direct measurements of turnover time could have likely reduced uncertainty within the PIC production values. Especially when considering the challenges around quantifying the oceanic calcium carbonate budget, any advancements in our ability to reduce error and uncertainty should be at the fore-front of our minds.
The author’s use of average coccolith/coccosphere calcite to calculate out the PIC inventory and PIC production rates could warrant more discussion on uncertainty and in general, the nuances of these assumptions. My concerns around this approach stem from the fact that Johns et al., 2023 clearly showed that the production and subsequent cycling of free coccoliths can be rapid and complex. Using a generalized [PIC] quota for each species of coccolithophore will not capture the dynamics of coccolith reabsorption, and I would suggest the authors acknowledge this. Additionally, in Lines 325-327, the authors state that “detached coccolith concentrations of… showed a significant positive relationship with their coccosphere cell concentrations, indicating that those detached particles were likely to have originated from living cells”. This statement is a bit misleading, as it supposes that free liths are only coming from cells that are no longer alive. I suggest modifying in consideration of the Johns et al 2023 paper, which shows that lith shedding is a dynamic process throughout the cell's life cycle. In general, I would like the authors to comment further on the dynamics of coccolith production and shedding (i.e. how it changes with dominant species shifts, how it may be depth dependent and subsequent consequences to different dissolution processes [particle vs gut], etc).
While the author's presentation of the environmental data and its influence on the coccolithophore community structure is certainly an interesting and impressive data set, I feel that it generally distracts the reader from the main goals of this manuscript. Given that the title is “Pelagic coccolithophore production and dissolution and their impacts on particulate inorganic carbon cycling in the western North Pacific”, I think the authors should stick to these research aims in the manuscript, and either try to weave the environmental aspects of the study more broadly throughout the manuscript, or consider presenting the environmental-drivers as a follow-on manuscript.
I appreciate the author's approach to assessing the role of metabolically driven dissolution, as this is an emerging mechanism impacting the shallow calcium carbonate cycle that has major implications for the field. I have some suggestions for interpreting the box-model dissolution output against previously reported observations of alkalinity regeneration (presented in Fig 10b). I was surprised the authors did not comment on the mismatch between the maxima of alkalinity regeneration between observations and model output. The observations show a clear peak alkalinity regeneration well above 500m, whereas the model output shows the peak occurring at or near 500m. I believe this mismatch is due to unique “metabolic” processes that occur at different depths in the water column. For example, micro- and macro-zooplankton primarily graze within that upper 500m, and therefore, could be the primary drivers of that shallow alkalinity regeneration due to calcium carbonate dissolution during ingestion and digestion processes. A recently published paper (Dean et al., 2024) titled provides experimental evidence that microzooplankton facilitate a substantial amount of coccolith calcite dissolution through grazing and digestion. I bring this up, because the model, and broader context of this discussion, is largely focused on particle associated metabolic dissolution. If the authors were to consider the different types of metabolic dissolution that occur at different depths in the water column, they may be able to directly comment on the maxima mismatches in Fig 10b and provide a stronger discussion for section 4.4. This is an active area of research which warrants a deeper discussion from the authors.
Overall, there is a lot of valuable and needed field data being presented in this manuscript which certainly contributes to our understanding of calcium carbonate cycling dynamics, especially as driven through coccolithophores. My overall suggestion to the authors is to really hone in the scope of this manuscript, and keep the focus to coccolith calcite production and dissolution. The environmental data, while incredibly interesting, could enhance the scope of this study if integrated into the discussion further, or would serve well being presented independently as a follow-on study.
Additional minor comments:
- I might suggest modifying the title to be “Coccolithophore production and dissolution and their impacts on pelagic particulate inorganic carbon cycling in the western North Pacific”, given that “pelagic” is really in reference to the depths of PIC cycling, and not coccolithophore PIC production, which is always pelagic.
- Line 59, could expand from just “zooplankton” to “micro and meso zooplankton”. Could also add a citation for my recent publication (Dean et al., 2024) which provides direct experimental evidence for microzooplankton facilitated dissolution, and shows that microzooplankton food vacuoles are acidic and eventually buffered from coccolith calcite dissolution.
- Citation for lines 65-66, regarding satellite PIC and measured/integrated PIC discrepancies?
- Suggestion is Neukermans et al., 2023 (see end of review for reference list)
- Line 119-120: Considerations around using an average coccolith calcite mass and potential error?
- While most other stated assumptions are followed with a comment on the uncertainty they bring, I found this assumption to be not acknowledged at the same level as others. Would suggest adding a statement which at least speculates on the uncertainty this is adding to your calculations. For example, different coccolithophore species have very differing growth rates, and consequently, different PIC production rates/turnover times. Given the marked differences in coccolithophore species composition reported in this study, I believe the authors should expand upon this section.
- Line 230-231: Clarification Question, is this in reference to all E. huxleyi associated calcite (i.e. coccospheres and liths)? Or just coccospheres?
- Fig 5.
- Suggestion for readability: Add labels to the pie charts to make it easiest to compare coccosphere to liths to total calcite
- Line 243: Why a depth of 150m for standing stock integration? I don’t think this was stated in the methods or results. Can you comment on the implications of setting the same depth for this integration?
- Fig 6b.
- I had never seen a violin plot before, so it was a bit difficult for me to interpret this figure on my first pass. I would suggest that the authors provide additional description in the caption to at least describe what the blue diamond, blue line, and shape of the plots represent.
- Line 292: RDA has not been previously defined. Only defined in Fig. 7 caption, so you may want to define it in the text. Additionally, I think the authors could provide additional justification for their choice in using RDA for this study. I would suggest adding this in the methods section, so that someone who is less familiar with multivariate analyses can understand why this is appropriate here.
- Fig 8a.
- Clarifying, does coccolithophore calcite = coccosphere and coccoliths? Or just coccosphere?
- Line 333: what is the “aggregation group”? Can you please explain further?
- Lines 339-341: “Thus, although E. huxleyi is one of the most abundant species in the ocean, larger coccolithophore species can also play an important role in CaCO3 export.”
- I would like the authors to provide some justification for this statement from the data presented in this manuscript. It seems that Fig 5. Could be modified to illustrate this point, perhaps by adding percentages and errors to the pie charts?
- I don’t think Fig 5 does a good job of illustrating this point, since they are grouped into the “other” category. I would suggest having a supplementary figure which shows the contributions of these larger species to the total inventory, so that you might reference them in this statement.
- Line 345-347: Does this need a citation?
- With respect to the DCM being deeper than 100m in subtropical gyres
- Lines 347-348: Can the authors provide a value or range of values for the underestimation here?
- Lines 369-372: Did the authors look at any metrics for the composition of the non-coccolithophore community members? Since there is a point here about competition, and the authors did such a thorough job of investigating the coccolithophore diversity, it would be interesting to extend that thought to the results of this study.
- Line 394: “high dynamics” seems vague to me, but perhaps I’m misinterpreting this?
- Line 401: I believe there is a typo in this sentence? “The results suggest that CaCO3 might start to dissolve in “setting” marine particles after sinking out of the…” Should “setting” be “settling”?
- Line 436-437: Could cite my recent publication (Dean et al 2024) to show that MZP also contributes substantially to dissolution in lab study.
References
Dean, C.L., Harvey, E.L., Johnson, M.J., Subhas, A.V. Microzooplankton grazing on the coccolithophore Emiliania huxleyi and its role in the global calcium carbonate cycle. Science Advances, 10 (2024).
Graziano, L.M., Balch, W.M., Drapeau, D., Bowler, B.C., Vaillancourt, R., Dunford, S. Organic and inorganic carbon production in the Gulf of Maine. Continental Shelf Research, 20 (2000).
Johns, C.T., Bondoc-Naumovitz, K.G., Matthews, A., Matson, P.G., Iglesias-Rodrigues, D.M., Taylor, A.R., Fuchs, H.L., Bidle, K.D. Adsorptive exchange of coccolith biominerals facilitates viral infection. Science Advances, 9 (2023).
Neukermans, G., Bach, L.T., Butterley, A., Sun, Q., Claustre, H., Fournier, G.R. Quantitative and mechanistic understanding of the open ocean carbonate pump - perspectives for remote sensing and autonomous in situ observation. Earth-Science Reviews, 239 (2023).
Citation: https://doi.org/10.5194/egusphere-2024-3492-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 107 | 29 | 5 | 141 | 32 | 2 | 0 |
- HTML: 107
- PDF: 29
- XML: 5
- Total: 141
- Supplement: 32
- BibTeX: 2
- EndNote: 0
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1