Extending planktic foraminiferal Mg/Ca palaeothermometry into polar temperature ranges: crust- and lamellae specific calibrations and non-thermal controls

Westgård, Adele; Ezat, Mohamed M.; Sykes, Freya E.; Meilland, Julie; Chalk, Thomas B.; Milton, J. Andy; Chierici, Melissa; Foster, Gavin L.

doi:10.5194/egusphere-2026-2415

Preprints

https://doi.org/10.5194/egusphere-2026-2415

Preprints

06 May 2026

| 06 May 2026

Extending planktic foraminiferal Mg/Ca palaeothermometry into polar temperature ranges: crust- and lamellae specific calibrations and non-thermal controls

Adele Westgård, Mohamed M. Ezat, Freya E. Sykes, Julie Meilland, Thomas B. Chalk, J. Andy Milton, Melissa Chierici, and Gavin L. Foster

Abstract. The rapidity of climate change in the polar regions underscores the need for improved understanding of its impacts on ocean circulation at both regional and global scales. Reconstructions of past polar ocean-cryosphere interactions can provide this context, but large uncertainties in existing proxies limit the utility of such studies. For instance, there are currently no low-temperature (<9 °C) culture-based Mg/Ca-calibrations for planktic foraminifera, a key tool for reconstructing past changes in ocean temperatures. There is also limited understanding of non-thermal influences on Mg/Ca in Neogloboquadrina pachyderma, the only modern polar planktic foraminifera. Moreover, this species exhibits considerable levels of heterogeneity in composition precipitating a thick lower-Mg/Ca outer crust over higher Mg/Ca inner lamellae calcite; specimens with predominantly, albeit variable crust–lamellae proportions, are thus thought to introduce substantial uncertainty into high-latitude palaeotemperature reconstructions. Here, we used N. pachyderma cultured across a range of temperatures, salinities, and carbonate chemistry conditions including experiments in which pH and [CO₃^2-] either covaried or were decoupled. By using a laser ablation approach to analyse crust and lamellae separately, we present new Mg/Ca-temperature calibrations for each component that extend culture-based calibrations in N. pachyderma down to the lower temperature-range (2–9 °C). The crust-specific calibration is of particular importance in high-latitude downcore records where N. pachyderma are commonly observed to preserve predominantly or only crust. Our results show significant carbonate chemistry influence on Mg/Ca with opposite influences from pH and carbonate ion concentration, when these variables changed in isolation. Additionally, we show that environmental conditions regulate crust-lamellar proportions, where increased salinity and temperature, and lower pH lead to less crust formation with implications for future ocean acidification and Arctic Atlantification, and for downcore reconstructions.

Received: 27 Apr 2026 – Discussion started: 06 May 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, 1637 KB)

Supplement (533 KB)

Download & links

Adele Westgård, Mohamed M. Ezat, Freya E. Sykes, Julie Meilland, Thomas B. Chalk, J. Andy Milton, Melissa Chierici, and Gavin L. Foster

Status: final response (author comments only)

RC1:
'Comment on egusphere-2026-2415', Lennart de Nooijer, 24 Jun 2026

Review Westgard et al., Biogeosciences (Egusphere-2026-2415)

Dear editor,

with great interest I read the manuscript of Westgård and co-workers. Although many results from the experiments were presented earlier, the elemental data on the cultured N. pachyderma are very interesting and form the basis of an important contribution to our field. I understand that culturing planktonic foraminifera is very challenging and this alone makes that the authors should be applauded for their work. Overall, the manuscript is well-written and the data is sound. However, I do have some issues with the translation from data to results and the use of statistics (or absence thereof). The annotated pdf contains a number of smaller suggestions, and some of them overlap with the major concerns, which are listed below. Another approach to the data may lead to different conclusions, so that some of the discussion may have to be rewritten (which I therefore did not always comment on in detail).
Sincerely,

Lennart de Nooijer

Statistics
To start with, there should be a separate section on the statistics used (lines 260-266). How exactly were the regressions calculated and was there (e.g.) a strategy to identify/ exclude outliers? Why were the uncertainty envelopes not calculated for the plots with the median Mg/Ca and D_Mg? Related: I am not a big fan of R or R² as a metric. It says something about the spread of the data, but not about the significance of the correlation (R can be very low, but still p<<< and vice versa). This means that for figure 1, I don’t see the reason for including a trendline if R is >0.5 even when p>0.1 (i.e. no significant trend).
Such a section would also have to explain, for example, why the chosen response model for the regression analysis is the most appropriate. For example, the choice for a linear relationship for the upper panels of figure 3: it seems that an asymptotic function would be just as appropriate (and wouldn’t result in negative Mg/Ca at very high pH). In short, please include a justification for the chosen function.

Maybe more importantly, the authors use the data for isolated analyses of Mg/Ca-environmental pairs. This is a shame, however, as for all treatments, all parameters are measured. I would argue that particularly a dataset like this would be best analysed by some sort of multivariate statistical approach.
Interpretation of trends
Another reason for my suggestion is that (table 2) although the ‘pH stable, varying [CO₃^2-] treatment, the pH is actually not that stable (varies from 7.83 to 8.12). This seems the case for many parameters, which are less stable than suggested in the text: this makes it hard to contribute changes in Mg/Ca to (e.g.) [CO₃^2-] only. In this particular case, see also the significant correlation between pH and [CO₃^2-] in Table S1. Rather than focusing on the trends per treatment (as done now), all of the data combined may provide a much more robust idea on the effect of all (calculated) parameters (and their interaction) on the Mg/Ca. This is important, as it matters a lot for most of the discussion (do pH and carbonate ion concentration have opposite effects on Mg incorporation?). A more rigorous analysis of the data may lead to more reliable statements of what is, and what is not correlated in this dataset.
Adjustment of the discussion: the section on biomineralization (4.5) makes perfect sense given the preceding sections, but given my comments on the relation between pH and [CO₃^2-] in this dataset, may need to be adjusted. Given the content of the discussion in this section, the analysis by Ries (2011), GCA 75: 4053 (http://dx.doi.org/10.1016/j.gca.2011.04.025), may be a good framework to place some of the discussion on this point in.
While looking into the numbers, I noticed that the data in figure 3 and in table 2 don’t match. I tried to do some statistics on the data, but I get different results: for example, the right panels in figure 3 have data for four [CO₃^2-], but table one only has three [CO₃^2-] (64, 195 and 206 µmol/kg). Please check carefully.
At various places, trends are named ‘significantly different’, but an actual statistical test to base this on, is lacking. For example, lines 310-311 and figure 2: Since the LC+CC should be an average of the LC and CC, but clearly isn’t, it is hard to believe that the trendlines for only CC and only LC are significantly different.
Miscellaneous
Lines 311-312 and figure 2: what does it mean that the median Mg/Ca is also correlated to temperature? I don’t see the added value of showing both the median Mg/Ca and the D_Mg, as they are essentially the same (due to the similar seawater Mg/Ca across this dataset).

Citation: https://doi.org/10.5194/egusphere-2026-2415-RC1
- AC1: 'Reply on RC1', Adele Westgård, 20 Aug 2026
  
  Response to reviewers
  Reviewer 1:
  Dear Lennart de Nooijer,
  Thank you for the thoughtful and constructive feedback, we (the authors) appreciate your thorough feedback and believe it will help us improve the manuscript. Overall, we agree with your input and see the need for clarification regarding statistical methods. For example, other regressions models and statistical tests than those described were tested, and the best fitting model was applied, however this was not explicitly stated in the manuscript. We also agree with your suggestion for some additional statistical tests to further the robustness of our output. These points will be rectified in the revised manuscript.
  See detailed responses to your comments below. As the comments in the manuscript file mostly overlap with comments discussed below, please see our responses there. The remaining comments are also responded to below.
  Sincerely,
  Adele Westgård
  On behalf of the authors
  
  Comment #1
  Statistics:
  To start with, there should be a separate section on the statistics used (lines 260-266). How exactly were the regressions calculated and was there (e.g.) a strategy to identify/ exclude outliers? Why were the uncertainty envelopes not calculated for the plots with the median Mg/Ca and D_Mg? Related: I am not a big fan of R or R² as a metric. It says something about the spread of the data, but not about the significance of the correlation (R can be very low, but still p<<< and vice versa). This means that for figure 1, I don’t see the reason for including a trendline if R is >0.5 even when p>0.1 (i.e. no significant trend).
  Such a section would also have to explain, for example, why the chosen response model for the regression analysis is the most appropriate. For example, the choice for a linear relationship for the upper panels of figure 3: it seems that an asymptotic function would be just as appropriate (and wouldn’t result in negative Mg/Ca at very high pH). In short, please include a justification for the chosen function.
  Response
  Data cleaning was performed using the two data processing scripts described in the methods “LabGrown” and “IDCrust” (Westgård et al., 2026), additional steps to identify and exclude outliers was not performed. Linear and non-linear regression models were tested for all Mg/Ca-environmental relationships and the best fit models (based on p-value, R², and t-statistics) were included in the manuscript. The authors will expand on statistical approaches in the revised manuscript.
  The authors understand your reluctance about R and R², particularly in regard to the Mg/Ca regressions, this is a metric that aided in the decision of best-fit models and is expected to be reported by the paleo-community, thus included. We have, however, only weighed the significance (p-value) in relation to Mg/Ca in our discussion. T-statistics were also calculated for each regression line, these will also be included in the revised manuscript, further indicating the significance of each correlation.
  Uncertainty envelopes were only reported for the regressions where all data was included, not also for medians and D_Mg. Given our treatments do not have the same number of data points (because the foraminifera did not grow equally in all treatments), reducing the data to treatment medians will weigh each treatment equally in a regular regression model. Given the large difference in Mg/Ca between chambers and individuals obtained in both this and other individual foraminifera analysis studies (Sykes et al., 2026; Westgård et al., 2026; Fehrenbacher et al., 2024; Hupp and Fehrenbacher, 2023, 2024; Groeneveld et al., 2019; Davis et al., 2020) fewer data-points means higher uncertainty, thus much higher uncertainty median-values. The uncertainty envelope therefore becomes quite large and with little additional value for treatment medians (and D_Mg). However, the regression models with all data included weights treatments with more data (thus also lower uncertainty) higher than treatments with less data. We therefore interpret these regressions as more accurate and precise than regressions based of medians and recommend these primarily in down-core reconstructions. Additionally, the regressions based on treatment medians were primarily included as comparison to traditional solution-based analysis where an average of the population Mg/Ca is produced. Whereas the D_Mg regression lines are interesting and useful to show, however not necessary in the main manuscript. These will be moved to supplements to make the figures cleaner. This will also be clarified in the revised manuscript.
  
  Comment #2
  Maybe more importantly, the authors use the data for isolated analyses of Mg/Ca-environmental pairs. This is a shame, however, as for all treatments, all parameters are measured. I would argue that particularly a dataset like this would be best analysed by some sort of multivariate statistical approach.
  Response
  We agree that the multivariate nature of the dataset provides an opportunity to evaluate the combined influence of the environmental parameters on Mg/Ca. We have therefore added partial least-squares (PLS) regression analyses to the revised manuscript, performed separately for the crust, lamellar, and crust+ lamellar datasets.
  The PLS analysis provides additional insight into the relative contribution of the different environmental parameters while also accounting for their covariance in the dataset. For example, in the crust dataset, a two-component PLS model including temperature, salinity, pH, and carbonate ion explains approximately 22% of the variance in crust Mg/Ca (R² = 0.22; RMSE = 1.48). Temperature and pH have the highest variable scores, while carbonate ion also contributes substantially; salinity is a comparatively weak predictor. The PLS-analysis indicate positive correlation for Mg/Ca-temperature, -carbonate ion, and -salinity, and negative Mg/Ca-pH correlation. These results are broadly consistent with the patterns identified from the individual Mg/Ca-environmental relationships, while also highlighting the substantial unexplained variability in the data. Equivalent analyses for the lamellar and crust+lamellar datasets show similar patterns and will also be included in the revised manuscript. We note, however, that a more comprehensive multivariate calibration, particularly one aimed at resolving interactions among environmental parameters, could benefit from additional experiments in which these parameters are systematically varied across a broader range of combinations e.g., varying carbonate chemistry and salinity independently across a range of temperatures to better resolve potential interactions among the environmental parameters. This limitation and the need for such experiments will be more explicitly discussed in the revised manuscript.
  Finally, our primary objective is to develop an empirically useful Mg/Ca–temperature calibration for N. pachyderma that can be readily applied in polar palaeoceanographic settings. In many palaeoceanographic applications, independent constraints on salinity and carbonate chemistry are not available, making a temperature calibration based on Mg/Ca particularly practical. We will clarify this distinction in the revised manuscript: while a fully constrained multivariate calibration is an important longer-term goal, the present dataset provides a basis for an Mg/Ca–temperature calibration while the PLS analysis provides a more integrated assessment of the environmental controls on Mg/Ca.”
  
  Comment #3
  Interpretation of trends:
  Another reason for my suggestion is that (table 2) although the ‘pH stable, varying [CO₃^2-] treatment, the pH is actually not that stable (varies from 7.83 to 8.12). This seems the case for many parameters, which are less stable than suggested in the text: this makes it hard to contribute changes in Mg/Ca to (e.g.) [CO₃^2-] only. In this particular case, see also the significant correlation between pH and [CO₃^2-] in Table S1. Rather than focusing on the trends per treatment (as done now), all of the data combined may provide a much more robust idea on the effect of all (calculated) parameters (and their interaction) on the Mg/Ca. This is important, as it matters a lot for most of the discussion (do pH and carbonate ion concentration have opposite effects on Mg incorporation?). A more rigorous analysis of the data may lead to more reliable statements of what is, and what is not correlated in this dataset.
  Adjustment of the discussion: the section on biomineralization (4.5) makes perfect sense given the preceding sections, but given my comments on the relation between pH and [CO₃^2-] in this dataset, may need to be adjusted.
  Response
  The authors agree with your point and will adjust our language regarding co-variance of treatments in the revised manuscript to reflect the values better. Following your above suggestion, we will also add PLS analysis to the revised manuscript. The PLS analysis suggests that the observation of opposing CO3 and pH influence on Mg/Ca is significant and accurate. However, we also see high variance in the data which implies that further experiments in (fully) decoupled salinity, pH, CO3 (and additional HCO3 experiments) conditions would further disentangle these relationships in the future. We will add results from the PLS analysis to the discussion in the revised manuscript.
  
  Comment #4
  Given the content of the discussion in this section, the analysis by Ries (2011), GCA 75: 4053 (http://dx.doi.org/10.1016/j.gca.2011.04.025), may be a good framework to place some of the discussion on this point in.
  Response
  Thank you for this suggestion. We will include Ries (2011)’s framework in the revised discussion.
  
  Comment #5
  While looking into the numbers, I noticed that the data in figure 3 and in table 2 don’t match. I tried to do some statistics on the data, but I get different results: for example, the right panels in figure 3 have data for four [CO32-], but table one only has three [CO32-] (64, 195 and 206 µmol/kg). Please check carefully.
  Response
  Thank you for pointing this out, there seems to have been a mistake when transferring data between documents, this will of course be rectified.
  
  Comment #6
  At various places, trends are named ‘significantly different’, but an actual statistical test to base this on, is lacking. For example, lines 310-311 and figure 2: Since the LC+CC should be an average of the LC and CC, but clearly isn’t, it is hard to believe that the trendlines for only CC and only LC are significantly different.
  Response
  The category CC+LC is not average of the CC and LC only categories. As not all laboratory-grown chambers were crusted, and in some cases, only the crust component was laboratory grown (see lines 250-253), the CC and LC categories do not have the same number of datapoints, and the CC+LC category only includes shells where both components were laboratory-grown. The LC and CC categories therefore have more datapoints (and differing numbers of datapoints) than the CC+LC, meaning the CC+LC does not equate to the median or mean of the two previous categories. As CC+LC therefore has fewer data points and is skewed by any differences in crust and lamellar calcite values, it can (and do) obtain different regression lines than the two components separately. This will be further clarified in the revised manuscript, and we will confirm that all use of “significantly different” and similar terms is based on statistical confirmations.
  
  Comment #7
  Miscellaneous
  Lines 311-312 and figure 2: what does it mean that the median Mg/Ca is also correlated to temperature? I don’t see the added value of showing both the median Mg/Ca and the D_Mg, as they are essentially the same (due to the similar seawater Mg/Ca across this dataset).
  Response
  See our response to comment #1.
  
  Comments from the manuscript file:
  The minor comments (e.g., syntax and spelling errors) will be revised appropriately in the updated manuscript. Some comments largely overlap with previous comments, e.g., regarding statistical methods and the combined crust and lamellar calcite category, see relevant responses above.
  
  Comment #8
  Related to Fig 2 and 3, and table 4:
  Thank you for the suggestions to make figure 2 & 3, their captions and related table 4 more readable, we will take them onboard.
  “There is a choice whether to plot the first column against pH (as done here) or against [CO32].” What you suggest here is plotted in Fig. S3 and discussed in section 4.2. The trend is negative also when plotted against CO3 and supports the observations made earlier. We will add a reference to this in the figure 3 caption for clarity.
  “Sometimes uncertainties are included (e.g. for the variables in teh Mg/Ca-temperature calibrations), but not for other (significant) correlations (all carbonate chemistry treatments). Why?” The uncertainties of the coefficients were only included for the temperature calibrations as that is the ones to apply in downcore reconstructions. However, we see that the others are just as relevant here and will be added in the revised manuscript.
  
  Comment #9
  Section 4.3: Regarding environmental influence on shell thickness and the proportion crust. We will perform additional statistical tests to check the robustness of these observations. Regarding the Atlantification (warming and increased salinity) and acidification discussion point we acknowledge that the true effect will be hard to quantify as increased salinity and acidification may have negative effects on shell thickness, but temperature may buffer it, which was also suggested by Manno et al. (2012). We will clarify in the revised manuscript.
  
  References
  Davis, C. V., Fehrenbacher, J. S., Benitez-Nelson, C., and Thunell, R. C.: Trace Element Heterogeneity Across Individual Planktic Foraminifera from the Modern Cariaco Basin, Journal of foraminiferal research, 50, 204–218, 10.2113/gsjfr.50.2.204, 2020.
  Fehrenbacher, J. S., Hupp, B. N., Branson, O., Evans, D., Foster, G. L., Glock, N., Thirumalai, K., and Wycech, J.: Individual foraminiferal analyses: a review of current and emerging geochemical techniques, Journal of Foraminiferal Research, 54, 312–331, 10.61551/gsjfr.54.4.312, 2024.
  Groeneveld, J., Ho, S. L., Mackensen, A., Mohtadi, M., and Laepple, T.: Deciphering the Variability in Mg/Ca and Stable Oxygen Isotopes of Individual Foraminifera, Paleoceanogr Paleoclimatol, 34, 755–773, 10.1029/2018PA003533, 2019.
  Hupp, B. N. and Fehrenbacher, J. S.: Geochemical differences between alive, uncrusted and dead, crusted shells of Neogloboquadrina pachyderma: Implications for paleoreconstruction, Paleoceanography and Paleoclimatology, 34, e2023PA004638, https://doi.org/10.1029/2023PA004638, 2023.
  Hupp, B. N. and Fehrenbacher, J. S.: Intratest trace element variability in polar and subpolar planktic foraminifera: insights into vital effects, ontogeny, and biomineralization processes, Journal of Foraminiferal Research, 54, 355–374, 10.61551/gsjfr.54.4.355, 2024.
  Manno, C., Morata, N., and Bellerby, R.: Effect of ocean acidification and temperature increase on the planktonic foraminifer Neogloboquadrina pachyderma (sinistral), Polar Biology, 35, 1311–1319, 10.1007/s00300-012-1174-7, 2012.
  Sykes, F. E., Ezat, M. M., Westgård, A., Foster, G. L., Meilland, J., Chalk, T. B., Milton, J. A., and Chierici, M.: Constraining environmental controls on the incorporation of Mg, Na and Sr in cultured Globigerina bulloides: implications for biomineralisation and high latitude palaeoceanography, Geochimica et Cosmochimica Acta, https://doi.org/10.1016/j.gca.2026.05.022, 2026.
  Westgård, A., Ezat, M. M., Sykes, F. E., Foster, G. L., Meilland, J., Chalk, T. B., and Milton, J. A.: Laboratory-grown crust in planktic foraminifera Neogloboquadrina pachyderma; insights into resolving inaccuracies in polar palaeotemperature estimates, Geochimica et Cosmochimica Acta, https://doi.org/10.1016/j.gca.2026.04.027, 2026.
  
  Citation: https://doi.org/10.5194/egusphere-2026-2415-AC1
RC2:
'Comment on egusphere-2026-2415', Anonymous Referee #2, 06 Jul 2026

The manuscript of Westgård and coauthors on 'Extending planktic foraminiferal Mg/Ca palaeothermometry into polar Temperature ranges: crust- and lamellae specific calibrations and non-thermal controls' presents a novel and interesting approach and valuable new data for better calibration of the Mg/Ca temperature proxy for paleoceangraphic and paleoclimate reconstruction. The paper merits publication in BG following some revision of the text, which is poorly written in places. In particular, the Discussion chapter would need to be written as a discussion, and not repetition of Results plus some references from the literatur. Having said this, most of the classical references on N. pachyderma a missing so far, and would add value to the topic discussed here when being cited. When rewriting the Discussion chapter in discussion style, while preserving the correct scientific content, the author may think about using AI. In addition to these more general comments, please see some more specific comments in the following.
To my impression, the term lamellae is often used in an unfortunate kind of way. For example, for the title, I would suggest: 'Extending planktic foraminiferal Mg/Ca palaeothermometry into polar temperature ranges for crust and ontogenetic calcite'; I would drop 'and non-thermal controls', because it could be misunderstood, and is not possibly important to be mentioned in the title. For the line 29 (abstract), I would suggest: '... ontogenetic shell calcite. Specimens with predominantly albeit variable crust–shell proportions, ...'. Please have a look at the other places in the text and consider accordingly.
The referencing is rather limited and should be extended. In line 52, at least the first Mg/Ca paper of Nürnberg et al. (1996) should be cited. Other papers on the population dynamics (Volkmann, Carstens, and coauthors), ecology (Spindler and coauthors), and species concept (Darling and coauthors) may be referred to, to put the finding of this paper into a wider context.
Line 66ff: In open waters, it can possibly be assumed that the life-cycle of N. pachyderma is four weeks long and adjusted to synodic lunar cycle similar to other planktic forams (Schiebel et al. 2017). This would be important for the understanding of the formation the ontogenetic shell calcite and the Mg/Ca signature.
Line 177: Please give details (light-dark cycles?, light intensity, etc) on the light programme, to make it reproducible.
Lines 271-273: Comparing pH with salinity gets me no-where. What are you aiming for here?
Lines 319-322, write straight: The weak positive relationship between Mg/Ca and salinity is not significant (p>0.1, Fig. 2). Due to the methods used to modify the seawater salinity, DIC, [HCO3-], and total alkalinity (TA) co-varied linearly with salinity (Table 2). The [CO32−] concentration varied non-linearly across the salinity treatments.
Lione 350: 2-9 C, and not 2-7 C !?
Line 358: give ref. for <0 to >10 C
Lines 381-382: 'A comparable...', modify sentence to make it a perspective for future work, to make it a positive statement.
Line 406: yes, well, CO32- and pH co-vary. Please re-write
Line 415-416: change to: 'As pH and [CO32-] co-vary, the driving parameter would be between these two.'
Line 420-435: Only saying that / if other peoples finding are the same or not does not make a good discussion.
From line 420 onward (at least), this is not a discussion but a rambing collection of own findings and selected references, often without any outcome. The discussion needs to be entirely rewritten to make it a convincing plea for the great findings detailed in the Results.

For example:
Line 442: Citing the Table 2, it may still be said what this should tell us.
Lines 444-447: 'Although...' Why is this mentioned? What can we learn?
Line 447-449: 'This difference...' Guessing doesn't provide any useful information. Better skip.
Lines 457-458: This is not a contradiction.
Line 460: Are all of these parameters not sign. at 95% confidence level?
Lines 457-469: For each of the stements in this section, figures and/or tables need to be referred.
Line 473: change to: 'To the temperature interval from 2–9C in addition to the interval from 9-12C (Davis et al. 2017).'
Lines 517-519: This sentence make no sense. Are both assumed to be driven by physiological processes such as calcification?
Lines 534-535: This would not make much sense, and would possibly invalid other assumption on similar processes in benthic and planktic foraminifers.
Lines 559-560: Please take out. This is another publication.

Wording and syntax:
Always add a comma before 'respectively'.
Check the entire manuscript for the correct use of tenses; e.g., line 271: 'observed'
Line 235: Porites sp.
Line 255: spell out 'percent'
Lines 78, 265, and others: delete 'ratio' as Mg/Ca is a ratio
Lines 277-280: change to: 'Based on tests where both laboratory-grown crust and ontogenetic calcite was analysed, the total wall thickness decreased with decreasing pH and temperature (|R|>0.5, p>0.05), and decreased with increasing [CO32-] (R= -0.8, p>0.05).
Figure 1, line 289: I see grey shading, not green. Line 290: Table in upper case.
Line 292, and following: distribution in lower case.
Figure 2: Salinity has no unit, not even psu. The grey and green color are difficult to tell apart. Make the green line maybe bright blue to also facilitate comprehension for the red green-blind men.
Line 359: change 'curve' to 'relationship'
Line 375: change 'comparable with' to 'similar to'
Line 411: change to '... on Mg/Ca. However, ...'
Line 419: change 'that to' to 'to that'
Line 461: %crust. This is not a term. Do you mean proportion of crust vs. ontogenetic calcite?
Line 493-496: Less impactful than where? This is a relative clause which should be closed.

Citation: https://doi.org/10.5194/egusphere-2026-2415-RC2
- AC2: 'Reply on RC2', Adele Westgård, 20 Aug 2026
  
  Dear reviewer,
  Thank you for your time, constructive comments, and suggestions for improvements to our manuscript.
  See detailed responses to your comments below.
  
  Sincerely,
  Adele Westgård
  On behalf of the authors
  
  Comment #1
  The manuscript of Westgård and coauthors on 'Extending planktic foraminiferal Mg/Ca palaeothermometry into polar Temperature ranges: crust- and lamellae specific calibrations and non-thermal controls' presents a novel and interesting approach and valuable new data for better calibration of the Mg/Ca temperature proxy for paleoceangraphic and paleoclimate reconstruction. The paper merits publication in BG following some revision of the text, which is poorly written in places. In particular, the Discussion chapter would need to be written as a discussion, and not repetition of Results plus some references from the literatur. Having said this, most of the classical references on N. pachyderma a missing so far, and would add value to the topic discussed here when being cited. When rewriting the Discussion chapter in discussion style, while preserving the correct scientific content, the author may think about using AI. In addition to these more general comments, please see some more specific comments in the following.
  Response
  We will revise our discussion accordingly, and in line with your further detailed comment below. However, we would like to note that our independently varied culture treatments have allowed us to detangle results that previous studies have not. The differences between this and previous studies are crucial to highlight to be able to understand why the results in said studies are different and thus to detangle Mg/Ca - environmental relationships. For example, decoupling the carbonate chemistry have in our study showed that pH and CO3 have opposing effects on Mg/Ca where previous studies have shown a wide range of Mg/Ca responses to pH and CO3. We also show that previous salinity correlations with Mg/Ca are likely a carbonate chemistry effect rather than salinity. Previous studies have made significant advances in our understanding of these relationships but limited experimental scale and/or environmental background data have led to incorrect assumptions that needs to be clarified and corrected in light of new observations. Therefore, although we understand that said comparisons may appear tedious to read, we believe they are crucial information to back up our conclusions. However, to ease the flow of reading, we will rephrase this section to better highlight the implications of said comparisons and how they relate to conclusions later in the discussion in relation to paleoclimate reconstructions and biomineralisation.
  
  Comment #2
  To my impression, the term lamellae is often used in an unfortunate kind of way. For example, for the title, I would suggest: 'Extending planktic foraminiferal Mg/Ca palaeothermometry into polar temperature ranges for crust and ontogenetic calcite'; I would drop 'and non-thermal controls', because it could be misunderstood, and is not possibly important to be mentioned in the title. For the line 29 (abstract), I would suggest: '... ontogenetic shell calcite. Specimens with predominantly albeit variable crust–shell proportions, ...'. Please have a look at the other places in the text and consider accordingly.
  Response
  We will clarify and define these terms early in the revised manuscript to minimize confusion. Ontogenetic and lamellar calcite are both commonly used in the literature for Neogloboquadrinids (e.g., Hupp and Fehrenbacher, 2024; Westgård et al., 2026; Davis et al., 2017; Jonkers et al., 2016). Although these terms refer to the same parts of the shell (the initially formed chambers and concurrent thickening), they have slightly different meanings. Ontogenetic refers to calcite formation prior to reproduction, lamellar refers to the structure of that part of the shell (Jonkers et al., 2016; Westgård et al., 2026; Kozdon et al., 2009). The crust is also part of the biologically calcified shell, however, has a different texture and forms later in the life cycle. As the distinction between the two components (crust and lamellar/ontogenetic calcite) in this study is based on texture, structure, and geochemical composition rather than the timing of formation, we prefer using the also correct, descriptive term lamellar calcite. This will be clarified in the revised manuscript.
  
  Comment #3
  The referencing is rather limited and should be extended. In line 52, at least the first Mg/Ca paper of Nürnberg et al. (1996) should be cited. Other papers on the population dynamics (Volkmann, Carstens, and coauthors), ecology (Spindler and coauthors), and species concept (Darling and coauthors) may be referred to, to put the finding of this paper into a wider context.
  Response
  Nürnberg et al. (1996) and other references will be added, and Spindler & Dieckmann (1986) is already cited.
  
  Comment #4
  Line 66ff: In open waters, it can possibly be assumed that the life-cycle of N. pachyderma is four weeks long and adjusted to synodic lunar cycle similar to other planktic forams (Schiebel et al. 2017). This would be important for the understanding of the formation the ontogenetic shell calcite and the Mg/Ca signature.
  Response
  We agree that understanding its life cycle is crucial for understanding shell formation and its subsequent geochemical signal. However, there isn’t any strong evidence for a lunar life cycle in N. pachyderma and recent studies have shown that a 4-week life cycle cannot be assumed and growth rate and patterns vary greatly between individuals for this species and some other planktic species (e.g., Sykes et al., 2024; Westgård et al., 2023; Meilland et al., 2022; Anglada-Ortiz et al., 2023). Thus, we have to keep in mind that the formation of the lamellar/ontogenetic calcite (chamber addition and concurrent thickening of earlier chambers) can occur sporadically throughout the specimen’s life cycle which can last for months (e.g., Westgård et al., 2023) Some individuals form many chambers rapidly within few days-weeks while others have longer hiatuses (many weeks-months) between chamber formations, a period in which growth conditions may have changed. The crust on the other hand forms at the end of the specimen’s lifetime and thus over a much shorter time span.
  Comment #5
  Line 177: Please give details (light-dark cycles?, light intensity, etc) on the light programme, to make it reproducible.
  Response
  Light conditions in the incubators were set to replicate the 24-hour daylight of Arctic summer with lower light intensity during the night. The incubators were programmed based on light conditions measured in the field giving values ranging from 30 to 160 photons/sec/cm2.
  This information will be added to the revised manuscript.
  
  Comment #6
  Lines 271-273: Comparing pH with salinity gets me no-where. What are you aiming for here?
  Response
  This is not intended as a comparison of the two parameters but rather highlights the endmembers of crust formation across all treatments, i.e., the lowest number of crusted specimens and the highest number of crusted specimens. This will be clarified in the revised manuscript.
  
  Comment #7
  Lines 319-322, write straight: The weak positive relationship between Mg/Ca and salinity is not significant (p>0.1, Fig. 2). Due to the methods used to modify the seawater salinity, DIC, [HCO3-], and total alkalinity (TA) co-varied linearly with salinity (Table 2). The [CO32−] concentration varied non-linearly across the salinity treatments.
  Lione 350: 2-9 C, and not 2-7 C !?
  Line 358: give ref. for <0 to >10 C
  Lines 381-382: 'A comparable...', modify sentence to make it a perspective for future work, to make it a positive statement.
  Response
  Referring to the four comments above: Thank you for pointing these out, we will edit accordingly.
  
  Comment #8
  Line 406: yes, well, CO32- and pH co-vary. Please re-write
  Response
  CO3 and pH do not co-vary in the experiments discussed in this section. As pH and CO3 were altered independently in two sets of experiments and co-varied in the third set of experiments, we are here showing that the decoupled experiments result in different relationships than the co-varying experiment, thus being able to conclude that pH is the primary driver observed when they co-vary (see also response to comment #2 from Lennart de Nooijer). This was not already established in literature on N. pachyderma and not possible to show without the decoupled experiments.
  We will clarify this in the relevant paragraph.
  
  Comment #9
  Line 415-416: change to: 'As pH and [CO32-] co-vary, the driving parameter would be between these two.
  Response
  We will change it to: “However, in these studies, as pH and [CO32-] likely co-vary (as is typical in nature), the driving parameter would be either of or between these two.”
  
  Comment #10
  Line 420-435: Only saying that / if other peoples finding are the same or not does not make a good discussion.” & “From line 420 onward (at least), this is not a discussion but a rambing collection of own findings and selected references, often without any outcome. The discussion needs to be entirely rewritten to make it a convincing plea for the great findings detailed in the Results.
  Response
  See our response above to Comment #1.
  
  Comment #11
  For example:
  Line 442: Citing the Table 2, it may still be said what this should tell us.
  Lines 444-447: 'Although...' Why is this mentioned? What can we learn? Line 447-449: 'This difference...' Guessing doesn't provide any useful information. Better skip.
  Response
  In this sentence we stated that all measured environmental parameters being mostly equivalent from two salinity experiments. We cited Table 2, where temperature, salinity, pH, and [CO32-] values are listed, so the reader can compare these values between the two experiments. We also followed this sentence with further discussion on what might have caused the Mg/Ca difference between the two “equivalent” experiments. This further discussion (Lines 444-447) describes fact differences between the two treatments which are important context. Even we are not able to provide a mechanistic explanation (Lines 447-449), we think this context (e.g., Lines 444-447) may provide insights for future studies.
  We believe these differences are important to highlight, showing that there are remaining uncertainties to be studied. However, we will rephrase the paragraph for clarity in the revised manuscript.
  
  Comment #12
  Lines 457-458: This is not a contradiction.
  Response
  OK.
  
  Comment #13
  Line 460: Are all of these parameters not sign. at 95% confidence level?
  Response
  Correct, as is clearly stated in this line 460. Significant relationships are also clearly indicated in figure 1 and R and p-values are reported in section 3.1.
  
  Comment #14
  Lines 457-469: For each of the stements in this section, figures and/or tables need to be referred.
  Response
  See line 459.
  
  Comment #15
  Line 473: change to: 'To the temperature interval from 2–9C in addition to the interval from 9-12C (Davis et al. 2017).'
  Response
  OK.
  
  Comment #16
  Lines 517-519: This sentence make no sense. Are both assumed to be driven by physiological processes such as calcification?
  Response
  One process (calcification rate, i.e., the rate of mineral growth) is a physical process interpreted in this study and other studies to be directly correlated to carbonate ion concentration, the other is a physiological process (e.g., correlated to increased energy-cost of proton transport with decreased pH) availability (e.g., Holland et al., 2017; Branson et al., 2025; Ziveri et al., 2025; Allen et al., 2016, and refs therein; Hall and Chan, 2004). These processes and our argumentation are explained in detail in lines 503-512 and 519-527.
  We will alter the passage in the revised manuscript for increased clarity: “We suggest that the Mg/Ca response to [CO32-] may be driven by calcification rate (a physical process), whereas that the Mg/Ca response to pH may be driven by a stress response of the cell, i.e., a physiological process.”
  
  Comment #17
  Lines 534-535: This would not make much sense, and would possibly invalid other assumption on similar processes in benthic and planktic foraminifers.
  Response
  Our evidence suggests the opposite response to the recent findings of François et al. (2025) for N. pachyderma. Thus, our options are to conclude that their theory is incorrect/faulted or assume that their study on benthic foraminifera doesn’t necessarily also apply to planktic foraminifera/Neogloboquadrinids. As there are several studies suggesting benthic and planktic, and high and low Mg/Ca foraminifera calcify through different processes (e.g., Branson et al., 2025; Bentov and Erez, 2006 and refs therein) the second option is best supported by literature, and we can only conclude that further studies are necessary.
  
  Comment #18
  Lines 559-560: Please take out. This is another publication.
  Response
  OK.
  
  Comment #19
  Wording and syntax:
  Always add a comma before 'respectively'.
  Check the entire manuscript for the correct use of tenses; e.g., line 271: 'observed'
  Line 235: Porites sp.
  Line 255: spell out 'percent'
  Lines 78, 265, and others: delete 'ratio' as Mg/Ca is a ratio
  Lines 277-280: change to: 'Based on tests where both laboratory-grown crust and ontogenetic calcite was analysed, the total wall thickness decreased with decreasing pH and temperature (|R|>0.5, p>0.05), and decreased with increasing [CO32-] (R= -0.8, p>0.05).
  Figure 1, line 289: I see grey shading, not green. Line 290: Table in upper case.
  Line 292, and following: distribution in lower case.
  Figure 2: Salinity has no unit, not even psu. The grey and green color are difficult to tell apart. Make the green line maybe bright blue to also facilitate comprehension for the red green-blind men.
  Line 359: change 'curve' to 'relationship'
  Line 375: change 'comparable with' to 'similar to'
  Line 411: change to '... on Mg/Ca. However, ...'
  Line 419: change 'that to' to 'to that'
  Line 461: %crust. This is not a term. Do you mean proportion of crust vs. ontogenetic calcite?
  Line 493-496: Less impactful than where? This is a relative clause which should be closed.
  Response
  Thank you for pointing these out, the text will be edited, and/or clarified appropriately.
  
  References
  Allen, K. A., Hönisch, B., Eggins, S. M., Haynes, L. L., Rosenthal, Y., and Yu, J.: Trace element proxies for surface ocean conditions: A synthesis of culture calibrations with planktic foraminifera, Geochimica et Cosmochimica Acta, 193, 197–221, 10.1016/j.gca.2016.08.015, 2016.
  Anglada-Ortiz, G., Meilland, J., Ziveri, P., Chierici, M., Fransson, A., Jones, E., and Rasmussen, T. L.: Seasonality of marine calcifiers in the northern Barents Sea: Spatiotemporal distribution of planktonic foraminifers and shelled pteropods and their contribution to carbon dynamics, Progress in Oceanography, 218, 103121, https://doi.org/10.1016/j.pocean.2023.103121, 2023.
  Bentov, S. and Erez, J.: Impact of biomineralization processes on the Mg content of foraminiferal shells: A biological perspective, Geochemistry, Geophysics, Geosystems, 7, https://doi.org/10.1029/2005GC001015, 2006.
  Branson, O., Chauhan, N., Evans, D., Foster, G. L., and Rickaby, R. E. M.: Geochemical tracers of biomineralisation processes, in: Treatise on Geochemistry (Third edition), edited by: Anbar, A., and Weis, D., Elsevier, Oxford, 177–235, https://doi.org/10.1016/B978-0-323-99762-1.00128-5, 2025.
  Davis, C. V., Fehrenbacher, J. S., Hill, T. M., Russell, A. D., and Spero, H. J.: Relationships Between Temperature, pH, and Crusting on Mg/Ca Ratios in Laboratory-Grown Neogloboquadrina Foraminifera, Paleoceanography, 32, 1137–1152, 10.1002/2017pa003111, 2017.
  François, D., Reichart, G.-J., and de Nooijer, L. J.: Open or closed: pH modulation and calcification by foraminifera, Science Advances, 11, eadq8425, doi:10.1126/sciadv.adq8425, 2025.
  Hall, J. M. and Chan, L. H.: Li/Ca in multiple species of benthic and planktonic foraminifera: thermocline, latitudinal, and glacial-interglacial variation 1 1Associate editor: D. Lea, Geochimica et Cosmochimica Acta, 68, 529–545, https://doi.org/10.1016/S0016-7037(03)00451-4, 2004.
  Holland, D. S., Speir, C., Agar, J., Crosson, S., DePiper, G., Kasperski, S., Kitts, A. W., and Perruso, L.: Impact of catch shares on diversification of fishers’ income and risk, Proceedings of the National Academy of Sciences, 114, 9302–9307, https://doi.org/10.1073/pnas.1702382114, 2017.
  Hupp, B. N. and Fehrenbacher, J. S.: Intratest trace element variability in polar and subpolar planktic foraminifera: insights into vital effects, ontogeny, and biomineralization processes, Journal of Foraminiferal Research, 54, 355–374, 10.61551/gsjfr.54.4.355, 2024.
  Jonkers, L., Buse, B., Brummer, G.-J. A., and Hall, I. R.: Chamber formation leads to Mg/Ca banding in the planktonic foraminifer Neogloboquadrina pachyderma, Earth and planetary science letters, 451, 177–184, 10.1016/j.epsl.2016.07.030, 2016.
  Kozdon, R., Ushikubo, T., Kita, N. T., Spicuzza, M., and Valley, J. W.: Intratest oxygen isotope variability in the planktonic foraminifer N. pachyderma: Real vs. apparent vital effects by ion microprobe, Chemical Geology, 258, 327–337, 10.1016/j.chemgeo.2008.10.032, 2009.
  Meilland, J., Ezat, M. M., Westgård, A., Manno, C., Morard, R., Siccha, M., and Kucera, M.: Rare but persistent asexual reproduction explains the success of planktonic foraminifera in polar oceans, Journal of Plankton Research, 10.1093/plankt/fbac069, 2022.
  Sykes, F. E., Meilland, J., Westgård, A., Chalk, T. B., Chierici, M., Foster, G. L., and Ezat, M. M.: Large-scale culturing of the subpolar foraminifera Globigerina bulloides reveals tolerance to a large range of environmental parameters associated to different life-strategies and an extended lifespan, Journal of Plankton Research, 46, 403–420, 10.1093/plankt/fbae029, 2024.
  Westgård, A., Ezat, M. M., Chalk, T. B., Chierici, M., Foster, G. L., and Meilland, J.: Large-scale culturing of Neogloboquadrina pachyderma, its growth in, and tolerance of, variable environmental conditions, Journal of Plankton Research, 10.1093/plankt/fbad034, 2023.
  Westgård, A., Ezat, M. M., Sykes, F. E., Foster, G. L., Meilland, J., Chalk, T. B., and Milton, J. A.: Laboratory-grown crust in planktic foraminifera Neogloboquadrina pachyderma; insights into resolving inaccuracies in polar palaeotemperature estimates, Geochimica et Cosmochimica Acta, https://doi.org/10.1016/j.gca.2026.04.027, 2026.
  Ziveri, P., Langer, G., Chaabane, S., de Vries, J., Gray, W. R., Keul, N., Hatton, I. A., Manno, C., Norris, R., Pallacks, S., Young, J. R., Schiebel, R., Zarkogiannis, S., Anglada-Ortiz, G., Bianco, S., de Garidel-Thoron, T., Grelaud, M., Lucas, A., Probert, I., and Mortyn, P. G.: Calcifying plankton: From biomineralization to global change, Science, 390, eadq8520, doi:10.1126/science.adq8520, 2025.
  
  Citation: https://doi.org/10.5194/egusphere-2026-2415-AC2

Adele Westgård, Mohamed M. Ezat, Freya E. Sykes, Julie Meilland, Thomas B. Chalk, J. Andy Milton, Melissa Chierici, and Gavin L. Foster

Supplement

https://doi.org/10.5194/egusphere-2026-2415-supplement

Adele Westgård, Mohamed M. Ezat, Freya E. Sykes, Julie Meilland, Thomas B. Chalk, J. Andy Milton, Melissa Chierici, and Gavin L. Foster

Viewed

Total article views: 471 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
392	58	21	471	23	20	26

HTML: 392
PDF: 58
XML: 21
Total: 471
Supplement: 23
BibTeX: 20
EndNote: 26

Views and downloads (calculated since 06 May 2026)

Month	HTML	PDF	XML	Total
May 2026	309	25	13	347
Jun 2026	39	8	3	50
Jul 2026	32	15	3	50
Aug 2026	12	10	2	24

Cumulative views and downloads (calculated since 06 May 2026)

Month	HTML	PDF	XML	Total
May 2026	309	25	13	347
Jun 2026	39	8	3	50
Jul 2026	32	15	3	50
Aug 2026	12	10	2	24

Viewed (geographical distribution)

Total article views: 441 (including HTML, PDF, and XML) Thereof 441 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 20 Aug 2026

Short summary

We present a novel approach to reconstructing past temperatures in the polar oceans using Mg/Ca in planktic foraminifera which accounts for their geochemically distinct shell components. We extend the lab-based temperature calibration from 9 to 2 °C and show that also pH and [CO₃^2-] influence Mg/Ca. These results may improve polar paleoclimate reconstructions. Additionally, our study shows that water conditions influence shell growth which has implications for ongoing climate change.


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