<em>Chaetoceros</em> resting spores record low diatom-bound nitrogen isotope values: Evidence from laboratory culture and marine sediment

Dove, Isabel A.; Bishop, Ian W.; Crosta, Xavier; Riedinger, Natascha; Kelly, R. Patrick; Robinson, Rebecca S.

doi:https://doi.org/10.5194/egusphere-2023-2564

Preprints

https://doi.org/10.5194/egusphere-2023-2564

Preprints

16 Nov 2023

| 16 Nov 2023

Chaetoceros resting spores record low diatom-bound nitrogen isotope values: Evidence from laboratory culture and marine sediment

Isabel A. Dove, Ian W. Bishop, Xavier Crosta, Natascha Riedinger, R. Patrick Kelly, and Rebecca S. Robinson

Abstract. The nitrogen isotopic composition of diatom frustule-bound organic matter (δ¹⁵N_DB) is often used to study changes in high latitude biological pump efficiency across glacial-interglacial cycles, but the proxy may be biased by species-specific effects. The genus Chaetoceros is of particular interest because of its abundance throughout ocean basins, its shifting biogeography during glacial periods, and many species’ ability to form heavily silicified resting spores. Here we investigate how Chaetoceros resting spores (CRS) record surface nitrate conditions in their nitrogen isotopic composition using both laboratory culture experiments and assemblage-specific sedimentary δ¹⁵N_DB measurements. We find that CRS record δ¹⁵N_DB values 7.0 ± 2.6 ‰ lower than vegetative Chaetoceros in culture and 5.6 ± 1.9 ‰ lower than non-CRS diatoms in sediment. Low values are attributed to assimilation of isotopically light ammonium, heavy silicification, and internal fractionation during sporulation. Despite the large δ¹⁵N_DB difference observed in CRS versus non-CRS diatoms, increased CRS relative abundance in open ocean glacial sediments does not significantly bias δ¹⁵N_DB records due to the spores’ small size.

Received: 31 Oct 2023 – Discussion started: 16 Nov 2023

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 preprint. The responsibility to include appropriate place names lies with the authors.

Download & links

Preprint (PDF, 1077 KB)

Supplement (416 KB)

Download & links

Isabel A. Dove, Ian W. Bishop, Xavier Crosta, Natascha Riedinger, R. Patrick Kelly, and Rebecca S. Robinson

Status: closed

RC1:
'Comment on egusphere-2023-2564', Anonymous Referee #1, 19 Dec 2023

Chaetoceros resting spores record low diatom-bound nitrogen isotope values: Evidence from laboratory culture and marine sediment
by Dove et al.
Summary
Dove et al cultured the diatom species Chaetoceros in the laboratory and analyzed both vegetative cells and resting spores for their opal-bound N isotopic composition, to study how they record nitrate N isotope values. In addition, they measured opal-bound N isotopes on 6 sediment samples from a sediment core close to the Antarctic margin on individual layers dominated by Chaetoceros resting spores + vegetative cells. They put their data into context with previously published opal-bound N isotope datasets across glacial-interglacial cycles with the goal to understand whether the changing abundance of Chaetoceros resting spores biases these d15N-DB records.
Although the manuscript is well written, I have substantial concerns regarding (i) the large claims the authors make based on a very small and noisy dataset, (ii) the planning and execution of the culture experiments, and (iii) the interpretation of their results, which I detail below. Throughout the manuscript, the authors are performing several calculations with a series of assumptions and corrections going into them, which is masking the fact that the actual data base is quite weak and inconsistent. Overall, I believe that this manuscript has been submitted prematurely on the basis of a very small and inconclusive dataset, and that additional, well-designed culture experiments are needed to support the statements the authors make. My recommendation therefore is to reject the manuscript but encourage re-submission once a more comprehensive and conclusive dataset can be presented.
Major comments
First of all, three culture experiments have been performed in triplicates each, of which the first two experiments were unsuccessful (in the sense that the authors have been unable to separate vegetative cells and resting spores of Chaetoceros and thus they have been unable to measure their individual d15N-DB). This leaves one successful culture experiment (no. 3), of which one of the triplicates did not yield enough diatoms for d15N-DB-CRS analysis. In another triplicate, the vegetative cells were “contaminated” with resting spores. To summarize, only two individual d15N-DB-CRS analyses have successfully been measured in the culture experiments, and in only one single culture sample/carboy, concurrent d15N-DB of vegetative cells and of resting spores have been determined. Thus, the “evidence from laboratory cultures that Chaetoceros resting spores record low d15N-DB values” as advertised in the title is very sparse.
Which brings me next to the evidence from marine sediments. In the sedimentary samples, in 4 out of the 6 samples, the d15N-DB of the bulk diatom fraction is indistinguishable from the d15N-DB of the Chaetoceros fraction (error bars in figure 4a are overlapping), only in samples no. 3 and 5 there seems to be an N isotopic difference. Sample no. 4 actually shows slightly lower d15N-DB-bulk values relative to d15N-DB-Chaetoceros, and similarly, in experiment 2, carboys 1 and 2 had higher d15N-DB in the mixed samples compared to the vegetative samples, suggesting the opposite effect in d15N-DB between CRS and vegetative cells. Thus, their statement (lines 297-298) that “Both culture experiments and sedimentary data suggest that CRS are characterized by low d15N-DB values relative to vegetative Chaetoceros and to other diatoms, (…)” is not supported by their data and will probably not withstand statistical tests.
Since the sedimentary samples are a mixture of Chaetoceros vegetative cells and resting spores, the authors try to “correct” for that, by assigning an offset of 7‰ (derived from the culture experiments) to the difference between the vegetative cells and the resting spores of the sedimentary samples (and by making a number of other assumptions, lines 320-324), making the resting spores eventually 5.6‰ lower than non-resting spore samples (lines 332-333). Clearly, this Delta-d15N-CRS offset is almost entirely due to the assumed difference of 7‰ between resting spores and vegetative cells, and there is no independent verification of that value from the sediment. Therefore, there is no conclusive evidence from sedimentary data that Chaetoceros resting spores record low d15N-DB values.
Regarding the culture conditions, the authors were initially unable to separate vegetative cells from the resting spores for N isotopic analysis, but came up with a method for their third culture experiment. So why did they then not repeat the first two experiments after finding a solution to their problem, which would have made this a much more comprehensive and reliable dataset?
Dissolved organic N concentrations were very high in the culture experiments (nearly as high as initial nitrate concentrations). In the text, this is attributed to bacterial remineralization, but the fact that concentrations were already high at the beginning of the experiments could also suggest that this comes from the natural seawater. This might be an issue for independent reproducibility of the experiments in other labs, especially if reduced N d15N values are thought to be an important driver for the d15N-DB-CRS values. Why not using artificial seawater for their experiments to avoid such high total reduced N (TRN) concentrations? With different N sources available to the diatoms (nitrate, ammonium, DON), it’s difficult/impossible to know which N source they were assimilating to what proportion at what point in time, and how this influenced the biomass and opal-bound N isotopic composition of the growing culture, and eventually, the d15N-DB of the resting spores.
Once N and Si became limiting in the cultures, were cultures kept illuminated? Continuous illumination while under nutrient limitation could have increased respiration, and with that induced15N-fractionation of amino acids. Is this an accurate representation of natural settings, where heavy silicification likely leads to the cells sinking to deeper layers where light is no longer present?
In terms of the interpretation of the low-d15N-DB signature of the CRS, I would like to see more information about the biology and physiology of the diatoms in question, and specifically, information on the timing of resting spore formation and the physiological changes that the cells undergo, as to better understand the processes that are integrated into d15N-DB-CRS. What is known about resting spore formation, its timing and the proteins involved that remain bound to the frustule? Do d15N-DB-CRS values represent a mixture of d15N-DB from the frustule bound proteins in vegetative cells and d15N-DBthat are newly synthesized upon resting spore formation (which in this case would be particularly negative)? Or are resting spores formed after multiple mitosis cycles after which half of the frustule (and with that the frustule bound proteins) is newly synthesized after each cycle? To what extent can the frustule bound proteins be exchanged/recycled in living diatoms? If additional harvests and d15N-DB-CRS analyses would have been planned and performed in between the sampling for the vegetative cells and the second sampling for the CRS, more information might have been gained about how the d15N-DB-CRS signal is acquired.

Citation: https://doi.org/10.5194/egusphere-2023-2564-RC1
- AC1: 'Reply on RC1', Isabel Dove, 22 Dec 2023
  
  We thank the referee for their thorough review and careful explanation of concerns with the experimental design. We understand that our dataset of direct measurements is small, but maintain that our conclusions are sound and based on justified assumptions.
  The experiments were time and labor intensive, as it was quite difficult to obtain sufficiently large samples from nitrate-limited cultures. Initial experimental results and concurrent sedimentary data hinted that Chaetoceros resting spores (CRS) record lower δ¹⁵N_DB values than non-CRS diatoms. Therefore, after successfully isolating CRS in two carboys from the third experiment and measuring consistently low values, we were convinced that this signal is real. A fourth experiment was not conducted due to time and resource constraints. We agree that more data is desirable, yet our results assuredly indicate that the δ¹⁵N_DB signature of CRS is lower than that of vegetative cells, which is novel and worth reporting.
  Regarding concern over the execution of culture experiments, we agree that high reduced nitrogen concentrations could be an issue. While reduced nitrogen concentrations are high, they increase throughout the experiment, suggesting that much of the nitrogen is not bioavailable and thus unlikely to influence our our δ¹⁵N_DB measurements. We also acknowledge that ammonium assimilation could contribute to low δ¹⁵N_DB values in CRS in culture, but show in the supplement that it cannot explain the difference observed in CRS-specific δ¹⁵N_DB values. More importantly, ε_DB values (δ¹⁵N_biomass – δ¹⁵N_DB) enable comparison of CRS- and vegetative-specific δ¹⁵N_DB values independent of effects from nutrient source. Field data show that δ¹⁵N_DB values do follow nutrient source δ¹⁵N, as low δ¹⁵N_biomass and δ¹⁵N_DB values were observed in the late growing season where the primary nutrient source was ammonium (Robinson et al., 2020). While our δ¹⁵N_biomass measurements in vegetative and CRS samples are similar, CRS-specific δ¹⁵N_DB values are much lower, reflected by higher ε_DB values. Distinctly different ε_DB values unequivocally show that, regardless of nutrient source, CRS-specific δ¹⁵N_DB values are lower than vegetative-specific δ¹⁵N_DB values.
  Cultures were kept illuminated throughout the entire experiment and we recognize that this is not representative of natural settings. This underscores why we also looked to marine sediments for corroboration.
  Anonymous Referee #1 brings up important concern over sparse data and the need for statistical tests. Focusing on sediment data, which avoids potential bias introduced by culture conditions, we find that CRS do record lower δ¹⁵N_DB values than non-CRS diatoms. Conducting a Wilcoxon signed rank test on isolated Chaetoceros samples versus non-CRS samples, we find that the Chaetoceros δ¹⁵N_DB values are lower (p=0.053). This test was conducted using δ¹⁵N_DB-bulk values adjusted only for varying CRS surface area, so are independent of assumptions dependent on culture results (calculated non-CRS δ¹⁵N_DB = 10.1, 11.3, 14.4, 8.7, 11.8, 8.9). The difference between Chaetoceros and non-CRS δ¹⁵N_DB values is then calculated to be 1.9 ± 2.3‰. This is obviously a smaller difference, but does not change our major finding that CRS do not bias downcore δ¹⁵N_DB records.
  We agree that Anonymous Referee #1’s questions about CRS formation are important and worth investigating. Our discussion is speculative because so little is currently known. While our focus is on paleo-implications of how CRS record δ¹⁵N_DB values, future work with carefully designed culture experiments can begin to answer remaining questions as to how CRS physiology impacts δ¹⁵N_DB values.
  As Anonymous Referee #1 noted, the goal of this work is to assess how changing CRS abundance over time impacts δ¹⁵N_DB records. We believe that, despite valid concerns over our culture experiment methodology, our conclusions that CRS record low δ¹⁵N_DB values but do not significantly bias published δ¹⁵N_DB records is sound and worthy of publication. Questions about species-specific effects have long been a concern for those using biogeochemical proxies (e.g. Jacot des Combes, 2008), so it is important to confirm that CRS are not a significant source of bias to δ¹⁵N_DB records. Additionally, this study sets the stage for future work to further investigate other species and brings up interesting questions about CRS biology and physiology.
  Again, we thank the referee for their feedback and for encouraging more thorough statistical validation of our results. The manuscript will be improved by adding discussion of potential bias introduced by culture conditions, more thorough explanation of ε_DB values, and emphasis on results from sediments.
  
  References:
  Robinson, R. S., Jones, C. A., Kelly, R. P., Love, A., Closset, I., Rafter, P. A., & Brzezinski, M. (2020). A test of the diatom‐bound paleoproxy: Tracing the isotopic composition of nutrient‐nitrogen into Southern Ocean particles and sediments. Global Biogeochemical Cycles, 34(10), e2019GB006508.
  Jacot Des Combes, H., Esper, O., de La Rocha, C. L., Abelmann, A., Gersonde, R., Yam, R., & Shemesh, A. (2008). Diatom δ13C, δ15N, and C/N since the Last Glacial Maximum in the Southern Ocean: Potential impact of species composition. Paleoceanography, 23(4).
  
  Citation: https://doi.org/10.5194/egusphere-2023-2564-AC1
RC2:
'Comment on egusphere-2023-2564 R2', Anonymous Referee #2, 19 Dec 2023

The Authors produced data from culture and sediments that Chaetoceros resting spores (CRS) have a lower d15N signature with respect to vegetative Chaetoceros. Given the variable abundance of CRS in sediments, they argue that is important to characterize sedimentary assemblages for paleo-d15N applications.
I am impressed by the amount of work that the authors produced. The causes of lower d15N in CRS remain speculative, while it is reasonable to assume more ammonium assimilation, the estimates are still based on a handful of datapoints and the authors conclude that ammonium alone cannot account for the total magnitude of difference. While this is new and valuable new data, the dataset of CRS direct estimates is really small, but I understand that this is due to the technical difficulties in both laboratory culturing and having to rely on indirect estimations from sediments. Regarding the latter, the calculation of d15N from sediments rely on assumptions of CRS abundance, but I am not sure how solid this is. Section 3 (Results) appears shorter than it should be: the explanation of what corrections have been applied (and the underlying assumptions) to estimate the d15N difference between Chaetoceros and CRS is not very clear, and should be better elucidated (or this should be done in the Supplement).
Overall, the authors present novel data and make a valid case for the importance of evaluating changes in diatom assemblages when studying diatom-bound d15N in sedimentary archives. However, if the lower d15N values of CRS are confirmed by further investigations, CRS would significantly affect paleo-reconstruction only in a very small number of cases, where their abundance is particularly high.

Smaller comments in the following:
Line 30: Word Nutrients is used twice in the same sentence: “nutrient- and CO2-rich waters to the surface, where nutrient…”
Line 78: Because CRS likely form under nutrient-depleted conditions (in which nitrate is enriched in 15N) and because they are preferentially preserved in marine sediments, we hypothesize that they bias records towards higher δ15NDB values, which signal a greater degree of nutrient utilization.
Line 125: why would 45 uM of nitrate already induce more CRS formation, if CRS are posed to form under nutrient stress?
Line 167: which amino acid standards were used? Are they international reference materials (RMs)? If so, how does their value compare to their international reference values. If this info exists, it would be informative to provide.
Line 173: “A mass balance calculation using nitrate concentrations and δ¹⁵N_NO3 values from un-oxidized samples yielded the nitrogen isotopic composition of 175 reduced nitrogen”. I do not understand this procedure. Does the δ¹⁵N_NO3of un-oxidized samples reflect the oxidation blank? How does this take into account the δ¹⁵Nof the organic blank? And how large was the oxidation blank relative to the size of your samples? This is quite important info.
Line 193: did you observe some refractory behavior of CRS during K2S2O8 dissolution and subsequent oxidation? In other words, can it be ruled out that there is not effect of partial dissolution of more strongly silicified CRS with respect to vegetative frustules?
Line 235 (and throughout the manuscript): use letters next to figure numbers to help the reader jump between the text and the figures (which already include letters).

Citation: https://doi.org/10.5194/egusphere-2023-2564-RC2
- AC2: 'Reply on RC2', Isabel Dove, 22 Dec 2023
  
  We thank the referee for their thoughtful review and especially appreciate their recognition of the amount of work that went into the culture experiments. As elaborated upon in our above comment to Anonymous Referee #1, we acknowledge that there are only two direct measurements from Chaetoceros resting spores (CRS) but remain confident that they record low δ¹⁵N_DB values based on data from mixed culture samples and from sediment.
  Regarding calculations of CRS-specific δ¹⁵N_DB values from sediment, we determined the relative abundance of both CRS and vegetative Chaetoceros spp. in the <63 μm bulk diatom assemblage using quantitative microscope slides (lines 226-228). These quantitative assemblage counts enabled us to determine the proportion of resting spores versus vegetative Chaetoceros, which is reported as CRS/Ch. in Table 2. From there, we assume that the proportion of CRS versus vegetative Chaetoceros remains constant in the <10 μm isolated Chaetoceros sample. We believe this is a reasonable assumption because we followed the microseparation technique outlined in Egan et al. (2012), which avoids fragmenting microfossils and clogging mesh.
  We appreciate that explanations of various calculations can be confusing. Much of the data required corrections, which we describe in the Discussion section. Therefore, the Results section is comparatively much shorter. A possible solution for increasing clarity is to combine the Results and Discussion sections. Regardless, revision will include more thorough and clear explanations.
  While 45 μM nitrate concentration is indeed high for the ocean, it is low enough to induce resting spore formation in culture. Our strains of C. socialis were maintained in f/2 media (882 μM nitrate concentration), so 45 μM nitrate concentration is comparatively quite low.
  We used the USGS65 glycine standard to ensure that all reduced N has been converted to nitrate during persulfate oxidation.
  Given surprising results from initial culture experiments, we realized the importance of tracking the concentration and isotopic composition of all forms of nitrogen, not just nitrate, throughout the third experiment. Total N (N_tot) concentration equals the sum of nitrate concentration and reduced N (ammonium and dissolved organic N) concentration. Our method of measuring reduced N concentration and isotopic composition relies on converting reduced N to nitrate (lines 166-167). Many samples were collected prior to complete nitrate consumption, so the oxidized samples we needed to investigate reduced N contained nitrate that was not previously in a reduced state. To account for this, we used the mass balance equation δ¹⁵N_tot= (δ¹⁵N_NO3)x(%NO₃) + (δ¹⁵N_reduced)x(1-%NO₃). Nitrate concentrations and isotopic compositions were known from measurements of the unoxidized samples (lines 159-164). Nitrate concentration in the oxidation blanks averaged 1.7 μM, while sample concentrations ranged from 31.9-70.8 μM. We agree that this information is important and will add clarification in revision.
  We have no reason to believe that CRS were not fully dissolved during the persulfate oxidation step. No particulate matter was observed after oxidation in culture samples or sedimentary samples. Sedimentary samples were larger and extremely rich in CRS, so if there were any issues with partial dissolution, one would expect to see it in those samples. Additionally, GLY standards were also used to ensure complete oxidation of culture samples. This clarification will be added to the text in revision.
  The smaller comments about lines 30, 78, and 235 are noted and will be addressed in revision. Thank you again to Anonymous Referee #2 - all comments are appreciated.
  
  References:
  Egan, K. E., Rickaby, R. E., Leng, M. J., Hendry, K. R., Hermoso, M., Sloane, H. J., Bostock, H., & Halliday, A. N. (2012). Diatom silicon isotopes as a proxy for silicic acid utilisation: A Southern Ocean core top calibration. Geochimica et Cosmochimica Acta, 96, 174-192.
  
  Citation: https://doi.org/10.5194/egusphere-2023-2564-AC2

Status: closed

RC1:
'Comment on egusphere-2023-2564', Anonymous Referee #1, 19 Dec 2023

Chaetoceros resting spores record low diatom-bound nitrogen isotope values: Evidence from laboratory culture and marine sediment
by Dove et al.
Summary
Dove et al cultured the diatom species Chaetoceros in the laboratory and analyzed both vegetative cells and resting spores for their opal-bound N isotopic composition, to study how they record nitrate N isotope values. In addition, they measured opal-bound N isotopes on 6 sediment samples from a sediment core close to the Antarctic margin on individual layers dominated by Chaetoceros resting spores + vegetative cells. They put their data into context with previously published opal-bound N isotope datasets across glacial-interglacial cycles with the goal to understand whether the changing abundance of Chaetoceros resting spores biases these d15N-DB records.
Although the manuscript is well written, I have substantial concerns regarding (i) the large claims the authors make based on a very small and noisy dataset, (ii) the planning and execution of the culture experiments, and (iii) the interpretation of their results, which I detail below. Throughout the manuscript, the authors are performing several calculations with a series of assumptions and corrections going into them, which is masking the fact that the actual data base is quite weak and inconsistent. Overall, I believe that this manuscript has been submitted prematurely on the basis of a very small and inconclusive dataset, and that additional, well-designed culture experiments are needed to support the statements the authors make. My recommendation therefore is to reject the manuscript but encourage re-submission once a more comprehensive and conclusive dataset can be presented.
Major comments
First of all, three culture experiments have been performed in triplicates each, of which the first two experiments were unsuccessful (in the sense that the authors have been unable to separate vegetative cells and resting spores of Chaetoceros and thus they have been unable to measure their individual d15N-DB). This leaves one successful culture experiment (no. 3), of which one of the triplicates did not yield enough diatoms for d15N-DB-CRS analysis. In another triplicate, the vegetative cells were “contaminated” with resting spores. To summarize, only two individual d15N-DB-CRS analyses have successfully been measured in the culture experiments, and in only one single culture sample/carboy, concurrent d15N-DB of vegetative cells and of resting spores have been determined. Thus, the “evidence from laboratory cultures that Chaetoceros resting spores record low d15N-DB values” as advertised in the title is very sparse.
Which brings me next to the evidence from marine sediments. In the sedimentary samples, in 4 out of the 6 samples, the d15N-DB of the bulk diatom fraction is indistinguishable from the d15N-DB of the Chaetoceros fraction (error bars in figure 4a are overlapping), only in samples no. 3 and 5 there seems to be an N isotopic difference. Sample no. 4 actually shows slightly lower d15N-DB-bulk values relative to d15N-DB-Chaetoceros, and similarly, in experiment 2, carboys 1 and 2 had higher d15N-DB in the mixed samples compared to the vegetative samples, suggesting the opposite effect in d15N-DB between CRS and vegetative cells. Thus, their statement (lines 297-298) that “Both culture experiments and sedimentary data suggest that CRS are characterized by low d15N-DB values relative to vegetative Chaetoceros and to other diatoms, (…)” is not supported by their data and will probably not withstand statistical tests.
Since the sedimentary samples are a mixture of Chaetoceros vegetative cells and resting spores, the authors try to “correct” for that, by assigning an offset of 7‰ (derived from the culture experiments) to the difference between the vegetative cells and the resting spores of the sedimentary samples (and by making a number of other assumptions, lines 320-324), making the resting spores eventually 5.6‰ lower than non-resting spore samples (lines 332-333). Clearly, this Delta-d15N-CRS offset is almost entirely due to the assumed difference of 7‰ between resting spores and vegetative cells, and there is no independent verification of that value from the sediment. Therefore, there is no conclusive evidence from sedimentary data that Chaetoceros resting spores record low d15N-DB values.
Regarding the culture conditions, the authors were initially unable to separate vegetative cells from the resting spores for N isotopic analysis, but came up with a method for their third culture experiment. So why did they then not repeat the first two experiments after finding a solution to their problem, which would have made this a much more comprehensive and reliable dataset?
Dissolved organic N concentrations were very high in the culture experiments (nearly as high as initial nitrate concentrations). In the text, this is attributed to bacterial remineralization, but the fact that concentrations were already high at the beginning of the experiments could also suggest that this comes from the natural seawater. This might be an issue for independent reproducibility of the experiments in other labs, especially if reduced N d15N values are thought to be an important driver for the d15N-DB-CRS values. Why not using artificial seawater for their experiments to avoid such high total reduced N (TRN) concentrations? With different N sources available to the diatoms (nitrate, ammonium, DON), it’s difficult/impossible to know which N source they were assimilating to what proportion at what point in time, and how this influenced the biomass and opal-bound N isotopic composition of the growing culture, and eventually, the d15N-DB of the resting spores.
Once N and Si became limiting in the cultures, were cultures kept illuminated? Continuous illumination while under nutrient limitation could have increased respiration, and with that induced15N-fractionation of amino acids. Is this an accurate representation of natural settings, where heavy silicification likely leads to the cells sinking to deeper layers where light is no longer present?
In terms of the interpretation of the low-d15N-DB signature of the CRS, I would like to see more information about the biology and physiology of the diatoms in question, and specifically, information on the timing of resting spore formation and the physiological changes that the cells undergo, as to better understand the processes that are integrated into d15N-DB-CRS. What is known about resting spore formation, its timing and the proteins involved that remain bound to the frustule? Do d15N-DB-CRS values represent a mixture of d15N-DB from the frustule bound proteins in vegetative cells and d15N-DBthat are newly synthesized upon resting spore formation (which in this case would be particularly negative)? Or are resting spores formed after multiple mitosis cycles after which half of the frustule (and with that the frustule bound proteins) is newly synthesized after each cycle? To what extent can the frustule bound proteins be exchanged/recycled in living diatoms? If additional harvests and d15N-DB-CRS analyses would have been planned and performed in between the sampling for the vegetative cells and the second sampling for the CRS, more information might have been gained about how the d15N-DB-CRS signal is acquired.

Citation: https://doi.org/10.5194/egusphere-2023-2564-RC1
- AC1: 'Reply on RC1', Isabel Dove, 22 Dec 2023
  
  We thank the referee for their thorough review and careful explanation of concerns with the experimental design. We understand that our dataset of direct measurements is small, but maintain that our conclusions are sound and based on justified assumptions.
  The experiments were time and labor intensive, as it was quite difficult to obtain sufficiently large samples from nitrate-limited cultures. Initial experimental results and concurrent sedimentary data hinted that Chaetoceros resting spores (CRS) record lower δ¹⁵N_DB values than non-CRS diatoms. Therefore, after successfully isolating CRS in two carboys from the third experiment and measuring consistently low values, we were convinced that this signal is real. A fourth experiment was not conducted due to time and resource constraints. We agree that more data is desirable, yet our results assuredly indicate that the δ¹⁵N_DB signature of CRS is lower than that of vegetative cells, which is novel and worth reporting.
  Regarding concern over the execution of culture experiments, we agree that high reduced nitrogen concentrations could be an issue. While reduced nitrogen concentrations are high, they increase throughout the experiment, suggesting that much of the nitrogen is not bioavailable and thus unlikely to influence our our δ¹⁵N_DB measurements. We also acknowledge that ammonium assimilation could contribute to low δ¹⁵N_DB values in CRS in culture, but show in the supplement that it cannot explain the difference observed in CRS-specific δ¹⁵N_DB values. More importantly, ε_DB values (δ¹⁵N_biomass – δ¹⁵N_DB) enable comparison of CRS- and vegetative-specific δ¹⁵N_DB values independent of effects from nutrient source. Field data show that δ¹⁵N_DB values do follow nutrient source δ¹⁵N, as low δ¹⁵N_biomass and δ¹⁵N_DB values were observed in the late growing season where the primary nutrient source was ammonium (Robinson et al., 2020). While our δ¹⁵N_biomass measurements in vegetative and CRS samples are similar, CRS-specific δ¹⁵N_DB values are much lower, reflected by higher ε_DB values. Distinctly different ε_DB values unequivocally show that, regardless of nutrient source, CRS-specific δ¹⁵N_DB values are lower than vegetative-specific δ¹⁵N_DB values.
  Cultures were kept illuminated throughout the entire experiment and we recognize that this is not representative of natural settings. This underscores why we also looked to marine sediments for corroboration.
  Anonymous Referee #1 brings up important concern over sparse data and the need for statistical tests. Focusing on sediment data, which avoids potential bias introduced by culture conditions, we find that CRS do record lower δ¹⁵N_DB values than non-CRS diatoms. Conducting a Wilcoxon signed rank test on isolated Chaetoceros samples versus non-CRS samples, we find that the Chaetoceros δ¹⁵N_DB values are lower (p=0.053). This test was conducted using δ¹⁵N_DB-bulk values adjusted only for varying CRS surface area, so are independent of assumptions dependent on culture results (calculated non-CRS δ¹⁵N_DB = 10.1, 11.3, 14.4, 8.7, 11.8, 8.9). The difference between Chaetoceros and non-CRS δ¹⁵N_DB values is then calculated to be 1.9 ± 2.3‰. This is obviously a smaller difference, but does not change our major finding that CRS do not bias downcore δ¹⁵N_DB records.
  We agree that Anonymous Referee #1’s questions about CRS formation are important and worth investigating. Our discussion is speculative because so little is currently known. While our focus is on paleo-implications of how CRS record δ¹⁵N_DB values, future work with carefully designed culture experiments can begin to answer remaining questions as to how CRS physiology impacts δ¹⁵N_DB values.
  As Anonymous Referee #1 noted, the goal of this work is to assess how changing CRS abundance over time impacts δ¹⁵N_DB records. We believe that, despite valid concerns over our culture experiment methodology, our conclusions that CRS record low δ¹⁵N_DB values but do not significantly bias published δ¹⁵N_DB records is sound and worthy of publication. Questions about species-specific effects have long been a concern for those using biogeochemical proxies (e.g. Jacot des Combes, 2008), so it is important to confirm that CRS are not a significant source of bias to δ¹⁵N_DB records. Additionally, this study sets the stage for future work to further investigate other species and brings up interesting questions about CRS biology and physiology.
  Again, we thank the referee for their feedback and for encouraging more thorough statistical validation of our results. The manuscript will be improved by adding discussion of potential bias introduced by culture conditions, more thorough explanation of ε_DB values, and emphasis on results from sediments.
  
  References:
  Robinson, R. S., Jones, C. A., Kelly, R. P., Love, A., Closset, I., Rafter, P. A., & Brzezinski, M. (2020). A test of the diatom‐bound paleoproxy: Tracing the isotopic composition of nutrient‐nitrogen into Southern Ocean particles and sediments. Global Biogeochemical Cycles, 34(10), e2019GB006508.
  Jacot Des Combes, H., Esper, O., de La Rocha, C. L., Abelmann, A., Gersonde, R., Yam, R., & Shemesh, A. (2008). Diatom δ13C, δ15N, and C/N since the Last Glacial Maximum in the Southern Ocean: Potential impact of species composition. Paleoceanography, 23(4).
  
  Citation: https://doi.org/10.5194/egusphere-2023-2564-AC1
RC2:
'Comment on egusphere-2023-2564 R2', Anonymous Referee #2, 19 Dec 2023

The Authors produced data from culture and sediments that Chaetoceros resting spores (CRS) have a lower d15N signature with respect to vegetative Chaetoceros. Given the variable abundance of CRS in sediments, they argue that is important to characterize sedimentary assemblages for paleo-d15N applications.
I am impressed by the amount of work that the authors produced. The causes of lower d15N in CRS remain speculative, while it is reasonable to assume more ammonium assimilation, the estimates are still based on a handful of datapoints and the authors conclude that ammonium alone cannot account for the total magnitude of difference. While this is new and valuable new data, the dataset of CRS direct estimates is really small, but I understand that this is due to the technical difficulties in both laboratory culturing and having to rely on indirect estimations from sediments. Regarding the latter, the calculation of d15N from sediments rely on assumptions of CRS abundance, but I am not sure how solid this is. Section 3 (Results) appears shorter than it should be: the explanation of what corrections have been applied (and the underlying assumptions) to estimate the d15N difference between Chaetoceros and CRS is not very clear, and should be better elucidated (or this should be done in the Supplement).
Overall, the authors present novel data and make a valid case for the importance of evaluating changes in diatom assemblages when studying diatom-bound d15N in sedimentary archives. However, if the lower d15N values of CRS are confirmed by further investigations, CRS would significantly affect paleo-reconstruction only in a very small number of cases, where their abundance is particularly high.

Smaller comments in the following:
Line 30: Word Nutrients is used twice in the same sentence: “nutrient- and CO2-rich waters to the surface, where nutrient…”
Line 78: Because CRS likely form under nutrient-depleted conditions (in which nitrate is enriched in 15N) and because they are preferentially preserved in marine sediments, we hypothesize that they bias records towards higher δ15NDB values, which signal a greater degree of nutrient utilization.
Line 125: why would 45 uM of nitrate already induce more CRS formation, if CRS are posed to form under nutrient stress?
Line 167: which amino acid standards were used? Are they international reference materials (RMs)? If so, how does their value compare to their international reference values. If this info exists, it would be informative to provide.
Line 173: “A mass balance calculation using nitrate concentrations and δ¹⁵N_NO3 values from un-oxidized samples yielded the nitrogen isotopic composition of 175 reduced nitrogen”. I do not understand this procedure. Does the δ¹⁵N_NO3of un-oxidized samples reflect the oxidation blank? How does this take into account the δ¹⁵Nof the organic blank? And how large was the oxidation blank relative to the size of your samples? This is quite important info.
Line 193: did you observe some refractory behavior of CRS during K2S2O8 dissolution and subsequent oxidation? In other words, can it be ruled out that there is not effect of partial dissolution of more strongly silicified CRS with respect to vegetative frustules?
Line 235 (and throughout the manuscript): use letters next to figure numbers to help the reader jump between the text and the figures (which already include letters).

Citation: https://doi.org/10.5194/egusphere-2023-2564-RC2
- AC2: 'Reply on RC2', Isabel Dove, 22 Dec 2023
  
  We thank the referee for their thoughtful review and especially appreciate their recognition of the amount of work that went into the culture experiments. As elaborated upon in our above comment to Anonymous Referee #1, we acknowledge that there are only two direct measurements from Chaetoceros resting spores (CRS) but remain confident that they record low δ¹⁵N_DB values based on data from mixed culture samples and from sediment.
  Regarding calculations of CRS-specific δ¹⁵N_DB values from sediment, we determined the relative abundance of both CRS and vegetative Chaetoceros spp. in the <63 μm bulk diatom assemblage using quantitative microscope slides (lines 226-228). These quantitative assemblage counts enabled us to determine the proportion of resting spores versus vegetative Chaetoceros, which is reported as CRS/Ch. in Table 2. From there, we assume that the proportion of CRS versus vegetative Chaetoceros remains constant in the <10 μm isolated Chaetoceros sample. We believe this is a reasonable assumption because we followed the microseparation technique outlined in Egan et al. (2012), which avoids fragmenting microfossils and clogging mesh.
  We appreciate that explanations of various calculations can be confusing. Much of the data required corrections, which we describe in the Discussion section. Therefore, the Results section is comparatively much shorter. A possible solution for increasing clarity is to combine the Results and Discussion sections. Regardless, revision will include more thorough and clear explanations.
  While 45 μM nitrate concentration is indeed high for the ocean, it is low enough to induce resting spore formation in culture. Our strains of C. socialis were maintained in f/2 media (882 μM nitrate concentration), so 45 μM nitrate concentration is comparatively quite low.
  We used the USGS65 glycine standard to ensure that all reduced N has been converted to nitrate during persulfate oxidation.
  Given surprising results from initial culture experiments, we realized the importance of tracking the concentration and isotopic composition of all forms of nitrogen, not just nitrate, throughout the third experiment. Total N (N_tot) concentration equals the sum of nitrate concentration and reduced N (ammonium and dissolved organic N) concentration. Our method of measuring reduced N concentration and isotopic composition relies on converting reduced N to nitrate (lines 166-167). Many samples were collected prior to complete nitrate consumption, so the oxidized samples we needed to investigate reduced N contained nitrate that was not previously in a reduced state. To account for this, we used the mass balance equation δ¹⁵N_tot= (δ¹⁵N_NO3)x(%NO₃) + (δ¹⁵N_reduced)x(1-%NO₃). Nitrate concentrations and isotopic compositions were known from measurements of the unoxidized samples (lines 159-164). Nitrate concentration in the oxidation blanks averaged 1.7 μM, while sample concentrations ranged from 31.9-70.8 μM. We agree that this information is important and will add clarification in revision.
  We have no reason to believe that CRS were not fully dissolved during the persulfate oxidation step. No particulate matter was observed after oxidation in culture samples or sedimentary samples. Sedimentary samples were larger and extremely rich in CRS, so if there were any issues with partial dissolution, one would expect to see it in those samples. Additionally, GLY standards were also used to ensure complete oxidation of culture samples. This clarification will be added to the text in revision.
  The smaller comments about lines 30, 78, and 235 are noted and will be addressed in revision. Thank you again to Anonymous Referee #2 - all comments are appreciated.
  
  References:
  Egan, K. E., Rickaby, R. E., Leng, M. J., Hendry, K. R., Hermoso, M., Sloane, H. J., Bostock, H., & Halliday, A. N. (2012). Diatom silicon isotopes as a proxy for silicic acid utilisation: A Southern Ocean core top calibration. Geochimica et Cosmochimica Acta, 96, 174-192.
  
  Citation: https://doi.org/10.5194/egusphere-2023-2564-AC2

Isabel A. Dove, Ian W. Bishop, Xavier Crosta, Natascha Riedinger, R. Patrick Kelly, and Rebecca S. Robinson

Supplement

https://doi.org/10.5194/egusphere-2023-2564-supplement

Data sets

Dissolved nutrients, cell counts, and nitrogen isotope measurements from Chaetoceros socialis culture experiments I. Dove https://doi.org/10.15784/601727

Sediment chemistry of ODP Site 1098 I. Dove https://doi.org/10.15784/601720

ODP Site 1098 deglacial diatom assemblage I. Dove https://doi.org/10.15784/601723

Isabel A. Dove, Ian W. Bishop, Xavier Crosta, Natascha Riedinger, R. Patrick Kelly, and Rebecca S. Robinson

Viewed

Total article views: 345 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
221	89	35	345	25	27	23

HTML: 221
PDF: 89
XML: 35
Total: 345
Supplement: 25
BibTeX: 27
EndNote: 23

Views and downloads (calculated since 16 Nov 2023)

Month	HTML	PDF	XML	Total
Nov 2023	13	3	0	16
Dec 2023	74	26	14	114
Jan 2024	25	6	2	33
Feb 2024	2	7	0	9
Mar 2024	12	10	0	22
Apr 2024	11	6	3	20
May 2024	7	8	4	19
Jun 2024	26	9	4	39
Jul 2024	12	1	5	18
Aug 2024	15	3	3	21
Sep 2024	13	3	0	16
Oct 2024	8	4	0	12
Nov 2024	3	3	0	6

Cumulative views and downloads (calculated since 16 Nov 2023)

Month	HTML	PDF	XML	Total
Nov 2023	13	3	0	16
Dec 2023	74	26	14	114
Jan 2024	25	6	2	33
Feb 2024	2	7	0	9
Mar 2024	12	10	0	22
Apr 2024	11	6	3	20
May 2024	7	8	4	19
Jun 2024	26	9	4	39
Jul 2024	12	1	5	18
Aug 2024	15	3	3	21
Sep 2024	13	3	0	16
Oct 2024	8	4	0	12
Nov 2024	3	3	0	6

Viewed (geographical distribution)

Total article views: 327 (including HTML, PDF, and XML) Thereof 327 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 21 Nov 2024

Short summary

The diatom-bound nitrogen isotope proxy is used to study how efficiently diatoms in the Southern Ocean help to remove CO₂ from the atmosphere, but may be biased by different diatom species. We examine a specific type of diatom, Chaetoceros resting spores (CRS), commonly preserved in Southern Ocean sediments. We find that CRS record surprisingly low δ¹⁵N_DB values compared to other diatoms, yet changes in their relative abundance over time does not significantly bias previously published records.


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