Glaciogenic iron transport pathways to the Kerguelen offshore phytoplankton bloom

Nalivaev, Alex; d'Ovidio, Francesco; Bopp, Laurent; Berta, Maristella; Rousselet, Louise; Azarian, Clara; Blain, Stéphane

doi:https://doi.org/10.5194/egusphere-2025-2145

Preprints

https://doi.org/10.5194/egusphere-2025-2145

Preprints

04 Jun 2025

| 04 Jun 2025

Glaciogenic iron transport pathways to the Kerguelen offshore phytoplankton bloom

Alex Nalivaev, Francesco d'Ovidio, Laurent Bopp, Maristella Berta, Louise Rousselet, Clara Azarian, and Stéphane Blain

Abstract. In contrast to the average low biological productivity across most of the Southern Ocean, the Kerguelen region is one of the few subantarctic regions to host massive phytoplankton blooms, extending hundreds of kilometers offshore. These blooms play a crucial role in the Southern Ocean carbon cycle and support a diverse ecosystem of patrimonial and commercial significance. The Kerguelen blooms are associated with a subsurface iron source that supplies surface waters both on the Plateau and offshore. The mechanisms of iron enrichment have only been partially elucidated. The resuspension of iron-enriched sediments over the Plateau, transported offshore by the Antarctic Circumpolar Current, is one mechanism that has been studied in the past years. However, the Kerguelen Islands host a glacier system, and two of the outlet glaciers of Kerguelen’s Cook Ice Cap are likely to provide iron enriched lithogenic material downstream to the coastal waters of the Golfe des Baleiniers. Whether the circulation is able to connect the glacier outlets to the open ocean, and how much of the offshore bloom extension can be reached by glaciogenic iron is not known. Using in situ and satellite data, including observations from the recent SWOT satellite mission, we reconstruct the horizontal advection of iron and show that glaciogenic iron supply reaches up to one third of the spatial extent of the offshore bloom onset. These findings have significant implications in the context of ongoing ice cap mass loss and glacier retreat observed on Kerguelen and other Southern Ocean islands under climate change.

Received: 07 May 2025 – Discussion started: 04 Jun 2025

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

Alex Nalivaev, Francesco d'Ovidio, Laurent Bopp, Maristella Berta, Louise Rousselet, Clara Azarian, and Stéphane Blain

Status: final response (author comments only)

RC1:
'Comment on egusphere-2025-2145', Thomas M. Holmes, 19 Jun 2025

EGUsphere-2025-2145 Review

General comments
The manuscript entitled “Glaciogenic iron transport pathways to the Kerguelen offshore phytoplankton bloom” by Nalivaev, et al. is a timely and interesting study of the effect of iron derived from glacial runoff on a massive annual phytoplankton plume in the Southern Ocean, where phytoplankton growth is generally chronically limited by extremely low iron concentrations. Given the rapid melt rate of these glaciers, it is very important to understand the biogeochemical processes that may be affected by changing supply of glacial runoff, which may have ramifications for marine ecosystems and carbon sequestration efficiency. The authors assess pathways for glacially derived iron from Kerguelen Island to reach the annual phytoplankton bloom over and downstream of the Kerguelen Plateau. They use a combination of satellite data, drifter observations and lagrangian modelling to show that iron from glacial melt on the island can fertilise a significant portion of the bloom.
The mechanisms discussed in this study will be important for informing forthcoming publications looking into in situ biogeochemical measurements collected on the recent MARGOCEAN cruise undertaken in this study region, and more broadly for future studies gauging the effects of retreating glaciers on Southern Ocean biogeochemistry. In my opinion, the scientific method utilised in this study is robust and the paper seems reasonably polished. I have minor suggestions on the scientific discussion and technical corrections. As such, this manuscript will be well suited for publication in Biogeosciences after minor revisions.
Specific comments
Figures: Either figures need to be slightly larger, or text in figures needs to be larger font. For multi-panel figures (2, 5, and 8) the authors could save space or enlarge figures by only having lat/lons on the outer axes (i.e. lats on the y-axis of panels a and c, lons on the x-axis of panels c and d).
Since there is quite a bit of discussion about the polar front, it might be good to show the mean polar front location (to the north of Kerguelen Island) in one of the figures.
Line 160 - 162: A little more information would be good here. At what rate does the iron content diminish? How is this calculated, or what is this based on? What assumptions are made? Is the reduction of iron to 0.1% after 60 days backed up by observational data? Could the authors be a bit more specific about the sensitivity test? Is this per day?
Line 170: Is this a fair assumption to make? Seems unlikely that there is no biological interaction with iron for 60 days during advection… Perhaps this needs rewording?
Line 173 – 177: So, the authors choose the threshold for defining the area of the offshore bloom based on the sensitivity tests? I would move the sentences ‘A statistical threshold of chlorophyll concentration is used to define the area of the offshore bloom. In addition, the iron plumes are horizontally extrapolated to the nearest n pixels (pixel size: 0.01° longitude and latitude).’ to after the sentence about the sensitivity tests if this is case. If it’s not the case, then how do the authors choose which percentile to use? Please explain the methodology a bit more clearly.
I think the SWOT component of this paper is interesting, but right now it seems a bit like a side-project that isn’t really too relevant to the results of the study. The authors could almost sum up the results in a single sentence that the increased resolution of SWOT didn’t really give different results due to the scale of the processes involved, and move the rest to the appendix. If the SWOT results are actually integral to this study, then it would be good to focus on/highlight that a bit more.
In the discussion, the authors correctly identify the limitation of this work as lack of high resolution in situ observations and assumption of the rate of iron removal. The authors state that removal rate is considered through the sensitivity tests they performed, however a comparison of the removal rates used in this study to results from previous studies is missing. It would be good to validate these assumptions by comparison with iron concentrations observed during previous voyages such as ANTARES 3 and KEOPS.
The discussion ends rather abruptly and could use a final sentence summarising the implications of this study in the context of the previous paragraphs.
Technical corrections
References: After checking the BG requirements, I think the format the authors have used for in-text citations is acceptable, except for where there is more than one reference together, where they should be separated by a semicolon, not a comma.
Figure 1: It might be nice for readers not familiar with this region to have an inset, or another panel showing where the plateau is in the Southern Ocean.
Line 54: Suggest replacing ‘englobing’ with ‘including’ or ‘surrounding’, and citing Figure 1 at the end of the sentence.
Line 59: Can now remove the final sentence of this paragraph as figure is cited above.
Line 66: Change ‘on’ to ‘in’.
Line 80: ‘North’ does not need to be capitalised here.
Line 81: ‘Golfe des Baleiniers’ is the name of the gulf, but as this paper is written in English, I think that ‘Golfe’ at the end of the sentence should just be ‘gulf’ (no capitalisation needed, as this is just the noun, not the name).
Line 121: The coordinates are not clear as currently written, suggest changing to ‘(66 – 90°E, 45 – 55°S)’.
Line 153: Please amend coordinates to match the format suggested at line 121.
Line 190: Same comment for ‘Golfe’ here as at line 81.
Line 192: Where does the threshold of 0.75 mg.m−3 come from? Please relate this to the previously mentioned percentiles. Also remove full stop after ‘0.75 mg.m−3’.
Line 198: Remove ‘for short’.
Line 200: Remove ‘…for methodological consistency. Indeed,…’ and replace with ‘as’.
Line 206: After ‘…(considered as control…’ add reference to grey shaded area in Figure 3d.
Line 207: I would add ‘used as control’ after ‘For the offshore bloom’ to make sure this is clear. Also what does the ‘grey curve’ relate to? Maybe just say ‘grey shaded area’.
Line 208 – 209: As with the previous comment, maybe change to ‘blue shaded area’ and ‘red shaded area’.
Figure 3: It could be good to have an annotation showing the location of the GdB on panel b. Change ‘blue/red/grey curve and area’ to just ‘blue/red/grey shaded area’ for consistency with previous comments. Make sure to use either ‘grey’ or ‘gray’ consistently throughout the manuscript. I would amend the final sentence of the caption to read ‘On graph c), uncertainties are shown as lighter shading and represent the interannual variability’. Lastly, this is purely aesthetic and to save a tiny bit of space, but could the authors align the tops of panels a) and b), and the bottoms of panels c) and d).
Line 221: Change ‘are’ to ‘were’.
Line 223: As per the BG instructions to authors, all figure citations should be in the format Fig. 4, unless at the start of a sentence.
Line 224: Please change ‘…Fig.4)’ to ‘(Fig. 4)’.
Line 245: Suggest moving ‘km²’ to after ‘106,100’.
Line 255: Fix figure citation.
Line 257: Remove full stop and replace ‘We calculate’ with ‘by calculating’.
Line 259: Fix figure citation.
Line 260: Fix figure citation.
Line 268: Would it be possible to show this data in the appendix?
Figure 5: Line 2 of caption, change ‘multisattelite’ to ‘multi-satellite’. I don’t think the hatching is the clearest way to delineate the 2000m bathymetry. Could the authors just show the 2000m contour in another colour, or perhaps finer scale hatching or shading in a lighter colour, or a combination of these? Also, what do the boxes represent in panels c) and d)?
Figure 6: Y-axis label, change ‘km2’ to ‘km²’. Error bars are a bit hard to see. Recommend adding markers to the end of the error bars, or removing grid lines.
Line 283: delete ‘in respect’.
Figure 7: Recommend adding markers to the end of the error bars, or removing grid lines. Perhaps even just making grid lines finer/lighter could help. Also, could grid lines be drawn under the bars, instead of over them?
Line 289: Please fix Bindoff reference.
Line 300: Fix figure citation.
Line 302: Consider replacing ‘remarkably’ with ‘closely’. Fix figure citation.
Line 305: Fix figure citation.
Line 331: Consider replacing ‘stirring’ with ‘mixing’ or ‘advection’.
Line 346: Suggest deleting ‘and soon to come papers’, as this is quite vague.
Line 347: Suggest changing the tone of this sentence, as ‘strong limitation of this study’ is quite negative and casts doubt on the strength of this study, which should not be the case as the authors have carried out a scientifically robust method. I would instead reword this to talk about the assumptions that the authors had to make, due to lack of data and scope of this work (which is outlined in following sentences).
Line 382: Remove comma after ‘scale’.
Figure C1: In the caption line 2, did the authors mean panels d to i, not e to i?

Citation: https://doi.org/10.5194/egusphere-2025-2145-RC1
- AC1: 'Reply on RC1', Alex Nalivaev, 29 Aug 2025
  
  We would like to thank Thomas M. Holmes for his thorough and constructive feedback, which we believe will help us to significantly improve our manuscript. We will ensure that all technical comments are dealt with in the revised version of the manuscript.
  Comments related to the scientific content :
  Iron removal processes during the advection experiments:
  This issue was also addressed by the second referee. From the referees’ comments, we deduce that there is a misunderstanding regarding the iron removal rates incorporated during Fe advection, a point that we will clarify in the revised version the manuscript.
  In this study, we apply a first-order exponential decay to the iron concentration over time, as in d’Ovidio et al. (2015) and Ardyna et al. (2017).
  C/C₀ = exp(-λt)
  In this formula, C₀ is the initial iron concentration at the iron sources (i.e. the glacier or plateau), which is undetermined (its estimation is one of the objectives of the MARGOCEAN cruise), t is the integration time (i.e. the time since the water parcel left the iron source), and C is the iron concentration at time t in the water parcel. We use the iron removal constant λ, which was estimated during the KEOPS2 cruise and reported in d’Ovidio et al. (2015) as λ=0.1 day^-1, including the contributions from biotic and abiotic iron removal processes.
  Although an approximation, this simple formulation enables us to consider the numerous and complex iron removal processes simultaneously. Singling out each of these processes (scavenging, complexation, etc) was considered to be beyond the scope of this study.
  Ardyna et al. (2017) term the variable C/C₀ “iron delivery”, which is what we refer to at lines 160–161 when we mention “a total reduction of the initial iron content to about 0.1%”. According to the formula above and to the estimation of λ=0.1 day^-1, a reduction of the initial iron content to 0.1% (i.e., C/C₀=0.001) corresponds to an integration time of approximately 69 days.
  t= -ln(C/C₀)/λ = -ln(0.001)/0.1 = 69 days.
  Due to this link between t and C/C₀, water parcels that left the iron source 60 days ago contain approximately 0.1–0.2% of the initial iron concentration at the source.
  The current state of scientific knowledge on iron limitation does not enable us to identify a clear iron delivery threshold below which the impact of iron supply on phytoplankton cells could be considered negligible. Therefore, we performed sensitivity tests on iron delivery thresholds, which is equivalent to performing sensitivity tests on the integration time t. We considered three different iron delivery thresholds, i.e. of 0.1%, 0.5%, and 1%, respectively, computing for each the match between the iron plume and the offshore bloom extension. Standard deviations of the overlaps (indicated by percentage variations in Fig.6, Fig.7 and Fig.9 of the first submitted version) are obtained from these sensitivity tests. In other words, we assumed that the water parcels after integration times of about 70, 55 and 45 days, respectively, did not supply enough iron to be considered significant to phytoplankton ecosystems.
  The biological consumption of iron along its transport pathways:
  In accordance with the referee’s comment, we would like to amend the statement made at line 170. When conducting iron advection experiments, we incorporated both biotic and abiotic iron removal processes via the parameter λ. As mentioned in d’Ovidio et al. (2015), the KEOPS2 cruise measurements allowed to estimate the biotic and abiotic removal constants (resp. λ_biotic ~ 0.06 day^-1 and λ_abiotic ~0.04 day^-1), according to the aforementioned first-order iron removal model. The KEOPS2 cruise was conducted during the same season as this study (early spring), justifying the use of the estimation of λ made by KEOPS2.
  Validation of the iron removal rates used:
  In our study, we explicitely use the iron removal rates (λ parameter mentioned above) estimated during KEOPS2 (d’Ovidio et al. 2015). In the previous version, this was not clearly stated but we will be sure to indicate this information in the revised version. We chose the estimation of KEOPS2 because the cruise was explicitely designed for relating removal rates to advective pathways. The sensitivity tests conducted on iron supply thresholds can also be viewed as sensitivity tests on the value of the λ parameter.
  For instance, consider a fixed advection time t of 45 days. An iron supply threshold of 1% (ie, 1% of the initial iron concentration remaining) is correlated with an iron removal rate λ of 0.1 day^-1, as previously stated. Changing the iron supply threshold from 1 to 0.1% while keeping the advection time fixed yields a change in the iron removal rate λ from 0.1 day^-1 to 0.15 day^-1 . This is coherent because, for the same advection time, a smaller fraction of the initial fraction remaining would entail stronger removal processes.
  Definition of the offshore bloom based on statistical thresholds:
  Our definition of the offshore bloom area is independent of the iron plume trajectories and the sensitivity tests performed on them. To compare the areas reached by iron plumes with the spatial extent of the offshore bloom, we need to convert both fields into binary grids. We applied a relative threshold of chlorophyll concentration, above which the area was considered to be impacted by a bloom. To alleviate our results' dependency on the choice of threshold, we chose three different thresholds (median, 60^th and 70^th percentiles) and performed a sensitivity test on our definition of the area of the offshore bloom. Using relative thresholds rather than absolute chlorophyll concentration thresholds is justified by the high interannual variability in the magnitude of the offshore bloom, given that our analysis in Results section 3.3 spans a ten-year period. Furthermore, we find this methodology to be consistent with our definition of iron plumes. As we have no knowledge of the amount of iron emitted by the sources or its interannual variability, applying the same absolute chlorophyll concentration threshold each year did not seem consistent. While we acknowledge that the choice of thresholds (median, 60^th and 70^th percentiles) is arbitrary, they appear to reproduce the reported spatial extent of the offshore bloom (e.g. in Bowie et al., 2015).
  SWOT component of the article:
  We fully agree with the referee’s remark on the results obtained using SWOT. As this study began when the first SWOT products became available, the authors were initially thrilled to apply SWOT data to biogeochemical studies for the first time. However, the fact that SWOT products did not significantly improve the fraction of the bloom reached by iron pathways with respect to conventional altimetry was rather unexpected. As suggested by both the first and second referees, we will consider displaying these results as Supplementary Material.
  
  Figure related comments and technical comments:
  As suggested by both the first and second referees, we will add the location of some of the ACC fronts to Figure 1, including the Polar Front (PF).
  On Figure 5, the boxes represent the areas where the statistics shown on Figures 6 and 7 are computed. They correspond to a fraction of the offshore bloom area. We will make sure this information is clearer in the revised version of the article.
  In accordance with the referee’s inquiry regarding line 268 of the manuscript, we will show in the appendix the results obtained when the areas of the offshore bloom and of the iron plumes are compared on seasonal average.
  
  Citation: https://doi.org/10.5194/egusphere-2025-2145-AC1
RC2:
'Comment on egusphere-2025-2145', Ryan Cloete, 26 Jun 2025

General Comments:
This study provides a novel and timely investigation into the contribution of iron sources to offshore phytoplankton blooms near the Kerguelen Islands, an area of high biological and biogeochemical interest. It addresses an under-explored iron source—glacial runoff from the rapidly melting Cook Ice Cap—and makes a strong case for its relevance in sustaining oceanic productivity in iron-limited HNLC regions. The integration of satellite data, Lagrangian modelling, and in situ drifter observations is innovative, multidisciplinary and represents a significant step forward in understanding fine-scale iron transport dynamics in the Southern Ocean. Furthermore, this study is relevant in terms of investigating the impact of changing climatic conditions on pristine environments and provides useful tools and methodologies for future studies addressing the influence of increasing glacial melting (and indeed other plume related nutrient inputs ) on coastal ecosystems.
I think this study is well suited to the themes of Biogeosciences. That being said, I think some some important clarifications are warranted first. Most of my concerns relate to biogeochemical transformations of Fe during advection which impact Fe loss terms and bioavailability to support offshore blooms. Furthermore, the study relies heavily on the use of chlorophyll as a proxy for Fe, the caveats of which need to be better explained.
Specific Comments:
Overlapping Sources: Plateau vs. Glaciogenic Iron:
The distinction between "Plateau" and "glaciogenic" iron sources requires clarification since the glaciogenic outflow region is nested within the Plateau area and could lead to the potential misattribution of the bloom area to glaciogenic sources that might instead be explained by sedimentary or resuspension-derived iron.
For the glacial discharge specifically, could you clarify whether this source is directly to the waters of the Golfe des Baleiniers. I ask this because from your reference to Verfaillie et al. (2021), and satellite images, it seems as though the various discharge outlets from the Cook ice cap flow first into glacial lakes, then to river systems and eventually to coastal waters. This would likely influence the Fe biogeochemistry and therefore its behaviour across the salinity gradients. If this is indeed the pathway of Fe from land to sea, are glacial vs fluvial Fe sources distinguished or grouped together as glaciogenic? In case there are marine terminating glaciers to the Golfe des Baleiniers, this would likely be a subsurface Fe source. What processes would then transport Fe to surface?
Fe speciation:
Throughout the manuscript, there is little focus on Fe speciation, chemical transformations and related bioavailability. Is the assumption that all sourced Fe is bioavailable? Van der Merwe et al., 2019 (and references in the introduction) showed that factors such as source rock composition and the geomorphology of weathering products had very different labilities and bioavailabilities. I understand a loss term of 0.1% was incorporated during Fe advection, however this should be validated more clearly with details of how this value was decided (see line comments for more). These factors should be discussed with implications for advected Fe and subsequent bloom importance. Additionally, could some of the observed bloom patterns arise from other mechanisms (e.g., vertical mixing, in situ regeneration)?
Advection of Fe in surface waters:
I think the use of CARTHE drifters is an exciting and multidisciplinary approach. If I understand correctly, these drifters accurately resemble the surface flow regime (0.6 m depth). Given the short residence time of Fe at the surface, its particle reactivity and that the drifters took on the order of days-weeks to reach the GdB outflow, it is likely that a significant portion of Fe is lost from the surface layer during advection (aggregation, particle settling, uptake, complexation etc). I think there needs to be some discussion on this, particularly how Fe is stabilised at surface and advected over the distance between source and GdB outflow.
Chlorophyll as a proxy for Fe:
The use of chlorophyll as a proxy for iron presence should be better justified. The potential for other environmental controls on bloom development (light – although this was mentioned a few times, grazing, temperature, nutrients, community composition) could be brought in.
In summary, I understand that there is a clear focus on Fe transport pathways in the manuscript, and that this contribution sets the scene for future papers focusing on in-situ data, however, I think the biological impact is inferred to a degree rather than measured. If supporting biogeochemical observations from the MARGOCEAN cruise cannot be used to reinforce the outcomes (because they are being saved for future papers), the processes affecting Fe flux magnitude, chemistry and transport should at least be discussed in more depth and related to bloom phenology.
SWOT satellite:
The use and incorporation of the SWOT data to analyse finer scale structures is innovative and I commend the authors for this effort. Of course, validating new data products is key to scientific advancement. I do however question it’s relevance here given that the conclusion is that physical circulation features around Kerguelen and the Plateau are large enough to be resolved by DUACS and that assessing the true value of SWOT is better suited to other areas with finer scaled physics. I think this part of the manuscript can by shortened and possibly moving Figure 8, 9 to the appendix.
Figures:
Figure 1: I appreciate the figure is already complex however labelling the water masses (SASW and WW) and showing the position of the Polar Front (this could also be shown in any of the figures 2-5) would provide needed oceanographic context. Consider adding a transparency filter over the ADT data or choosing a different colour scheme as currently the colour gradient is overpowering the figure detail e.g. the bathymetry lines (which could be bolder). Some wider geographical context as to the relevant position of the study area will also be valuable (i.e. including an inset map).
References:
Check in-text references. I am not familiar with the style used and it doesn’t appear to be the suggested style from the journal.
Technical/line by line comments:
Line 16: ‘Phytoplankton are…. and serve….’
Line 33: What are typical timescales between bloom initiation and termination?
Line 42: You can add Krause et al., 2024 too (https://doi.org/10.5194/tc-18-5735-2024).
Line 91: First use of SWOT should be typed out in full.
Line 106: Figure captions say 2004-2023. Please check.
Line 135: …were conducted in the Golfe…
Line 155: Is there a reference for the 60 day duration required for glaciogenic and Plateau iron to reach the region of the offshore bloom?
Line 160-163: Please clarify, so Fe concentrations are being reduced to 0.1% of initial concentration during 60 days of advection? Or being reduced by 0.1%? From Line 161 and Figure C1 it seems Fe was reduced by 0.1% (with sensitivity testing including 0.5 and 1% Fe loss). Is this Fe reduction applied constantly over the 60 day period? This is relevant because I would think most of the Fe, particularly the glaciogenic Fe, would be lost through scavenging and aggregation on entry into the saline coastal waters (similar to how the majority of riverine derived Fe is lost in estuaries before discharge to coastal waters). Some discussion of the relevance of the 0.1% Fe decrease, and the associated assumptions, is warranted.
Line 165: Please check dates throughout as they do not align with figure captions in some cases.
Line 170-171: This assumption requires further details. Figure 2 climatologies show at least some less intense and more isolated areas of biological activity. What is sustaining this? Is it all recycled Fe? Are other Fe loss terms included during this advection period?
Line 189: Referring to Figure 3b before 3a makes the reader feel as if they’ve missed some text. Suggest rearranging figure order.
Line 193: What is the justification for using a chl threshold of 0.75mg.m−3?
Line 199: Referring to Figure 3d before 3c.
Line 211: I agree that there appears to be a difference in bloom magnitude across the assessment areas. However, the interannual variability (shaded areas) are overlapping over most of the integration period. I think some statistical testing here would help strengthen this statement.
Line 224: missing single bracket. Same for elsewhere (e.g. line 259, 260, 263).
Line 297: ‘space’ before reference Raiswell.
Line 300-301: I would tone this statement down because Fe is inferred not measured.
Line 367: I’m not sure how many people are familiar with the term ‘fortiori’, consider using a more lay term.
Line 385: The last sentence ends as if it were mid-thought. Perhaps ending with something along the lines of Subantarctic islands like Kerguelen can be viewed as important and pristine environments where climate change impacts can be studied. Also that this study provides new tools and multidisciplinary methods to use in future studies looking at plume related nutrient sources.

Citation: https://doi.org/10.5194/egusphere-2025-2145-RC2
- AC2: 'Reply on RC2', Alex Nalivaev, 29 Aug 2025
  
  We would like to thank Ryan Cloette for his relevant comments on the iron biogeochemistry and iron transformation issues discussed in this article. We will consider the referee's suggestions for clarifying the issues addressed, particularly those concerning the processes affecting iron during advection, with the aim of improving the accuracy of the article from a Fe chemistry perspective. We would also like to thank Ryan Cloette for his thorough reading of the article and his precise suggestions for technical improvements.
  Comments related to the scientific content :
  Overlapping Sources: Plateau vs. Glaciogenic Iron:
  We fully agree with the referee’s remark on the distinction between ‘Plateau’ and ‘glaciogenic’ iron sources, which was one of the many challenges of this study. The area labelled as a 'glaciogenic' iron source is contained within the Northern Plateau and most likely benefits from two iron sources: a resuspension-derived supply and a glaciogenic supply, as stated in Results section 3.1 (lines 210–213) and in the Discussion (lines 327–328). However, our study is independent of iron concentrations at the source (in fact, this is one of the main limitations of the article). Instead, this article is based on a comparison between the advective pathways of water coming from potential iron sources and the observed spatial extension of the spring bloom. We do not claim to demonstrate the extent to which glaciogenic iron contributes to the development of the offshore bloom (from a biogeochemical perspective), but rather to estimate the fraction of the offshore bloom that can be reached by a potential glaciogenic iron source. Therefore, the existence of one or more iron sources overlapping over the area labelled as 'glaciogenic source' does not influence our results.
  Note that, as presented in the Discussion (lines 328-330), other glaciogenic iron sources may also be present in the ‘Plateau’ contribution, notably iron supply from glaciers located on the southern and western coasts of Kerguelen Islands.
  The land-terminating glacier:
  As the referee pointed out, the glacier is land-terminating (Verfaillie et al., 2021). Therefore, we agree that we use the term « glaciogenic » as a short cut, as what is called « glaciogenic » Fe was indeed also transported by riverine systems along its land-to-sea pathway. A more accurate description of the iron studied here would be a formulation such as « Fe inputs influenced by glacial processes » or « Fe containing a glaciogenic contribution ». However, these formulations were considered too long.
  Instead, we propose clarifying the origin of Fe at the beginning of the revised manuscript to increase its accuracy from a biogeochemical point of view, and use the term « glaciogenic » as a short cut in the rest of the manuscript.
  We chose not to detail the processes influencing Fe biogeochemistry on its land-to-sea pathway in the manuscript as it was not the scope of the paper. These processes are being thoroughly studied by other scientists who participated in the MARGO cruise.
  Fe bioavailability and speciation:
  The objective of our article is not to quantify the biogeochemical contribution of the glacier to the phytoplankton bloom, but to demonstrate that the region where glaciogenic iron is injected into the ocean is connected to the offshore region where the phytoplankton bloom is observed. That said, the lack of consideration of the speciation of iron of glacial origin and its bioavailability in the study is currently a critical knowledge gap that hinders a quantitative link between glaciogenic iron and offshore biogeochemical budgets. Ongoing MARGO studies are helping to improve our current knowledge on these two aspects. For example Thoppil et al. (2025) found that the material of glacial origin (MGO) stimulates bacterial activity, a first step possibly making MGO available for phytoplankton.
  Fe removal processes:
  This issue was also raised by the first referee. From the referees’ comments, we deduce that there is a misunderstanding regarding the iron removal rates incorporated during Fe advection, a point that we will clarify in a revised version the manuscript.
  In this study, we apply a first-order exponential decay to the iron concentration over time, as in d’Ovidio et al. (2015) and Ardyna et al. (2017).
  C/C₀ = exp(-λt)
  In this formula, C₀ is the initial iron concentration at the iron sources (i.e. the glacier or plateau), which is undetermined (its estimation is one of the objectives of the MARGOCEAN cruise), t is the integration time (i.e. the time since the water parcel left the iron source), and C is the iron concentration at time t in the water parcel. We use the iron removal constant λ, which was estimated during the KEOPS2 cruise and reported in d’Ovidio et al. (2015) as λ=0.1 day^-1, including the contributions from biotic and abiotic iron removal processes.
  Although an approximation, this simple formulation enables us to consider the numerous and complex iron removal processes simultaneously. Singling out each of these processes (scavenging, complexation, etc) was considered to be beyond the scope of this study.
  Ardyna et al. 2017 term the variable C/C₀ “iron delivery”, which is what we refer to at lines 160–161 when we mention “a total reduction of the initial iron content to about 0.1%”. According to the formula above and to the estimation of λ=0.1 day^-1, a reduction of the initial iron content to 0.1% (i.e., C/C₀=0.001) corresponds to an integration time of approximately 69 days.
  t= -ln(C/C₀)/λ = -ln(0.001)/0.1 = 69 days.
  Due to this link between t and C/C₀, water parcels that left the iron source 60 days ago contain approximately 0.1–0.2% of the initial iron concentration at the source.
  The current state of scientific knowledge on iron limitation does not enable us to identify a clear iron delivery threshold below which the impact of iron supply on phytoplankton cells could be considered negligible. Therefore, we performed sensitivity tests on iron delivery thresholds, which is equivalent to performing sensitivity tests on the integration time t. We considered three different iron delivery thresholds, i.e. of 0.1%, 0.5%, and 1%, respectively, computing for each the match between the iron plume and the offshore bloom extension. Standard deviations of the overlaps (indicated by percentage variations in Fig.6, Fig.7 and Fig.9 of the first submitted manuscript) are obtained from these sensitivity tests. In other words, we assumed that the water parcels after integration times of about 70, 55 and 45 days, respectively, did not supply enough iron to be considered significant to phytoplankton ecosystems.
  Biological consumption of Fe during its advection:
  In accordance with the referee’s comments, we would like to emphasize that the iron removal processes accounted for in the formula above include biological consumption. As reported in d’Ovidio et al. (2015), the KEOPS2 cruise measurements allowed an estimation of biotic and abiotic removal rates (resp. λbiotic ~ 0.06 day^-1 and λabiotic ~0.04 day^-1) during the spring season (same season as in our study), according to the aforementioned first-order iron removal model. We will make sure that this aspect is clarified in the revised version of the manuscript.
  Other mechanisms influencing the observed bloom patterns:
  In this article, we studied the extent to which horizontal advection pathways connect potential iron sources to the offshore bloom. We agree that several other mechanisms may also provide iron to the blooming region. These mechanisms include vertical mixing, which we addressed in the discussion, as well as regeneration, which we will cite in the revised version. As also pointed out in the Discussion, these other mechanisms may explain part of the remaining 60% of the offshore bloom that was not reached by horizontal advective plumes in our experiment.
  Advection of Fe in surface waters:
  In our methodology, we assumed that iron was distributed within the mixed layer. Both in situ measurements and model outputs indicate that the mixed layer is several tens of meters deep, that is, much deeper than the 0.60 m reached by the CARTHE drifters. For this reason, we subtracted the Ekman velocity from the total velocity of the CARTHE drifters in order to obtain a 'pseudo in situ' measure of the geostrophic velocities, which, according to Lehahn et al. (2018), are more representative of the motion within the entire mixed layer.
  With regard to iron removal from the mixed layer, we acknowledge that a significant proportion of iron is lost during advection due to various mechanisms, which we have attempted to account for in our simplified model of iron removal. However, we acknowledge that our rationale has a limitation in that we are unable to single out the impact of each of the processes listed by the referee. This will be addressed in forthcoming MARGOCEAN articles and is beyond the scope of the article.
  We acknowledge that our consideration of iron removal in this study was unclear to readers, and we thank the referees for highlighting this issue. We intend to address this issue in the revised version of the article.
  Chlorophyll as a proxy for Fe:
  As mentioned by the referee, it is possible that there are regions supplied in iron containing glacial contribution, but where the bloom did not develop, due to light or Si limitation, or grazing. However, the opposite (i.e., a bloom without an iron supply) is not possible. Therefore the area we defined as “GdB outflow” and considered as a glaciogenic iron source in Results section 3.3 can be seen as a lower bound of the area impacted by iron containing glacial contribution.
  SWOT component of the article:
  We agree with the referee’s remark, and will display the results obtained with SWOT as Supplementary Material.
  Technical/line-by-line comments:
  Period of the Lagrangian experiments (line 106 of the manuscript):
  The surface chlorophyll concentration data used in Results section 3.1 were indeed extracted over the period 2004-2023 (Copernicus GlobColour multi-satellite data). However, interannual analyses of iron advection using DUACS altimetry products were performed over a 10-year period (2014-2023), hence the misunderstanding.
  60 day duration required for glaciogenic and Plateau iron to reach the region of the offshore bloom (line 155 of the manuscript):
  This choice was motivated by the Lagrangian advection experiments performed. We found that 60days was the order of magnitude of the duration required for iron particles originating from iron sources to reach the area of the offshore bloom where our statistics are computed (this area is shown as black boxes in Fig. 5 and Fig.8, panels c) and d)). We note that the same duration was observed by d’Ovidio et al. (2015) in the framework of the KEOPS2 cruise (see in particular Figure 8 in d’Ovidio et al. 2015). We considered the possibility of extending the advection period to, e.g., 90 days or more. However, as discussed in the Material and Methods section as well as in the responses above, after a duration of 60days of advection, the iron concentration of water parcels represents about 0.1-0.2% of the initial concentration at the source (following our iron removal model and using the iron removal rates estimated by KEOPS2). Given the even smaller fraction of the initial iron content that would remain in the water parcels after a longer advection, we considered limiting the advection experiments to a duration of 60days to be a reasonable choice.
  Statistical significance of the comparison of bloom magnitudes over the GdB outflow area and the Plateau (Results section 3.1, line 211 of the manuscript):
  We agree with the referee’s remark. We will consider performing a Brunner Munzel statistical test (as done in Lévy et al. 2025) in order to confirm whether the magnitude of the bloom over the GdB outflow area is stronger than the average bloom magnitude over the Northern Plateau.
  
  Citation: https://doi.org/10.5194/egusphere-2025-2145-AC2
RC3:
'Comment on egusphere-2025-2145', Yoav Lehahn, 03 Jul 2025

This a very interesting and timely paper that provides, for the first time, robust quantification and mechanistic understanding on the contribution of glaciogenic iron supply from the Golfe des Baleiniers to offshore phytoplankton blooms in the Southern Ocean, where productivity is limited by iron availability. The work described in this paper relies on synergy between multiple scientific approaches, combining information from satellite remotes sensing (including the relatively new SWOT mission), in-situ observations and Lagrangian modeling.
Using this comprehensive dataset the authors show that about one third of the early offshore bloom is reached by glacial iron supply pathways, supporting the rationale that the development of phytoplankton blooms in the Kerguelen region strongly depends on glaciogenic iron supply routes. The authors also show the by which the phytoplankton-rich plume comprising the bloom is structured by fine scale circulation patterns, providing a mechanistic understanding on the iron-enrichment process.
The paper is very well written and easy to follow. The author describe in a clear and convincing way the rational of the research, the methodology that was used and the obtained results. I am sure this paper will be of great interest to a broad readership, and I recommend accepting it for publication in Biogeosciences in its current form.

Citation: https://doi.org/10.5194/egusphere-2025-2145-RC3
- AC3: 'Reply on RC3', Alex Nalivaev, 29 Aug 2025
  
  The authors would like to thank Yoav Lehahn for his encouraging review of our manuscript and for his interest in our study.
  
  Citation: https://doi.org/10.5194/egusphere-2025-2145-AC3

Alex Nalivaev, Francesco d'Ovidio, Laurent Bopp, Maristella Berta, Louise Rousselet, Clara Azarian, and Stéphane Blain

Viewed

Total article views: 624 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
521	81	22	624	14	34

HTML: 521
PDF: 81
XML: 22
Total: 624
BibTeX: 14
EndNote: 34

Views and downloads (calculated since 04 Jun 2025)

Month	HTML	PDF	XML	Total
Jun 2025	164	42	11	217
Jul 2025	68	16	2	86
Aug 2025	144	19	7	170
Sep 2025	145	4	2	151

Cumulative views and downloads (calculated since 04 Jun 2025)

Month	HTML	PDF	XML	Total
Jun 2025	164	42	11	217
Jul 2025	68	16	2	86
Aug 2025	144	19	7	170
Sep 2025	145	4	2	151

Viewed (geographical distribution)

Total article views: 624 (including HTML, PDF, and XML) Thereof 624 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 14 Sep 2025

Short summary

The Kerguelen region hosts a phytoplankton bloom influenced by several iron sources. In particular, glaciers supply iron to the coastal waters. However, the importance of glacial iron for the bloom is not known. Here we calculate iron transport pathways from glaciers to the open ocean using in situ and satellite data, showing that one third of the offshore bloom is reached by glacial iron. These results are important in the context of the melting of the Kerguelen ice cap under climate change.


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