Dear Authors,

This is a novel and insightful study that contributes to our understanding of the biogeochemical cycle of iron, particularly in hydrothermal regions. The manuscript will be supported for publication once the following issues are addressed. I would like to thank the authors this effort to have all the amount of analysis for the paper from that cruise.

We thank the reviewers for their helpful comments.

We have responded to each comment in italics below.

Minor Comments:

- Please correct the term "physiochemical" to "physico-chemical" throughout the manuscript.

Corrected, thank you.

- Lines 97–102: This information may be omitted from the introduction as it is repeated in the experimental section.

Omitted.

- The introduction should include an overview of existing incubation studies found in the literature.

Thank you for this suggestion. To our knowledge the only other study like this is that of Lough et al., 2017. We have now highlighted this in the introduction.

- Lines 196–198: This sentence is redundant with previous content and could be removed.

Removed.

Technical Issues:

- The authors should justify the use of two different artificial ligands for LFe determination. Since both reverse and forward titration methods can use the same ligand, this would allow for direct comparison of conditional stability constants, which are ligand-dependent.

Although both ligands could indeed be used for forward and reverse titrations, we chose to use what we thought would be the optimal approach for each. SA has been vetted in intercalibration exercises and widely applied to forward titrations of Fe speciation in seawater. Forward titration results with SA are also incorporated in the GEOTRACES Data Products, allowing synthesis and direct comparison of our experimental work with water column results also achieved through forward titrations. Similarly, NN has been well characterized for reverse titrations of Fe speciation and applied previously to reverse titrations of hydrothermal plume samples, allowing our experimental results here to be directly compared with previously conducted field work with this method. 

Cheize et al. (2012) demonstrated that NN is suitable for low-salinity solutions. However, in this incubation study, samples could be diluted using UV-irradiated seawater or NaCl to maintain constant ionic strength. Altering ionic strength significantly influences physico-chemical processes and iron speciation.

Cheize, M., et al. (2012). Iron organic speciation determination in rainwater using cathodic stripping voltammetry. Analytica Chimica Acta, 736, 45–54.

We agree that clean UVSW could be an ideal diluent for the samples. Here we applied the NN method following the recommendation of Hawkes et. at. 2013 for hydrothermal samples and used ultrapure water for the sample dilutions. The Hawkes 2013 study did test different dilution factors (20x, 10x, and 5x), and found the salinity difference to have no effect on the KFeL values, but lower salinities did cause a slight overestimate of L concentrations. We used a dilution factor of 5x so have added this potential caveat to the method section (line 232).  

Hawkes, J. A., Connelly, D. P., Gledhill, M., & Achterberg, E. P. (2013). The stabilisation and transportation of dissolved iron from high temperature hydrothermal vent systems. Earth and Planetary Science Letters, 375, 280-290.

- Please justify the selected detection window with SA and explain why an alternative was not used that might allow the identification of weaker ligands.

We were especially interested in the role of strong ligands on mediating physico-chemical exchange of Fe in hydrothermal plumes. This was one of the reasons we chose to take speciation samples from the soluble fraction and why we employed a 25 µM SA analytical window. As noted above, we also chose this method with SA to allow synthesis with GEOTRACES field studies.

- Why is the equilibration time with SA limited to only one hour? Justification is needed, as most studies. including those by van den Berg, typically use overnight equilibration.

The equilibration with SA is quite fast when added after the Fe additions in the titration (Rue and Bruland 1995), so although overnight equilibration times can be applied, they are not necessary. Both shorter and overnight SA equilibrations provide the same conditional stability constants for the SA determined from competition with EDTA (Rue and Bruland 1995; Buck et al. 2007; Abualhaija and van den Berg 2014), and a recent assessment of the two equilibration times showed excellent agreement in L and K results for samples (Mahieu et al. 2024). We also have found that the shorter equilibration time with SA provides more accurate results in studies of known ligands (e.g. desferrioxamine B) as model ligands added to seawater (Rue and Bruland 1995; Buck et al. 2010; Mahieu et al. 2024) than has been reported with overnight equilibration (e.g. Gerringa et al. 2020) – likely because overnight equilibration times can exacerbate any issues with vial conditioning (Mahieu et al. 2024). 

- What is the concentration of Fe(II) in the samples? This is important because the voltammetric method used detects only Fe(III). If Fe(II) is present in significant and stable amounts, LFe values could be overestimated.

There was some Fe(II) present at the start of all the reducing plume incubations. Initial Fe(II) concentrations were ~0.8 nM at Lucky Strike, ~70 nM at Rainbow near-field, and ~10 nM at TAG. We will add these details to the results section. However, by the time the samples were frozen, thawed, equilibrated, and analyzed back at the University of South Florida months later, we expect that any Fe(II) would be completely oxidized to Fe(III). 

González-Santana, D., Lough, A. J., Planquette, H., Sarthou, G., Tagliabue, A., & Lohan, M. C. (2023). The unaccounted dissolved iron (II) sink: Insights from dFe (II) concentrations in the deep Atlantic Ocean. Science of the Total Environment, 862, 161179.

- In the forward titrations, are the samples diluted? If so, was salinity corrected accordingly?

Dilution was only used on samples where dissolved Fe concentrations were exceedingly high (>100 nM), and thus they were only employed in a subset of the reverse titrations. No dilutions were conducted for any forward titrations.  We will clarify this in the Methods sections (lines 233). We will also indicated the diluted samples in the data tables in supplementary information. 

- How did the pH evolve during the incubations? pH plays a key role in Fe(II) oxidation and associated physico-chemical processes.

Potentiometric  pH measurements were taken along this cruise and some opportunistic measurements were made in a few of these incubations. We didn't observe any large changes in pH, with the largest difference being ~0.1 units from the environmental samples Gonzalez-Santana et al., 2023, cited above). That said, we buffered all titrations to pH 8.2 prior to competition with the competitive ligand so our results are all from the same pH conditions.

- In Figures 2c and 2e, dFe (and consequently dLFe) concentrations increase on Day 1. How is this explained? Were replicate measurements performed?

We found this unusual as well, and did make replicate measurements of this sample. The total Fe within this incubation was extremely high (>5,000 nM), with an increase in soluble Fe as well as H2S concentrations (Supplemental Figure 7a) just before day 1 (sampling was limited for Fe(II) and H2S as environmental samples were prioritized for these time sensitive measurements). Thus, we think the day 1 increase could reflect initial dissociation of pyrite particles (perhaps from temperature or pressure changes from in-situ to incubation conditions) originally formed at the vent-ocean interface followed by the rapid formation of colloids (presumably Fe(oxy)-hydroxides) observed only in the 24 hour time point before aggregating into the particulate phase. Sulphides are also known ligands for Fe(II) but we do not expect reduced species (Fe(II) or sulphides) to be preserved in the sample to analysis and their binding strengths are much weaker than ligands measured in these samples. Its possible organic ligands co-precipitated with the reduced Fe particles were also liberated at this time, leading to the increase in ligands, and were subsequently scavenged with the oxidized particles. This interpretation is speculative, so was not included in the current manuscript.

- The manuscript refers to differences in logK values; however, these do not appear to be significant in the figures. Please provide a statistical analysis to determine whether the variations are significant.

Thank you for this suggestion. We performed a t-test/one-way anova to test the significance of differences highlighted in the text and to add context to these statements. These statistics have now been added to the manuscript. For some experiments there was a significant difference between soluble and dissolve phase ligands but in others there were no statistical significance. We have added statements on in the results sections.

- In Figure 5b, how do the authors explain that sFe concentrations exceed cFe concentrations?

This likely reflects analytical uncertainty in the measurements, with this sample having a dFe concentration of 6.78 ± 0.52 and a sFe concentration of 7.95 ± 0.53. We interpret these results to reflect 100% of the dissolved Fe in the soluble phase, and the separation driven either by a slight underestimate of dFe or overestimation of sFe.

I hope this constructive revision will help the authors to improve the manuscript.

Thank you!

Mellett et al studied the temporal evolution of Fe speciation and fractionation at Atlantic HT plumes using incubation experiments. The study very neatly shows physicochemical evolution of Fe and ligands - their initial introduction in the near systems, quick floccolation / scavenging and particularly the downrange introduction of complexed Fe.

The manuscript is a pleasant read, with little to remark on layout and typography.

We thank the reviewer for their supportive feedback and helpful comments on the manuscript. Our responses are included below. 

This reviewers main concern is where part of the discussion posits strong ligands, particularly siderophores, having a role in the down-plume stabilisation of Fe. This part of the discussion is hard to reconcile with one of the main findings of the incubation experiments being an overall weaker ligand pool. These sections (towards the ends of both sections 4.1 and 4.2), require better justification. In the opinion of this reviewer the identification of families known for putative siderophore pathways in the face of an overall weaker ligand pool needs more convincing arguments for a meaningful role of siderophores in the speciation in HT systems.

Thank you for this comment. We have attempted to describe the presence of two distinct ligand pools, one present initially near the vent itself (the weaker ligand pool) and one that is produced in-situ by the microbial community (the stronger ligand pool). Without siderophore samples from these specific experiments, we are relying on a few pieces of evidence to support the interpretation of siderophores contributing to a portion of dFe stabilization within plume systems:

1. Siderophores were found within plumes from environmental samples taken from the same cruise (Hoffman et al., 2024).

2. Siderophores should fall within the soluble fraction of the ligand pool. and the soluble ligand pool increased at later stages of many of the unfiltered incubations.

3. The fractionation of dFe isotopes in the later stages of unfiltered treatments at TAG and Rainbow near-field. Particle scavenging would be a process observed in both the filtered and unfiltered treatments, particularly towards the later stages of these incubations where particle concentrations were at their highest. To our knowledge there is no experimental data that shows that siderophore cycling and active uptake by microbes causes isotopic enrichment of dFe, but the combination of such mechanisms has been speculated in the surface waters of the Southern Ocean (Ellwood et al., 2020, Sieber et al., 2021). 

4. 16S data showed that the microbial community at the Family level contained putative siderophore pathways, and these microbial groups increased in abundance towards the end of many experiments. This is by no means the strongest evidence, but the best that can be done in the absence of RNA (transcriptomic) data.

We think these lines of evidence support siderophore production over time in the experiments, but we do agree that the caveats could be communicated more clearly and transparently in sections 4.1 and 4.2. For example, the signal observed in these experiments are likely amplified due to the closed system nature of the experiments and the increased incubation temperatures. But we still believe it is an important mechanism to consider in these systems. We added explicit statements that show this logical deduction more clearly and highlight the uncertainties in these observations. 

Specific remarks

physiochemical used throughout the manuscript, authors probably mean to say physicochemical.

Thank you for catching this. This has been changed throughout the manuscript.

figures 2-5: a lot is happening in the first days, suggest making the horizontal axes nonlinear to better show the quick successions taking place initially.

Thank you for this suggestion. We did try plotting these figures using log-scale x-axis. The transformation was not useful on longer incubations (Figures 4 and 5) and ended up compressing these datasets, but did provide some benefits to spreading out the condensed early sampling in the short-term incubations (Figure 2 and Figure 3). However, in Figure 3 this visualization ended up compressing data on the other end of sampling time. We did notice that the log plots appeared very similar to physiochemical breakdown the figures  (a, b) due to the even spacing of the bar plots. Thus, we ultimately decided to maintain consistency amongst all plots for simplicity, and hope that the physicochemical components in the first panel captures the quick succession of Fe in early sampling experiments. We have changed the scale of the y-axes in Figure 2a to more clearly illustrate the changes observed. Thank you for the thoughtful suggestion.

587-589 Please provide additional justification for the link between Fe-limited Southern Ocean surface water processes to the present findings.

This sentence was attempting to draw comparisons between the dFe isotope data in this study to others that point to signatures of active biological cycling. We added additional context and caveats to this statement in the manuscript to highlight the uncertainty. 

590-592 Please provide additional justification of Hoffman et al. (2024) finding strong ligands in HT systems relating to the present findings.

Hoffman et al. (2024) found siderophores in water column samples from the same cruise, at the same locations as the experiments that were done in this study. While we were not able to measure siderophores in these experiments due to volume constraints, we think there is strong evidence that the increase in soluble and dissolved organic ligands towards the end of these experiments could be due to the production of siderophores along with the other pieces of evidence outlined above.