Unprecedented Fe delivery from the Congo River margin to the South Atlantic Gyre

Rivers are a major supplier of particulate and dissolved material to the ocean, but their role as sources of bio-essential dissolved iron (dFe) is thought to be limited due to rapid, efficient Fe removal during estuarine mixing. Here, we use trace element and radium isotope data to show that the influence of the Congo River margin on surface Fe concentrations is evident over 1000 km from the Congo outflow. Due to an unusual combination of high Fe input into the Congo-shelf-zone and rapid lateral transport, the Congo plume constitutes an exceptionally large offshore dFe flux of 6.8 ± 2.3 × 108 mol year−1. This corresponds to 40 ± 15% of atmospheric dFe input into the South Atlantic Ocean and makes a higher contribution to offshore Fe availability than any other river globally. The Congo River therefore contributes significantly to relieving Fe limitation of phytoplankton growth across much of the South Atlantic.

The manuscript entitled "Unprecedented Fe delivery by the River Congo to the South Atlantic Gyre" by Vieira et al. synthesizes important results regarding the role of the Congo River in delivering iron, radium, and other trace elements to the Atlantic Ocean. The authors demonstrate that river discharge, river-dominated coastal sediments, or submarine groundwater discharge in a riverdominated area combine to ensure that globally, the Congo is the most significant riverine source of iron to the oceans. The topic of the paper is of high interest to the majority of chemical oceanographers engaged in modern research, as the debate surrounding the sources of iron to the ocean is one of the most currently contested. The authors employ state of the art techniques and the work is compliant with GEOTRACES -going a long way in ensuring that analyses and quality control was likely meticulous and sufficient. I do have some questions regarding the data interpretation in several areas of the manuscript, but am excited with the larger implications of the paper, which demonstrate the river is important with respect to synoptic scale ocean processes (alleviation of Fe-limitation in the Western Atlantic), but also mechanistically speaking (that an organic/Fe rich river system may provide a disproportionately large contribution of Fe to the oceans). As far as relative importance in the field, rivers have been discounted relative to atmospheric dust, and more recently sediments and hydrothermal vents. Much of this stems from early assumptions that iron mostly flocculates and is therefore not worth considering on global scales in models, as well as the fact that the "world average river" from which estimates are primarily derived do not consider smaller rivers that may drain more organic/Fe-rich coastal plain sediments. For this reason, the number of studies focusing on river sources of Fe have been relatively limited in number in the last couple of decades (in my opinion). It is nice to see a GEOTRACES data set that focuses specifically on riverine discharge, which to my knowledge have also been limited beyond at least the Amazon (nothing yet for the Mississippi River for example). Sure enough, the results illustrate the river is important -with observed concentrations far exceeding expectations based on river flow alone -with the overall conclusions that rivers cannot be ignored in global ocean iron budgets.
I provide comments below, with moderate (albeit thought-provoking comments) comments/concerns/areas for improvement addressed by line number. I recommend publishing after consideration of the these points.
1. The title, "Unprecedented Fe delivery by the River Congo to the South Atlantic Gyre" is a bit misleading in that while the river is ultimately responsible for the Fe delivery to the gyre, it is likely first delivered to the sediments first, so there is this physical disconnect. I don't have the perfect title either, but something that conveys the idea: "…by the River Congo-dominated ocean margin to the…" Overall, I think the authors should approach this as a point of pride -they are demonstrating that river plumes can influence larger scale oceanographic processes if not directly, then through a continental margin "intermediate", that receives TE and is then culprit in supporting fluxes to the water column.
2. Line 17: I would change to "The Congo dFe outflow therefore contributes significantly to relieving Fe-limitation…", because there is no data in the manuscript that demonstrates the river is the sole or primary source of Fe to the South Atlantic Gyre, as it currently (could) read. 3. Line 29: Is this 3% figure, for the riverine delivery of new ocean Fe, obtained before or after accounting for flocculation in estuaries? Ambiguous language in the literature on this topic have led to a lot of inconsistencies in the community, so I recommend clarifying for the reader. 4. Line 33: "Unique" is an adjective describing the equatorial plume itself, but the conditions in the first half of the same sentence refer to conditions about the river and the shelf. How is the plume itself unique? I think this is too vague. 5. Lines 40-42: Please add Radium references here. My own question is whether or not we would expect to see additional Radium released from continental margin sediments hosting mineral dissolution processes, e.g. microbial iron reduction? Later on in the paper (line 140) you seem to suggest this is possible. So, I am concluding yes, but would appreciate some more details as it is critical to interpreting the work. 6. Line 45: Please clarify this figure and mark very clearly, on the diagram itself perhaps, the 3 groups of samples that are discussed: the "Congo Shelf Zone", the "Coastal Transect", and the "Offshore Plume". While I think all the info can be gathered from the text as it stands, it did take me a couple of reads to really nail these down. 7. Line 48: Add "offshore" before "3 degrees S", and perhaps one line that says this was the seasonal high-flow period. 8. Line 67-78: I would move a version of the sentence "low removal primarly reflects…" to the beginning of this pagagraph (e.g. before "Extrapolating" in line 57), because otherwise it appears as if you are trying to state as the primary point of this section that only 50% of the iron is removed. It sets the reader up for confusion. 9. Line 59: Saying that you are extrapolating to a "zero salinity endmember" is also confusing, because this would generate a that technically occurs prior to any estuarine flocculation processes (that happen at salinities > 0). Instead, extrapolation of the data points in Figure 2A actually reveals the estuarine endmember that should have already undergone mixing and hosted flocculation processes. I would support language, "extrapolating to an estuarine endmember" 10. Line 74: Are these ratios specific to the Congo? Please state if so. Otherwise, I don't think this is really a valid exercise given how variable river concentrations can be. 11. Line 75: Add "relative to seawater background" before "in the Congo"? 12. Line 78 and others: I went to the cited reference because I was questioning the logic here, and I don't think the cited paper supports the argument in this manuscript: "The Congo River plume may be classified as surface-advected (Yankovsky & Chapman, 1997); buoyant river inflow remains on top of shelf water forming a thin layer that, decoupled from bottom stress, spreads many hundreds of kilometres offshore." Now, just because turbulent benthic interactions are minimal offshelf, this doesn't mean upwelling cannot be significant off-shelf. In fact, the cited paper proceeds to state in the next paragraph: "An estuarine type circulation has been identified here, where surface outflow is compensated at depth by a residual up-canyon transport of saline water (Denamiel, Budgell, & Toumi, 2013;Eisma & Van Bennekom, 1978)." I would remove or rephrase line 78 -I don't think it supports the argument. Here is where I hope to ignite some thought: the Congo River is the only large river in the world still connected to its deep sea fan via a turbidite canyon, and because of these turbidity flows, the Congolobe expedition observed incredibly high rates of accumulation of organic/iron minerals (>2 cm/year) and high rates of iron solubilization from shelf-break down to deep-sea fan sediments (consistent with those in Fe-rich estuaries) up to 800 km offshore (Beckler et al. 2016;Taillefert et al. 2017;Rabouille et al. 2017 and other references therein). Under conditions of intense irondiagenesis in sediments, a corresponding flux of Fe and TE typically also occurs, so I suspect that these processes, i.e. upwelled slope/deep-sea waters previously in contact with these sediments, may also be contributing to plume enrichments (in addition to shelf sediments as stated in line 140). The actual offshore transect in this manuscript processed to the North of the Congo River canyon. I wonder if during other times of year when the plume may be situated more southward, if we may expect even a higher contribution of sediment-derived Fe within the plume, and overall contributing to the gyre? 13. Some comments/alternative explanations relating to differences in Fe content, but similar salinities, for the CSZ and off-shelf sample sets ( Figure 2): a. Could the northern freshwater plumes be from a different river source and thus contain a different amount of Fe altogether? b. Are the northern and southern sample plume water masses both from the Congo but from different time periods, as weekly variations in flow may result in various concentrations of iron in the river water? 14. Line 99: I would remove this comparison, because the utility of the "zero salinity" end-member is virtually meaningless now that you have established that the dominant factor is indeed shelf input. This really had me confused. 15. Line 140: I suspect based on other thoughts throughout the paper that you believe here that explanation (ii) is the most logical. I do feel the atypical sedimentary environment of the Congo margin can explain this phenomenon.

-Jordon Beckler
Reviewer #3 (Remarks to the Author): Review "Unprecedented Fe delivery by the River Congo to the South Atlantic Gyre" by L.H. Vieira This study estimates fluxes of dissolved trace elements (Fe, Co, and Mn) based on 228Ra data from the Congo shelf. The results of this study will be useful for the scientific community; however, I think that the paper needs to be significantly improved before considering it for publication. I am also wondering if Nature Communication Journal is the right place for this study, because the method applied is not new, the authors presented a flux of trace elements from the Congo River that is new information but not original, and the short length of this Journal seemed to have prevented the authors to completely discuss the results. I have three main points that I think the authors should address: 1) I noted some odd points in the method that I describe in details below. One point that needs to be addressed on the method is that the authors did not correct the 228Ra average activity in the Congo-shelf-zone from offshore 2228Ra activities (if they did, it is not clearly written). This would affect the 228Ra flux, which is then used for the TE fluxes. Therefore, the fluxes of the shelf would likely need to be re-calculated.
2) More comparisons are needed regarding the fluxes of Ra228, Fe, Co, and Mn. An estimation of dust deposition should be included, especially after the recent study published by Mendez Barraqueta et al. 2019 based on data from the same cruise showing significant atmospheric deposition. The high dFe concentration on the Congo shelf might well be the results of the combination of high dust deposition, shelf sediment inputs, and large Congo River, rather than attributing the Fe in sweater solely to the Congo River inputs. Comparison of these fluxes with other study areas should also be included to answer the question: Is the Congo shelf a unique supplier of TE to the ocean? 3) In my opinion, the title, abstract, and intro needs to be re-written a little. I personally don't like the catchy Nature forced title "unprecedented Fe delivery". An alternative title could be in the lines of "the Congo shelf represents a unique (or major) supplier of TE to the South Atl..." I am confused by the title/Abstract/intro that focus on the Congo River. The TE fluxes are estimated from Ra228, which is a sediment tracer (except if the Congo river has exceptionally high 228Ra activity, which hasn't been demonstrated in this study). The fluxes are estimated using Ra for the "Congo-shelf-zone" and on an "off-shelf transect", which include inputs from both the river and from sediment. The salinity range is pretty high, mainly above 32, which suggest less than 20% of freshwater contribution in the study zone. The author estimates the Congo river flux (discharge times concentration in the river), which is not necessarily comparable to the Ra flux because the latter is estimated for a specific area and not all the river freshwater discharge actually makes it into this restricted zone. I think there are a lot of discussion that has been cut probably because of the very short format of the journal, but I feel that it would improve the paper to provide more details regarding what these fluxes actually represent. Methods: -The second count for Ra224 was performed 6 weeks later whereas the second count for Th228 is usually done 2 weeks after collection. Was it really done after 6 weeks? Then some Bateman correction should probably be applied.
-After measurements of short-lived on the RaDeCc, the authors said that the fibers were ashed and then leached. Did the author leached the ashes? I am aware of two methods: ash (Charette et al. 2001) or leach/co-precipitate (Moore method) the fibers but I have never seen both processes applied one after the one. Is it a typo by the authors or was it the real procedure? -The Congo-self-zone flux of 228Ra was estimated as inventory/residence time. The authors used the excess 228Ra in the flux calculation. The excess Ra228 should be the average 228Ra concentration in the study area corrected from the offshore 228Ra concentration. This is to account for mixing with offshore waters that have a background 228Ra. If the authors had used 224Ra, there would be no need to consider the excess 224Ra because offshore waters have negligible amount of 224Ra. The authors should use an excess 228Ra of 14.5 -4 = 10.5 based on figure 2, the 228Ra is about 4 dpm/100L outside of the congo shelf zone. Or 8.6 dpm/100L considering the average offshore activities from the offshelf transect on Other comments: -What is the input of atmospheric deposition of Fe to the study area? What is the contribution of dust the high Fe concentration in the Congo River plume? Dust is a significant source of Fe to the ocean (Jickells et al., 2005), therefore, the contribution of dust-derived Fe in the surface concentration cannot be ignored. Especially considering a recent publication from Menzel Barraqueta et al. (2019) showing that the study area receives one of the highest atmospheric deposition rate to the Atlantic Ocean calculated from Al data from the same cruise GA08 (~8-10 g/m2/y). A rough estimate of the flux of Fe from dust can be calculated using the average atm. deposition, the concentration of Fe if measured on aerosol during the cruise or the composition of Fe in crust, and the solubility factor. Similar calculations can be done for Mn, and Co shown in the suppl. info. A more precise location can also be done since the deposition rate is estimated for the same stations. The author could thus perform a mass balance and potentially correct the Fe concentration that is due to dust.
-The comparison of the Fe:Mn ratio in the river to the Fe:Mn ratio in the shelf-zone sounds a bit off to me, but I might be wrong. The lower Fe:Mn ratio in the shelf could simply reflect a different behavior of Fe and Mn. Fe is removed faster than Mn, thus the faster removal of Fe would result in a lower Fe:Mn ratio in the shelf compare to the Fe:Mn ratio in river without involving additional source as suggested by the authors.
-Can the authors discuss the three hypothesis that could explain the high Ra flux from the Congoshelf-zone. 1) exceptionally high Ra in the Congo, 2) high Ra diffusion from sediment, 3) SGD. Which one is the most likely? Why Ra would be much higher in the Congo River compare to other riverine system. -Another approach to estimate the Congo river input could be used. The authors used the TE concentrations measured in the river and the river discharge. An estimation of the amount of freshwater actually entering the Congo-shelf zone can be done using the average salinity and applying conservative mixing of salinity to calculate the volume of freshwater instead of using the river discharge. Doing so, the source of TE from the freshwater Congo river will be directly comparable to the sediment derived estimate of Ra (since the mol/y unit implies a specific surface considered).
-The authors ruled out the benthic supplies of Ra and TE because of the shallow Congo river plume. But we don't really know what is the water column depth along the transect. Since the plume is moving along shore, it could be hugging the coast in shallow water. Thus, it would be useful to know the water column depth compare to the 15 m of river plume. Similarly, the odv plot figure 4 in suppl. info has no bathymetry, how steep is the shelf? Maybe adding the bathymetry to the Figure 1 would already give some information.
-Can the author discuss the difference of Fe flux between the Congo shelf zone and the off shelf transect? Does this provide information on the scavenging rate or biological uptake of Fe in between the two zones?
-Can the author discuss the exceptionally high dFe concentration 600 km from the Congo river mouth? Is the Fe 600 km away from the mouth river from the river or does it has another origin? That is just a side comment that I am curious about. Comment on figures: - Figure 1: I would add a large scale map of Africa for example in a subplot to better situate the study area, to be more visual. Add the units for latitude and longitude. Could the authors show the salinity underway data to better visualize the actual plume during the survey? Should be better than satellite image, especially because the authors acknowledge a difference in salinity between satellite and in situ data. Add the bathymetry to see the width and steepness of the shelf. -In general, I think that surface plots would be useful and could be added as subplots to the figure 2 for example since the cruise track is composed by an along shore transect and an offshore transect.
- Figure 2: switch planels c and d to show Fe in a) and c), and Ra in b) and d) - Figure 3: the Congo data do not stick out, I would exchange the symbol between Congo and Mississippi to highlight the data from this study in red. Based on this graph, the 228Ra are consistent with other estuaries. Could this help in the discussion of the 3 hypotheses about the high flux of 228Ra I mentioned earlier?
We would like to thank the reviewers for their constructive comments and suggestions. We followed the majority of the reviewers' suggestions as described below. Our responses to the reviewers' comments are written in italic, and transcriptions from the revised manuscript are in bold with initial lines in brackets.

Reviewer #1
NCOMMS-19-18100 review This paper describes a set of measurements of salinity, dissolved Fe and Ra isotopes to quantify the input of dFe from the River Congo to the SE Atlantic Ocean. The data seem to be quite accurate, as demonstrated by their analysis of the GEOTRACES consensus reference seawater samples (and other CRMs). The text is well written and their logic towards calculating the dFe flux is easy to follow. However, there is nothing in the paper about seasonality (in all aspects: river flow rate, river concentrations, coastal mixing, offshore transport, etc.). These factors all affect the fluxes that are calculated, and they are all based on a single set of samples collected over a short period of time. Including additional variance due to seasonality would enlarge the relative standard deviations of the fluxes. This should be addressed.
Thank you for the suggestion. A discussion about seasonal changes has been added to the revised manuscript as follows:

Congo, but seasonal variations of the river flow (mean annual range between 35 and 60 x 10 3 m 3 s -1 39 with a ca. twofold seasonal variation in dFe concentrations (Supplementary Table 1)) can strongly affect the Congo River plume dispersion 40 , and potentially the delivery of riverderived materials to the SE Atlantic Ocean. The seasonal variation in benthic supply to overlying plume waters of 228 Ra and TEs on the Congo shelf is unconstrained." References in the revised manuscript.
I have only a few specific comments, keyed to line number: 127-128: This last sentence is circular logic. The previous sentence assumes conservative mixing to yield an effective riverine end-member for Ra-228, so it could simply be stated that this assumes no additional inputs of Ra-228.
Confusing wording has been revised. See the changes in the revised manuscript:

[Line 115] "Conservative mixing between the Congo-shelf-endmember and offshore waters (Fig. 2b) indicates a riverine 228 Ra effective-zero-salinity-endmember concentration of 85 ± 4 dpm 100 L -1 , including both dissolved and desorbed Ra. Together with the river discharge (1.3 x 10 12 m 3 yr -1 ) 31 , this suggests a fluvial 228 Ra flux of 4.8 ± 0.4 x 10 21 atoms yr -1 , which is similar to the 228 Ra flux estimated for the Congo-shelf-zone (3.4 ± 0.9 x 10 21 atoms yr -1 ). If the assumption of conservative mixing behavior is correct, the effective-zero-salinity-endmember would be similar to actual river Ra concentrations, which are not known."
130-138: This text is unnecessary. It is sufficient to say what has been reported for other rivers (around 20 dpm/100L), and that the apparent riverine end-member for Ra-228 (85 dpm/100L on line 124, but 79 dpm/100L on line 140!!) is much higher than that. Then list the possible reasons why, starting on line 138. Can you decide which reason is most likely? Why do you state two different end-member concentrations? There is no value in calculating atom fluxes that would be expected if the river were closer to 20 dpm/100L.

The activity of 85 dpm/100 L takes into account the desorbed and dissolved Ra fractions, while
79 dpm/100 L corresponds to only the dissolved Ra, estimated by the difference between the total activity (85 dpm/100 L) and the desorbed Ra activity (6 dpm/100 L). By doing this, we can compare the potential dissolved Ra activity in the Congo River with other systems (~20 dpm/100 L).

The most likely explanation cannot be number (i), because rivers do not show large variation in
Ra activity, and the Ra concentration in the Congo would have to be higher by factor of 4 to support the observed flux. Numbers (ii) and (iii), or a combination of both, may be more likely. 228 Ra diffusion from shelf sediments has a large variability (e.g. Vieira et al., 2019Vieira et al., (doi.org/10.1016Vieira et al., /j.marchem.2018) and the existence of another source of Ra such as submarine groundwater discharge in a large river system such as the Congo is rather possible (e.g., Moore, 1997Moore, , doi.org/10.1016 . We do not add more information about this issue as that would only be speculation. These two mechanisms are challenging to distinguish based on what data we have from an understudied region. See the text in the revised manuscript:

[Line 132]"This suggests that either (i) the dissolved 228 Ra concentration in the Congo River is exceptionally high compared to other large rivers (~79 dpm 100 L -1 , vs <20 dpm 100 L -1 elsewhere); (ii) 228 Ra diffusion from shelf sediments in this region is anomalously high compared to other regions globally; or (iii) there is another source of Ra such as submarine groundwater discharge 34, 35 . Based on observations elsewhere of large variability of 228 Ra diffusion from shelf sediments 36 and SGD input 14 , (ii) and (iii), or a combination of both, are most likely."
In addition, we must do this calculation to show that the Congo fluxes we estimate are unusual.
The flux calculation using the global "typical" river value allows us to show that if the Congo River were like other major river systems, then the benthic fluxes would be unusually large.
143-147: Remind the readers that these TE fluxes are based on the unusually high Ra-228 flux (because of the unusually high riverine end-member concentration) and include some text about how seasonal variations would affect these fluxes.
High Ra fluxes were due to the large inventory, as well as high Fe fluxes are due to high concentrations. A discussion about seasonal variability has been added, see comment above.
164: Explain how aerosol Fe solubility was measured or estimated. Include an error analysis on the statement that the dFe flux is 40% of the atmospheric flux. I get 40+/-20%. If you add the Congo river dFe flux to the atmospheric flux, then the Congo flux is 29+/-13% of the total dFe flux to the entire South Atlantic. That is still significant, but the error propagation should be included.
Atmospheric flux (14.3 x 10 8 mol/yr, with no stated uncertainties) was estimated based on data from Duce and Tindale, 1991(doi.org/10.4319/lo.1991.36.8.1715. The uncertainty of 40% (25% to 55%) is added to the revised manuscript based on the uncertainties of our estimate. No comparison with "total" flux has been made because we do not have a complete budget.
Nonetheless, we make a comparison with sedimentary flux reported elsewhere. See also our answer to reviewer 3 related to atmospheric deposition.

Reviewer #2
The manuscript entitled "Unprecedented Fe delivery by the River Congo to the South Atlantic Gyre" by Vieira et al. synthesizes important results regarding the role of the Congo River in delivering iron, radium, and other trace elements to the Atlantic Ocean. The authors demonstrate that river discharge, river-dominated coastal sediments, or submarine groundwater discharge in a river-dominated area combine to ensure that globally, the Congo is the most significant riverine source of iron to the oceans. The topic of the paper is of high interest to the majority of chemical oceanographers engaged in modern research, as the debate surrounding the sources of iron to the ocean is one of the most currently contested. The authors employ state of the art techniques and the work is compliant with GEOTRACES -going a long way in ensuring that analyses and quality control was likely meticulous and sufficient. I do have some questions regarding the data interpretation in several areas of the manuscript, but am excited with the larger implications of the paper, which demonstrate the river is important with respect to synoptic scale ocean processes (alleviation of Fe-limitation in the Western Atlantic), but also mechanistically speaking (that an organic/Fe rich river system may provide a disproportionately large contribution of Fe to the oceans). As far as relative importance in the field, rivers have been discounted relative to atmospheric dust, and more recently sediments and hydrothermal vents.
Much of this stems from early assumptions that iron mostly flocculates and is therefore not worth considering on global scales in models, as well as the fact that the "world average river" from which estimates are primarily derived do not consider smaller rivers that may drain more organic/Fe-rich coastal plain sediments. For this reason, the number of studies focusing on river sources of Fe have been relatively limited in number in the last couple of decades (in my opinion). It is nice to see a GEOTRACES data set that focuses specifically on riverine discharge, which to my knowledge have also been limited beyond at least the Amazon (nothing yet for the Mississippi River for example). Sure enough, the results illustrate the river is important -with observed concentrations far exceeding expectations based on river flow alone -with the overall conclusions that rivers cannot be ignored in global ocean iron budgets.
I provide comments below, with moderate (albeit thought-provoking comments) comments/concerns/areas for improvement addressed by line number. I recommend publishing after consideration of these points.
1. The title, "Unprecedented Fe delivery by the River Congo to the South Atlantic Gyre" is a bit misleading in that while the river is ultimately responsible for the Fe delivery to the gyre, it is likely first delivered to the sediments first, so there is this physical disconnect. I don't have the perfect title either, but something that conveys the idea: "…by the River Congo-dominated ocean margin to the…" Overall, I think the authors should approach this as a point of pride -they are demonstrating that river plumes can influence larger scale oceanographic processes if not directly, then through a continental margin "intermediate", that receives TE and is then culprit in supporting fluxes to the water column.
Thank you, we accepted the suggestion and the title has been changed. It now reads: "Unprecedented Fe delivery from the River Congo margin to the South Atlantic Gyre." 2. Line 17: I would change to "The Congo dFe outflow therefore contributes significantly to relieving Fe-limitation…", because there is no data in the manuscript that demonstrates the river is the sole or primary source of Fe to the South Atlantic Gyre, as it currently (could) read.

Changed as suggested:
[Line 16] "The River Congo therefore contributes significantly to relieving Fe-limitation of phytoplankton growth across much of the South Atlantic." 3. Line 29: Is this 3% figure, for the riverine delivery of new ocean Fe, obtained before or after accounting for flocculation in estuaries? Ambiguous language in the literature on this topic have led to a lot of inconsistencies in the community, so I recommend clarifying for the reader.
This 3% from Raiswell refers to after estuarine removal. Raiswell guesses something around 90% estuarine removal, although this actually varies widely from 60-99% (even within the Boyle 1977 reference usually cited as demonstrating 90% removal there is pronounced variation between rivers at different times) .

[Line 28] "Whilst riverine Fe concentrations are 3-5 orders of magnitude greater than those in seawater 8 , rivers provide only ~3% (after estuarine removal) of the new Fe delivered annually
to the oceans 9 ." 4. Line 33: "Unique" is an adjective describing the equatorial plume itself, but the conditions in the first half of the same sentence refer to conditions about the river and the shelf. How is the plume itself unique? I think this is too vague.
It is unique in the sense that it is the only major river system with a plume entering an into an eastern boundary upwelling region. The text has been changed for clarification in the revised manuscript:

[Line 33] "The Congo is the second largest river on Earth by discharge volume 11 , and is the only major river to discharge into an eastern boundary ocean region with a narrow shelf 12unique characteristics for a near-equatorial river plume subject to low Coriolis forces 11 ."
5. Lines 40-42: Please add Radium references here. My own question is whether or not we would expect to see additional Radium released from continental margin sediments hosting mineral dissolution processes, e.g. microbial iron reduction? Later on in the paper (line 140) you seem to suggest this is possible. So, I am concluding yes, but would appreciate some more details as it is critical to interpreting the work.
In line 140, we mean that the diffusion flux can vary depending on sediment type (i.e., Th concentrations and nature of the sediment) As per an earlier comment, high 228Ra diffusion from shelf sediments in this region and/ or an addition source of Ra such as submarine groundwater discharge are the most likely sources. See answer to reviewer 1 and 3.

[Line 132] " This suggests that either (i) the dissolved 228 Ra concentration in the Congo River is exceptionally high compared to other large rivers (~79 dpm 100 L -1 , vs <20 dpm 100 L -1 elsewhere); (ii) 228 Ra diffusion from shelf sediments in this region is anomalously high compared to other regions globally; or (iii) there is another source of Ra such as submarine groundwater discharge 34, 35 . Based on observations elsewhere of large variability of 228 Ra diffusion from shelf sediments 36 and SGD input 14 , (ii) and (iii), or a combination of both, are most likely."
6. Line 45: Please clarify this figure and mark very clearly, on the diagram itself perhaps, the 3 groups of samples that are discussed: the "Congo Shelf Zone", the "Coastal Transect", and the "Offshore Plume". While I think all the info can be gathered from the text as it stands, it did take me a couple of reads to really nail these down.
Thank you for the suggestion. Figure 1 has been changed, and these points are now addressed.
See also the legend of the figure. 7. Line 48: Add "offshore" before "3 degrees S", and perhaps one line that says this was the seasonal high-flow period.

Line 48 was the legend of the figure 1 in the first version of the manuscript. Figure 1 and its legend has been changed. A line about the seasonal high-flow period has been added, as
suggested.

[Line 145] "Sample collection occurred during the high discharge season of the Congo (…)."
See also response to reviewer 1 about seasonal variability.
8. Line 67-78: I would move a version of the sentence "low removal primarily reflects…" to the beginning of this paragraph (e.g. before "Extrapolating" in line 57), because otherwise it appears as if you are trying to state as the primary point of this section that only 50% of the iron is removed. It sets the reader up for confusion.
The text has been changed for clarification.

3,150 nM, similar to limited previous measurements (~9,000 nM) 2 . Extrapolating the linear regression line of dFe vs. salinity in the Congo-shelf-zone (Fig. 2a) to zero salinity provides an effective-zero-salinity-endmember concentration of 3,910± 610 nM (R 2 = 0.76), indicating that only ~50% of dFe is removed during estuarine mixing processes. This is consistent with prior work 2 , but notably limited compared to other river systems where 90-99% is typically stripped
from the water column 2, 6 ." 9. Line 59: Saying that you are extrapolating to a "zero salinity endmember" is also confusing, because this would generate a that technically occurs prior to any estuarine flocculation processes (that happen at salinities > 0). Instead, extrapolation of the data points in Figure 2A actually reveals the estuarine endmember that should have already undergone mixing and hosted flocculation processes. I would support language, "extrapolating to an estuarine endmember" Indeed, we used this approach exactly to account for all the estuarine processes causing nonconservative behavior at salinity > 0 (as discussed in the classical application/derivation of the approach: Officer 1979, Boyle et al 1974, Liss 1976, Hydes and Liss 1977. The 'estuarine' endmember, for example, would be very different because of those processes. Therefore, we prefer to retain our original wording.
10. Line 74: Are these ratios specific to the Congo? Please state if so. Otherwise, I don't think this is really a valid exercise given how variable river concentrations can be.
Those ratios were calculated from our Congo data. The text has been revised for clarification. Fig. 1)

. An additional TE source in the Congo-shelf-zone is also evident in the lower Fe:Mn (6.3 ± 6.0) and Fe:Co (525 ± 490) ratios compared to the Congo River (Fe:Mn = 71.2 ± 37.5; Fe:Co = 4.29 ± 2.34). These ratios reflect how dFe is removed relative to the other elements, which is unclear from fluxes alone. These ratios suggest that River Congo dFe is removed by a factor of 10, whereas the fluxes (Table 1, discussion below) indicate that the net removal is only a factor of 2. In summary, a multi-element approach also corroborates significant dTE inputs into the Congoshelf-zone other than River Congo water."
11. Line 75: Add "relative to seawater background" before "in the Congo"?
The word "elevated" in the text refers to the surrounding water. We think that "background" is a little ambiguous and difficult to specifically define, so we prefer to keep the original wording.
12. Line 78 and others: I went to the cited reference because I was questioning the logic here, and I don't think the cited paper supports the argument in this manuscript: "The Congo River plume may be classified as surface-advected (Yankovsky & Chapman, 1997); buoyant river inflow remains on top of shelf water forming a thin layer that, decoupled from bottom stress, spreads many hundreds of kilometres offshore." Now, just because turbulent benthic interactions are minimal offshelf, this doesn't mean upwelling cannot be significant off-shelf. In fact, the cited paper proceeds to state in the next paragraph: "An estuarine type circulation has been identified here, where surface outflow is compensated at depth by a residual up-canyon transport of saline water (Denamiel, Budgell, & Toumi, 2013;Eisma & Van Bennekom, 1978)." I would remove or rephrase line 78 -I don't think it supports the argument.

We argued that benthic input supplies Ra and TEs between the river mouth and the Congo-shelfzone, but conservative Ra mixing behavior indicates no additional inputs beyond the Congoshelf-zone. Nevertheless, the line has been removed as suggested.
Here is where I hope to ignite some thought: the Congo River is the only large river in the world still connected to its deep sea fan via a turbidite canyon, and because of these turbidity flows, the This process could indeed be important, but it is not the dominant feature that explains our observations, because this upwelling source would likely bring high Fe, but probably not high Ra, or Co, etc. The key point here is that we see elevated levels of Fe, Ra, and other trace elements (e.g. Mn and Co) and they all co-vary, indicating a common source. We appreciate the reviewer's thoughtful input. Additional text has been added to the revised manuscript to reflect these points. Fig. 2b; Supplementary Fig. 1) suggest they have a common source, likely shelf-sediments 3, 19,

2017) which may also contribute to the relatively high dFe concentrations in the Congo-shelfzone."
13. Some comments/alternative explanations relating to differences in Fe content, but similar salinities, for the CSZ and off-shelf sample sets ( Figure 2): a. Could the northern freshwater plumes be from a different river source and thus contain a different amount of Fe altogether?

The Congo is the dominant river in the region (approximately 4 times larger than the combined discharge of the small rivers to the north; Milliman and Farnsworth 2011), so we consider that the contribution of other rivers to a low salinity plume observed in the study region on the
Western African shelf is minimal. We appreciate the comment and a text has been added to address this point.

In addition, as the Congo is the dominant river in the region (approximately 4 times larger than other more northern rivers combined 30 , we consider that the contribution of other rivers to the low salinity plume observed in the study region is minimal."
b. Are the northern and southern sample plume water masses both from the Congo but from different time periods, as weekly variations in flow may result in various concentrations of iron in the river water? Variation in the Fe endmember is possible/likely, but large short-term variation is unlikely due to coherence of the observed mixing patterns (e.g., Mn, Co, Ra) over very large spatial scales.

Ra signals observed, so variation in the riverine Fe endmember would have limited effect on the estimated fluxes.
14. Line 99: I would remove this comparison, because the utility of the "zero salinity" endmember is virtually meaningless now that you have established that the dominant factor is indeed shelf input. This really had me confused.
We use this comparison to demonstrate the effect of estuarine processes leading to nonconservative mixing behavior. The effective zero salinity endmember (EZSE) calculation is necessary to make this comparison (see earlier response to reviewer 1 about the underlying theory).
15. Line 140: I suspect based on other thoughts throughout the paper that you believe here that explanation (ii) is the most logical. I do feel the atypical sedimentary environment of the Congo margin can explain this phenomenon.

We suspect that the most likely explanation is number (ii) or number (iii), or a combination of both. The unusual sedimentary environment is certainly interesting and potentially a
contributing factor to high benthic inputs, but cannot explain the elevated concentrations of all dTEs measured-Please see earlier comment and response to the reviewer 1 above.

Reviewer #3
Review "Unprecedented Fe delivery by the River Congo to the South Atlantic Gyre" by L.H.

Vieira
This study estimates fluxes of dissolved trace elements (Fe, Co, and Mn) based on 228Ra data from the Congo shelf. The results of this study will be useful for the scientific community; however, I think that the paper needs to be significantly improved before considering it for publication. I am also wondering if Nature Communication Journal is the right place for this study, because the method applied is not new, the authors presented a flux of trace elements from the Congo River that is new information but not original, and the short length of this Journal seemed to have prevented the authors to completely discuss the results. I have three main points that I think the authors should address: The method may not be new, but it was not the focus of the paper. The results we present have major implications for the biogeochemical cycles in the Congo-influenced SE Atlantic Ocean.
We think this is an important contribution to our understanding of the role of rivers in marine systems because it is fundamentally different from any other river system investigated globally.
Although word count is limited, there is ample space within the limits to address the reviewer comments.
1) I noted some odd points in the method that I describe in details below. One point that needs to be addressed on the method is that the authors did not correct the 228Ra average activity in the Congo-shelf-zone from offshore 228Ra activities (if they did, it is not clearly written). This would affect the 228Ra flux, which is then used for the TE fluxes. Therefore, the fluxes of the shelf would likely need to be re-calculated. A new paragraph has been added related to the dust deposition into our study region. As shown in the revised manuscript, atmospheric deposition contributes very little to the dFe concentration in our study region and, as noted, the flux calculated is likely an over-estimate as it assumes Aluminum in the region of interest arises exclusively from dust, whereas in reality a fraction likely also originates from the river. Also, we showed that Fe correlates with salinity, which indicates that dust could only have a minor effect on Fe concentration compared to riverassociated sources. Furthermore, all elements have a common source, including Ra, which also indicates that this source is probably not dust. The new paragraph reads as: 13-4900 µmol m -2 yr -1 ( a minimum and maximum limit given the contribution of dAl from   non-atmospheric sources in this zone), or 0.01-4.93 x 10 7 mol yr -1 , 1-3

orders of magnitude lower than our estimated dFe fluxes from the Congo-shelf-zone to the same region (Table 1)."
See references in the revised manuscript.
3) In my opinion, the title, abstract, and intro needs to be re-written a little. I personally don't like the catchy Nature forced title "unprecedented Fe delivery". An alternative title could be in the lines of "the Congo shelf represents a unique (or major) supplier of TE to the South Atl..." I am confused by the title/Abstract/intro that focus on the Congo River.
We appreciate the suggestion; other reviewers supported the title with slight changes. It now reads: "Unprecedented Fe delivery from the River Congo margin to the South Atlantic Gyre".
The TE fluxes are estimated from Ra228, which is a sediment tracer (except if the Congo river has exceptionally high 228Ra activity, which hasn't been demonstrated in this study). The fluxes are estimated using Ra for the "Congo-shelf-zone" and on an "off-shelf transect", which include inputs from both the river and from sediment. The salinity range is pretty high, mainly above 32, which suggest less than 20% of freshwater contribution in the study zone. This is true, and most of this is discussed in the manuscript. We argue that there must be source (s) of 228Ra between the Congo River mouth and the Congo-shelf-zone. As noted in our response to the other reviewers, we show that either (or both) high 228Ra diffusion from shelf sediments in this region and/ or an additional source of Ra such as submarine groundwater discharge are the most likely sources; the shelf-influenced-river waters are transported offshore. The fact that such high Ra, Fe and other dTE concentrations are found at such high salinities is a key point of the study as such concentrations are not observed so far offshore in other systems.
The author estimates the Congo river flux (discharge times concentration in the river), which is not necessarily comparable to the Ra flux because the latter is estimated for a specific area and not all the river freshwater discharge actually makes it into this restricted zone. I think there are a lot of discussion that has been cut probably because of the very short format of the journal, but I feel that it would improve the paper to provide more details regarding what these fluxes actually represent.

For any Ra work it is necessary to define an area of interest, and in the context of an
oceanographic study this is the shelf region corresponding to our lateral transect, which was run in order to calculate offshore transport. The point of calculating both the Congo River flux and the Ra Congo-Shelf-Zone flux is to see if they are comparable and thus to what extent the River flux alone can explain concentrations on the shelf. Obviously, in addition to freshwater being laterally transferred west, the plume also disperses along the shelf generally in a northward direction. But they key finding of our study is that unusually large concentrations of dTEs are found at considerable distances off-shelf, and thus we focus on this lateral transport and not the fate of river water itself which isn't exported off-shelf.

River flux alone can explain the 228 Ra concentrations on the shelf. Nonetheless, our study focuses on the transport of the Congo shelf-influenced-river waters enriched in 228 Ra and TEs to offshore regions."
Methods: -The second count for Ra224 was performed 6 weeks later whereas the second count for Th228 is usually done 2 weeks after collection. Was it really done after 6 weeks? Then some Bateman correction should probably be applied.
The fibers were first counted on-board the ship to determine the total activities for the shortlived Ra isotopes. The total activity of a daughter (e.g., 224

[Line 235] "The fibers were counted onboard and aged for six weeks, in order to allow excess 224 Ra to completely decay. They were then recounted to determine 228 Th concentrations and thus correct the total 224 Ra for the supported activity."
-After measurements of short-lived on the RaDeCc, the authors said that the fibers were ashed and then leached. Did the author leached the ashes? I am aware of two methods: ash (Charette et al. 2001) or leach/co-precipitate (Moore method) the fibers but I have never seen both processes applied one after the one. Is it a typo by the authors or was it the real procedure?
This was indeed the actual procedure. Additional text has been added to clarify the methodology:

150), the ashes were subsequently leached followed by co-precipitation with BaSO 4 . Ashing the fibers before leaching produced a more homogeneous material that was easier to handle.
Although atypical, this method has been used elsewhere 64 Continental Shelf Research, 105, 95-100, doi:10.1016/j.csr.2015.05.014, 2015 As per earlier comments, we corrected the 228Ra average activity in the Congo-shelf-zone.
Other comments: -What is the input of atmospheric deposition of Fe to the study area? What is the contribution of dust the high Fe concentration in the Congo River plume? Dust is a significant source of Fe to the ocean (Jickells et al., 2005), therefore, the contribution of dust-derived Fe in the surface concentration cannot be ignored. Especially considering a recent publication from Menzel Barraqueta et al. (2019) showing that the study area receives one of the highest atmospheric deposition rate to the Atlantic Ocean calculated from Al data from the same cruise GA08 (~8-10 g/m2/y). A rough estimate of the flux of Fe from dust can be calculated using the average atm.
deposition, the concentration of Fe if measured on aerosol during the cruise or the composition of Fe in crust, and the solubility factor. Similar calculations can be done for Mn, and Co shown in the suppl. info. A more precise location can also be done since the deposition rate is estimated for the same stations. The author could thus perform a mass balance and potentially correct the Fe concentration that is due to dust.

34). These ratios reflect how dFe is removed relative to the other elements, which is unclear from fluxes alone. These ratios suggest that River Congo dFe is removed by a factor of 10, whereas the fluxes (Table 1, discussion below) indicate that the net removal is only a factor of 2. In summary, a multi-element approach also corroborates significant dTE inputs into the Congo-shelf-zone other than River Congo water."
-Can the authors discuss the three hypothesis that could explain the high Ra flux from the Congo-shelf-zone. 1) exceptionally high Ra in the Congo, 2) high Ra diffusion from sediment, 3) SGD. Which one is the most likely? Why Ra would be much higher in the Congo River compare to other riverine system. This point is now clarified in the text as also suggested by reviewer 1 and 2 (see above). As mentioned previously, the most likely explanation are numbers (ii) and (iii), or a combination of both, because 228 Ra diffusion from shelf sediments has large regional variability and the existence of another source of Ra such as submarine groundwater discharge in a large river system such as the Congo is plausible. We can't exclude number (i), but agree with the reviewer that it is unlikely because global rivers do not show large variation in Ra activity, and the Ra concentration in the Congo would have to be higher by factor of 4 to support the observed flux. Our dFe flux, for example, from the Congo shelf zone is an order of magnitude higher than reported elsewhere (e.g., Sanial et al., 2018;Charette et al., 2016). The total off-shelf Fe flux, is consistent with Sanial et al., 2018;Charette et al., 2016, because (Table 1)  -Another approach to estimate the Congo river input could be used. The authors used the TE concentrations measured in the river and the river discharge. An estimation of the amount of freshwater actually entering the Congo-shelf zone can be done using the average salinity and applying conservative mixing of salinity to calculate the volume of freshwater instead of using the river discharge. Doing so, the source of TE from the freshwater Congo river will be directly comparable to the sediment derived estimate of Ra (since the mol/y unit implies a specific surface considered).
The method suggested by the reviewer would likely work in a more completely sampled river plume. Our limited sampling coverage in the Congo-shelf-zone (due to restrictions by Angola and local oil/gas platform operators) would lead to unacceptably high uncertainty.
Unfortunately, satellite-derived surface salinity data shown in the first version of this manuscript doesn't match well with the measured salinity over the cruise track (Fig. 1) as mentioned in the manuscript. Uncertainties in the plume thickness (e.g., Fig S4) would have to be considered as well.
-The authors ruled out the benthic supplies of Ra and TE because of the shallow Congo river plume. But we don't really know what is the water column depth along the transect. Since the plume is moving along shore, it could be hugging the coast in shallow water. Thus, it would be useful to know the water column depth compare to the 15 m of river plume. Similarly, the odv plot figure 4 in suppl. info has no bathymetry, how steep is the shelf? Maybe adding the bathymetry to the Figure 1 would already give some information.
Thank you for the suggestion. Figure 1 has been changed and the bathymetry is now added. The previous figure 1 has been moved to the supplementary material.
-Can the author discuss the difference of Fe flux between the Congo shelf zone and the off shelf transect? Does this provide information on the scavenging rate or biological uptake of Fe in between the two zones?
The biogeochemical processes controlling the distribution of dFe in our study region are still unclear (but will include scavenging and biological uptake). This point will hopefully be clarified in future publications about the GA08 cruise when full Fe speciation data is available and biological activity and particle fluxes for the region are better constrained -Can the author discuss the exceptionally high dFe concentration 600 km from the Congo river mouth? Is the Fe 600 km away from the mouth river from the river or does it has another origin?
That is just a side comment that I am curious about.
There is a strong indication that Fe is associated with the Congo outflow, as dFe correlates strongly with both Ra and with salinity. As discussed above, other sources (e.g., dust) appear to be more than two orders of magnitude too low to account for the source. We cannot think of any other hypothesis which would be consistent with observed distributions of both Ra and Fe over the transect. We added a text about that.

Comment on figures:
- Figure 1: I would add a large scale map of Africa for example in a subplot to better situate the study area, to be more visual.
Could the authors show the salinity underway data to better visualize the actual plume during the survey? Should be better than satellite image, especially because the authors acknowledge a difference in salinity between satellite and in situ data. Add the bathymetry to see the width and steepness of the shelf.
Thank you for the suggestion. Figure 1 has been changed as suggested -In general, I think that surface plots would be useful and could be added as subplots to the figure 2 for example since the cruise track is composed by an along shore transect and an offshore transect. This will be useful, but as we need to show the mixing line on scatter plots for quantitative discussion, we prefer not to add qualitative plots. The sharp changes in dTE concentrations due to freshwater 'pockets' offshore also make interpolated surface plots undesirable due to the artificial gradients created between such data points.
- Figure 2: switch planels c and d to show Fe in a) and c), and Ra in b) and d) We numbered the graphs according to the order they are mentioned in the text. So, to follow the text organization, we prefer to keep the current format.
- Figure 3: the Congo data do not stick out, I would exchange the symbol between Congo and Mississippi to highlight the data from this study in red. Based on this graph, the 228Ra are consistent with other estuaries. Could this help in the discussion of the 3 hypotheses about the high flux of 228Ra I mentioned earlier?
The reviewer is referring to figure 3 of the supplementary info. The figure has been changed as suggested. This is true that the 228Ra are consistent with other estuaries in high salinity (and this is why we don't believe in hypothesis number (i) (see discussion above and text below)).
Nonetheless, without directly sampling the river endmember and lower salinities, it is not possible to conclusively rule out the possibility of an unusually high river Ra endmember concentration.