|Page and line numbers correspond to those in the “bg-2017-299-author_response-version1.pdf|
The authors have made improvements in this revised version that helps to clarify their meaning, and I understand better now what they were intending. I also think that the new additions discussing budget uncertainties are useful. I think that a lot of the two-way misunderstanding between the authors and me have to do with differences in terminology, and their intended meaning and mine were not as different as we each thought. Nonetheless, I still found some sections confusing, mainly to do with terminology used to describe different fractions of particulate iron.
For example, the authors seem to use “lithogenic” or “terrigenous” to indicate poorly exchangeable abiotic mineral particles in general, because they are initially interested in distinguishing these types of particles from biogenic particles, whereas I use “lithogenic” and “terrigenous” to specifically mean weathered aluminosilicate material that have elemental ratios consistent with crustal averages, and further separate the general class of abiotic mineral particles into “lithogenic” (i.e. aluminosilicates) vs “authigenic and scavenged” classes based on elemental ratios of sample particles compared to crustal averages. There are, of course, terrigenous and lithogenic particles that have elemental ratios different than crustal averages, so it’s not that my version is necessarily better, but I think that if the authors explicitly and clearly define their use and meaning of the words “lithogenic”/ “terrigenous” early in the manuscript, this will help to reduce confusion amongst the general readership about what they are talking about.
Another source of confusion was the use of the word “scavenged”. This could refer to surface adsorbed Fe, which would probably be readily exchangeable (i.e. bioavailable), but could also refer to authigenic precipitates ranging from rather labile oxyhydroxides that would be accessible by a weak leach to more aged hydroxides and oxides that are not. At various points in the manuscript, the higher Fe/Al ratios they see in suspended particles are referred to as scavenged Fe (e.g. section 3.1.2 lines 412-418) or lithogenic Fe (e.g. section 3.2.2 lines 548-550), depending on the way they are talking about these terms in each of these sections. They should pick a definition of “scavenged” and “lithogenic” that can be used consistently throughout the manuscript.
I also think it would help if the authors could come up with terminology to more clearly differentiate leachable particulate metals from their 25% HAc leach of their SAPS samples (LP) compared to the leachable particle metals from their total dissolvable seawater samples (LPun). They do define this in section 3.1.1, but in referring to these fractions in the text, they are both sometimes referred to as “leachable” particulate X, even though the former leach is poorly correlated with total particulates, but the latter leach is well correlated and used as a stand-in for total particulates. A more differentiated set of terminology to refer to each of these will further help to clarify.
Below I highlight specific places where I think additional clarifications will help.
Figure 2 caption: recommend replacing “corresponds to the concentration lables of LPunFe” with “corresponds to the same axis as LPunFe” to clarify the meaning.
Lines 370-372 and Figure 2: The authors state that LPun concentrations “were slightly lower than the particulate fraction”, but there are a few instances when LPun Fe exceeds PFe (e.g. deep stn 13, surface stn 18). As the authors note, these are different sampling approaches and pore sizes, so it’s not unexpected, but the authors should acknowledge this with language to indicate that the LPun were *usually* lower than P.
Lines 380-383: Ok, I understand now what the authors are doing by using LPun as an *indicator* of PFe. I transcribed the data from the tables and made a scatter plot of LPun Fe vs PFe and fit a regression to convince myself that they are indeed related (and they are), since it’s not always obvious from looking at profiles, but this is a plot that they should include in the supplementals, reporting slope and R2, to justify the usage of LPun to indicate PFe.
Lines 396-418: the clarifications added by the authors in the revised version help. I think these paragraphs could be made even clearer by stating at the outset that there are three potential origins to particles: lithogenic, biogenic, and authigenic (including scavenged), and then proceed to argue why the ratios indicate the importance of authigenic precipitation (or scavenging). Note that this is partly a matter of terminology, but what the authors call scavenged, or surface adsorbed, Fe and Mn may not be surface adsorbed to lithogenic or other particles in the traditional sense, but rather exist as discrete Fe and Mn oxyhydroxide mineral particles formed by authigenic precipitation.
I think it would be helpful in this and other discussions if the authors made a table summarizing the Fe/Mn, Fe/Al, and Mn/Al ratios for the particle pools that they talk about (crustal, sediments, suspended particles, fecal pellets, and phytoplankton).
Fe/Mn Fe/Al Mn/Al
crustal 58.00 0.20 0.00345
sediments 51.50 0.34 0.00660
suspended 68.00 1.25 0.01838
fecal pellet 70.50 0.48 0.00681
I don’t think the additions regarding titanium are necessary (lines 403-405), since the authors already have Al data, which is an adequate lithogenic tracer. I would, however, add a discussion of the ratio of PMn/PAl as it relates to crustal averages (which shows that PMn is enriched relative to PAl in suspended particles compared to both sediments and the typical crustal average), as this illuminates what is going on with Mn too and gives context for what it means to normalize Fe to Mn, which I still don’t really understand.
I am also still puzzled as to why the correlations are so good between PFe, PMn, and PAl (Figure 3). I would not expect scavenging (oxidation, precipitation, hydrolysis, etc.) behavior to be the same for these three elements, so why do these correlate so well?
Line 429: I suggest saying “enriched with surface bound and authigenic Fe”
Lines 457-464: The authors conclude that variability in the PFe/PAl ratio in various pools is explained by different contributions of biogenic material and Fe scavenged onto particles surfaces. The biogenic contribution is likely negligible, since Fe concentrations in phytoplankton is so much lower than in lithogenic and authigenic particles (lines 408-409). I agree that variations in the amount of scavenged and authigenic Fe is important for the PFe/PAl ratio. Note that there are large variations in the PMn/PAl ratio too between the different particle pools (see table above), indicating that there is also variation in the scavenged/authigenic Mn, which makes interpreting the Fe/Mn difficult.
The authors conclude that similar PFe/PMn in faecal pellets compared to suspended particles in the water column indicate that “Fe in krill faecal pellets was predominately associated with terrigenous material”. Firstly, do they mean terrigenous as opposed to biogenic? If so, which terrigenous? From the molar ratios for all the pools (see table above), the ratios in faecal pellets seem most similar to sediments rather than to suspended particles, so it may be worth highlighting the similarity to sediments instead of comparing to suspended particles.
Lines 548-550: The authors conclude that the particulate trace metals were “mainly incorporated in lithogenic material”, since the leachable fraction from a 25% HAc leach was so small. What do the authors mean here by “lithogenic material”, since elsewhere in the text, the authors argue that scavenging of DFe is responsible for increasing the Fe/Al and Fe/Mn in suspended particles compared to sediment and crustal averages (lines 412-418), and that scavenged Fe is exchangeable with DFe in the water column (lines 559-565). 25%HAc is a fairly weak leach, and there can still be authigenic minerals (hydroxides and oxides) that are not accessed by this leach, so do they mean that lithogenic material includes some of these hydroxides and oxides, or do they mean that the suspended particles have a lithogenic (aluminosilicate?) origin that has a much higher Fe/Al and Fe/Mn than the sediments and crustal averages? This point, though seemingly picky, is an example of how it would help to define their use of “lithogenic” more precisely.
Lines 574-583: the description of the faecal pellet Fe flux is clear, but it would be useful to add in total particulate Fe excreted by krill (faecal pellet mass and % of total Fe that is leachable, so giving an indication of total Fe would complete the picture).
Typos, wording, etc.
P.10, line 309: add *also*: “Unfiltered surface seawater samples were *also* collected and dispensed...”