Review of manuscript bg-2020-466 (submitted to Biogeosciences)

Title: Mercury mobility, colloid formation and methylation in a polluted fluvisol as affected by manure application and flooding-draining cycle.

Authors: Lorenz Gfeller, Andrea Weber, Isabelle Worms, Vera I. Slaveykova, Adrien Mestrot

Reviewer: Jan G. Wiederhold (German Federal Institute of Hydrology, BfG)

This manuscript reports the results of laboratory experiments in which two Hg-contaminated soil samples were incubated with and without addition of manure for six weeks over a controlled flooding-draining-flooding cycle. Soil solution samples were collected repeatedly during the experiments through open-pore suction cups, followed by measurements of colloidal and dissolved Hg, methyl-Hg and many other parameters using a variety of analytical techniques. The topic of the study is suitable for Biogeosciences and the results are novel and relevant for the large research community interested in mercury biogeochemistry.

The experiments were well-designed, the methods are described and validated in great detail, and the quality of the analytical data is high. The interesting results are presented and discussed in a detailed manner and supported by nice figures and tables (only with slightly too small font sizes for my taste in some cases). Some of the main findings include the association of Hg with Mn during the mobilization into soil solution after flooding, the lower Hg mobility in the manure-treated soils, a detailed characterization of Hg-bearing colloidal particles, the relatively large differences between the first and the second flooding period, and the inference that Hg methylation was limited by microbial activity/uptake rather than bioavailable Hg during the experiments. I congratulate the authors to their very interesting study and I recommend that the manuscript should be published in Biogeosciences after moderate revisions considering the following comments and suggestions.

My only general comment refers to the direct comparison of the results with previous related studies and therewith a more concise identification of the new insights generated by this study and their implications for Hg biogeochemistry in contaminated floodplains. I acknowledge that the authors present a thorough literature review in the introduction (a few additional recent studies are listed below), but the later parts of the discussion and conclusion sections could maybe still be improved by highlighting the similarities and differences of the new results with previous studies investigating contaminated soils from other field sites. Despite the detailed soil characterization and previous work at the site, the main binding form(s) of Hg in the soils at the studied contaminated site still remains somewhat unclear, making a direct transfer of the results to other contaminated legacy sites more difficult. Anyway, this is just an appeal to try carving out the specific new findings of the study and their implications to a larger extent than what is already done in the well-prepared manuscript. I look forward to seeing the final product in print.

l12: I would use “ecosystems” instead of “eco-systems”.

l19: I am not sure whether the term “control soils” is helpful in this context. The same two soils were used in all experiments, once with manure addition and once without manure addition. The experiments without manure addition could be denoted as control experiments to assess the effect of the manure addition, but the soils are not “control soils” in my opinion.

l21: I don´t think that you were able to monitor “methylation of Hg in the soil solution”. You measured MeHg levels in the soil solution, but it´s not clear that the methylation process also took place in the soil solution.

l22: “lower” instead of “lowest”?

l25-26: What do you mean by “proportional increase”? Do you refer to a higher fraction of colloidal Hg relative to total Hg in the manure vs. the non-manure experiments? Do the percent values indicate relative or absolute values? Maybe “higher relative” instead of “proportional”? Please rephrase to clarify.

l27: “Net Hg methylation” is not the same as “MeHg/Hg”, but it could be maybe described as “increase of MeHg/Hg relative to the initial condition” if no absolute MeHg values can be compared.

l47: Hg is not “found as FeS” but can be associated with this mineral phase. HgS could be both cinnabar or metacinnabar.

l55: The term “immediate decrease” is not really clear in my view. A release of Hg into soil solution first causes a concentration increase. Maybe “relatively rapid” instead of “immediate”?

l62: Here and throughout the manuscript: If a publication is cited with the author name in the text, then the year should be in brackets (here: “(2013)”.

l72: I understand that Hg(II) binding to thiol-rich NOM is thermodynamically favored but I am not sure about the term “larger”. Do you refer to molecular mass/size and can you give a reference to support this statement?

l89: “has” instead of “had”

l96: The charge of sulfate is “2-“.

l102: Weber et al. (2009) did not study Hg. Some additional Hg studies on temperate floodplain soils include for example Wallschläger et al. (1998, doi: 10.2134/jeq1998.00472425002700050009x) and Lazareva et al. (2019, doi: 10.1007/s12665-019-8253-9).

l104: You may also refer here to the recent studies on Hg dynamics in similar experimental systems with biochar additions (e.g., Beckers et al., 2019, doi: 10.1016/j.scitotenv.2019.03.401 and 10.1016/j.envint.2019.03.040; Wang et al. 2020, doi: 10.1016/j.envpol.2020.115396 and 2021, doi: 10.1016/j.chemosphere.2020.127794). Concerning similar experimental studies on other types of Hg-contaminated material, the recent studies by Zhu et al. (2018, doi: 10.1016/j.gca.2017.09.045) and Eckley et al. (2021, doi: 10.1016/j.envpol.2020.116369) could be of interest as well.

l106: “studies” or “researchers” but not “researches”

l110: Did you have an initial hypothesis on how the addition of manure would influence the system? If yes, it might be useful to present such a hypothesis here and then get back to it in the discussion/conclusion sections.

l118: Maybe better use the term “waste water releases” instead of “emissions” to clarify the pathway of the contamination. I think that many people primarily think about atmospheric pathways in the context of “emissions”.

l118: The company did (and still does) not only produce acetaldehyde but also many other chemicals. Mercury was also used in several other processes including e.g., production of vinyl chloride and chlor-alkali electrolysis (see cited historical report by Glenz&Escher, 2011).

l137: through

l138: Maybe better “Hg level” instead of “pollution”. There could be also other pollutants present.

l147: add “and” after “soil”

l180: I think that it should be “Table 2” instead of “Table 1” here (change numbers if this is mentioned first).

l185: In my opinion, there is no need to capitalize mineral names.

l191: I suggest adding the information which relative fraction of the total solution phase was withdrawn via sampling during the experiments. Could the lower water level already have had an influence on the results for the later sampling points?

l202: DOC concentrations are later reported as mg/L, so I suggest using the same unit here for the blank value.

l246: This section does not only describe Hg dynamics but also many additional parameters.

l251: Here and in the following: I suggest clarifying early on in the manuscript what the indicated “+/-“ values represent. I assume 1SD of the triplicate experiments?

l288: delete “but”?

l342: “suggests” instead of “suggest, “

l343: There could be also other relevant Hg(II) binding sites in NOM even if all the thiol groups are saturated. Is there any indication in the literature that Hg(II) binding to Mn oxide phases would be preferred relative to, for example, Hg(II) binding to carboxyl groups in NOM or binding to Fe oxide phases? Anyway, I certainly agree that your interesting data suggests that Mn oxides play an important role for Hg cycling in the studied system.

l368: The spelling of “sulfate/sulphate” and “sulfide/sulphide” should be consistent.

l386: Add “of “ after “formation”

l415: Can you specify the approximate proportion of mobilized Hg relative to total soil Hg over the course of the experiment?

l439: words/values are missing after “up to”

l455: delete “A”

l470: from

l473: Please explain how the sampling could have influenced the element concentrations in the remaining soil solution. As written before, I suggest describing the water level changes in the microcosm during the experiment and its potential effects on the investigated parameters.

l477: Is chloride really an important component of inorganic fertilizers? I thought that most crop plants don´t like elevated chloride levels. And even though chloride forms could potentially form stable complexes with Hg(II) in soil solution, binding of Hg(II) to DOC (or generally NOM) is probably still dominant.

l481: I suggest that you try specifying the observed “distinct effect” of the manure addition. You could potentially come back here to initially defined hypotheses (see comment above) and conclude whether you have verified or falsified them. I could imagine that such an approach might be helpful in further highlighting the novelty of the findings compared with previous work. This is a carefully conducted and well-described experimental study, but I believe that it might be possible to identify more clearly which specific insights on Hg cycling in contaminated soils were generated and how these findings could be relevant to other field sites and future work.

l489: suggests

l489: Which changes in redox conditions do you refer to here? Higher/lower redox potential or do you mean that fluctuating redox conditions in general (irrespective of the direction) increase Hg methylation?

l490: Maybe better “is removed from the soil” instead of “declines from the soil”?

l492: add “of” after “changes”

l492: Wording: Are the “temporal changes” really limited by “microbial activity”? Or rather “controlled by the extent of microbial activity”?

l493: Maybe “stimulated” instead of “facilitated”?

l497: It´s nice if your findings are supported by earlier studies, but I suggest highlighting the novelty of your findings (e.g., important role of Mn redox dynamics? decreased mobility due to manure addition? etc.).

l498: How does this finding compare with other studies in which organic amendments were added to Hg contaminated soils (see e.g., references listed above)?

l499-500: In my view, the sentence on “more work is needed” is superfluous. This is always the case.

l510-514: Please make sure that each sentence contains a verb.

l511: Stephane

l514: “advice” instead of “advises”

l582: Historische

Figure 1: I suggest increasing the font size in the Table. This will be very small in a printed article.

Figure 2: This is a well-designed figure containing a lot of information. You could consider removing all the x-axes except the lowest one to make it a bit less busy. What about PFe (did you see a significant fraction of Fe colloids)? The “-1” in the y-axis caption of panel g should be superscript.

Figure 3: I suggest pointing out in the figure caption that Hg concentrations are shown here in ng/L instead of µg/L  in Figure 2.

Figure 4: y-axis caption “colloid”

Figure 5: y-axis caption “Fluorescence”, legend “Composition” and “dissolved”

Figure 6: I suggest that all y-axis ranges should start at zero to avoid a wrong impression of relative changes between the treatments. For the MeHg/Hg ratio, I suggest that you consistently use either percent or permil throughout the manuscript text and in figures and tables.

Table 1: I suggest adding “Relative” before “Particulate” in the second last line.

Table 2: Please clarify the origin of the SD values (I assume based on triplicate experiments?).

S3, l5: from

S3, l8: have

S3, l25: Merck

S3, l26: subscript “3”

S4, l25: define abbreviation DCM (dichloromethane)

S4, l29: add “to” after “transferred”

S9: I suggest clarifying in the figure caption that not only the map but also the high-resolution Hg concentration data was taken from the DUS report.

Summary

The study titled “Mercury mobility, colloid formation and methylation in a polluted fluvisol as affected by manure application and flooding draining cycle” aims to identify the release dynamics of Hg in two soils under two conditions (with and without manure) over two flooding periods. Two soils were characterized and incubated in laboratory microcosms with synthetic rainwater and with/without manure over two flooding cycles. Pore water was documented at numerous points over the two flooding periods, and measured for total Hg, metals, anions, cations, DOC, and pH and Eh. Colloids were collected at 3 time points during each of the 2 flooding periods, and AF4 measurements determine the size distribution and some elemental composition (Hg, Cu, Fe, Mn, carbon). Methylmercury was quantified in the soil a 4 time points between t=0 and t=final conditions.

Overall, the study documents some nice results from the incubation experiment that test the effects of soil properties and manure addition. The study design and methods are well done, and I agree with the majority of the conclusions. However, my main comments are about the presentation of the work and ways to improve the clarity in presentation. I have itemized general comments and specific comments that should be addressed by the authors before considering this work for publication. The authors are encouraged to edit the manuscript thoroughly for editorial clarity. I did not identify all the sentences and statements that were unclear, but have listed some editorial comments in the Specific Comments section below.

General Comments:

• The importance of sulfate reduction should be revisited in this paper, as inorganic sulfide will scavenge pore water Hg(II) and result in authigenic formation of β-HgS. There is very little to no discussion of the decrease in sulfate concentrations in the microcosms, which indicates sulfate reduction and is a key biogeochemical transformation that can result in Hg partitioning back to the soils. Figure S7 in the SI shows very high levels of sulfate at the start of the experiments (150-1000 mg/L) and drastic decreases in concentration with flooding time.
• For the presentation of the Microcosm results, and figure presentation, I recommend the authors (1) use the redox ladder to guide the initial presentation, (2) consider discussing the Hg release dynamics in terms of “stages” or periods of time describing trends in the concentrations, and (3) detail the release dynamics of the other metals separately. Regarding item 1 of the redox active elements, in Lines 247-255, there is no mention of Fe or sulfate and all pertinent constituents (nitrate, Mn, Fe, sulfate) should be presented together in a single figure (at present, the reader has to look to the SI and main text figures). The observation that reductive dissolution of Fe wasn’t observed in Flooding period 1 is still a result that needed to be stated, and there is no mention of the decrease in sulfate from ~1000mg/L to 500 mg/L in flooding period 1 of the HMLC incubations. Regarding item #2 of the Hg release dynamics, on Lines 274-278, you may consider revising to describe the release dynamics in ‘stages’. “Concentrations of Hg were low between X-X days (phase 1), increased to a maximum at 4 days (phase 2), and decreased between 4 and 14 days (Phase 3).” These same ‘stages’ could be references when describing the colloidal data. Regarding the third item on other metal contaminants, the study presents data on diverse metals (Cu, and all metals in Figure S8), but Cu is the only metal discussed. The authors need to discuss the data they present in all figures, otherwise it is unclear why those data are presented in the first place. I commend the authors for a nice study and recognize that presenting the various non-metal metal data is challenging.
• Have the authors considered including an analysis that estimates if the thiol content of the DOM was exceeded in their experiments, to contrast with the soil analysis (Lines 339-345)? The strong binding site capacity of DOM has been quantified to be ~5 nmol/mg DOM (https://pubs.acs.org/doi/10.1021/es025699i). They can assume DOC is 50% of the DOM, and compare the strong binding site capacity of the DOM to the total Hg concentration < 0.02 um. Further, they could estimate how the addition of manure changed the Hg binding state (saturated vs unsaturated)in both the soil and pore water.
• As a reader, it will improve clarity if you spell out the various soils and treatments. I had to go back repeatedly to the methods to remind myself what the various acronyms meant (“HMLC”). This is important because of the two soils and two treatments (control vs manure).

Main Text Specific Comments:

Line 28 – consider deleting “again”.

Line 30 – consider “formation and aggregation” of …

Line 46-47 – There is “cinnabar” in the environment, but in the form of mineral deposits or associated with mercury mining activities. In the vast majority of environments, mercuric sulfide is present as authigenic nanoparticulate metacinnabar (β-HgS nano).

Line 59 – consider deleting “e.g.”.

Line 62-63 – consider expanding to include the microbial process. At present, it reads awkward because Hofacker 2013 and 2015 are referenced but it is somewhat unclear what the 2015 study contributed.

Line 66 – The first reports of DOM facilitating the dissolution of HgS were by https://pubs.acs.org/doi/10.1021/es9804058. This study should be cited.

Line 66-68; here you reference “altering the reaction kinetics of HgS(s) formation”, in which case you should cite Ravichandran et al., 1999 and Deonarine and Hsu-Kim, 2009.

Line 72-74; Ligand exchange is important, but in the vast majority of systems there is an excess of strong thiols binding sites in the DOM for all the Hg(II). This was first documented by https://pubs.acs.org/doi/10.1021/es025699i and should be integrated into this sentence.

Lines 66-70 and 430-431; the authors need to cite primary literature that document how DOM controls the nanocrystalline structure of β-HgS particles, which is a key property influencing the bioavailability of mercury under sulfidic conditions. Two paper that should be considered due to conditions that closely represent natural conditions include https://pubs.acs.org/doi/abs/10.1021/es201837h and https://pubs.acs.org/doi/10.1021/acs.est.7b02687.

Line 88; the authors may consider also looking at a recent paper on OM amendments to mine tailings. https://www.sciencedirect.com/science/article/pii/S0269749120370585

Line 92-93; The authors are encouraged to revise this sentence, as it could be improved to highlight the various environments where methylation is prominent (riparian zones, saturated soils etc) because of the redox conditions.

Lines 108-2213; somewhere in this paragraph it should be mentioned that “microcosm experiments” were carried out.

Line 112 – revised to “0.02 and 10 µm” or an equivalent term. At present, “0.02/10 µm” is a fraction and doesn’t make sense to me.

Line 132: First, this sub-header should read “Microcosm Experiments”. It is confusing to call them “incubations” when later you refer to them as microcosms – please be consistent and in all instances state “microcosms”. Second, in a section below you detail the “Incubation experiment blanks” but those are not detailed in this section, and they should be.

Line 135-136 – revised to “After the initial incubation period soils were used in the flooding and draining experiments, which were conducted in 1 L borosilicate glass aspirator bottles (Fig. S2).” It would appear Fig. S2 should be called out.

Line 137- revise to “Microcosm experiments were performed in experimental triplicate….”

Line 141-142; revise “were” to “was”; rainwater is singular.

Line 146; I presume you mean “remove any remaining air bubbles…”

Lines 192-195; consider revising to “At each sampling time, sample splits were preserved without further filtration (<10um) and filtered at 0.02 µm (add filter details). Additionally, at 2,5 and 9 days an additional sample split was filtered at 0.45 µm for colloid characterization.” What was the filter type for the 0.45 um filter?

Line 102 – DOC should be reported in units of mg/L, for consistency with incubation results.

Line 206 – revise to “filtered” fractions. And, it is not common to use “suffix” to describe a subscript, which is what is presented for each term.

Line 231 – revise to “0.5% HCl and 1.0% HNO₃”.

 Line 249 and 275 – consider revision to “NO₃^- depletion” or “exhausted”.  

Line 260262; this sentence doesn’t make sense and needs revision.

Line 261 – when describing concentrations in the text, the same units should be used as presented in the figure. Figure 2 presents Mn in units of mg/L.

Line 270 – The reader probably won’t remember the “cornfield soil” is the HMLC soils. See my comment above to just spell out the soil type. Consider revising “throughout the experiment” to “over both flooding periods”.

Figure 4 caption; it is entirely unclear what is meant by “Details on the deconvolution procedure are provided in the supplement”.

Figure 4 – should the y-axis label for the top panel indicate “particulate” and should state “total Hg”. 

Section 3.2 – Consider finding locations in this section to emphasis that you’re looking at time points across the two flooding events. Visually, the size proportion of Hg species data look interesting as they show trends in the first flooding period and little change after that. 

Section 3.3; the sub-header title should specify this is for the ‘soil’. 

Line 329-330; I don’t agree with this conclusion regarding the association of Hg to particulate Mn. In looking at Figure 2, the relative proportion of particulate Mn and Hg decreases with flooding time, but their overall concentration is still low. It is more likely that Hg is release from the soil. The decomposition (and solubilization) of OC in the soils can also release Hg. The pore water DOC concentration is reflecting both release and utilization of DOC, so may not necessarily capture the role of DOC on the Hg mobilization due to carbon mineralization.

Lines 346-349 – can you discount the possibility that soil heterogeneity could have contributed to the observed variability?

Lines 367-368 – the pore water data strongly support that sulfate reduction is occurring, which show drastic decreases in the concentration of sulfate with increased flooding time. In microcosms of this nature, several biogeochemical processes are occurring simultaneously and the Eh of the system isn’t sufficient to assess if sulfate reduction is or is not occurring. My assessment is that it is more likely that sulfate reduction resulted in the re-association of porewater Hg with soils, compared to the NOM complexation.

Line 378 – Poulin et al 2016 shows distinct Hg(0) formation in contaminated soil incubations, and should be cited here.

Lines 380 – one would need citations for the sentence on abiotic vs biotic reduction.

Lines 436-438 – one interpretation is that the soils had been subject to period soil flooding that contributed to mercury methylation.

Lines 444-445; could the higher microbial activity be the result of addition of labile carbon? The author should consider highlighting here the diversity of microbial communities that can methylate Hg, as is provided in the Introduction. Sulfate reducing bacteria, metal reducing bacteria, and fermenters are possible contributors to mercury methylation.

Figure 2 – the y-axis labels are very hard to read, and would be nearly impossible to read in print form. Consider re-working the figures as I suggest above, where all terminal electron acceptor processes and DOC are included in a single figure, then all Hg measurements, then all other trace metals.

SI Specific Comments:

Line 7 – this figure should be revised to state “total Hg” when total Hg is measured. This needs to be fixed in all cases in text and figures, in the main text and SI.