Multiple Sulfur Isotope Records of the 3.22 Ga Moodies Group, Barberton Greenstone Belt

Round 1

Reviewer 1 Report

Although many modifications have been made, this new version of the manuscript maintains a biased approach in the discussion and presentation of the data.

CapD33S values are interpreted considering 'sizeable' or 'meaningful' values without any reasonable scientific basis. This concern was already raised in the previous version of the manuscript, but it has not been properly addressed.

Terms of comparison are not used, so that when it is said 'small' or 'big CapD33S values' the meaning of the sentence results misleading.

In the discussion, processes imparting variation in S isotope signatures are evaluated in a superficial way. For example, decrease in d34S is associated with BSR since some CapD33S values are negative. The same patter could have been caused by many other processes such as TSR, or precipitation at variable pH-Eh conditions. To argue that BSR played a role, more convincing evidence, such as the presence of biological remains, should have been provided.

Another example of superficial interpretation is the sentence in L.427-428. "fast-accumulated sediments may have substantially diluted sulfur derived from an atmospheric source during their deposition."
There are no references in support, no is process explained or model provided. This appear to be a pure speculation.

Also, previous comments have not been addressed properly.

For example, when I pointed out that:
''(17) L.341: “Archean sediments are generally characterized by sizeable S-MIF”. This statement is too generic and not backed by a statistical analysis. Firstly, it is not clear what represents a “sizeable” S-MIF signature. Also, according to the data of the Global Sedimentary Sulfur Isotope Database provided by Selvaraja et al. (2017; http://www.cet.edu.au/research-projects/special-projects/gssid-global-sedimentary-sulfur-isotope-database) 2746 analyses over a total of 4776 (of Archean sediments older than 2.4Ga) have a CapD33S between +1 and -1‰, which has been considered in this manuscript as “unclear S-MIF”. I recommend the authors to access the data from Selvaraja et al. (2017), revise, and discuss more in detail this statement."

The answer from the authors has been:
"New L.415: See our responses to general comments (1) by the reviewer. The reviewer’s major concern about this statement is due to our unclear definition of S-MIF in the former manuscript. In the revised manuscript, we followed the S-MIF thresholds proposed in LaFlamme et al.
(2018). We judged that the present statement is unnecessary here and thus removed it.'"

No, my concern here was not the threshold definition, but about your interpretation of sizeable MIF-S, which, although no more contained in this sentence, it is still utilised as a recurring concept in the manuscript. I find this answer inadequate and dismissive.

Other comments below.

In reference to the responses to the reviewer 1:

(1) LaFlamme et al. (2018) MIF-S thresholds have been included, but Farquhar and Wing’s (2003) are still used. This makes the figures and the data interpretation confusing. Remove Farquhar and Wing (2003) thresholds, or justify them under a physical point of view.

Comments on the new version of manuscript:

L.25 - “a typical Archean S-MIF signal”

It is unclear what a typical MIF-S signal is supposed to be. From the Global Sedimentary Sulfur Isotope Database (Selvaraja et al., 2017) the CapD33S median values of sediments older than 2.6Ga is +0.16 per mill, and the mean is 0.64 per mill. Are this the values which are considered as typical?

L.27-29 and 427-428 - “Fast accumulation (~0.1-1 mm/year) of the Moodies sediments might have substantially diluted sulfur derived from an atmospheric source during their deposition [...]”

It is unclear how the rate of sedimentation could have affected S-isotope signatures. There are no references regarding this alleged process, nor any physical explanation or modelling.

L.71 - “The Archean S-MIF has been conventionally defined by a sizeable (> 0 ± 0.2‰) ∆ 33 S value [16]”

The convention cited in this line refers to an information box in Farquhar and Wing (2003) which states: ‘We note that the definition of CapD33S used in this paper specifically refer to the terrestrial distribution of sulfur isotopes and CapD33S values within ± 0.2‰ of zero can arise through the normal operation of mass-dependent fractionation processes.’

This explanation is not anymore adequate to define the MIF-S thresholds, after 17 years of research and new data published the current understanding of MIF-S threshold has changed as discussed in many papers, among which LaFlamme et al. (2018).

Please justify the motives for using Farquhar and Wing (2003) MIF-S thresholds. It is my opinion that Farquhar and Wing (2003) thresholds do not have anymore a coherent justification, therefore they should not be used.

L.125 - “(sample 12-007-1: 25°42'4.3" S, 31°04'46.1" E)”

In my previous comments I asked ‘not to report sample coordinates in the text, but provide a sample list as a Electronic Supplementary Material (ESM) specifying coordinates, sample names, sample types, and any other relevant information useful to characterise the samples’ A sample list is still missing and coordinates are still within the main text. Modify accordingly.

L.237 - Figure 5

The mixing area drawn in Figure 5 is wrong. When BSR is involved mixing must account for non-atmospheric MIF-S. Revise Figure 5 following the indications provided in Ono et al. (2006), Aoyama and Ueno (2017, DOI: 10.1111/gbi.12268).

L.311 Discussion (general comments)

“a small MIF signature (Δ 33 S < 1.0‰)”

Relative comparisons must have a term of comparison. Small/big compared to what?? Marin-Carbonne et al. (2014, http://dx.doi.org/10.1016/j.epsl.2014.02.009) documented zoned sedimentary pyrite nodules with both positive and negative CapD33S signatures, consider that sedimentary sulfur being dissolved and remobilised, it would be 100% sedimentary with an overall CapD33S proximal to zero. How can the author argue that since the gold deposits have a a small MIF signature, thus little sedimentary sulfur was involved. The term of comparison, in this part of the discussion is pivotal, as a consequence the current interpretations are entirely speculative.

“Role of BSR”

In order to suggest that bacterial sulfate reduction played a role in the modification of the original sulfur isotope signatures more robust evidence must be provided. Thermochemical sulfate reduction could have been active too, variation in pH during deposition could have caused similar shift in d34S. Many other processes and imparted a shift in d34S are reasonable candidates for explaining the d34S variation presented in this manuscript. For example, traces of biological remains should be provided in order to argue that BSR played a role. Otherwise this interpretation is conjectural.

Author Response

Please see attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

This is the second time that I have reviewed this manuscript. It has been significantly modified and reworked in comparison with the original version. The new manuscript is well-written and makes a good job of addressing the comments made during the first review stage. I am satisfied with the changes that have been made in answer to my original comments. The manuscript can be accepted for publication in close to its current form.

In reviewing the new version, one additional potential problem came to light. The authors argue that the negative capdelta33S values on Figures 5 and 6 might be affected by mixing with sulfide derived from MSR and sulfide derived from a juvenile source, possibly a hydrothermal one. It is important to rule out that this mixing process was not due to mixing of sample within the analysis volume of the SIMS. This could potentially arise if different zones in a heterogeneous pyrite were measured in the sample analysis spot. An additional sentence to address this in the revised manuscript would be a good idea.

Author Response

Please see attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

I am happy with almost all the modifications made to the previous version of this manuscript. I am now keen to recommend this manuscript for publication after a single minor revision.

In line 74 is reported: "The Archean S-MIF has been conventionally defined by a significant (>0 ± 0.2‰) ∆33S value [16]". As discussed in previous comments to this manuscript, this threshold  is not any more meaningful and supported by any physical evidence or modelling. I do not believe that it is useful to report outdate information in newly published papers. For this reason, I recommend to remove this sentence.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

This manuscript discusses the S-isotope signatures of selected samples from the Moodies Group (Barberton Greenstone Belt). This work provides novel data of interest to the scientific community, and as such it is worth of being published. However, I believe that major revisions are required to address a number of relevant issues, among which:

(1) The data has been interpreted applying thresholds for “meaningful” CaD33S which are not supported by neither physical nor analytical justifications. As detailed in the comments below, it is unclear why the thresholds of +/-1.0‰ has been chosen. This decision strongly affects the interpretation of the data, as it artificially disguises the input of Archaean atmospheric sulfur in the sedimentary rocks studied in this work.

(2) S-isotope analyses performed by SIMS are interpreted as reflecting internal isotopic homogeneity. However, according to supplementary data, there are samples characterised by intragrain variations, such as sample 11-196 and sample 14-145. Isotopic variations within grains should be discussed more, and the location of those analyses represented in pictures either within the manuscript or as supplementary material.

(3) Data transparency. The sample list is missing, as a result the coordinates and the most relevant features of each sample are not clearly identifiable. The S-isotope data has been provided as word document, this makes it difficult to reviewers and future readers to utilise and re-elaborate the data. Please provide the data in a spreadsheet format, or in a database format.

Detailed Comments.

L.124: I recommend not to report sample coordinates in the text, but to provide a sample list as a Electronic Supplementary Material (ESM) specifying coordinates, sample names, sample types, and any other relevant information useful to characterise the samples.

L.152: I recommend to include in the caption the sample names represented in Figure 3.

L.195-197: “[…] (data not shown).” I recommend to include the data mentioned in the ESM. Data transparency is pivotal, otherwise this claim cannot be accepted.

L.226: I recommend to convert the S-isotope data (in the ESM) in a spreadsheet format. Text documents are not suitable for data as they are difficult to read and re-elaborate by fellow researchers.

L.226: I recommend not to provide “averaged values”, but mode values instead. As CapD33 values in this work are both positive and negative, the mean tends to provide an inaccurate statistical representation skewed toward zero.

L.237-238: “The CapD33S values are mostly <+0.5‰ (Figure 6b) and show no undisputable S-MIF signature (>+1.0‰), [...]”. I disagree with this claim which is fundamental for the following data interpretations. Specifically, I find the motivations for setting the thresholds that define S-MIF signatures at +/-1.0‰ unclear. What are the physical or analytical reasons for choosing +/-1.0‰, instead of +/-0.5‰ or +/-2.0‰? The analytical techniques utilised have uncertainties of 0.1‰ and 0.01‰ on CapD33S for SIMS and IR-MS respectively, so the thresholds of +/-1.0‰ for identifying undisputable S-MIF signature seems unjustified. Under a physical point of view, the thresholds for “true” S-MIF signatures should be considered variable as a function of the d34S values, as some fractionation processes (different from Archaean photodissociation) can have lamda values different from 0.515 and therefore generate “false” S-MIF signatures. This is well explained in LaFlamme et al. (2018, https://doi.org/10.1016/j.gr.2018.02.005). So, I recommend to apply LaFlamme et al. (2018) thresholds and reconsider the statistical distribution of the CapD33S values accordingly.

L.261-262: Regarding the absence of a D33S-D36S trend reflecting the ARA ratios. The data displayed in Figure 8 show a negative relationship between D33S and D36S values which is parallel to the ARA ratios and shifted to more negative D36S values of about 0.15‰. Although the data so not overlap the ARA trend, it follows it very neatly with a slight shift to more negative D36S values. The data presented in this study is not scattered around zero D33S-D36S, and do show non-zero D36S values. I recommend to reconsider the D33S-D36S distribution, noting the presence of anomalous D36S values parallel to the ARA trend. Also a linear regression line would be helpful to better identify the presence (or absence) of a trend.

L.273-274: I recommend to include in the caption the sample names represented in Figure 9 and Figure 10.

L.279: “pyrites” correct to PYRITE as minerals names are uncountable, for the plural form use PYRITE GRAINS or synonyms.

L.278-281: Regarding the statement on the absence of “clear S-MIF”, I recommend the authors to consider the work of LaFlamme et al. (2018) for a more accurate definition of the thresholds. Also, there are a number of analyses from rounded pyrite which definitively have S-MIF, such as 14-145@33, 14-145@34, 14-145@28, 14-145@22, 14-145@15, 14-145@14, 14-145@13, 14-145@12, 14-145@8, 11-196@17.

L.281-282: “Pyrite grains with no clear S-MIF may have been derived from surrounding intrusive rocks [...]”. This is an oversimplification, I recommend to discuss more thoroughly the ingrain S-isotope variability. Sample 11-196 which is considered as the “exceptional” grain having a “clear S-MIF” signature also has a large number of analyses with no S-MIF. Similarly, sample 14-145 has some negative “clear S-MIF” signatures (between -0.37 and -0.59‰) and many with no S-MIF. Such cases deserve to be discussed, also there is no mentioned of the negative S-MIF signatures of sample 14-145.

L.311-312: “Several gold deposits are known in the northern central BGB, some in close special proximity to the analyzed Moodies samples”. I recommend to substitute the generic terminology of “close special proximity” with numerical references of the distance between the Au deposits and the samples of this studies. Also, it should be explained what features make “special” the proximity between the samples of this study and the Au-deposits.

L.335-336: “Marin-Carbonne, unpublished data”. The use of opaque references is not acceptable in modern scientific works. Data transparency is pivotal for building solid and credible advances in our field. Please provide the data as Electronic Supplementary Material, or remove the comparison to these data.

L.341: “Archean sediments are generally characterized by sizeable S-MIF”. This statement is too generic and not backed by a statistical analysis. Firstly, it is not clear what represents a “sizeable” S-MIF signature. Also, according to the data of the Global Sedimentary Sulfur Isotope Database provided by Selvaraja et al. (2017; http://www.cet.edu.au/research-projects/special-projects/gssid-global-sedimentary-sulfur-isotope-database) 2746 analyses over a total of 4776 (of Archean sediments older than 2.4Ga) have a CapD33S between +1 and -1‰, which has been considered in this manuscript as “unclear S-MIF”. I recommend the authors to access the data from Selvaraja et al. (2017), revise, and discuss more in detail this statement.

L.342: I recommend to rephrase this sentence avoiding the terminology of “dampened down S-MIF”, as it suggests that the anomalies were previously higher and subsequently decreased. However, this has only been speculated in the previous parts of the discussion, and cannot be taken as certain. I’d suggest to just use “low S-MIF”.

L.355: “S-MIF is immune to modern weathering”. I recommend the authors to reword this statement in a more balanced fashion considering that the most effective process that erases S-MIF is mixing, and that mixing of sulfur with contrasting signatures could have happened in Archaean weathering scenarios, as well as modern weathering scenarios.

Reviewer 2 Report

The manuscript by Saitoh et al. presents valuable new multiple S isotope data on sedimentary rocks from the Moodies Group, Barberton Greenstone Belt. This study area is important since it represents, at 3.22 Ga, one of the first successions of well-preserved nearshore shallow-water and terrestrial rocks in the geological record. Previous studies have suggested that such near surface geological environments are more likely to preserve multiple S isotope heterogeneities in the output products from atmospheric photolysis than deep water marine sediments (currently the bulk of the literature data) where primary variability is more likely to be lost due to mixing in the oceans. This study, which is one of the first in the Moodies Group is thus an important test to look for potentially more primary atmospheric signals in the Paleoarchean S isotope record. It is thus an interesting and unexpected result that the variability in the isotope record presented here is so limited. 

The paper will be of interest to the readers of Geosciences and should be accepted for publication. I have some suggestions for revisions below which will help to improve the manuscript.

General comments

1. I am not convinced that metasomatism is the only process that can explain the relative homogeneity of the pyrite multiple S isotopes. It is certainly a good candidate, but there is no clear direct evidence for this presented in the paper. Are there other supporting data for metasomatic reworking available that you can add? The fact that at least some of the pyrites show complex internal growth structures, with inherited detrital cores, suggests that metasomatism did not completely dissolve and reprecipitate the pyrite populations. It would be important to show how extensive the inherited core samples are in the data set in order to further assess this. Late stage metasomatic pyrite is often internally homogenous on EMP maps. in the case of the pyrite shown in Fig 9, the outermost rim could potentially be a metasomatic stage of growth, whilst the inner parts are related to earlier metasomatic, diagenetic or detrital stages. In principle you should be able to resolve parts of this on the SIMS (I realize that spatial resolution is limiting here). The limited range in multiple S isotopes could also be a function of the low S concentrations and efficient mixing of S from photolytic sulfate and elemental S sources, prior to and/or during incorporation in pyrite. I agree that you cannot argue for or against microbial reworking with this data set, but in the absence of metasomatic overprinting this might have also removed any pre-exisiting heterogeneity in the source S pools.

2. How representative do you think your data are for the Moodies Group as a whole? This is an important point because the abstract suggests that you are making conclusions for the whole group. The data are taken from the mid- to upper part of MdQ1 from subtidal through to coastal floodplain environments. Could it be possible that much great heterogeneity could be found in other parts of the stratigraphy or more distant from the zone of high detrital input/ metasomatic overprinting? Looking further at new fresh drill core samples (line 357) will certainly help, but it might be more important to simply extend sampling to other parts of the greenstone belt. I think you need to be clearer about this in the abstract and in the discussion.

3. The data shown in the paper are for pyrite, AVS and CRS. Were there sulfide minerals other than pyrite present in the rocks? This could be an important issue when discussing the origin of the sulfides. The sulfide mineralogy associated with Au mineralization in this area is much more diverse than that found in background sedimentary environments.

4. I'm not sure that a high depositional rate would automatically lead to a suppressed MIF signal. This would depend on the relative amounts of detrital sulfide versus atmospheric S input. Since atmospheric input is likely to be episodic, and weathering could sample both sulfide-rich and sulfide-poor source rocks, the relationship is less clear. Isn't it also possible that rapid burial could better preserve primary atmospheric variability due to the suppression of mixing/reworking in the near surface environment? 

5. Please report the raw isotope ratios that were measured in the supplementary material. This should ideally be accompanied by absolute and relative S ion yields. Equivalent data for the calibration and secondary standards should also be reported- I think showing standard data should be routine in multiple S isotope papers

Specific comments

Line 17- sulfides? or pyrites?

Line 27- typo 'from'

Line 30- typo 'are supporting' should be replaced by 'support'. But see general comment 1 above

Line 33- 'dampened down' is better replaced with ' suppressed'

Line 40- add 'in the geological record' after 'unit'

Line 63- rewrite as 'the Moodies rocks have however not yet been addressed in previous studies'

Line 144- typo 'cross bedding'

Line 145- typo 'laminae'

Line 147- replace 'contained' with 'present'

Caption to Fig 4- rewrite 'The depositional environments are based on interpretations from..'

Figure 5- how significant are the negative capdelta33S values on this figure? is there a hint of more primary sulfate recorded here?

Figure 9- the labels and axis numbers are small and difficult to read

Figure 10- are the 2 sigma errors standard errors or standard deviations?

Line 288- how do you reach this conclusion about reworking? reference?

Lines 302-306- see general comments above. Would you expect multiple stages of precipitation for metasomatic pyrite? How would you identify this from diagenetic pyrite? What do you mean by 'these processes do not exclude one another?' I agree that they could both operate, but without further evidence they are speculations

Lines 308-309- see general comments above. Was this previously proposed? if so please back up with references. This part seems speculative.

Lines 307-327- the link to the Sheba and Fairview data is not so convincing in my opinion. Yes, there are similarities in the ranges in multiple S isotopes, but there is no other evidence presented here to support fluid migration further into the Saddleback and Eureka synclines. Do such data exist?

Lines 348-349- see general comment above

Line 350- I don't really get the rinsing out argument. Sulfate would have been soluble, but elemental S not. Sulfide would have been highly reactive with Fe and possible other cations or molecules. In the absence of an oxidative part to the S cycle, S transport might have been quite limited. Unfortunately there is an absence of strong evidence to firm up these arguments at the moment

Line 355- the wording 'S-MIF are immune to modern weathering' is awkward. Weathering can of course lead to mixing that eliminates a MIF signal so please rewrite this. What I think you mean is 'modern weathering does not produce mass independent fractionation'

Line 373- typo 'similarity'

Lines 377-378- this is not really very extensively discussed in the manuscript, so it would be better to move this from the conclusions to the discussion