pXRF Measurements on Soil Samples for the Exploration of an Antimony Deposit: Example from the Vendean Antimony District (France)

Round 1

Reviewer 1 Report

The manuscript has been significantly improved in response to the comments.  Most of the major points have been addressed. 

Reviewer 2 Report

This manuscript describes the application of pXRF for locating elemental anomalies in soil covering veins of Sb-Au mineralization in Vendée, France. It promises rapid availability of results, dynamic sampling design, decrease in sample number and a small sampling footprint. The authors demonstrate that by using transects perpendicular to known veins, anomalies of Sb, Mn and As can be detected by this method. The results are improved by using statistical processing and evaluation of the pXRF data. For quality control, the pXRF data are compared to laboratory results of aqua regia digestion and ICP-ES and ICP-MS, which show good correlation with the pXRF.

As I reviewed a previous Version of the manuscript, there are no additional comments.

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

The paper presents a potentially useful study to help expand understanding of the strengths and limitation of pXRF for the setting and device used.

The objectives of the paper are not clear. Is it a study of the capacity of pXRF to detect low-level Sb anomalies in soil?  Is it a comparison of PCA applied to raw geochemical data (which we know has major problems) versis clr-transformed data? Is it a study on dispersion of Sb and related elements through cover?

There are substantial gaps in the explanation of the sites (such as depth of soil or regolith cover), the way PCA (or factor analysis) was applied. It is also unclear why multivariate methods were used, given the objective seems to be looking for Sb anomalies above known Sb-bearing veins, and why things like factor scores were not plotted. Why did the authors not look at Q-Q or fractal plots to give more objective assessment of background and anomalous population thresholds?

In other cases there is material in the paper that seems to have little relevance to the study (e.g. Fig 16)

Given these issues, it is difficult to properly assess the validity of the conclusions.

The data looks good (including assessment of QC) but paper need major work to take advantage of such data.

Further comments in the annotated manuscript.

 

 

Comments for author File: Comments.pdf

Reviewer 2 Report

see attached file

Comments for author File: Comments.pdf

Reviewer 3 Report

Summary

This manuscript describes the application of pXRF for locating elemental anomalies in soil covering veins of Sb-Au mineralization in Vendée, France. It promises rapid availability of results, dynamic sampling design, decrease in sample number and a small sampling footprint. The authors demonstrate that by using transects perpendicular to known veins, anomalies of Sb, Mn and As can be detected by this method. The results are improved by using statistical processing and evaluation of the pXRF data. For quality control, the pXRF data are compared to laboratory results of aqua regia digestion and ICP-ES and ICP-MS, which show good correlation with the pXRF.

Broad comments

The manuscript provides an interesting and appropriate application of pXRF. Since the area was studied before, the success of this work could be demonstrated in general. However, the manuscript lacks detailed information on the measurement procedure as well as some background information on the used methods, which makes it in part difficult to understand their interpretation. A scientific publication should be written in a way that it is, in general, understandable for a scientist without any further literature study and repeatable with the given information. Furthermore, the significance of data processing in the lab should be discussed in more detail, since this approach is supposed to be a quick field method. Then, this manuscript should be a useful research contribution.

Specific comments

• l. 36 “cost-effective exploration”: Strictly speaking, is it exploration or rather reconnaissance/prospecting what is meant? In L. 389 it is finally called reconnaissance.
• Are there any other studies dealing with pXRF and specifically Sb that might be worth mentioning in the introduction?
• ll. 93 ff: Only the geology of Les Brouzils is mentioned, what about the two other deposits? Are they similar, or do they belong to the group? Please explain their relation.
• ll. 143 ff: There is only little information given on the pXRF device. Please add information about the Niton: target, acceleration energy (or energies?), used filters, calibration and quantification method, etc. Which lines are used for the measurements? Are there overlaps for Sb with other elements, e.g. Ca at 3.6 keV or did you use the Kα-line?
• Please add more information on the sample preparation: how much material was used? What was the sample depth? Was the porosity of soil considered in the calibration? Were the matrices of the samples and the standards comparable? These are some factors that influence the quantification and might help to explain the bias mentioned later.
• Are the readers of Minerals supposed to know about CoDa, CA and Cochran? Instead of just mentioning the literature in 2.4, giving at least a few explanatory sentences in the introduction on why these methods were selected and how they work should be helpful. Moreover, there is only one short sentence on Fig. 6 and 7 showing the results of Cochran’s C-test; here one explanatory sentence might have space, as well.
• Why are table 1 and 2 identical?
• l. 203: Why use half a page for table 3 if “no meaningful trend is observed”? Does this result not completely undermine the whole concept of on-site sample selection and dynamic sampling if there is no correlation of the raw data and complex data processing is necessary? See also comment on L. 365.
• In the first reading, it was not clear, that 3.1 and tab 1 refer to the pXRF data, since 2.3 is about the AQ250 method. Please give a short pointer to the pXRF at the beginning of 3.1. Maybe it might be even helpful to move paragraph 3.3 on quality control up before the actual “field”-XRF-data.
• l. 297: Does vegetation or land use have any effect on the data? There are three types of land use described (grassland, maize, and forest). Should groups be treated differently? At least for Fig. 8., since you assume additional Ca is used on the agriculturally soil and on others not.
• ll. 226 and 259: “Aitchison 1986”: Format reference.
• l. 286: Horizons were treated like independent data sets. Are they? How are they related? What effect does sampling of both horizons have on the data interpretation?
• ll. 290 ff: Why As and Mn? What is the relation to Sb? Are there any explanations besides the statistically anomalous pattern before?
• l. 297: Should there a % added to the Y-Axis in Fig. 8?
• l. 298: “We suspect that it reflects surficial enrichment.” Should this kind of interpretation be in the results section?
• 2.1: Why not sorting the paragraphs according to the figures? It will save the reader some trouble searching for the right location.
• 2.2: There is no text about the As anomalies.
• l. 350: What might be the reason for the underestimation of Sb and As by pXRF? Maybe sample preparation or porosity?
• l. 351: The large scatter of Sb might be due to the calibration with samples < 175 ppm Sb and the extrapolation to for the sample > 400 ppm Sb. However, there is no information on how the instrument was calibrated and which method was used (except for Soil mode).
• l. 357: It is not clear what “Using a lower concentration profile for Sb” means. Please explain.
• Please correct the formatting of paragraph ll 355-359. L. 359 contains the caption of Fig. 11, is it supposed to be within the parentheses?
• l. 365: Mn is a pathfinder element, but how is it connected to Sb? Could this relationship be also applied to other sites? LL. 406f says the pathfinder signature needs to be determined prior to the survey. Does it mean data processing and statistical analysis in the lab? How is then optimized sample selection in the field possible? Would you have been able to identify Mn as a pathfinder without the PCA?
• Is data processing (e.g. CoDA) in the field possible for optimized sample selection. Please discuss in more detail, how the results would help in the field without processing?
• ll. 370 – 373: There seems to be another strong Sb anomaly in Les Brouzils in the northern part. Was it checked? Is this what is referred to in LL. 375f? It would be a huge gain to the publication, if actually a yet unknown vein was located.
• l. 14 we revisited a historic…
• l. 98: richly mineralised
• l. 109 mineralized: American English?