Geochemical Assessment of Mineral Resource Potential in a Hg-Sb-Pb-Zn Mining Area: The Almadén and Guadalmez Synclines (South-Central Spain)

Round 1

Reviewer 1 Report

This MS is focusing on the exploration geochemistry of soil in a mining area in South-Central Spain. However, (1) the original data is lack with only statistical parameters and some key elements are also lack; (2) the sample sites are in the mining area mostly and no comparative area are sampled and the sampling density, 6 samples per 100 km2, is too low relative to mineral exploration; (3) the data processing methods are common and the processing results is dependent on your data indeed which may be changed with the increasing area.

 

Some revision suggestions are as the followings.

In Table 1, how to calculate the VC? Is the average or mean used?

In Table 2, the major components should be listed in major oxides. The items of Na2O TiO2 P2O5 (maybe change P to P2O5) MnO (maybe change Mn to MnO) and Total are lack. The Total may be helpful to assess the statistical parameters.

Move the SOM data into Table 2, and SOM should be incorporated into the Total to assess your geochemical data.

In Figure 4, major oxides are suggested rather than elemental species. Are some blank areas reliable on your interpolated values?

In Figure 6, F1 and F2 are similar in Al Ni P? Some elements are locating in different factors.

Author Response

Reviewer-1

This MS is focusing on the exploration geochemistry of soil in a mining area in South-Central Spain. However, (1) the original data is lack with only statistical parameters and some key elements are also lack;

We agree and we have included a complete table as supplementary material. But the authors think that a complete explanation about dataset is provided, based on distribution maps or violin plots, formats more easily understandable than raw data.

(2) the sample sites are in the mining area mostly and no comparative area are sampled and the sampling density, 6 samples per 100 km2, is too low relative to mineral exploration;

We have made an effort to ensure representativeness not only of mines and mineral insights, but also concerning all the lithologies in the two synclines, as well as local background areas. To accomplish this, we have selected sampling sites into each grid cell with this criterion, aimed to get data about all lithological units, and about local background areas.

(3) the data processing methods are common and the processing results is dependent on your data indeed which may be changed with the increasing area.

We have proposed the use of common statistical techniques to take advantage of large soil geochemistry datasets for mineral exploration. As the authors have explained in discussion section (lines 304-356), anthropogenic pollution is considered as a “masker” of geogenic anomalies, suggesting that only advanced techniques (fuzzy logic, random forest...) can solve this problem. Here we propose and try to demonstrate that common statistical techniques, properly applied, can provide useful information on the large new datasets (FOREGS, among others).

Some revision suggestions are as the followings.

In Table 1, how to calculate the VC? Is the average or mean used?

VC, also called Relative standard deviation, is the variation coefficient, a standardized measure of dispersion of a probability distribution or frequency distribution. It is often expressed as a percentage and is defined as the ratio of the standard deviation to the mean (or its absolute value) – Taken from Wikipedia; it is a well-known basic statistical parameter, aimed to indicate the variability of an analytical parameter in percentage, instead of in absolute value, which is shown by the SD, standard deviation. Concerning mean and average, we have used average.

In Table 2, the major components should be listed in major oxides. The items of Na2O TiO2 P2O5 (maybe change P to P2O5) MnO (maybe change Mn to MnO) and Total are lack. The Total may be helpful to assess the statistical parameters.

We are sorry, but the analyses carried out generally for soil geochemistry are not aimed to calculate the total composition, as for rock geochemistry; the soil analyses are carried out after acid dissolution of the elements, and not after total digestion. That is why in soil geochemistry the elements, including the major ones, are not expressed as oxides, as verifiable, for example, in the FOREGS geochemical atlas of Europe.

Move the SOM data into Table 2, and SOM should be incorporated into the Total to assess your geochemical data.

As answered in the previous question, the total sum is not a proper calculation for samples analyzed after dissolution with Aqua Regia. Therefore, changing SOM from Table 1 to Table 2 is not needed, and it is confusing since Table 1 corresponds to the pedological parameters.

In Figure 4, major oxides are suggested rather than elemental species. Are some blank areas reliable on your interpolated values?

Yes, we have used IDW to interpolate data with no trends analyze that can create fake blank areas.

In Figure 6, F1 and F2 are similar in Al Ni P? Some elements are locating in different factors.

Yes, this statistical treatment allows that one element belongs to more than one group.

Author Response File: Author Response.docx

Reviewer 2 Report

1. Symbol  between the numbers of thousand and houndred should be ",", not ".". Please check the text throughout. 

2. The sampling numbers for the three synclines should be provided.

3. The text in Figures 4 and 5 is unclear. You should modify them.

Author Response

Reviewer-2

1. Symbol between the numbers of thousand and hundred should be ",", not ".". Please check the text throughout.

Done.

2. The sampling numbers for the three synclines should be provided.

Done. The sentence is as follows: “Due to access restrictions, instead of the forecasted 150 samples, it was only possible to take a total of 129 samples, from the Almadén syncline (N=40), the Guadalmez syncline (N=55) and the Alcudia anticline (N=18) (Figure 1).

3. The text in Figures 4 and 5 is unclear. You should modify them.

Done. The sentence has been changed to: “Figure 4. Spatial distribution maps and violin plots (B and C graphs) of major elements (all data are expressed as a percentage).” and “Spatial distribution maps and violin plots (B and C graphs) of trace elements (expressed as mg kg-1 for Pb, As and Sb and ng g-1 for Hg).”

Reviewer 3 Report

This is an interesting study in an interesting region of the world. Long known for its mineral deposits, this area has a long history of mining and is globally significant. The study examines the results of a soil survey but falls somewhat short in the analysis. Much of the focus is on cluster analysis and factor analysis. 

Initial comments include that many indicators are suggesting that the lithology control is extensive. But there is no comparison with the chemistry of the rocks of the area. There are some signs that black shales in particular could be controls on the regional analysis. There is however no analysis or reference to lithology samples. 

In the sampling methods, how/which horizon soil samples were collected was not discussed. 

Other than the application of the two statistical techniques there is no discussion about how they are seeing through the anthropogenic signal. 

In the maps presented, the known deposits are not showing up as anomalous areas. If this approach is to guide us to additional resources, why some are highly anomalous and some deposits are not, needs to be examined. 

Some discussion of these deposits may be warranted given their importance. Connecting them to the clusters and factors needs some work. given the strong structural overprint in the area and a common orientation to most of the geology, some of the plots should have been forced in that direction. Given the wide-spaced nature of the samples, a purely grid-focused analysis seems limited as an interpretation. 

Not clear why the elements that were analyzed were chosen. It would be interesting to see how gold is reacting in this area.

The abstract actually has a better conclusion than the conclusions do. 

Comments for author File: Comments.pdf

Author Response

Reviewer-3

This is an interesting study in an interesting region of the world. Long known for its mineral deposits, this area has a long history of mining and is globally significant. The study examines the results of a soil survey but falls somewhat short in the analysis. Much of the focus is on cluster analysis and factor analysis.

Many thanks for these observations. Concerning the criticism, as answered to Rev. #1, with this work we propose and try to demonstrate that common statistical techniques, properly applied, can provide useful information on the large new datasets (FOREGS, among others).

Initial comments include that many indicators are suggesting that the lithology control is extensive. But there is no comparison with the chemistry of the rocks of the area. There are some signs that black shales in particular could be controls on the regional analysis. There is however no analysis or reference to lithology samples.

Thanks very much for this comment. Undoubtedly, the question of the influence of lithology in the presence of this kind of mineralizations is crucial, as stated by Gumiel (1982) (reference included in the manuscript). However, that is a metallogenic question, not considered in this study, centered in soil geochemistry. The study has been centered on the advantage of large datasets of soil geochemical samples to mineral prospection, but the link with lithogeochemistry is interesting for future manuscripts, thanks for the suggestion.

In the sampling methods, how/which horizon soil samples were collected was not discussed.

Done. A sentence has been added: “Samples were taken from the A horizon of poorly developed soils (mainly Inceptisols);...”

Other than the application of the two statistical techniques there is no discussion about how they are seeing through the anthropogenic signal.

The mining activity in Almadén syncline was active since roman times and has produced a huge Hg dispersion in the environment. Although this anthropogenic pollution, some clues about Hg deposits have been revealed by the statistical approach. In Guadalmez syncline, Sb mining works were smaller, and with a much more restricted dispersion halo in terms of space, but the statistical treatment has found successfully trends to find new probable Sb deposits. A sentence in the discussion section has been edited to complete this topic: “The FA has unveiled the possibility of discovering relationships between the data that were not detectable in the cluster analysis (Ouchchen et al., 2022), and avoiding the interference of anthropogenic pollution in an extensively mined area for Hg, Sb, Pb and Zn.”.

In the maps presented, the known deposits are not showing up as anomalous areas. If this approach is to guide us to additional resources, why some are highly anomalous and some deposits are not, needs to be examined.

Yes, the statistical treatment has not considered outlier values and some of the known deposits has been mined and produced outliers, and some of such deposits remain unexploited and have produced smaller dispersion haloes, not sufficient as to produce outliers on their proximity.  This can be the explanation, especially in Almadén syncline and related with Hg deposits.

Some discussion of these deposits may be warranted given their importance. Connecting them to the clusters and factors needs some work. given the strong structural overprint in the area and a common orientation to most of the geology, some of the plots should have been forced in that direction. Given the wide-spaced nature of the samples, a purely grid-focused analysis seems limited as an interpretation.

Thanks for these suggestions. However, we consider that including these questions should enlarge too much the manuscript, with data which can be found in other publications related directly with the geology and metallogeny of the area or needing new research.

Not clear why the elements that were analyzed were chosen. It would be interesting to see how gold is reacting in this area.

We have analyzed, using ICP-AES, all the elements with atomic number > 10. Then, we have selected to include in the study the ones usually related to Sb mineralizations and available in our analytical data. In the discussion section we have explained that the selection of elements is very important in these statistical treatments, and that in many cases (as for our manuscript) the bias introduced by the analytical technique used is more decisive than other factors (lithological and geochemical). We agree that this is the most important limitation of this approach, as other references have been explained before (Reimann et al., 2002). Concerning gold, this element has not been analyzed by us; besides, this is an area intensively studied by Minas de Almadén y Arrayanes S.A. (MAYASA), a mining company owner for many years of the mining exploration ranges, and we warranty they explored this area for gold, as much as for other elements, without finding any indication of the presence of this element.

The abstract actually has a better conclusion than the conclusions do.

Done, we have added to the conclusions the sentence: “Some of these areas coincide with the discovery of mineralized zones, specifically in the area SE of a derelict Sb mine, confirming the usefulness of these datasets and statistical tools in areas with re-cent mining activity.”.

Responses to comments from this reviewer included in the PDF file:

Line 85: So this is important but not realy discussed in this paper it seems.

Of course, this is an important concern. But in this study, we have pointed the attention on the geogenic contributions, aimed to discover anomalies not directly related with the mining (anthropogenic) activity. Besides this, anthropogenic pollution is superimposed on geogenic “pollution”, and so, it would be very difficult to study at a scale of a soil geochemistry grid.

Line 96: would not the soil types also be important.

Done, we have added information about the soil types (mainly Inceptisols).

Line 105: I dont recall seeing much on W in this paper.

Yes, true. But these are separate questions: the AUREOLE project refers to both elements, and we are studying both in other areas. But here in the area there is no presence of W, and our work centers on Sb and locally related elements.

Line 124: How were the samples collected? was there a standard soil horizon? how deep were the holes? are they surface samples?

Done: combined with an answer to Rev. 1, we have added the sentence: “…from the A horizon (0-15 cm. after eventually cleaning the litterfall) of poor developed soils (mainly Inceptisols);” in the last paragraph of 2.1 Sampling section.

Line 132: is this not going to change your si content?

Well, of course it is well known that milling with agate may introduce some Si contamination. But for sure this is a very minor contribution, as compared with the high Silica contents of our samples.

Line 153: so what qa/qc protocol was undertaken? are there field duplicates? any standards inserted, blanks? etc?

Done. A new sentence has been added: “...with recovery ratios between 97 and 100% (ICP-AES). Duplicate samples were taken at random locations within a radius of 500 meters in the same sites. Larger data dispersions are observed at major elements (SiO2, among others), and significant dispersions of some trace elements has been observed, especially Pb, As, S and Zn (>25% Relative standard deviation), which represent local anomalies in the geological materials (Precambrian).”

Line 157: this is good in a non trending area but you have said there is a pronounced fold trend to the region. The maps should have used a trended analysis.

What has not been described is sorting out the lithologicial signal, the anthropogenic signal from the anomolous mineral deposit signal.

Some interesting work has been done in British Columbia sorting out different signals.

https://www.geosciencebc.com/projects/2016-028/

Thanks very much for this information. We will consider it for future research, but at present we do not see it as reasonable to include these AI techniques as an important part of this our study.

Figure 2:

Done. We have changed this denomination using the abbreviation ASQ, corresponding to “alternations of shales and quartzites” in the whole manuscript.

Line 172: not surprising in an agate mill.

As indicated in a previous comment, we do not consider the Si contamination produced by grinding as significative.

Line 184: This hints at something interesting but it is really not expanded on in the examination of the results.

We agree with this comment, but the manuscript is too long and we needed to summarize some questions like this.

Line 184: Could use a bit more explanation why this conclusion works.

Done. A new sentence has been added: “The reference values described by Jimenez-Ballesta et al. (2010) belongs to a more restricted soil samples dataset in terms on sample number but from the entire Castilla-La Mancha region, including siliceous and carbonated materials; therefore, the lower values presented on this work can be consider as baseline values for siliceous units.”

Line 188: This is a confusing sentence. So Pb-Zn are common in the Deposits but not the bedrock? or not exclusively the bedrock?

Done. The sentence has been changed to: “The trace elements of the second group are common in the mineral paragenesis of the Pb-Zn-Ag vein deposits of the Alcudia Valley, mainly hosted in the Precambrian substrate, but with presence also in Almadén and Guadalmez synclines materials.”

Line 202: In this paper, the mobility of elements is important. The pH is an important indigcator of how mobile some of your elements might be. In some cases if you have a alkaline to neutral circumstance, Pb will move differntly than Zn. This may result in the seperation of the Anomolies. Instead the black shales should weather quite acidically and Pb and Zn should move together. There is little discussion regarding the Ec, pH and organic contents. These are important for moving and traping more than just Hg.

We agree with this comment. The approach has various limitations, as it has been briefly explained in the discussion and conclusions sections, exemplified with the behavior of Hg, a volatile element with very different dispersion transfer roots and mobility rates.

Line 203: I am not sure what you are trying to say Sb has a low mobility so the natural dispersion should be small. Instead with due to anthropogenic disturbance there is a broader halo?

The authors have previous experience with Sb dispersion around derelict mines and Sb mobility rates are very low in a semiarid climate, with oxic conditions and negligible role of bacteria in Sb mobilization. But this is the topic of another manuscript already in preprint (https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4165229 )

Line 212: So you are suggesting the list of elements have been leached out of the soils.

No, the authors suggest that differences in mineralogy in parents' materials implies differences in mobility. We have changed the sentence as follows: “This may be due to the predominant lithology of the area, generally deficient in these elements, and their higher mobility rates from parent materials.”

Line 228: Is the silica number valid when you are using an agate mill?

Yes, the cross-contamination can be considered as negligible as Si is a major element, expressed in %.

Figure 4: if the x axis is different between the two synclines it becomes hard to compare the two synclines. Perhaps put both synclines violin plots on the same graph so the differences are more obvious.

Thanks for the suggestion. We have improved the quality of these figures. And we have tried to use the same x axes for the same elements in the two synclines; but for some of them the variability is too marked to get readable violin graphs.

Line 282: Any other elements associated with these deposits? any telurium?

We are sorry, we do not have analytical Te results.

Line 290: Is it showing fractionation?  or a differentiation.

Done. The sentence has been changed to: “The violin graphs also show a contrasting differentiation for this element, with the anomalous values in the quartzite-sandstone levels and the alternations in Guadalmez and in the slates in the Almadén syncline.”

Line 296: what is the mineralogy of the Sb deposits. Sulphosalts might be common.

Done. A sentence has been added in the text: “The primary mineral samples obtained at the La Balanzona mine have been identified by XRD as stibnite (Sb2S3), with bindehimite (Pb2Sb2O6(O,OH)) and stibiconite (Sb3+(Sb5+)2O6(OH)) as the main oxidized phases.”

Line 320: this is not a lithology... give a proper description.

Done. See response to comment of Figure 2.

Line 358: Are there no chemical analyses of bedrock material? Comparing the results to a broad soil data is interesting but why not compare with the material that the soil was derived from. Some of the rocks will have distinct chemistry yes and that can be removed from your soil data set. Then what is left is actually anomalous.

Thanks very much for this suggestion, but we do not have geochemical analysis of bedrock material.

Line 373: the trend that should be used to biase the invese distance plots.

Sorry but we think that introducing a bias in the maps based on lithology/structural trends masks the information provided by the statistical treatment. For this reason, we have selected an interpolation method aimed to visualize factors and not to detect trends (kriging). Our approach is more statistical than geostatistical because we think that scale of sampling grid is a serious limitation to apply geostatistical treatment. These large datasets, like FOREGS or national soil geochemistry research, do not have enough sampling density to apply kriging without creating false areas and even false anomalies.

Line 375: In your factor 4 plot the pb zn distribution does line up best with the Black shale trace. This is not a surprise as black shales can be quiet metalliferous. It is note worthy that they do not line up with the mines as well. This suggest that there are different controls on the anomalous soil results.

Yes, we agree with the reviewer, but the sampling design had the criteria to take baseline samples and to avoid anthropogenic pollution. We think that this could be the reason for this, combined with the sampling density.

Figure 6: most of the antimony mines seem to be outside the elevated factor areas. a lot of the map presented has no data either and should be blanked. including data points would demonstrate the limits of the interpolations.

Yes, this is a fact but is derived by the criteria of sampling design, searching for baseline samples and avoiding mining areas, and for the low mobility of Sb in a semiarid climate, creating small haloes around mineral outcrops.

Line 410: Both factor 4 and factor 5 were poor at predicting the existing mines. Some mines are associated with elevated factors but more are associated with low factor values.

What explains this. Especially as the mines should have contaminated the soils. Given the abundance of Hg, it should have fallen into a more specific factor. to have it tied up in F2 would need to be explained.

Yes, in fact the porpoise of our approach is not to discover the known mines, but to search in baseline elemental contents trends to discover unknown ore deposits. Probably, if we consider structural information, lithological data and geochemical data, and we run a random forest treatment, we could prove that this treatment was working discovering the existing mines. This could be an interesting approach for the next manuscript. Thanks!

With the mixture of elements that were analysed, why are some important elements ie Au not in the data set? Were detection limits too low?

Unfortunately, we do not have data about Au in our datasets and we do not have possibilities to study this.

Since this is meant to point to the best targets in the area. should there not be an image that does identify them?  The title suggest that you are looking at the resource potential of the whole area. But mainly it is an interesting comparison of the two synclines.

We agree, the main limitation of this statistical treatment is the element selection, derived from the analytical technique selected and the elemental package; but we consider that the objectives included in the title and mentioned by the reviewer have been satisfactorily reached with our study.

Author Response File: Author Response.pdf

Reviewer 4 Report

The authors did good research. The manuscript is interesting. Also, the results, English level and writing of manuscript is good.

I suggest acceptance after minor revision.

Some of my comments are in following:

1. Abstract is too long and similar to conclusions. Some same statements are inserted in both parts. I suggest the authours, please revise the abstract part and also, make the text short.

2. Introduction part is too long. I suggest the authours to revise the introduction part significantly and delete the unnecessary parts.

3. The Fig. 4 and Fig. 5 are not clear, the numbers and words are not readable. I suggesr the authors to insert new high quality images for these two figures.

Author Response

Reviewer-4

The authors did good research. The manuscript is interesting. Also, the results, English level and writing of manuscript is good.

Thanks very much for these general comments.

I suggest acceptance after minor revision.

Some of my comments are in following:

1. Abstract is too long and similar to conclusions. Some same statements are inserted in both parts. I suggest the authours, please revise the abstract part and also, make the text short.

We are sorry, we have revised the abstract and we do not find much to eliminate. We consider that this ‘tells the complete story’ of the research, and it is not out of the Journal’s norms.

2. Introduction part is too long. I suggest the authours to revise the introduction part significantly and delete the unnecessary parts.

We have made some minor changes.

3. The Fig. 4 and Fig. 5 are not clear, the numbers and words are not readable. I suggest the authors to insert new high quality images for these two figures.

Done.

Round 2

Reviewer 1 Report

1  Change mean to average as you above reply.

2  Valid numbers are necessary in Tables (some in current version are too numbers).

Author Response

Dear Editor,

These are our responses to the concerns by Reviewer 1:

1  Change mean to average as you above reply.

Done.

2  Valid numbers are necessary in Tables (some in current version are too numbers).

We have simplified numbers to one decimal. In the tables, but also in the text.

We acknowledge these suggestions.