Luminescence Sensitivity of Rhine Valley Loess: Indicators of Source Variability?

Round 1

Reviewer 1 Report

see attached file

Comments for author File: Comments.pdf

Author Response

General comments:

1. The use of sensitivity and how it was calculated need better description in the methods section. We are told that the values come from the signals of the first test dose after measurement of the natural luminescence signal, but what part of that signal is used? The initial signal? Initial minus background? Integration of the whole signal? If initial signal, then we need to know what portion of the signal was considered initial? First 0.5s, first 1s? What was the intensity of the light source? Duration of signal measurement? Portion of the signal used for background subtraction?

We have now provided this additional detail; the light sources are standard for the Risoe measurement equipment used, and for which we cite the published specifications.

 

2. For the poly-mineral fines and k-spar sand samples, which measurement step was used to calculate the sensitivity? The initial low-temperature step, or the post-IR elevated temperature stimulation step? Is there a difference in down-section trends in sensitivity between those two measurements?? It might be best to compare the IR50 measurement step in the processed samples with the unprocessed bulk sediment. The figure legends suggest that a post-IR signal is being used for feldspar sand and poly-mineral fine-silt fractions, while the IR50 is being used for the bulk unprocessed sediment samples.

For post-IR measurements, we integrated the signals from the elevated temperature (290°C) measurements, since both 200°C and 50°C preheats were used and we wished to ensure consistency between those protocols. We have clarified this in the text. The IRSL50 signal was calculated only for the unprocessed samples. We decided against comparing the unprocessed IRSL50 with the preheat-IRSL50 signal from K-spar sands since the size of the aliquots would be different and direct comparison would not be valid.

 

3. Clarify and align the units presented for the sensitivity data. In some plots the x axis is reported as just ‘test dose sensitivity (a.u.)’, elsewhere ‘test dose sensitivity (cnts/Gy/s)’. In the text there are also examples of counts/s/Gy used. Be consistent throughout the text and supplemental data.

We have made all units consistent throughout the text and figures (cts/Gy/s).

 

4. Related to above. I am not sure of the meaning of the use of ‘seconds’ in the units for sensitivity. ‘counts/second’ suggests a rate term, but instead I think the authors are presenting an absolute value of signal per unit dose. Perhaps the ‘seconds’ comes from the use of the first second of the signal decay curve? If background is subtracted, then this ‘time’ term is removed? I think that is the methods section describes the interval used for the calculation of the ‘signal’ (and background), then the ‘seconds’ can be removed from the unit for sensitivity.

This is a valid point and we have removed the time component from the units.

 

 

5. The apparent sensitivity of a sample is not only affected by the inherent luminescence properties of the mineral being analyzed, but also the number of grains contributing to the signal (think of the difference between single-grain and large aliquot signals). The authors are correct in normalizing sensitivity by unit dose given to the samples, and using a similar test dose value for each assessment, but the paper does not state the amount of sediment analyzed for each of the mineral and grain-size fractions. It is expected that more grains will be contributing a signal in the very fine silt and very fine sand fractions in comparison to the fine sand fractions. I do note that on line 136 the paper states that ‘small aliquots of 50-100 grains’ were analyzed for the quartz very fine (63-100 μm) and fine (100-200 μm) sand fractions from the RP1 profile. Is it correct to assume that the finer grains had the larger number of grains mounted on each disc? This difference in number of contributing grains should be accounted for in the sensitivity calculations.

This is a very valid point and one which unfortunately cannot be rectified, since the aliquots measured have since been disposed of. We have now included mention of this as a limitation of the study, although our aliquot-loading method is assumed to produce uniform masses of grains.

 

6. A similar correction/normalization should be made for the very fine polymineral silt (4-11 μm) and unprocessed samples – OR to keep comparisons of these less easy to quantify samples to themselves (apples and apples). As these samples make up a majority of the data presented, it should be NOTED that differences in the magnitude of sensitivity between the coarser grained samples and the fine-grained polymineral silt and unprocessed samples are influenced by the number of contributing grains to the signal.

We agree with this statement and since it is not possible to normalize the sensitivity with regard to aliquot mass, we have now noted that the differences in absolute sensitivity are influenced by the number of contributing grains to the signal.

 

7. Related to above and thinking about the use on ‘unprocessed’ sediment, Im curious if there is any relationship of the IR50 signals from the unprocessed samples and the sediment grain-size of the loess/soil samples within the profile? Relationship to soil carbonate (inorganic carbon_ content (assuming it might dilute the mineral volume in carbonate soil horizons)?

We do not observe any relationship between IR50 and soil carbonate, and have stated this in the discussion; the grain-size data available to us prevents more detailed comparison.

 

The % clay content is presented in Fig 4 and 6, but the values are relatively constant between 10% and 15% throughout the core in the intervals of interest. Clay (<2 microns) might be assumed to contribute little luminescence signal if it is composed of authogenic clay minerals and not small mineral fractions of quartz and feldspar. Is the composition of this clay fraction known?

No, we have not been able to analyse the composition of the clay; it is unclear if it contains solely clay minerals or clay-sized quartz and feldspar grains.

 

What about the other grain-sizes that will likely contribute to the luminescence signal from an unprocessed sample? It appears that this loess section contains silt and sand fractions to allow processing and analysis of sediment up to 200 μm. It would be assumed that unprocessed samples from coarser/finer intervals would have different volumes of coarser/finer sediment contributing to the bulk luminescence signal. Making corrections for this change in grain size could be difficult, but it seems like an important variable when discussing changes in bulk sediment feldspar sensitivity. Other studies have used changes in grain-size distribution in loess sequences to suggest changes in wind strength or sediment source changes, so this is another important variable to consider (e.g. Peng et al 2016, Quaternary Research 85, 290-298 and Ben-Israel et al. 2015. Quaternary Research 83, 105-115 as the first couple hits on a Google search).

This is a very good point. Unfortunately, since we do not have suitable grain-size data with which to compare, we cannot investigate this idea further at this stage. However, we will certainly bear in mind the suggestion for future work!

 

8. Linkage of Fld sensitivity to provenance is difficult to assess because of the clear to primary loess vs soil horizons. The paper points out that the bulk IR50 data produces a similar trend to Si/Al ratio data from the same core and suggest that this is a more established indicator of source change and ref #83 (Vinnepand et al in review) is cited – Please cite published paper that establish this proxy as an established tool for provenance. I have not heard of this provenance indicator before and a brief Google Search found the Profe etal 2016 paper cited in the manuscript where Geochem work on Schwalbenberg loess was used to indicate weathering and source area changes. The Profe et al. paper suggests that Si/Al is not a clear proxy for source area and it can be affected by grain-size changes. Other source-area indexes are available for the core site, why not provide those data for comparison? IF Si/Al is kept as in the paper, then provide citations for support of its use as a established source-area indicator. Alternately, if Si/Al ratios provide additional information about grain size or weathering, those inferences should be discussed with relationship to how they may relate to changing IR 50 sensitivity.

We appreciate this comment and would like to state that this was a miscommunication on the part of the co-authors. We have cited the relevant reference and provided a more nuanced explanation for what this ratio can be taken to mean.

 

9. The relationship between bulk IR50 sensitivity and soil/loess units may reflect changes in source sediment, but given the nature of the clay-rich soils it is more likely that in-situ weathering that produced the clay to form the B horizons of these soils was directly linked to weathering of feldspars, leading to reduced feldspar content, or perhaps their decreased sensitivity. Given its significance, the influence of weathering should be discussed in the paper. Other evidence for weathering and provenance change that can support the current assessment of this paper should be presented and folded into the discussion. In the end, this is a pilot project, so it is fine if some uncertainty and questions remain – it just provides fuel for further research looking into Qtz and feldspar sensitivity as an indicator of source area and/or weathering.

In light of our new statistical analyses, we have restructured the discussion to show that weathering, source change, or both are possible influences on the IR50 sensitivity which we observe. There remain a lot of questions and there is plenty of scope for future research.

 

10. This paper ends without a Conclusions section. It would be great if the main points, observations and interpretations of the paper are summarized in a short Conclusions sections.

We have now included a concluding paragraph as a separate section.

 

Specific comments:

Line 32 – the use of ‘downslope’ here is confusing without additional information. I would add ‘downslope of the core site’ to clarify.

We have changed the wording of this sentence.

Line 33-35 – Im not sure that one can say that there is an inverse relationship between the Qtz and Feld sensitivities given the limited number of Qtz data and large uncertainty in that data. I would skip making this comment in the Abstract and stick to the main observations and interpretations of the paper linked to changes within soil intervals and potential links to source sediment changes

We agree and have removed statements about the quartz results, particularly since there are insufficient data points.

Line 35-36 – here the Abst states that Fld sensitivity INCREASES during periods of soil formation, but the data suggest the opposite. Please check thoughout the paper that this observation and interpretation are clearly presented and correctly stated, as I had thought I saw a similar statement about higher sensitivity in the soils elsewhere too.

The statement was incorrect and has been amended

Line 37-38 – please clarify which ‘more established indicators of provenance’ you are referring to. Note that I would recommend not basing this interpretation on just one index (Si/Al) and mention affects of weathering (soils) and grain-size on all proxies (source area, weathering and IR sensitivity)

We have edited the abstract substantially to reflect the data more accurately – particularly in light of the statistical analyses

Line 45 – is ‘permeable’ an important descriptor for loess for this paper?

Removed

Line 108 – perhaps ‘geochemistry’ should be used here instead of chemistry, which brings aqueous reactions to my mind. Geochemistry more broadly reflects the chemical and isotopic make-up of the sediments.

Amended

Line 136 – here it is stated that small aliquots of quartz sand were used for sensitivity and De measurements for samples from the RP1 site. The aliquot size should also be provided for the feldspar fractions from samples from both sites and the quartz fractions from the REM3B core. This is important for normalizing the sensitivity between sites/samples.

Included, but this had to be an estimate for the reasons detailed above

Line 152 – see comment above about providing more detail on how the sensitivity signals were calculated, intensity of the stimulating light source, duration of stimulation, what part of the signal was used and if background was subtracted. Discuss normalization between coarse grained aliquots and differences with fine-grained and bulk samples leading to non-uniform comparison between different data types.

The light source is described in the references cited on the measurement equipment. We have now included the detail on the signal integration and background subtraction.

Lines 168, 181-182 and throughout – There is some confusion in the text and figures as to what grain-sizes were used for different minerals and locations. On line 168 it says that 63-100 μm sediment processed to purify K-feldspar from the REM3B core, but on line 181 it mentions 63-90 μm, then on the next line it mentions 63-100 μm. I know 10 microns is not much, but it would be good to keep the text descriptions consistent. Legends on figures suggest 63-100 μm was used for the RP1 profile and 63-90 was used for the REM3B core, this should be confirmed.

The latter observation is correct – 63-100 was measured on the RP1 profile because of sieving equipment available in Mannheim lab (measured before the Mainz lab was set up), whereas 63-90 was measured on the REM3B samples. We have endeavoured to clarify this in the text.

Line 168-169 – the density of heavy liquid used for k-spar separation should be stated – this along with other details can be put in the supplemental material, which currently only have extra figures and no added text on methods details.

We have cited references describing the processing approaches used.

Line 169 – the mention of the difference in minerals analyzed in the fine-sand vs the fine-silt fraction is an important point and should be brought up again when discussing the sensitivity results. I assume that the same comment would apply to silt and bulk sample data from the REM3B core.

Correct

Lines 209-211 – provide more detail on what the ‘signal’ was, how calculated (background subtraction?), aliquot sizes for each mineral/grainsize and normalization. For the feldspar and polymineral samples – was the sensitivity calculated from the initial IR measurement, or the post-IR step?

We have provided more detail

Lines 223-223 – cite a figure to direct the reader to where they can see that sensitivity from the pIR200/290 method produced higher scatter than with lower first temp step.

We have no access to these data due to an unfortunate lack of communication with one of the co-authors. We hope very much to get hold of it in time for the publication of the complete age dataset from the core.

Line 260-264 – discussion here about the higher De values for the k-spar IR results compared to the Qtz De values should include acknowledgement that this is expected – due to higher expected dose rates for k-feldspar due to internal dose from potassium. The 50% higher De values are in line with a similar magnitude for increased dose rate for feldspar over quartz.

Now included

Line 265 – it should be noted that inverted qtz De values at ~4.5m depth in the REM3B core could also be due to changes in dose rate at that depth, which would affect the De but could retain a correct stratigraphic age-depth profile. These data also have large uncertainties and all qtz De data from 4-7m depth are within error of each other. I suggest not giving inversions in De values much discussion due to the uncertainty in the data and possible influences on dose rate.

This is a valid observation, and we infer the reason to be because of chemical (and therefore dose rate) changes due to the presence of the Eltville tephra at this depth.

Line 308-309 – Im not sure the comment that minimal variation in the sensitivity from the pIR data from can be supported given the limited number of samples analyzed and large stratigraphic gaps between the samples. This comment was directed at the data shown in figure 3, but the same can be said for the data presented in Figure 4. This difference in the sampling resolution should be noted if this comment remains.

We agree, and have removed unnecessary discussion on the pIR results in light of the more robust statistical analysis on the IR50 samples.

Line 309-310 – change wording here to describe as the sensitivities are ‘higher or lower’, not increase or decrease, which suggests a trajectory of change the way it is currently worded. Note that this same observation with regard to sensitivity in soils and loess units cannot be assessed with the Qtz or processed feldspar samples due to sample bias to only the most loess-rich intervals (not soil horizons).

Corrected

Line 328-329 – this observation of sensitivity being higher in the loess units and lower in the clay-rich soil intervals should be followed by a comment regarding possible dilution of the luminescence-deriving feldspars with increased clay content. Weathering and removal of k-spar in the soil horizons is also a possibility as it is rapidly broken down to clay during chemical weathering.

This is certainly possibly the case and we discuss it now extensively in the revised discussion

Lines 332-340 – this paragraph is somewhat repetitive with previously stated information in the geologic setting section. I would move any new information to section 2.1 Sample site.

Moved

Line 343-344 – information on what is plotted in Fig 4 should be included in the caption for that figure. It can be restated here for clarity, but the reader should be able to determine that the data are reported as the median and one standard deviation by looking at the caption (or legend).

We have clarified this

Line 365 – I would use geochemistry here instead of chemistry

Corrected

Lines 395-396 – it seems that sensitivity changes due to dilution of feldspar (from greater percentage of clay) or reduced feldspar content (from weathering in a soil horizon), especially within soils characterized by illuvial clay sourced from weathering should be mentioned here.

Line 427 – the use of ‘partly stagnant conditions’ is a bit oddly wording and confusing. Perhaps you mean ‘regions under of standing water’?

Line 448 – the description of the ‘threshold for IR50 bulk sensitivity’ is a bit confusing here. Why not just state what the threshold value is that you are picking?? One Fig 6 it looks like the red vs blue shading is switching back and forth from the +1-sigma to the -1-sigma side of this threshold, why make this arbitrary call. Instead, I think that the plot would be just as clear and illustrative if the red-vs-blue shading followed the mean crossing the chosen threshold. Remind the reader the number of data points used to create the bulk IR50 data and plot the 1-sigma lines to provide guidance to the uncertainty of the data.

Lines 455-456 – the similarity in the trend with the IR50 and Si/Al data is visually apparent in some portions of the core. Why not provide a statistical assessment of this, since this leads to the conclusion of the paper related to sediment source differences? There are other proxies available for this core, do other data support the sediment source conclusions? Alternatively support a link to weatherings?

460-462 – I do not follow the reasoning in the comment that the fluvial sediments are not affected by pedogenesis and still have low sensitivity. Im not sure the comment about the sensitivity of the fluvial sediments not being affected by pedogensis supports the statement being made. I would think that the fluvial sediments would display a different source area than what might be the more far travelled loess sediments and if bulk IR50 sensitivity is an indicator of sediment source that is might support the main conclusions of the data… if the results are normalized based on grain size and number of grains contributing to the bulk signal.

Lines 476-481 – here the inverse relationship between bulk IR50 sensitivity and clay content is pointed out. Given that the clay-rich intervals are soils, and pedogenic clay accumulation is through weathering largely of feldspars through hydrolysis, there should be mention of the potential influence of weathering on the sensitivity data. A closer look at the mineral content of the clay-rich soil horizons could help determine if feldspars are being removed.

Regarding the comments relating to lines 395-481: We have borne these suggestions in mind with our restructuring and revised emphasis of the discussion. The individual comments cannot be directly corrected for, but are addressed.

Line 482 – use of ‘anti-correlation’ is odd and confusing – do you mean inverse correlation, or no correlation?

Correct – inverse correlation. We have corrected this term.

 

Comments on Figures

Figure 2 – I have a hard time seeing the difference in the green and orange to brown colors. There is important information in these data but the choice of similar colors makes these differences difficult to see and therefore discussions in the text difficult to follow. I suggest using more dramatic color differences, perhaps even different symbol shapes to separate out the Qtz, k-spar and bulk sediment data.

This is a good suggestion and we have changed the symbol shapes accordingly.

 

Figure 3 – update units for the sensitivity data (don’t use a.u if you have real units). Consider using updated colors/symbols for the sensitivity data to improve identifying the different data types – use same as chosen for an updated Fig 2. Question: Why plot the quartz sensitivity here as box plots? I would keep systematic between figures. It is hard to see the ‘box’ in these plots due to how skinny they are when plotted on the x-axis.

We have now updated the units and changed the shapes of all symbol categories to be consistent between the figures. We have also now presented all sensitivity data in the same way.

 

Figure 4 – Provide information about how the sensitivity data are shown in the caption or legend (median and standard deviation). The Blue and green dots for the Qtz data and darker data points for the feldspar data look to similar for me to clearly pick them out on the figure – increase color difference, perhaps change symbols – keep consistent with Fig 2 and 3. Update the units for the test dose sensitivity (note currently listed as cts/Gy/s while elsewhere cts/s/Gy). Im not sure seconds is needed here, consider normalizing for aliquot size (number of contributing grains)

We have now included information about the presentation of sensitivity data, changed the symbols for the different measurement protocols and size fractions, and corrected the sensitivity units.

 

Figure 5 – It is not clear what the blue arrows represent, or if they are pointing out specific features. Update the units of the sensitivity data (do not use a.u. without describing why). Describe what is plotted on the right side of panel A – why is the inner bar so thick, is this a box-plot? I would stick to the same plotting style as used for the left side (you can keep as a linear scale if you want). Panel B: Why is the test-dose sensitivity now plotted as a connected line now? IT seems best to keep the individual data points and uncertainty/spread here, you can keep the line over those data to show the trends you are pointing out. Blue horizontal bar matches the gelic soil but the yellow bar only aligns with the lower part of the overlying loess section, why?

We have clarified the meaning of the blue arrows in the caption, and corrected the sensitivity units.

We have removed Panel B now since we have redirected our focus to more robust statistical, rather than visual, correlations. The inner bar is the thickness of the soil package, which is thicker in RP1 than in the core REM3.

 

Figure 6. This is an important figure and summarizes a lot of important information and related data from the site/region. The lines plotted with Sensitivity data are hard to see. The orange solid and dotted lines (mean and +/- 1 SD) blend in with the red fill making it hard to see. I suggest using black line color

here. I also suggest using the mean of the data to pick the fill color based on it crossing the chosen ‘threshold’ value and then having the + and – 1-sigma value dotted line fall outside/inside the colored portion of the curve. The figure caption should state what the ‘threshold’ is for plotting the curve fill as red vs blue. The figure caption should also reference the reader to Fig 4 to see the legend for the colors presented in the stratigraphic column on the far left.

We have made all suggested corrections.

 

 

Reviewer 2 Report

This study presents the first steps towards exploring the potential of feldspar luminescence sensitivity as possible indicators of provenance change within sedimentary sequences on the example of German loess/paleosol sequences. The topic is of high interest as the application of such an approach is pioneering in European loess sequences and as such well worth of acceptance and publication in the journal. The paper is well-written and recommendable for publication in the present form with some minor add-ons. I just wish to make some notes that may help improve the manuscript and make some questions, problems raised in the paper more clearcut for readers.

Using different symbols (not just filled circles) besides coloring would be great for different measurement types on all figures.

Lines 267-269 two of three samples are mentioned in the text in connection with the figure, however, it is hard to see, which too. So perhaps assigning some numbering to the referred samples on the figure would help.

Lines 307-309 "We observe minimal 307 variation in the sensitivities derived from the samples measured using pIR200IR290. By contrast, there 308 is significant variability in the IR50 measurements"- considering the width of the error bars IR50 values seem to have lower variability compared to IR290 values as in 2 cases the error bars are quite large (except for the middle sample) If you consider the mean values to be relevant the TDS values of the IR290 samples remain relatively constant (at least in sample 2 and three),  while in case of the IR 50 samples there is a high variability.

Lines 309-310 " On the whole, IR50 sensitivities increase within the 309 primary loess units, and decrease in the more clay-rich, pedogenic horizons."-it is true that values are lower in the clay-rich horizons compared to the loess. But within the loess, the general increase in sensitivity is not valid. There are parts where there is an increase (lowermost loess in unit D and the lowermost part of the topmost loess in unit E) but in the rest, we see a decreasing trend. As notable in the figure, the lower values in loess seems to appear in areas where there are smaller or larger peaks in the clay (so there is a higher clay content) similarly to the paleosols. It is perhaps worth mentioning. Nevertheless, the reasoning seems correct, that wherever you have lower clay content sensitivity increases. Where you have higher clay content sensitivity decreases.

Lines 328-330: same issue as in the previous comment. It might be worth adding the information that there is a downward increase in the clay content with which sensitivity decreases parallelly

Lines 351-354: As I see it there is a downward decrease in sensitivity in unit E in general parallel with a general overall increase in the clay content just like in Fig.3. To me, minor variations seem to be again correlated with peaks in clay. The same correlation is observable in the topmost loess at 14 m in Unit D, clay content decreases signal sensitivity increases. Also in the loess at 20 m in Unit C. Even in the case of the paleosols sensitivity values appear to be higher in areas with lower clay content (e.g. 14-15m, 16.5 m, 22 m). Perhaps adding this to the referred part would be beneficial

Would be great to include a plot of the clay content of the two profiles on Fig5A too if available to see if subtle changes here too correlate with the clay content. Also, some grain-size proxy perhaps and an accumulation rate inferred from 14C data

Are the observed sensitivity changes truly an artefact of dust source changes or changes in the degree of weathering, perhaps the rate of dust input to the site? There is a negative correlation with the clay content-high lower sensitivity, low higher sensitivity. It would be nice to add perhaps one or two grain-size proxies (especially those free of pedogenic clay) to see if this correlation is positive or negative with them also an accumulation rate curve based on 14C data if available.

Detailed suggestions:

24 in the source 25 index that is the product of the interplay 32 the comparison 37 show a correlation 99 the comparison 103 focuses 121 to the REM3A 122-123 cores spanning the  150 a signal 161 were derived 162 study since 187 on the processing 192 We, therefore, 217 We, therefore, 262 bleaching or 263 However, 266 Unfortunately, 279 Consequently,  282 fact that the quartz ( it had two "the" before the quartz) 284 a/the medium component 289 and luminescence and 306 sensitivities since 310 units and 354 units and 374 exception and 387-388 We, therefore, 295 the source 425 Furthermore, 438 Likewise, 440 Nevertheless, 462 Nevertheless,

Author Response

Comments and Suggestions for Authors

This study presents the first steps towards exploring the potential of feldspar luminescence sensitivity as possible indicators of provenance change within sedimentary sequences on the example of German loess/paleosol sequences. The topic is of high interest as the application of such an approach is pioneering in European loess sequences and as such well worth of acceptance and publication in the journal. The paper is well-written and recommendable for publication in the present form with some minor add-ons. I just wish to make some notes that may help improve the manuscript and make some questions, problems raised in the paper more clearcut for readers.

We thank this reviewer for their positive comments.

 

Using different symbols (not just filled circles) besides coloring would be great for different measurement types on all figures.

We have amended the figures accordingly.

 

Lines 267-269 two of three samples are mentioned in the text in connection with the figure, however, it is hard to see, which too. So perhaps assigning some numbering to the referred samples on the figure would help.

We have added a box around the two samples referred to, and have made a reference to this in the text.

 

Lines 307-309 "We observe minimal 307 variation in the sensitivities derived from the samples measured using pIR200IR290. By contrast, there 308 is significant variability in the IR50 measurements"- considering the width of the error bars IR50 values seem to have lower variability compared to IR290 values as in 2 cases the error bars are quite large (except for the middle sample) If you consider the mean values to be relevant the TDS values of the IR290 samples remain relatively constant (at least in sample 2 and three),  while in case of the IR 50 samples there is a high variability.

This may be an oversight on the side of the reviewer here. We used a log10 scale to present the sensitivities in the REM3B samples here, which leads to an appearance that there is less variability in the IR50 than might initially appear. Their observation, which supports our statement, is nevertheless valid: that the IR50 is substantially more variable than the pIR290.

 

Lines 309-310 " On the whole, IR50 sensitivities increase within the 309 primary loess units, and decrease in the more clay-rich, pedogenic horizons."-it is true that values are lower in the clay-rich horizons compared to the loess. But within the loess, the general increase in sensitivity is not valid. There are parts where there is an increase (lowermost loess in unit D and the lowermost part of the topmost loess in unit E) but in the rest, we see a decreasing trend. As notable in the figure, the lower values in loess seems to appear in areas where there are smaller or larger peaks in the clay (so there is a higher clay content) similarly to the paleosols. It is perhaps worth mentioning. Nevertheless, the reasoning seems correct, that wherever you have lower clay content sensitivity increases. Where you have higher clay content sensitivity decreases.

We agree that there is some flexibility in the interpretation here, and we agree that this could be more precise. To that end we have now included statistical analysis comparing the degree of correlation between clay percentage and IR50 sensitivity, which makes a clearer point.

 

Lines 328-330: same issue as in the previous comment. It might be worth adding the information that there is a downward increase in the clay content with which sensitivity decreases parallelly

See above – we believe that the statistical analysis addresses this concern now.

 

Lines 351-354: As I see it there is a downward decrease in sensitivity in unit E in general parallel with a general overall increase in the clay content just like in Fig.3. To me, minor variations seem to be again correlated with peaks in clay. The same correlation is observable in the topmost loess at 14 m in Unit D, clay content decreases signal sensitivity increases. Also in the loess at 20 m in Unit C. Even in the case of the paleosols sensitivity values appear to be higher in areas with lower clay content (e.g. 14-15m, 16.5 m, 22 m). Perhaps adding this to the referred part would be beneficial

We agree that the observations made by the reviewer here appear to be  the case, and believe that the best way to investigate this more objectively is through the statistical analyses of correlations, now added to the revised manuscript.

 

Would be great to include a plot of the clay content of the two profiles on Fig5A too if available to see if subtle changes here too correlate with the clay content. Also, some grain-size proxy perhaps and an accumulation rate inferred from 14C data

We did compare the grain-size data run in Fischer et al. 2021 to our data and found there to be little correlation with our trends. In part this may be due to the range of mechanisms influencing grain size variability which haven’t yet been fully investigated; whereas here, we are hoping to test for one parameter only (change in source through time). To that end, to avoid unnecessary complexity in our empirical investigation, we chose to work with the clay percentage (as an indicator of weathering/soil formation), potassium concentration (as an indicator of clay and feldspar concentration) and Si/Al (discussed in Fischer et al. as an indicator of source change). We nevertheless contend that the statistical analyses added in the revision strengthen our discussions.

As yet, accumulation rates based on the 14C dates for the profile RP1 have not been produced. However we do now include the dates, published in Fischer et al. 2021, in a revised version of Figure 3.

 

Are the observed sensitivity changes truly an artefact of dust source changes or changes in the degree of weathering, perhaps the rate of dust input to the site? There is a negative correlation with the clay content-high lower sensitivity, low higher sensitivity. It would be nice to add perhaps one or two grain-size proxies (especially those free of pedogenic clay) to see if this correlation is positive or negative with them also an accumulation rate curve based on 14C data if available.

We believe that our new statistical regressions testing the degree of correlation between the clay (as weathering indicator) and Si/Al (as source change indicator) provide the most robust possible test of whether the IR50 sensitivity is an artefact of dust source or weathering. Unfortunately as yet we don’t have 14C-based accumulation rates, nor a suitable, reasonably fool-proof grain size proxy curve with which to test those correlations.

 

Detailed suggestions:

24 in the source 25 index that is the product of the interplay 32 the comparison 37 show a correlation 99 the comparison 103 focuses 121 to the REM3A 122-123 cores spanning the  150 a signal 161 were derived 162 study since 187 on the processing 192 We, therefore, 217 We, therefore, 262 bleaching or 263 However, 266 Unfortunately, 279 Consequently,  282 fact that the quartz ( it had two "the" before the quartz) 284 a/the medium component 289 and luminescence and 306 sensitivities since 310 units and 354 units and 374 exception and 387-388 We, therefore, 295 the source 425 Furthermore, 438 Likewise, 440 Nevertheless, 462 Nevertheless,

Unfortunately the meaning of these suggestions was lost in our version of the reviewers’ comments. Perhaps this refers to suggestions to clarify the text; we have now reviewed the text and edited it for clarity where relevant.

Reviewer 3 Report

Thank the editor for inviting me to evaluate the manuscript – “Luminescence sensitivity of Rhine valley loess: indicators of source variability?”.

In this paper the authors investigate luminescence sensitivity in loess-palaeosol sequences in the Schwalbenberg loess profiles (Germany Rhine valley) as possible tool for identifying changes in source. They find an overall inverse relationship between quartz and feldspar sensitivity, as well as variability in sensitivity between different quartz grain sizes. Thy observed increasing of feldspar sensitivity during periods of soil formation down the core. The authors believe that measurements of IR50 sensitivity on unprocessed sediments may suggest source variability over millennial timescales.

I have read carefully the manuscript that I found interesting. However, as I am not a specialist in luminescence research, not everything was clear to me. Also for the authors, the results of their experimental study were not fully conclusive. The article is clearly methodological in nature, and the analysis of the results in a pelaeogeographical context is somewhat lacking. For this it is necessary to take into account several important items of literature, indicated below.

I hope that the authors will soon confirm their preliminary results in the study of further loess profiles.

In conclusion, I think that the article can be published, despite the comments. It will certainly accelerate the process of verification of the proposed results.

Missing literature:

Antoine, P., Rouseau, D-D., Moine, O., Kunesch, S., Hatte, Ch., Lang, A., Tissoux, H., Zöller, L., 2009. Rapid and cyclic aeolian deposition during the Last Glacial in European loess: A high-resolution record from Nussloch, Germany. Quat. Sci. Rev. 28, 2955–2973. 

Antoine, P., Rouseau, D-D., Degeai, J-P., Moine, O., Lagroix, F., Kreutzer S., Fuchs, M., Hatte, Ch., Gauthier, C., Svoboda, J., Lisa, L., 2013. High-resolution record of the environmental response to climatic variations during the Last Interglacial-Glacial cycle in Central Europe: The loess-palaeosol sequence of Dolní Vestonice (Czech Republic). Quat. Sci. Rev. 67, 17–38.
Moine, O., Antoine, P., Hatté, Ch., Landais, A., Mathieu, J., Prud’homme, C., Rousseau, D-D., 2017. The impact of Last Glacial climate variability in west-European loess revealed by radiocarbon dating of fossil earthworm granules. PNAS 114(24), 6209-6214.

Pańczyk, M., Nawrocki, J., Bogucki, A.B., Gozhik, P., Łanczont, M., 2020. Possible sources and transport pathways of loess deposited in Poland and Ukraine from detrital zircon U-Pb age spectra. Aeol. Res. 45, 100598

Rousseau, D.D., Chauvel, C., Sima, A., Hatté, Ch., Lagroix, F., Antoine, P., Balkanski, Y., Fuchs, M., Mellett, C., Kageyama, M., Ramstein, G., Lang, A., 2014. European glacial dust deposits: Geochemical constraints on atmospheric dust cycle modeling. Geophys. Res. Lett., 41, 7666–7674.

Author Response

Comments and Suggestions for Authors

Thank the editor for inviting me to evaluate the manuscript – “Luminescence sensitivity of Rhine valley loess: indicators of source variability?”. In this paper the authors investigate luminescence sensitivity in loess-palaeosol sequences in the Schwalbenberg loess profiles (Germany Rhine valley) as possible tool for identifying changes in source. They find an overall inverse relationship between quartz and feldspar sensitivity, as well as variability in sensitivity between different quartz grain sizes. Thy observed increasing of feldspar sensitivity during periods of soil formation down the core. The authors believe that measurements of IR50 sensitivity on unprocessed sediments may suggest source variability over millennial timescales.

I have read carefully the manuscript that I found interesting. However, as I am not a specialist in luminescence research, not everything was clear to me. Also for the authors, the results of their experimental study were not fully conclusive. The article is clearly methodological in nature, and the analysis of the results in a pelaeogeographical context is somewhat lacking. For this it is necessary to take into account several important items of literature, indicated below.

We appreciate the suggestion by this reviewer to include the literature items listed. However, we would like to note that the focus of this manuscript was not on the palaeogeographic context, which is investigated in great depth in the paper by Fischer et al. (2021; Catena), and which indeed cites most of the above, but rather on whether we can use various luminescence signals to indicate changes in source. As an empirical study, we (and our colleagues in the field) need to undertake more detailed investigations of the possible mechanisms affecting feldspar signals, which are more complicated than quartz since the former is a solid solution. The approach is therefore not yet so developed as, for example, the detrital zircon and geochemical methods cited above. We have reviewed our reference list in light of the suggestions but prefer to retain it as it is (73 references).

 

I hope that the authors will soon confirm their preliminary results in the study of further loess profiles.

This certainly is part of our plans for the coming years, thank you!

Reviewer 4 Report

I was asked to review the work entitled “Luminescence sensitivity of Rhine valley loess: indicators of source variability?”, by Kathryn E. Fitzsimmons, Zoran Perić, Maike Nowatzki, Susanne Lindauer, Mathias Vinnepand, Charlotte Prud’homme, Aditi K. Dave, Andreas Vött  and Peter Fischer.

The article focused on variations in the luminescence parameters (mainly luminescence sensitivity) of quartz and feldspar grains in loess-palaeosol sequences (LPS) as possible proxy for changes in source of deposits.

Line-by-line comments are in the attached file; below, I would like to remark my overall impression, also proposing some general comment and suggestion. Of course, authors may, or may not, agree with these notes.  

- The work considers LPS in a deeply investigated area, previously subject of several studies, some led on by the same authors, however proposing a large unpublished dataset (supplementary materials), an interesting approach, and a provocative title. As far as I was able to verify, the paper proposes new elements, integrating what has already been published. The former data are used, in the intention of the authors at least, exclusively for comparison. In this way, the work proposes the progresses of an ongoing research. The expectations created in the reader, however, are progressively disappointed in the text, up to the discussions, where the initial provocation remains more a working hypothesis than an interpretation supported by the results. This, unfortunately, is in the words of the authors, who do not seem to take a clear direction (e.g.: lines 422-442).

- I am not a native English, so I am not the right person to propose revision of language and style. I do not like direct first-person speech ( “we made”, “we describe” etc.) in a scientific paper, and indirect form in sentences should be preferable. Anyway, this is just a matter of personal preferences. The English is plain and readable; the reading difficulties depend on the organization and choices in the presentation of the data rather than on language and style. The reader gets lost between new and old data, sections lacking stratigraphic constraints, basic information dispersed through the text, detailed analytical protocols, and so on. Please consider to re-organize the text.

- Starting with “Introduction” and “Sampling site” sections (the latter should be better part of Introduction or Geological Settings), it is not always clear where are the new data, and what already derives from literature (e.g.: lines 127, 135, 400). The reader who is unfamiliar with the study area, and with the former literature, has to continually scroll the text back and forth in order not to get lost and find all the information. This becomes more marked in Results, and also reflects in organization of figures. In Figure 1 (inspired to Figure 1 in [54])  all the sites mentioned or analyzed in the text should be reported, even if just for comparison. A figure showing the correlation between the different stratigraphic columns, and with the reference succession, would also be useful, proposing all the useful information that is currently dispersed in the text, even suddenly introduced (e.g.: lines 119-120, 208, 245, 400). This would allow to focus attention in the following figures on the new data and their position through the REM3b core section (or the composite REM column as well). All general information should still be provided from the beginning, even with a section dedicated to the description of the site (2.1), and not introduced directly into Results or Discussion. Finally, methods, results and discussion are not always clearly separated, and some sentences should be moved through the text, or rephrased.

- The methodological part is in my opinion too extensive and detailed, and hinders reading a bit. The reader has the sensation of suddenly plunging into a methodological work. For those who are not very familiar with the techniques used in this research, the results obtained and their implications are more relevant. On the other hand, excessive methodological detail is superfluous for experts. As reference is made to analytical protocols widely used and codified in the literature, this part could be significantly reduced. Otherwise, methods should be detailed in supplementary materials.

- In the discussion (I refer particularly to paragraph 4.2), attention is not adequately focused on what (according to the title) should be the central topic of the article (e.g. the use of luminescence sensitivity). There is a feeling that the authors are the first to be convinced of the interpretation provided, as if it were the reader who had to convince himself. I would have expected solid arguments, or at least a discussion of the deductions supporting the use of the method as a source proxy, and an overview of alternative options that make that choice the most satisfactory. In my opinion, this part should be reconsidered and enlarged.

- Please check Figures, particularly colors and legend.  Also the order can be modified: for example, Figures 3 and 4 should be probably inverted.

I think the paper could be considered for publication in Quaternary, but it cannot be accepted in its present form. In my opinion, the manuscript still needs some work by the aothors before publication.

I hope my comments could be somehow helpful, anyway.

Best Regards

Comments for author File: Comments.pdf

Author Response

Comments and Suggestions for Authors

I was asked to review the work entitled “Luminescence sensitivity of Rhine valley loess: indicators of source variability?”, by Kathryn E. Fitzsimmons, Zoran Perić, Maike Nowatzki, Susanne Lindauer, Mathias Vinnepand, Charlotte Prud’homme, Aditi K. Dave, Andreas Vött  and Peter Fischer. The article focused on variations in the luminescence parameters (mainly luminescence sensitivity) of quartz and feldspar grains in loess-palaeosol sequences (LPS) as possible proxy for changes in source of deposits. Line-by-line comments are in the attached file; below, I would like to remark my overall impression, also proposing some general comment and suggestion. Of course, authors may, or may not, agree with these notes.  

Many thanks; we have responded to individual comments in detail in the accompanying PDF.

 

- The work considers LPS in a deeply investigated area, previously subject of several studies, some led on by the same authors, however proposing a large unpublished dataset (supplementary materials), an interesting approach, and a provocative title. As far as I was able to verify, the paper proposes new elements, integrating what has already been published. The former data are used, in the intention of the authors at least, exclusively for comparison. In this way, the work proposes the progresses of an ongoing research. The expectations created in the reader, however, are progressively disappointed in the text, up to the discussions, where the initial provocation remains more a working hypothesis than an interpretation supported by the results. This, unfortunately, is in the words of the authors, who do not seem to take a clear direction (e.g.: lines 422-442).

We disagree that we raise false expectations in the reader and suggest only a working hypothesis in the manuscript. The question mark in the title is intended to introduce a new concept/application of luminescence to the field, and suggests that we explore this concept in the study. This, we have done. From the dataset which we produce, the best resolution derives from the higher sampling resolution measurements of the IR50 signal down the REM3B core. We explore the potential of that signal more thoroughly since it is very rapidly measured and, if found to be of promise, could routinely be applied to any down-profile situations to investigate source variability. Our IR50 results are now supported by multi-linear regressional statistics against the stratigraphy, clay percentage (largely an indicator of weathering) and Si/Al ratio (taken as an indicator of source change). We find that the IR50 signal correlates with Si/Al to a certain degree – therefore it is conceivable that this signal is reflecting changes in source through time. However, there is also correlation with the stratigraphy and clay percentage, and therefore we cannot rule out the possibility that post-depositional weathering and alteration of the feldspar crystal structures affected the IR50 signal. We are transparent about these potential uncertainties, and about the fact that this is an empirical study which cannot provide insight into the mechanisms driving IR50 sensitivity. However, our results are compelling enough and we very clearly explore the statement in the title, even if, unfortunately, we could not conclusively provide a new provenance tool.

 

- I am not a native English, so I am not the right person to propose revision of language and style. I do not like direct first-person speech ( “we made”, “we describe” etc.) in a scientific paper, and indirect form in sentences should be preferable. Anyway, this is just a matter of personal preferences. The English is plain and readable; the reading difficulties depend on the organization and choices in the presentation of the data rather than on language and style. The reader gets lost between new and old data, sections lacking stratigraphic constraints, basic information dispersed through the text, detailed analytical protocols, and so on. Please consider to re-organize the text.

We have endeavoured to clarify any potential confusion in the text and its structure.

 

- Starting with “Introduction” and “Sampling site” sections (the latter should be better part of Introduction or Geological Settings), it is not always clear where are the new data, and what already derives from literature (e.g.: lines 127, 135, 400). The reader who is unfamiliar with the study area, and with the former literature, has to continually scroll the text back and forth in order not to get lost and find all the information. This becomes more marked in Results, and also reflects in organization of figures. In Figure 1 (inspired to Figure 1 in [54])  all the sites mentioned or analyzed in the text should be reported, even if just for comparison. A figure showing the correlation between the different stratigraphic columns, and with the reference succession, would also be useful, proposing all the useful information that is currently dispersed in the text, even suddenly introduced (e.g.: lines 119-120, 208, 245, 400). This would allow to focus attention in the following figures on the new data and their position through the REM3b core section (or the composite REM column as well). All general information should still be provided from the beginning, even with a section dedicated to the description of the site (2.1), and not introduced directly into Results or Discussion. Finally, methods, results and discussion are not always clearly separated, and some sentences should be moved through the text, or rephrased.

See above – we have tried to clarify the structure and delineate the new and old data more clearly. We have also introduced a new section on the regional setting, which hopefully removes some of this confusion.

 

- The methodological part is in my opinion too extensive and detailed, and hinders reading a bit. The reader has the sensation of suddenly plunging into a methodological work. For those who are not very familiar with the techniques used in this research, the results obtained and their implications are more relevant. On the other hand, excessive methodological detail is superfluous for experts. As reference is made to analytical protocols widely used and codified in the literature, this part could be significantly reduced. Otherwise, methods should be detailed in supplementary materials.

Other reviewers commented in the reverse. We have now tried to strike a balance between sufficient detail and detracting from the results. The supplementary provides more detail.

 

- In the discussion (I refer particularly to paragraph 4.2), attention is not adequately focused on what (according to the title) should be the central topic of the article (e.g. the use of luminescence sensitivity). There is a feeling that the authors are the first to be convinced of the interpretation provided, as if it were the reader who had to convince himself. I would have expected solid arguments, or at least a discussion of the deductions supporting the use of the method as a source proxy, and an overview of alternative options that make that choice the most satisfactory. In my opinion, this part should be reconsidered and enlarged.

Since this is really the first time anyone has investigated the IR50 signal in feldspars as a possible tool for source variability, and there remain a number of uncertainties with respect to mechanism and the interplay between source variability and post-depositional alteration, we prefer not to be any more conclusive than we have been (see comments above and in our responses to individual comments in the PDF provided by reviewer 4). The approach shows promise; we as the authors needed to be convinced of that first; then the onus is on us to lead the reader through that logical progression in the empirical study.

 

- Please check Figures, particularly colors and legend.  Also the order can be modified: for example, Figures 3 and 4 should be probably inverted.

We have revised the figures as requested, but retain the order of figures; arguments regarding this decision are detailed in the PDF.

 

I think the paper could be considered for publication in Quaternary, but it cannot be accepted in its present form. In my opinion, the manuscript still needs some work by the aothors before publication.

I hope my comments could be somehow helpful, anyway.

 

We thank all four reviewers for their constructive comments and trust that our revised manuscript is now sufficiently improved for acceptance to publication.

 

Kathryn Fitzsimmons, on behalf of the authors

24.11.2021

Author Response File: Author Response.pdf

Round 2

Reviewer 4 Report

Dear Authors,

I noted your work on improving the text, and I think you answered to the comments. Thus, I believe the manuscript could be accepted in the present form.  

In this second review, I only found minor points, which you may directly consider in the editing phase (proofs):

• Line 77: rather than "metamophosis" should be "metamorphism".
• Line 130: space lacking between "OIS" and number (please check it throughout the text).
• Line 221: space lacking between the two sentences.
• Figure 4: if possible, in the keys, check the horizontal alignement of text and objects.

Best Regards

The reviewer

Author Response

I noted your work on improving the text, and I think you answered to the comments. Thus, I believe the manuscript could be accepted in the present form.  

We thank the reviewer for this positive feedback.

• Line 77: rather than "metamophosis" should be "metamorphism".

We have corrected this mistake.

• Line 130: space lacking between "OIS" and number (please check it throughout the text).

This is now corrected throughout the text.

• Line 221: space lacking between the two sentences.

We have corrected this mistake.

• Figure 4: if possible, in the keys, check the horizontal alignement of text and objects.

We have made this amendment.

We have now responded to all changes requested and look forward to hearing from you regarding acceptance of the manuscript.

Kind regards

Kathryn Fitzsimmons

on behalf of the co-authors

Author Response File: Author Response.pdf