Composition of Modern Dust and Holocene Aeolian Sediments in Archaeological Structures of the Southern Levant
, Amir Sandler 2, Kim André Vanselow 1, Hendrik J. Bruins 3, Nizar Abu-Jaber 4, Rupert Bäumler 1, Naomi Porat 2 and Paula Kouki 5
Yury Chendev
Round 1
Reviewer 1 Report
My overall impression after reading is that of a very well prepared, written, and documented study. The main subject is on loess deposits and their archeological significance that is outside my competence. Thus I can not evaluate the manuscript in any details.
I have a just a few minor comments:
The appendices should be deposited as supplementary material. There is absolutely no evaluation of the accuracy and precision of the XRF analyses. There is not even any explanations as to how they date were obtained. Is there any possibilities that some of the variability in some compositional data should be due to dust/aerosol from burning of organic material such as wood? The graphs are in general difficult to read, particularly the PC analyses. Any possibility of improving this by say guide lines etc? Would it be possible to normalizing the elemental/oxide compositions to a common soil such as the average crust/soil? This could greatly improve the presentation.Author Response
Dear editor, dear reviewers,
Thank you very much for your efforts to review our manuscript. We addressed and implemented all issues.
Figures 6 and 13 were revised as recommended by the editor: sample names were replaced by sample numbers and the letter size increased. The use of sample numbers made it possible to legibly display all samples in the diagrams. In order to further improve legibility and to facilitate identification of sample groups, we used different colors for the sample types, which are explained in a legend and consistent with the colors attributed to the statistically modeled deposit types.
The sample numbers were added in an additional column to all tables displaying data in order to support sample identification by their number. In addition, now all citations were numbered according to their appearance in the text as required by the journal. The added numbers and changed reference style were not marked by the tracking tool, but all changes of the text and captions are highlighted.
Regarding individual comments:
#Reviewer 1:
We follow the journal guidelines to present the additional data either as supplementary files or appendix. We inquired beforehand and as an appendix seems the usual practice, prepared the material accordingly.
The accuracy and precision of the XRF analyses had been explained in the detailed method description in the appendix. We present the relative error of major and trace element measurements. In order to make this clearer, we extended the description of the XRF method in the main text.
We carefully checked the detailed method description in the appendix again and think that it should be clear how the presented data were obtained. As this issue was not mentioned by the other reviewer, we prefer to keep the method explanation in the main text short and refer to the extensive text in the appendix, which is rather long as it explains all details. If needed, we can of course further extend the method description in the main text.
We are certain that the data are not affected by aerosols from burning of organic material. Wood is scarce in the southern Levant and there is absolutely no indication that burning contributed to current dust samples. In archaeological soils, ash deposits can occur. If present in larger amounts, they can be detected macroscopically. They were present in the monastery garden (Jh site 2, samples 51-57), were probably burnt excrements were added as fertilizer (Lucke et al., submitted). In addition, ash could be observed in the paleosol buried below the site of Abu Suwwan (sample no. 58, Abu Suwwan below nw). These have been described in the detailed profile description in the appendix, which was extended in case of Abu Suwwan.
This ash apparently caused strongly elevated magnetic susceptibility at Abu Suwwan (but only this one sample), which is now explicitly mentioned in the pedogenesis results (chapter 4.5). With regard to grain sizes or geochemical composition, however, there is no difference between samples that apparently contain ash and those that do not. This is now outlined in the discussion of pedogenesis (chapter 5.4) and in line with earlier observations on the role of ash in archaeological soils (Lucke, 2008). We consider the role of burnt material negligible for the overall question of the paper.
We do not think that normalizing the data, e.g. to average crust, would improve the presentation. On the contrary, it would make it more difficult to directly compare the obtained data to results from other soil studies in the southern Levant. Normalizing to e.g. UCC could make sense if only dust samples were compared, but the main thrust of our paper is to connect dust results with soil properties. We think it is important that the properties of the studied archaeological soils and dust can directly be compared with soils in the area, and the presentation of our material fits the literature on the region. We therefore prefer to keep the data in the present format.
#Editor:
The lower right panel in Figure 7 was fixed, as well as the typo in the subheader of Table 7. We are happy to make the data files for the statistical analyses available as supplementary material.
With regard to the formation of aggregates in dust, we added three citations to the references and some lines addressing the issue in the discussion of the role of remote sources (chapter 5.8).
We do not think, however, that such aggregates are the reason for divergent results of grain size analysis by laser and sedigraph devices. Both procedures aim to completely disperse any aggregates. As the sizes of such aggregates may randomly vary due to processes of adherence and dissolution, most procedures of grain size analysis require maximum dispersal.
Sodium hetamexaphosphate was applied in pretreatment for both wet sieving with sedigraph analysis and laser measurements. In addition, all laser measurements were preceded by at least 5 minutes of ultrasonic treatment, or until shadowing remained constant, which means that no further dispersal could be achieved (described in detail in the methods description in the appendix).
It is a well-known issue that higher clay contents will be obtained with sedigraph analysis, and the vast literature on the problem suggests that insufficient pre-treatment of laser samples is not the reason. Various correction factors have been proposed (see summary in Lucke et al., 2019), and we decided to follow one proposed for similar dust samples, which showed satisfactory results in our systematic methods comparison test (presented in Lucke et al., 2019). Of course, the question of "over" or "under"-estimation is a matter of point of view: results from both devices are correct, but there is a small and systematic difference that needs to be addressed by application of a correction factor.
We would appreciate publication of the revised article.
Best regards
Bernhard Lucke on behalf of all authors
Cited references:
Lucke, B. Demise of the Decapolis. Past and Present Desertification in the Context of Soil Development, Land Use, and Climate. Verlag Dr. Müller: Saarbrücken, Germany, 2008; ISBN 978-3639006131.
Lucke, B., Roskin, J., Vanselow, K., Bruins, H., Abu-Jaber, N., Deckers, K., Lindauer, S., Porat, N., Reimer, P., Bäumler, R., Erickson-Gini, T., Kouki, P. Character, rates, and environmental significance of dust accumulation in archaeological hilltop ruins in the southern Levant. MDPI geosciences 2019, 9(190), 1-60, doi:10.3390/geosciences9040190.
Lucke, B., Birk, J., Hunt, C., Laparidou, S., Keilholz, P., Kouki, P., Abu-Jaber, N., Bäumler, R., Fiedler, S., Deckers, K., Abudanah, F. The terraces of Petra, Jordan: environmental archives of a lost agricultural hinterland. Geoarchaeology, submitted.
Reviewer 2 Report
The article has novelty and strong significance. It contains very important results, which have been received by using of modern methods and approaches. Different mathematical and statistic connections were presented. Aeolian sedimentation in the study area, its rates and spatial distribution are discussed. Representative quantity of key areas and soils have been studied. I recommend the article for publication but, it seems to me, figures 6 and 13 are too complicated for understanding (due to big quantity of names as well as small size of letters and words).
Author Response
Dear editor, dear reviewers,
Thank you very much for your efforts to review our manuscript. We addressed and implemented all issues.
Figures 6 and 13 were revised as recommended by the editor: sample names were replaced by sample numbers and the letter size increased. The use of sample numbers made it possible to legibly display all samples in the diagrams. In order to further improve legibility and to facilitate identification of sample groups, we used different colors for the sample types, which are explained in a legend and consistent with the colors attributed to the statistically modeled deposit types.
The sample numbers were added in an additional column to all tables displaying data in order to support sample identification by their number. In addition, now all citations were numbered according to their appearance in the text as required by the journal. The added numbers and changed reference style were not marked by the tracking tool, but all changes of the text and captions are highlighted.
Regarding individual comments:
#Reviewer 1:
We follow the journal guidelines to present the additional data either as supplementary files or appendix. We inquired beforehand and as an appendix seems the usual practice, prepared the material accordingly.
The accuracy and precision of the XRF analyses had been explained in the detailed method description in the appendix. We present the relative error of major and trace element measurements. In order to make this clearer, we extended the description of the XRF method in the main text.
We carefully checked the detailed method description in the appendix again and think that it should be clear how the presented data were obtained. As this issue was not mentioned by the other reviewer, we prefer to keep the method explanation in the main text short and refer to the extensive text in the appendix, which is rather long as it explains all details. If needed, we can of course further extend the method description in the main text.
We are certain that the data are not affected by aerosols from burning of organic material. Wood is scarce in the southern Levant and there is absolutely no indication that burning contributed to current dust samples. In archaeological soils, ash deposits can occur. If present in larger amounts, they can be detected macroscopically. They were present in the monastery garden (Jh site 2, samples 51-57), were probably burnt excrements were added as fertilizer (Lucke et al., submitted). In addition, ash could be observed in the paleosol buried below the site of Abu Suwwan (sample no. 58, Abu Suwwan below nw). These have been described in the detailed profile description in the appendix, which was extended in case of Abu Suwwan.
This ash apparently caused strongly elevated magnetic susceptibility at Abu Suwwan (but only this one sample), which is now explicitly mentioned in the pedogenesis results (chapter 4.5). With regard to grain sizes or geochemical composition, however, there is no difference between samples that apparently contain ash and those that do not. This is now outlined in the discussion of pedogenesis (chapter 5.4) and in line with earlier observations on the role of ash in archaeological soils (Lucke, 2008). We consider the role of burnt material negligible for the overall question of the paper.
We do not think that normalizing the data, e.g. to average crust, would improve the presentation. On the contrary, it would make it more difficult to directly compare the obtained data to results from other soil studies in the southern Levant. Normalizing to e.g. UCC could make sense if only dust samples were compared, but the main thrust of our paper is to connect dust results with soil properties. We think it is important that the properties of the studied archaeological soils and dust can directly be compared with soils in the area, and the presentation of our material fits the literature on the region. We therefore prefer to keep the data in the present format.
#Editor:
The lower right panel in Figure 7 was fixed, as well as the typo in the subheader of Table 7. We are happy to make the data files for the statistical analyses available as supplementary material.
With regard to the formation of aggregates in dust, we added three citations to the references and some lines addressing the issue in the discussion of the role of remote sources (chapter 5.8).
We do not think, however, that such aggregates are the reason for divergent results of grain size analysis by laser and sedigraph devices. Both procedures aim to completely disperse any aggregates. As the sizes of such aggregates may randomly vary due to processes of adherence and dissolution, most procedures of grain size analysis require maximum dispersal.
Sodium hetamexaphosphate was applied in pretreatment for both wet sieving with sedigraph analysis and laser measurements. In addition, all laser measurements were preceded by at least 5 minutes of ultrasonic treatment, or until shadowing remained constant, which means that no further dispersal could be achieved (described in detail in the methods description in the appendix).
It is a well-known issue that higher clay contents will be obtained with sedigraph analysis, and the vast literature on the problem suggests that insufficient pre-treatment of laser samples is not the reason. Various correction factors have been proposed (see summary in Lucke et al., 2019), and we decided to follow one proposed for similar dust samples, which showed satisfactory results in our systematic methods comparison test (presented in Lucke et al., 2019). Of course, the question of "over" or "under"-estimation is a matter of point of view: results from both devices are correct, but there is a small and systematic difference that needs to be addressed by application of a correction factor.
We would appreciate publication of the revised article.
Best regards
Bernhard Lucke on behalf of all authors
Cited references:
Lucke, B. Demise of the Decapolis. Past and Present Desertification in the Context of Soil Development, Land Use, and Climate. Verlag Dr. Müller: Saarbrücken, Germany, 2008; ISBN 978-3639006131.
Lucke, B., Roskin, J., Vanselow, K., Bruins, H., Abu-Jaber, N., Deckers, K., Lindauer, S., Porat, N., Reimer, P., Bäumler, R., Erickson-Gini, T., Kouki, P. Character, rates, and environmental significance of dust accumulation in archaeological hilltop ruins in the southern Levant. MDPI geosciences 2019, 9(190), 1-60, doi:10.3390/geosciences9040190.
Lucke, B., Birk, J., Hunt, C., Laparidou, S., Keilholz, P., Kouki, P., Abu-Jaber, N., Bäumler, R., Fiedler, S., Deckers, K., Abudanah, F. The terraces of Petra, Jordan: environmental archives of a lost agricultural hinterland. Geoarchaeology, submitted.
