Round 1

Reviewer 1 Report

The study examined the effects of tree disturbances (tree-uprooting) on the spatial variability of soil chemical properties in Europe. 23 soil chemical properties were detected in both the surface and subsurface layer. Geostatistics analysis were used to quantify the spatial autocorrelation of soil chemical properties, and correlation and redundancy analyses were used to assess the indirect effects of soil disturbance regimes on the variability of these soil chemical properties. The autocorrelation distance for each property were presented, and the results confirmed the significant effect of treethrow density on soil properties. The study was very interesting, but I lost myself in the paper for several times.

First, the data tables were too large, and the captions for the tables were too simple, also what’s the meaning for each abbreviation Table 4 (soil unit). Please show the important results, or reorganize the results in the table.

Second, how did treethrow density acquired. For each square grid, maybe several up-rooting events would happen, and they may belong to different decomposed levels (am I right?). Thus, how to evaluate this square grid belong to which decomposed level, since in Table 5 each plot has its treethrow decomposed level. Moreover, the total plot in table 5 didn’t equal to 309 as described in sampling method. Please show the data of treethrow density in the paper.

Third, there was a lack of in-depth and comprehensive discussion on the experimental results. The results were repeatedly described in the discussion. Why the effect of treethow density changed for different soil properties and in different decomposed level? Treethow depth also correlated with several soil properties, but no discussion was given. Also, the explanation power for treethow density and depth were small (about 3-6%), what’s the other main factor?

 

Some specific comments in this paper are as follows:

1. What does soil richness mean?
2. What’s the purpose for the correlation analysis among the 23 soil properties?
3. P11, 1545 sites or 309 sites?
4. Maybe give some introductions about the two type soil properties (external sources and pedogenic sources) in the introduction part.

Author Response

We would like to thank the four reviewers for all comments to our manuscript. Their feedback is highly appreciated and encouraged us to significantly improve the quality of this study. Wherever possible we tried our best to carefully incorporate all comments and suggestions from the four reviewers. 

All in all, we are confident that this revised manuscript is significantly. Below we address all reviewers’ comments (indicated in blue) point by point (please, see the attached document)

 

 

Reviewer #1

The study examined the effects of tree disturbances (tree-uprooting) on the spatial variability of soil chemical properties in Europe. 23 soil chemical properties were detected in both the surface and subsurface layer. Geostatistics analysis were used to quantify the spatial autocorrelation of soil chemical properties, and correlation and redundancy analyses were used to assess the indirect effects of soil disturbance regimes on the variability of these soil chemical properties. The autocorrelation distance for each property were presented, and the results confirmed the significant effect of treethrow density on soil properties. The study was very interesting, but I lost myself in the paper for several times.

Thank you very much for your suggestions. We have addressed all the comments.

First, the data tables were too large, and the captions for the tables were too simple, also what’s the meaning for each abbreviation Table 4 (soil unit). Please show the important results, or reorganize the results in the table.

We have modified the table 1. We have reduced the table 1 and the original table 1 is moved to Supplementary material.

 

Also, the table 4 has been modified. We have moved the table 4 to Supplementary material and we have left in the manuscript the values of the table 4 with a significant relations p<0.001, p<0.05 and p < 0.1.

 

Finally, we modified the table 5, we removed the treethrow depth and fresh treethrow because they are not significant in the results. The original table 5 is in Supplementary material.

 

The soil unit in Table 4 has been specified.

 

Second, how did treethrow density acquired. For each square grid, maybe several up-rooting events would happen, and they may belong to different decomposed levels (am I right?). Thus, how to evaluate this square grid belong to which decomposed level, since in Table 5 each plot has its treethrow decomposed level. Moreover, the total plot in table 5 didn’t equal to 309 as described in sampling method. Please show the data of treethrow density in the paper.

We have included the treethrow density in the figure S3 and the original table 5 in the Supplementary material.

In each plot we can have one, two, three or none treethrow decomposed level. We have calculated the treethrow density of each treethrow decomposed level in each plot. The total of plots in the table 5 are not 309 because in some plots there are different level of decomposition or are empty of treethrow.

Third, there was a lack of in-depth and comprehensive discussion on the experimental results. The results were repeatedly described in the discussion. Why the effect of treethow density changed for different soil properties and in different decomposed level? Treethow depth also correlated with several soil properties, but no discussion was given. Also, the explanation power for treethow density and depth were small (about 3-6%), what’s the other main factor?

We have modified the discussion. We have removed from discussion every reference to results and we have focused only to discuss the results (Line 669-707).

“In a previous study on the old-growth temperate forest at Zofin, [15] detected a spa-tially non-random effect of tree uprooting on soil taxonomy. Here, we assessed the effects of treethrow disturbances on the vertical and spatial patterns of soil chemical properties, which are one of the drivers of soil spatial complexity.

We found that soil chemical properties are most strongly affected by totally decom-posed treethrows, both in the surface and subsurface layers of Entic Podzols, but particu-larly in the subsurface, where treethrow density explained more variability in the soil chemical properties of pedogenic sources. We also found some evidence of the effect of tree disturbances of partially decomposed treethrows on soil chemical properties in the sub-surface layer. In general, the negative correlation between treethrow density and almost all soil chemical properties of pedogenic sources in the surface layer (MnKCl, Alox, Mnox, Feox, Siox, Aldit and Fedit) indicates a loss of these soil chemical properties in the surface as they are translocated to the subsurface layer. Consequently, soil chemical properties of pedo-genic sources in the subsurface layer were positively correlated with treethrow density (table 4). These findings indicate that the influence of the density of totally decomposed treethrows are particularly important in podzolisation at our site. In addition, these effects can be explained due to the greater soil dynamics in the surface layer than in the subsur-face layer, which promotes this podzolisation. This relationship between treethrow den-sity and soil chemical properties may also be interpreted as a linkage between the spatial pattern of disturbance regimes and soil patterns. No specific patterns were found in the relationship between treethrow depth and soil chemical properties”

Some specific comments in this paper are as follows:

1. What does soil richness mean?

We have modified this term of the manuscript, Line 288 “soil types richness”

1. What’s the purpose for the correlation analysis among the 23 soil properties?

The purpose is that we are interested in analysing the relationships between soil chemical properties both in the surface than in the subsurface layer.

Also, it is necessary to examine the distribution of each variable in the response matrix for RDA analysis (in this case soil chemical properties) against other variables in its own matrix. If the relationships are linear, the effect of outliers is reduced.

1. P11, 1545 sites or 309 sites?

We have sampled 1545 soil profiles or sites. In the section “Soil sampling and analyses” is specified (line 115)

 

1. Maybe give some introductions about the two type soil properties (external sources and pedogenic sources) in the introduction part.

We have added in the introduction the two types of soil properties: Line 77

“We define soil properties from external sources as those that are from complexes with organic matter or from parent material. We define soil properties from pedogenic sources as those that are formed by soil processes through the profile, e.g. those involved in podzolisation process”

 

Author Response File: Author Response.pdf

Reviewer 2 Report

## General comments

The study is dedicated to a very interesting topic that is of relevance to forest ecologists: How variable is soil formation due to typical disturbances within forests. The sampling methodology is sound and I have the impression that the authors know their trade as soil ecologists. Also, it seems that the authors conducted the statistical tests with care, but here comes my first however: I am not quite sure if the authors accounted for spatial autocorrelation in all their analyses appropriately. Some kind of spatial analysis was performed, but the core research question (how forest disturbance affect soil properties) was assessed in a Principal Component Analysis (which the authors erroneously entitled as Redundancy Analysis) with post-hoc fitting of environmental variables. I could not see that the spatial dependency between their sampling plots, which the authors even accredited themselves to be present, is reflected in this analysis. The sampling plots in this study do NOT reflect independent samples, because they are all from adjacent grid squares. This has to be improved urgently! Further, as a reader, I drowned in information about the multitude of p-values of diverse tests that were reported in the results section. I had the impression that the authors were on a p-value hunt, instead of trying to deliver pointed information regarding the main research question. This brings me to another point: I think that the manuscript is overloaded due to the attempt to address several research questions (Q1-Q3 at the end of the introduction) at once. However, the title and the introduction ultimately lead only to Q3, how forest disturbances affect soil formation and soil properties. Focusing on this Q3 would also be a neat story on itself, it would make the manuscript much clearer, and it is also the questions most interesting to a journal such as “forests”. Q1 and Q2 are in my eyes “only” exploratory pre-analyses leading to the core research question. Instead of throwing all results of this pre-analysis at the reader, the job of the authors would be to guide the reader through the key results in order to make their case on Q3. So, I also recommend focusing and trimming the manuscript in many places, for which I will highlight only some examples in the specific comments. Further, the sentence structure and formulations were often quite wordy and confusing. I also recommend a very (!) thorough English editing. It is obvious that no native English speaker had a check on this manuscript before, which I really find an inconvenience as a reviewer.

 

 

## Specific comments

 

L8 “belongs among issues” is an example for the strange wording in the manuscript

 

L19 this sentence is an example for the strange wording and for the fixation of the authors on reporting statistical results instead of reporting ecological observations. This sentence is actually not meaningful. Everything is an explanatory variable of variability in other things’ properties. Absolutely redundant information! What do the authors want to transmit with sentences as this? Like I said, this is just one of many examples that the authors have to refine before I would accept this manuscript!

 

L54 The sentence structure in the manuscript often lacks any connectivity! “They”, who is this? Cambisols, Podzols, organo-mineral complexes… ? Absolutely not clear here! Again, this is just one example.

 

L57 Which soil richness, where does this refer to?

 

L65 Tree disturbance or rather forest disturbance? Yes, a tree is falling over, but it actually is a disturbance in the forest ecosystem, or for the soil -> soil disturbance.#

 

L71 I think Q3 is the only “true” research question from a (forest) ecological perspective. Q1 and Q2 are rather methodologies to get to answering Q3.

 

L103 What does this mean, “where soil occurred”? What was in the other places, the void? Why now 23 meter diameter circles? Why does a square become a circle now?

 

Figure 1 The captions and the figure labeling is generally poor. For example, in the map it says “Czech Republic”, while all other countries are indicated as abbreviations without explanation. Indeed, it might be difficult to understand this map for non-European readers. In the figure caption it is redundant to say that this is a map. This is obvious. Also, it seems that the figures and tables were not thought through to the end. The elevation legend is very poor and I doubt that the authors have a DEM down to the accuracy of mm elevation (3 digits!). Avoid false accuracy.

 

Figure 2 Although the elements are given in the legends, it seems poor practice to put only units on the x-axes. The caption: “Some soil chemical properties” sounds like a random selection. The explanation of the abbreviations in the figure legend is missing. Apart from elements (Ca, Mg, C, …) why not spelling out in the legend itself? What is Aldit/Fedit/Mndit/Sidit? You are addressing forest ecologists here. I even doubt that such a graph and such abbreviations could come without explanation in a soil ecology journal.

 

LL165ff The whole description of the statistical procedure is inflated. The information given is essential, yes, and I can see that much thought was put in the analysis. However, the content could be described more precise and less wordy and I guess half of the words could suffice to do the job.

 

LL179ff When first reading through this I actually thought the geospatial analysis was conducted in order to explore spatial dependency in the data and to account for this in the later actual data analysis (the relationship between disturbance and soil properties). However, I am quite disappointed that this was not done, was it?

 

LL208ff “an analysis … was tested”? Please avoid such strange wording throughout the manuscript.

 

L210 “soil units” is “soil type”??

 

L211 Meaningless sentence, redundant and grammatically terrible. It was used to do what? Instead of telling us that the p-value was used (however this works), information on the definition of the significance threshold in this study is lacking!

 

L215 you mean "multivariate" with "multidirectinal" or what should this be?

 

L215 I actually doubt that you performed a RDA, from what is described here. You probably used the function "rda" in vegan, but you only used the dependent variables in this first step when setting up this ordination. So this makes it a "simple" principal component analysis PCA. You only fitted the environmental variables afterwards! The RDA would be the constrained case of the PCA, where you directly feed the constraining (environmental) variables into the ordination right from the start. Please read the statistics manuals carefully first, before performing and reporting statistical analyses!!!

 

LL220ff Why was that done, to what purpose? Why was H0 that the variation is larger than random? This should then be Ha (alternative)!

 

LL225f Redundant!

 

LL226ff Already interpretation!

 

LL232ff Not quite sure what authors want to say with this. Also unsure whether this belongs here anyhow!

 

LL247ff Is this comparison actually fair? I struggle to understand how many Cambisol samples there were actually in the data! Were the soils not mostly Podzols?

 

Table 1: I am very sorry, but the table looks terrible. Hard to draw any conclusion from it, when looking at it. Why not using boxplots/histograms/whatever more creative graphical solution than this horrid table.

 

Figure 3 Very poor figure labelling. Where is the explanation for the abbreviations in the Figure? Why not simply spelling out in the figure. There is enough space to do so!

 

L306 Regarding the authors’ understanding of statistical significance: A relationship is either significant or not (p > 0.05, if you choose to define so). Reporting of significance levels makes no sense! A p value is not a good representative of the effect size! Colors and r values indicate goodness of correlation in this graph!

 

LL309ff Redundant/does not belong into results section! Please only take these as examples, there are many more parts like this that have to be improved.

 

Table 3 what "paired difference test" was actually applied? T-Test?

 

LL347 Redundant and not meaningful information…

 

L377 What is the definition of “interesting results” for the authors? This sentence stands symbolic for what I think about the results section in general: It feels like the authors try to wake interest in their study by uniting as many reports of (highly/most/...) significant correlations in the results section. With all these mentionings of significances, the authors completely fail to actually deliver to me what is interesting from the perspective of a forest ecologist. I like to know about the ecological (windthrow -> soil properties) relationships that were found in the study. This information is completely lost behind a mass of purely "statistical" reporting.
I feel that there is no kind of red thread leading through the results. Rather, the readers have to suffice with uninformative and redundant sentences in the style of "interesting results are presented in this table/figure... go and check yourself, we did not care to do the job". That is a pity, because I feel that the results of this very interesting study deserve more!

 

Table 4 Where are the explanations of the abbreviations?

 

LL402ff This paragraph is actually illustrating why I dislike the authors focussing so overly much on the strength of significance. The effect sizes are actually quite low (a few percent explained variance), but the authors retain "highly significant" results, likely due to their relatively large sample size. The p-value is highly dependent on sample size, so it is really disturbing that approx. half of the sentences in the results section deal with the size of the significance, while a clear display of the actuall effect sizes is mostly neglected!

 

LL403ff Again, example for redundant information. Why actually is the effect called "indirect"? I do not understand, because I think treethrow has a very direct (dramatically direct, in a way) influence on soil disturbance!

 

LL418f This is very crucial: It seems the authors did not (!) account for spatial autocorrelation in their analysis here. This might lead to inflated significances, which would not be so dramatic if the authors would not permanently highlight these. How was spatial autocorrelation and dependency accounted for in the analysis?

 

Table 5 I feel that the treethrow/disturbance narrative is the pivotal part of this study, which could be backed by a nice (!) representation of the ordination (PCA!) analysis. The table is somewhat dissapointing. Why no graphical representation of the PCA? Nicely explained, it is a very powerful tool to deliver your message!

 

LL446ff The authors want to reach out to modelers by describing general patterns of podzolisation. However, I wonder if their results can be generalized. Indeed a lot of samples were taken, but they are all representative for the very study region, which is locally very confined, and thus only descriptive for the area instead of globally valid!
In contrast, I feel that the whole study is a neat case study on how disturbance might (not) alternate pedogenesis. As such, I can really appreciate the merit of the study. I think the authors should focus on this aspect of their study instead of trying to sell a second story about general soil formation processes, which is in truth rather a backing story for the question on disturbance-soil formation relationships.

 

LL481ff I absolutely agree and I was wondering throughout the whole manuscript why the authors ever made the very strong case of dividing into external vs pedogenic factors.

 

LL484 I find this reasoning (here and also in other parts of the manuscript) very circular! A podzol is defined as a soil with (more) advanced pedogenesis (compared to Cambisol, for example). So it is not surprising to find that the soil formation is more advanced when you compare Podzol and Cambisol.

 

L490 Another example of redundant wording in the manuscript.

 

LL502f This could be at least translated into ecological meaning, when stated in the discussion. What does "higher autocorrelation distance" and "higher CV in the surface than in the subsurface" actually mean? For example, does higher autocorrelation distance in the surface actually mean that the soil properties are more similar there across larger distances/areas?

 

L516 What are "driver factors"? Please make sure to very much improve the English wording in the manuscript. -> "This indicates that the main factors for the formation of the two layers are different". Also, I think that this finding is not surprising nor novel, by definition. The B horizon is a subsurface horizon which is more related to the geological substrate (influence "from below"), while the A horizon is strongly influenced from humus incorporation "from above". So, by definition, different factors drive their development.

 

L566 You mean loss of concentrations? How can a soil loose chemical properties? Even the absence of any elemental concentration would be a property, theoretically.

 

LL581-586 Not a conclusion, but repetition of methods

 

LL587-596 to at least 90% a repetition of the results, not a conclusion… and I doubt the novelty in L587

 

LL597-606: Mostly repetition of results in “statistics speech”.

 

LL604-605 The overall goal of this whole study was to show a podsolization process? Isn’t the classical analysis of soil profiles in the field a less laborious way?

 

LL607ff This seems to come closer to a conclusion now.

 

L619f “The differences … are statistically significant” is a bad-practice formulation in the results. Even worse is its prominent placement in the last sentence of the manuscript.

 

LL634ff The first author wrote the manuscript and reviewed/edited it also by himself. I feared this when reading the manuscript. This could have been a job at least for the second author, before throwing this manuscript at the peer-reviewers.  

 

References list

Some article titles are in upper case letters. Please check the whole references list again by yourself, I will not provide details now.

 

Supplementary

Figure S5 Rather use a histogram here. A boxplot for a one-level factor…?

Author Response

We would like to thank the four reviewers for all comments to our manuscript. Their feedback is highly appreciated and encouraged us to significantly improve the quality of this study. Wherever possible we tried our best to carefully incorporate all comments and suggestions from the four reviewers.

All in all, we are confident that this revised manuscript is significantly. Below we address all reviewers’ comments (indicated in blue) point by point. (please, see the attached document)

 

Reviewer #2

## General comments

The study is dedicated to a very interesting topic that is of relevance to forest ecologists: How variable is soil formation due to typical disturbances within forests. The sampling methodology is sound and I have the impression that the authors know their trade as soil ecologists. Also, it seems that the authors conducted the statistical tests with care, but here comes my first however: I am not quite sure if the authors accounted for spatial autocorrelation in all their analyses appropriately. Some kind of spatial analysis was performed, but the core research question (how forest disturbance affect soil properties) was assessed in a Principal Component Analysis (which the authors erroneously entitled as Redundancy Analysis) with post-hoc fitting of environmental variables. I could not see that the spatial dependency between their sampling plots, which the authors even accredited themselves to be present, is reflected in this analysis. The sampling plots in this study do NOT reflect independent samples, because they are all from adjacent grid squares. This has to be improved urgently! Further, as a reader, I drowned in information about the multitude of p-values of diverse tests that were reported in the results section. I had the impression that the authors were on a p-value hunt, instead of trying to deliver pointed information regarding the main research question. This brings me to another point: I think that the manuscript is overloaded due to the attempt to address several research questions (Q1-Q3 at the end of the introduction) at once. However, the title and the introduction ultimately lead only to Q3, how forest disturbances affect soil formation and soil properties. Focusing on this Q3 would also be a neat story on itself, it would make the manuscript much clearer, and it is also the questions most interesting to a journal such as “forests”. Q1 and Q2 are in my eyes “only” exploratory pre-analyses leading to the core research question. Instead of throwing all results of this pre-analysis at the reader, the job of the authors would be to guide the reader through the key results in order to make their case on Q3. So, I also recommend focusing and trimming the manuscript in many places, for which I will highlight only some examples in the specific comments. Further, the sentence structure and formulations were often quite wordy and confusing. I also recommend a very (!) thorough English editing. It is obvious that no native English speaker had a check on this manuscript before, which I really find an inconvenience as a reviewer.

Thank you for reviewer comments. Below we address all reviewer’ comments (indicated in blue) point by point. We have replied to the general comment in each of the specific comments below.

 

Furthermore, as far as possible we have tried to avoid a statistical meaning for example "statistically significant" in our manuscript. We have included the statistical observations in the results section, but omitted them from the discussion and conclusions.

 

An English reviewer, David Hardekopf, has reviewed our manuscript.

 

## Specific comments

L8 “belongs among issues” is an example for the strange wording in the manuscript

We have changed this phrase

Line 14 “is an issue that has not been comprehensively resolved.”

L19 this sentence is an example for the strange wording and for the fixation of the authors on reporting statistical results instead of reporting ecological observations. This sentence is actually not meaningful. Everything is an explanatory variable of variability in other things’ properties. Absolutely redundant information! What do the authors want to transmit with sentences as this? Like I said, this is just one of many examples that the authors have to refine before I would accept this manuscript!

We have changed this phrase

Line 25“The treethrow density partially explained the variability of soil chemical properties”

L54 The sentence structure in the manuscript often lacks any connectivity! “They”, who is this? Cambisols, Podzols, organo-mineral complexes… ? Absolutely not clear here! Again, this is just one example.

We have changed this phrase

 

Line 61 “Cambisols”

L57 Which soil richness, where does this refer to?

We have removed this term of the manuscript

L65 Tree disturbance or rather forest disturbance? Yes, a tree is falling over, but it actually is a disturbance in the forest ecosystem, or for the soil -> soil disturbance.#

We have used the following terms “tree disturbance or treethrow disturbance” based on our previous several studies in this area. For example:

Šamonil, P., Daněk, P., Schaetzl, R. Vasickova, I., Valtera, M. (2015). Soil mixing and genesis as affected by tree uprooting in 743 three temperate forests: Soil mixing and evolution as affected by tree-throw. European Journal of Soil Science. 66. 744 10.1111/ejss.12245.

 

L71 I think Q3 is the only “true” research question from a (forest) ecological perspective. Q1 and Q2 are rather methodologies to get to answering Q3.

We have modified this part of the introduction: Line 72 “The aim of this work is therefore to analyse the effect of tree disturbances produced by treethrows on the spatial variability of soil chemical properties in both the surface and subsurface layers and on soil formation to explain the soil spatial pedocomplexity in one of the oldest forest reserves in Europe”

L103 What does this mean, “where soil occurred”? What was in the other places, the void? Why now 23 meter diameter circles? Why does a square become a circle now?

We have modified the text in the manuscript:

Line 112 “Then, in each square (309 in total), we identified a circle 23 meters in diameter around the centre of each square. In these circles, five soil profiles were described in terms of soil morphology…”

Figure 1 The captions and the figure labeling is generally poor. For example, in the map it says “Czech Republic”, while all other countries are indicated as abbreviations without explanation. Indeed, it might be difficult to understand this map for non-European readers. In the figure caption it is redundant to say that this is a map. This is obvious. Also, it seems that the figures and tables were not thought through to the end. The elevation legend is very poor and I doubt that the authors have a DEM down to the accuracy of mm elevation (3 digits!). Avoid false accuracy.

This figure 1 and its caption have been changed in the manuscript following the suggestions of the reviewer.

Figure 2 Although the elements are given in the legends, it seems poor practice to put only units on the x-axes. The caption: “Some soil chemical properties” sounds like a random selection. The explanation of the abbreviations in the figure legend is missing. Apart from elements (Ca, Mg, C, …) why not spelling out in the legend itself? What is Aldit/Fedit/Mndit/Sidit? You are addressing forest ecologists here. I even doubt that such a graph and such abbreviations could come without explanation in a soil ecology journal.

The figure 2 and the caption has been modified

LL165ff The whole description of the statistical procedure is inflated. The information given is essential, yes, and I can see that much thought was put in the analysis. However, the content could be described more precise and less wordy and I guess half of the words could suffice to do the job.

The statistical information, especially at the beginning of the description, section “2.3 Statistical data analysis” has been removed and shortened.

LL179ff When first reading through this I actually thought the geospatial analysis was conducted in order to explore spatial dependency in the data and to account for this in the later actual data analysis (the relationship between disturbance and soil properties). However, I am quite disappointed that this was not done, was it?

The spatial dependency is dependent on distance between sampling points in the field, not only on data itself.

Our study explores the spatial autocorrelation in the soil chemical properties with geostatistical techniques. Then, independently, we assessed the effect of treethrow disturbances in soil.

The spatial dependence of soil chemical properties can be assessed by the nugget/sill ratio. Following the study of Cambardella et al. (1994), a ratio <25 indicates a strong dependence, 25-75 shows a moderate dependence and >75% a weak dependence.

In the following table, we have included the spatial dependence (last column). Spatial dependence in the surface layer was high for most of the soil chemical properties between sampling points in the study area (ratio nugget/sill <25), except for Ca+2 and Na+, which have a moderate spatial dependence (ratio nugget/sill between >27-75%). Regarding to subsurface layer, the spatial dependence was high for most of the soil proper-ties, except for pH, Ca+2, and Al+1, which have a moderate spatial dependence, and CEC with a weak spatial dependence (ratio nugget/sill >75%)

 

Surface layer							Subsurface layer
	Model	AIC	Effective  range	Sill	Nugget	Nugget/Sill		Model	AIC	Effective range	Sill	Nugget	Nugget/Sill
External
Cox log	Exponential	-110.68	52.53	0.05	0.00	0.00	Cox log	Circular	-34.24	80.57	0.50	0.11	0.22
pH	Exponential	-137.91	59.25	0.00	0.00	0.00	Ph log	Exponential	-179.88	165.30	0.02	0.01	0.46
CEC	Exponential	343.53	42.08	345.43	0.00	0.00	CEC log	Circular	-84.67	254.49	0.06	0.07	1.20
EA	Exponential	329.66	39.85	584.07	0.00	0.00	EA log	Spherical	-121.38	41.98	0.06	0.00	0.00
Mg +2 log	Exponential	-76.66	60.57	0.09	0.00	0.00	Mg +2 log	Gaussian	-5.11	25.83	0.44	0.00	0.00
Ca +2 sqrt	Gaussian	-157.16	132.05	0.03	0.02	0.65	Ca2+ log	Circular	-187.18	159.06	0.02	0.01	0.41
Na + log	Gaussian	-25.55	102.826	0.171	0.122	0.71	Na+ log	Exponential	-19.98	70.48	0.27	0.00	0.00
K + log	Exponential	-98.13	50.16	0.07	0.00	0.00	K+ log	Circular	-59.44	33.68	0.20	0.00	0.00
Al(X)+1	Exponential	381.161	50.304	670.103	0.000	0.00	Al+1 log	Gaussian	-101.05	86.10	0.04	0.02	0.37
Al(Y)+2 log	Exponential	-94.75	45.52	0.07	0.00	0.00	Al+2 log	Exponential	-31.25	61.11	0.23	0.00	0.00
Al+3	Exponential	556.43	46.27	20068.31	0.00	0.00	Al+3 log	Exponential	-111.91	27.46	0.06	0.00	0.00
Pedogenic
AlKClsum	Exponential	568.00	44.58	25403.97	0.00	0.00	AlKClsum log	Circular	-120.12	41.93	0.06	0.00	0.00
AlKCl	Exponential	576.26	47.06	29209.92	0.00	0.00	AlKCl log	Circular	-126.81	42.35	0.05	0.00	0.00
MnKCl log	Gaussian	103.22	99.74	2.50	0.60	0.24	MnKCl log	Exponential	73.38	88.15	1.56	0.00	0.00
FeKCllog	Exponential	0.53	48.38	0.45	0.00	0.00	FeKCl log	Exponential	23.83	90.65	0.60	0.00	0.00
AlOx log	Exponential	-109.73	59.53	0.06	0.00	0.00	AlOx log	Exponential	-40.08	47.97	0.21	0.00	0.00
MnOx log	Exponential	94.73	102.25	2.31	0.00	0.00	MnOx log	Exponential	78.67	98.80	1.70	0.00	0.00
FeOx log	Exponential	-72.03	53.11	0.11	0.00	0.00	FeOx log	Exponential	-90.83	34.69	0.09	0.00	0.00
SiOx	Exponential	507.16	72.99	15894.54	2607.06	0.16	SiOx log	Exponential	7.11	44.49	0.53	0.00	0.00
Aldit	Exponential	771.62	55.64	1175527.62	0.00	0.00	Aldit sqrt	Exponential	-35.38	50.12	0.22	0.00	0.00
Mndit log	Exponential	249.49	80.20	2.40	0.37	0.15	Mndit sqrt	Exponential	30.84	75.61	0.71	0.00	0.00
Fedit	Exponential	910.02	52.69	17233684.54	0.00	0.00	Fedit	Exponential	-110.34	40.78	0.06	0.00	0.00
Sidit sqrt	Exponential	98.72	31.73	3.28	0.00	0.00	Sidit sqrt	Exponential	26.41	54.64	0.71	0.00	0.00

 

LL208ff “an analysis … was tested”? Please avoid such strange wording throughout the manuscript.

This phrase was changed in the manuscript Line 264“we analysed the relationships between treethrow density,…”

L210 “soil units” is “soil type”??

Yes

L211 Meaningless sentence, redundant and grammatically terrible. It was used to do what? Instead of telling us that the p-value was used (however this works), information on the definition of the significance threshold in this study is lacking!

We have added the following information in the text. Although, this information was indicated in the table 4.

Line 267 “We used the Pearson correlation coefficient (r) to evaluate the relationship between variables and the significant relations are indicated with asterisks according to their significance level: ***p<0.001; **p < 0.05; *p<0.1.”

L215 you mean "multivariate" with "multidirectinal" or what should this be?

Yes, we mean “multivariate”. This term has been modified in the manuscript

L215 I actually doubt that you performed a RDA, from what is described here. You probably used the function "rda" in vegan, but you only used the dependent variables in this first step when setting up this ordination. So this makes it a "simple" principal component analysis PCA. You only fitted the environmental variables afterwards! The RDA would be the constrained case of the PCA, where you directly feed the constraining (environmental) variables into the ordination right from the start. Please read the statistics manuals carefully first, before performing and reporting statistical analyses!!!

We have modified this paragraph because it is confusing (Line 274). We performed RDA in our analysis. Conceptually, RDA is a multivariate (meaning multiresponse) multiple linear regression followed by a PCA. We use both soil chemical properties (response variables) and tree disturbances (explanatory variables) data as input.

The statistics books followed:

Zuur, Alain & Ieno, Elena & Smith, Graham. (2007). Analysing Ecological Data. Springer International Publishing. 10.1007/978-0-387-45972-1.

Borcard, D., Gillet, F., & Legendre, P. (2018). Numerical ecology with R (2nd ed.). Springer International Publishing.

Roeland Kindt (2020). Redundancy analysis with vegan and BiodiversityR as an alternative to Discriminant Analysis of Principal Components for the analysis of genetically structured populations. https://rpubs.com/Roeland-KINDT/706490

 

LL220ff Why was that done, to what purpose? Why was H0 that the variation is larger than random? This should then be Ha (alternative)!

We have modified this paragraph Line 280“To test the null hypothesis that the explained variation in the dataset is produced by random chance, a Monte Carlo test with 10000 permutations was applied. The null hypothesis is rejected if the p value is equal to or smaller than the predefined significance level (α=0.05).”

LL225f Redundant!

We have removed this phrase in the text.

LL226ff Already interpretation!

We have removed this phrase in the text

LL232ff Not quite sure what authors want to say with this. Also unsure whether this belongs here anyhow!

We have moved this phrase to the section “2.2 Soil sampling and analyses” Line 130 “. In the soil profiles studied we distinguished and assessed two aspects of soil properties based on expectations of their behaviour in soils and field measurements of soil chemical concentrations”

LL247ff Is this comparison actually fair? I struggle to understand how many Cambisol samples there were actually in the data! Were the soils not mostly Podzols?

Most of the soils are Podzols but we have compared the mean values between soil units.

Table 1: I am very sorry, but the table looks terrible. Hard to draw any conclusion from it, when looking at it. Why not using boxplots/histograms/whatever more creative graphical solution than this horrid table.

We have modified table 1 in the manuscript and we have moved the original table 1 to the Supplementary material. In the manuscript we left the mean, SD and CV and we have divided the information in two tables. Also, in the supplementary material we have added the boxplots of some examples of soil chemical properties that have considerable differences (Figure S5).

Figure 3 Very poor figure labelling. Where is the explanation for the abbreviations in the Figure? Why not simply spelling out in the figure. There is enough space to do so!

We have completed the information of the abbreviations in the figure 3.

L306 Regarding the authors’ understanding of statistical significance: A relationship is either significant or not (p > 0.05, if you choose to define so). Reporting of significance levels makes no sense! A p value is not a good representative of the effect size! Colors and r values indicate goodness of correlation in this graph!

Thank you. We have modified both the figure 3 and the figure caption. “Figure 3. Pearson correlation coefficients for relationships between soil chemical properties for sur-face and subsurface layers. Positive correlations are displayed in blue and negative corre-lations in red. The color intensity and the size of the circle are proportional to the correla-tion coefficients. Below side of the correlogram, the legend color shows the correlation co-efficients and the corresponding colors. Correlations with p-value>0.05 are considered in-significant, and the correlation coefficient values are left blank. Abbreviations: organic carbon (Cox), soil reaction (pH), exchangeable base cations of magnesium, calcium, sodium and potassium (Mg2+, Ca2+, Na+, K+), aluminium cations: Al (X)1+, Al (Y)2+, Al3+, exchangeable acidity (EA), cation exchange capacity (CEC), exchangeable aluminium, manganese and iron (AlKCl, MnKCl, FeKCl), aluminium, manganese, iron and silicon oxides (Alox, Mnox, Feox, Siox and Iron, Manganese and Silicon dithionite (Aldit, Mndit, Fedit, Sidit).”

LL309ff Redundant/does not belong into results section! Please only take these as examples, there are many more parts like this that have to be improved.

We have removed this phrase in the results section.

Table 3 what "paired difference test" was actually applied? T-Test?

The test applied was non parametric (Wilcoxon test). Previously, it was indicated in the manuscript (line 259, 459) but we also indicated this information in the table 3.

LL347 Redundant and not meaningful information…

We have removed this information in the manuscript

L377 What is the definition of “interesting results” for the authors? This sentence stands symbolic for what I think about the results section in general: It feels like the authors try to wake interest in their study by uniting as many reports of (highly/most/...) significant correlations in the results section. With all these mentionings of significances, the authors completely fail to actually deliver to me what is interesting from the perspective of a forest ecologist. I like to know about the ecological (windthrow -> soil properties) relationships that were found in the study. This information is completely lost behind a mass of purely "statistical" reporting. I feel that there is no kind of red thread leading through the results. Rather, the readers have to suffice with uninformative and redundant sentences in the style of "interesting results are presented in this table/figure... go and check yourself, we did not care to do the job". That is a pity, because I feel that the results of this very interesting study deserve more!

We mean that the results showed in the table are the most significant. For example, we have only showed the results in the soil units in the totally decomposed treethrows. We don´t show the results of the soil units in the rest of the decomposition degree because they are not significant.

We have moved non-significant tables to the supplementary material.

We have already published an article in Forests Journal, which relates windthrow and soil. For an ecological relationship, please read the following manuscript:

Vašíčková, I.; Šamonil, P.; Kašpar, J.; Román-Sánchez, A.; Chuman, T.; Adam, D. Dead or Alive: Drivers of Wind Mortality Initiate Multiple Disturbance Regime in a Temperate Primeval Mountain Forest. Forests 2021, 12, 1599. https://0-doi-org.brum.beds.ac.uk/10.3390/f12111599

Now the focus of our manuscript is other. We try to demonstrate our aim in a statistically way.

Table 4 Where are the explanations of the abbreviations?

We have included the abbreviations in the caption of the table 4.

LL402ff This paragraph is actually illustrating why I dislike the authors focussing so overly much on the strength of significance. The effect sizes are actually quite low (a few percent explained variance), but the authors retain "highly significant" results, likely due to their relatively large sample size. The p-value is highly dependent on sample size, so it is really disturbing that approx. half of the sentences in the results section deal with the size of the significance, while a clear display of the actuall effect sizes is mostly neglected!

In the results section we showed the statistical results obtained in our work. Previously, the number of samples was indicated in table 5. Also, we have included the number of samples in the results section “3.4 Effects of soil disturbance regimes on soil chemical properties” and in the conclusion, we have indicated “moderate but significant” (line 718).

We can see in our data that the effect of partly decomposed treethrows is significant for subsurface layer and the number of samples is 33. However, in the fresh treethrow neither of the explanatory variables were significant and the number of samples is 102.

LL403ff Again, example for redundant information. Why actually is the effect called "indirect"? I do not understand, because I think treethrow has a very direct (dramatically direct, in a way) influence on soil disturbance!

We have removed this information.

In order to avoid confusion, we removed “indirect” from our study.

LL418f This is very crucial: It seems the authors did not (!) account for spatial autocorrelation in their analysis here. This might lead to inflated significances, which would not be so dramatic if the authors would not permanently highlight these. How was spatial autocorrelation and dependency accounted for in the analysis?

In this part (line 556) we are highlighting the model significance.

On the other hand, most of soil chemical properties showed spatial autocorrelation (figure 3)

Table 5 I feel that the treethrow/disturbance narrative is the pivotal part of this study, which could be backed by a nice (!) representation of the ordination (PCA!) analysis. The table is somewhat dissapointing. Why no graphical representation of the PCA? Nicely explained, it is a very powerful tool to deliver your message!

We have added one figure more in our manuscript to show the RDA analysis (figure S8). The description of this figure is:

“Figure S8: Redundancy analysis (RDA). Ordination diagram (triplot) showing the significant relationships of the table 5. Sites (denoted by points), explanatory variables (treethrow density and treethrow depth; red arrows), and response variables (soil properties; blue arrows). First axis is horizontal, second axis is vertical. The angles among arrows denote the degree of correlation between the individual variables, and the smaller the angle, the larger the correlation. In addition, positively correlated variables are shown as arrows pointing in the same direction, negatively correlated variables pointing in opposite directions.”

LL446ff The authors want to reach out to modelers by describing general patterns of podzolisation. However, I wonder if their results can be generalized. Indeed a lot of samples were taken, but they are all representative for the very study region, which is locally very confined, and thus only descriptive for the area instead of globally valid! In contrast, I feel that the whole study is a neat case study on how disturbance might (not) alternate pedogenesis. As such, I can really appreciate the merit of the study. I think the authors should focus on this aspect of their study instead of trying to sell a second story about general soil formation processes, which is in truth rather a backing story for the question on disturbance-soil formation relationships.

In a future work we are going to develop a soil formation model including the factors that intervene in Podzolisation. Also, we believe the results of this study can be generalised to temperate forests around the world with the same conditions

Šamonil, P., Daněk, P., Schaetzl, R. Vasickova, I., Valtera, M. (2015). Soil mixing and genesis as affected by tree uprooting in three temperate forests: Soil mixing and evolution as affected by tree-throw. European Journal of Soil Science. 66. 744 10.1111/ejss.12245)

 

LL481ff I absolutely agree and I was wondering throughout the whole manuscript why the authors ever made the very strong case of dividing into external vs pedogenic factors.

In order to explain the spatial variation of soil chemical properties in surface and subsurface layer using geostatistical techniques, we differentiated external vs pedogenic sources.

LL484 I find this reasoning (here and also in other parts of the manuscript) very circular! A podzol is defined as a soil with (more) advanced pedogenesis (compared to Cambisol, for example). So it is not surprising to find that the soil formation is more advanced when you compare Podzol and Cambisol.

We discussed in this paragraph (4.1 Vertical patterns of external and pedogenic sources) the findings of vertical distribution of soil chemical properties. At the end of this section, we indicated that in the Podzols studied here there are signs of advanced pedogenesis in comparison with the Cambisols.

L490 Another example of redundant wording in the manuscript.

We have removed this phrase in the manuscript

LL502f This could be at least translated into ecological meaning, when stated in the discussion. What does "higher autocorrelation distance" and "higher CV in the surface than in the subsurface" actually mean? For example, does higher autocorrelation distance in the surface actually mean that the soil properties are more similar there across larger distances/areas?

“Higher autocorrelation distance” is that the soil properties are more similar across larger distances. We have modified this paragraph with an ecological meaning. Line 650 “we found that most of the soil properties of pedogenic sources are related across a greather distance in the surface than in the subsurface layer”

Using geostatistical analysis, we have found that the distance of autocorrelation of soil chemical properties of external source is higher in the subsurface than in the surface. Also, the distance of autocorrelation of soil chemical properties of pedogenic source is higher in the surface than in the subsurface. These trends are opposite to CV.

L516 What are "driver factors"? Please make sure to very much improve the English wording in the manuscript. -> "This indicates that the main factors for the formation of the two layers are different". Also, I think that this finding is not surprising nor novel, by definition. The B horizon is a subsurface horizon which is more related to the geological substrate (influence "from below"), while the A horizon is strongly influenced from humus incorporation "from above". So, by definition, different factors drive their development.

We have improved the English of our manuscript thanks to David Hardekopf

We want to explain in this paragraph that the difference in the variogram parameters in the surface and subsurface horizon is related with different factors for the formation of the two layers.

L566 You mean loss of concentrations? How can a soil loose chemical properties? Even the absence of any elemental concentration would be a property, theoretically.

We mean in this phrase to the podzolisation process. Line 658 “these soil chemical properties are translocated to subsurface layer”

LL581-586 Not a conclusion, but repetition of methods

We have removed this part of the conclusion

LL587-596 to at least 90% a repetition of the results, not a conclusion… and I doubt the novelty in L587

We have removed part of this paragraph

LL597-606: Mostly repetition of results in “statistics speech”.

We have removed this part of the conclusion

LL604-605 The overall goal of this whole study was to show a podsolization process? Isn’t the classical analysis of soil profiles in the field a less laborious way?

We have removed this part of the conclusion

LL607ff This seems to come closer to a conclusion now.

Thank you

L619f “The differences … are statistically significant” is a bad-practice formulation in the results. Even worse is its prominent placement in the last sentence of the manuscript.

We have removed this phrase of the manuscript.  As far as possible we have tried not to add "statistically significant" in our manuscript. We have included the statistical observations in the results section, but omitted them from the discussion and conclusions.

LL634ff The first author wrote the manuscript and reviewed/edited it also by himself. I feared this when reading the manuscript. This could have been a job at least for the second author, before throwing this manuscript at the peer-reviewers.

It is a mistake. All authors have contributed in the paper in this point “writing—review and editing”

References list

Some article titles are in upper case letters. Please check the whole references list again by yourself, I will not provide details now.

We have changed them

Supplementary

Figure S5 Rather use a histogram here. A boxplot for a one-level factor

We have changed the boxplot. We have created a histogram

Author Response File: Author Response.pdf

Reviewer 3 Report

Review for Forests-1642630

Overall, this is an interesting paper which studies the soil spatial pedocomplexity and analyse the indirect effects of soil disturbance regimes of three uproot. I found it well written and with adequate methods. Here you can find general and specific comments in order to improve the manuscript.

General

• I suggest to review the title. For me it only covers one of the three parts that the manuscript is presenting. It misses the differences between the sources (external vs. pedogenic) of soil chemical properties and the spatial variation of soil chemical properties
• I found the “Descriptive statistics” sections really long, with big tables and difficult to follow. It was not clear for me if there is a novelty in this section or you found everything as expected. If there is no novelty, I suggest to move big part of that section to the supplementary material, so it does not hide the real contribution of your manuscript in the following sections.
• Improve the tables: they are too many, too long and in several cases do not add to your results or discussion rather than saying that there is not enough correlation or not significant. I suggest to improve the presentation of some of those tables and leave only the main information in the manuscript and the additional information move it to the supplementary material.
• In your results you showed that the threethrows (decomposed) explain 4-6% of the variation in soil chemical properties. Is this enough to generate a difference between surface and subsurface? I believe it is significant, but please justify why.
• Regarding that in ~60% of the threes, roots depths are up to 55 cm, why do the three-uprooting only affect the surface (0-15cm)? Please discuss
• Please check that all the tables and figures have the same symbols as referred in the text (For example see legend in Figure 2).
• Could you please mention: when were the 1545 samples collected?
• Additionally, how long does it take the decomposition of the tree?
• If the soil samples and the creation of the database (threethrows) were collected in different time: How can this affect your results/conclusions?
• I think your work can benefit if you add a Figure (in the supplementary material?) showing the treethrow density in your study area.

Specific

Please see the attached document, which content other comments in different parts of the manuscript.

Comments for author File: Comments.pdf

Author Response

We would like to thank the four reviewers for all comments to our manuscript. Their feedback is highly appreciated and encouraged us to significantly improve the quality of this study. Wherever possible we tried our best to carefully incorporate all comments and suggestions from the four reviewers.

All in all, we are confident that this revised manuscript is significantly. Below we address all reviewers’ comments (indicated in blue) point by point. (please, see the attached document)

 Reviewer #3

Overall, this is an interesting paper which studies the soil spatial pedocomplexity and analyse the indirect effects of soil disturbance regimes of three uproot. I found it well written and with adequate methods. Here you can find general and specific comments in order to improve the manuscript.

Thank you very much for your positive comment. We have addressed all comments and suggestions below and in the manuscript attached.

General

• I suggest to review the title. For me it only covers one of the three parts that the manuscript is presenting. It misses the differences between the sources (external vs. pedogenic) of soil chemical properties and the spatial variation of soil chemical properties

 

We have modified the title in accordance with this suggestion “The effect of tree-uprooting on the soil spatial complexity in an old-growth temperate forest, Central Europe”

 

Also, following the suggestions of the 4th reviewer, we have modified the objectives and focused on analysing the effect of treethrow on the spatial variability of soil chemical properties to explain soil spatial pedocomplexity. According to the 4th reviewer, we have considered the source of soil properties and the application of geostatistics as methodologies to achieve the main objective. Therefore, in line 72 we have written:

 

“The aim of this work is therefore to analyse the effect of tree disturbances produced by treethrows on the spatial variability of soil chemical properties in both the surface and subsurface layers and on soil formation to explain the soil spatial pedocomplexity in one of the oldest forest reserves in Europe. To achieve this goal, we differentiate soil chemical properties according to their source (external and pedogenic) and use geostatistical and multivariate techniques. We define soil properties from external sources as those that are from complexes with organic matter or from parent material. We define soil properties from pedogenic sources as those that are formed by soil processes through the profile, e.g. those involved in the podzolisation process”

 

I found the “Descriptive statistics” sections really long, with big tables and difficult to follow. It was not clear for me if there is a novelty in this section or you found everything as expected. If there is no novelty, I suggest to move big part of that section to the supplementary material, so it does not hide the real contribution of your manuscript in the following sections.

We have modified the section “Statistics and data analysis”. We have moved to Supplementary material the original tables 1, 4 and 5. Also, we have removed part of the text of the manuscript regarding to basic statistical data analysis. Line 223.

 

• Improve the tables: they are too many, too long and in several cases do not add to your results or discussion rather than saying that there is not enough correlation or not significant. I suggest to improve the presentation of some of those tables and leave only the main information in the manuscript and the additional information move it to the supplementary material.

We have moved part of the tables with not enough correlation or not significant to Supplementary material (tables1,4 and 5).

• In your results you showed that the threethrows (decomposed) explain 4-6% of the variation in soil chemical properties. Is this enough to generate a difference between surface and subsurface? I believe it is significant, but please justify why.

In table 5, in the section of the results “3.4 Effects of soil disturbance regimes on soil chemical properties” ,Line 543, and in the discussion “4.3. Impacts of tree uprootings on vertical and spatial patterns of soil chemical properties”,Line 669, we showed that both the surface and subsurface the results are very significant (p<0.001) in the case of totally decomposed but the percentage of explained variability is higher in the subsurface than in the surface. On the other hand, the results are only significant in the subsurface (p<0.05) in the case of partially decomposed treethrow. The treethrows explained the 4 and 6% of the variation in soil chemical properties in the surface and in the subsurface respectively.

We have modified part of the discussion (“4.3. Impacts of tree uprootings on vertical and spatial patterns of soil chemical properties”) to explain better this significance. Line 669-690

• Regarding that in ~60% of the threes, roots depths are up to 55 cm, why do the three-uprooting only affect the surface (0-15cm)? Please discuss

In the previous draft we wrote that “the treethrow depth affected both the surface layer (0-15 cm) and the subsurface layer (~15-60 cm)”. Line 478

• Please check that all the tables and figures have the same symbols as referred in the text (For example see legend in Figure 2).

We have changed the symbols

 

 

• Could you please mention: when were the 1545 samples collected?

We have added this information in the manuscript “The samples were collected between 2010 and 2015.” Line 122

• Additionally, how long does it take the decomposition of the tree?

The age of tree decomposition is studied in several studies. Following these studies, the decay age is depending on the macroclimatic categories. The decay age is between 57 and 106 years for silver and 81 years for spruce.

Přívětivý, Tomáš & Janík, David & Unar, Pavel & Adam, Dušan & Král, Kamil & Vrška, Tomáš. (2016). How do environmental conditions affect the deadwood decomposition of European beech (Fagus sylvatica L.)?. Forest Ecology and Management. 381. 177-187. 10.1016/j.foreco.2016.09.033.

Přívětivý, Tomáš & Adam, Dušan & Vrška, Tomáš. (2018). Decay dynamics of Abies alba and Picea abies deadwood in relation to environmental conditions. Forest Ecology and Management. 427. 10.1016/j.foreco.2018.06.008.

 

• If the soil samples and the creation of the database (threethrows) were collected in different time: How can this affect your results/conclusions?

Previously, we specified in our manuscript he following statement Line 167 : “Tree disturbances have occurred at Zofin during the Holocene period from 1700 years to the present day (Samonil et al., 2020).”

 

• I think your work can benefit if you add a Figure (in the supplementary material?) showing the treethrow density in your study area.

We have added a figure with treethrow density in supplementary material (Figure S3)

Specific

Please see the attached document, which content other comments in different parts of the manuscript.

We have replied to these comments in the same document (please, see the attached document with corrections in the manuscript)

 

Author Response File: Author Response.pdf

Reviewer 4 Report

The manuscript is well written in English and it reports the effects of tree disturbances produced by treethrows on the spatial variability of soil chemical properties and on soil formation to explain the soil spatial pedocomplexity in one of the oldest forest reserves in Europe. Regarding the introduction, methods, results and discussion they are well documented. I have no specific comments to the authors.

Author Response

Thank you very much for these extraordinary comments about our work.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

I appreciate the efforts of the authors and the new English proofreader. I still have some recommendations which the authors could improve.

 

In Figure 1 I advised to change the labels IN THE figure itself, instead of giving long explanations in the caption. You could, at least “cheaply” modify Figure 1 in Powerpoint or whatever simple program you prefer to improve this in some few minutes.

 

Table 1 Already in the last review round I noted on false accuracy and by example I stated the elevation legend in Figure 1. However, I see that the authors modified or erased the specific examples I gave, but my advice was not taken to other parts of the manuscript. Now again, I see that the values in Table 1 are stated to a precision of two decimal digits. Again, just an example in this example: Did you really measure the Aluminum content to a precision of 0.01mg per kg?? I expect the authors to check their manuscript independently of this example for meaningful numbers and precision.

 

Results section, only by way of example LL372-376 (in the track-changes version): I guess that I already commented on this in the last revision round: A result is significant or not below/above p = 0.05, which is now nicely stated by the authors in the methods section. Still, the results section is full of p-value reporting. I give a specific example now and I would recommend to apply this to the whole results section and manuscript. In LL372-376, why not write: “Al(X)+1 and calcium were negatively related and the relationship was stronger in the surface than in the subsurface layer (Table/Figure… or r = … vs. r = …). Magnesium was strongly related to AL(Y)+2 and magnesium was stronger negatively related to AL+3 in the surface than in the subsurface layer.” Not that a p-value is not a measure of the strength of a relationship. The relationship is either significant or not and the strength (of a correlation) is given by the correlation coefficient. This is just one example, please check the rest of the text independently!

 

Results section, only by way of example LL360f: Still redundant information. Please remove and reference to the Figure in brackets behind the respective formulations. This is just one example, please check the rest of the text independently!

 

 

L471 a bracket (closing) missing?

 

 

Author Response

We would like to thank reviewer 2 for all the comments to our manuscript in the second round. We have significantly improved the study. We have carefully incorporated all comments and suggestions, which are indicated in blue below and in the track-changes manuscript.

 

Reviewer #2

 

I appreciate the efforts of the authors and the new English proofreader. I still have some recommendations which the authors could improve.

Thank you very much for this positive comment.

 In Figure 1 I advised to change the labels IN THE figure itself, instead of giving long explanations in the caption. You could, at least “cheaply” modify Figure 1 in Powerpoint or whatever simple program you prefer to improve this in some few minutes.

We have modified the figure 1.

 

Table 1 Already in the last review round I noted on false accuracy and by example I stated the elevation legend in Figure 1. However, I see that the authors modified or erased the specific examples I gave, but my advice was not taken to other parts of the manuscript. Now again, I see that the values in Table 1 are stated to a precision of two decimal digits. Again, just an example in this example: Did you really measure the Aluminum content to a precision of 0.01mg per kg?? I expect the authors to check their manuscript independently of this example for meaningful numbers and precision.

We have modified the Table 1 in the manuscript and in the Supplementary material. We have specified only one decimal digit. Also, in the manuscript we have modified all reference related with this suggestion. This was a mistake in the previous draft. 

Results section, only by way of example LL372-376 (in the track-changes version): I guess that I already commented on this in the last revision round: A result is significant or not below/above p = 0.05, which is now nicely stated by the authors in the methods section. Still, the results section is full of p-value reporting. I give a specific example now and I would recommend to apply this to the whole results section and manuscript. In LL372-376, why not write: “Al(X)+1 and calcium were negatively related and the relationship was stronger in the surface than in the subsurface layer (Table/Figure… or r = … vs. r = …). Magnesium was strongly related to AL(Y)+2 and magnesium was stronger negatively related to AL+3 in the surface than in the subsurface layer.” Not that a p-value is not a measure of the strength of a relationship. The relationship is either significant or not and the strength (of a correlation) is given by the correlation coefficient. This is just one example, please check the rest of the text independently!

We have modified the results section according your suggestion. Line 331-348

Results section, only by way of example LL360f: Still redundant information. Please remove and reference to the Figure in brackets behind the respective formulations. This is just one example, please check the rest of the text independently!

We have removed all redundant information related with tables, figures in the manuscript.

 L471 a bracket (closing) missing?

Yes, thank you very much

 

Author Response File: Author Response.pdf