Recommendation of an interesting analysis of ontogenic and adaptive variations in local wood properties in European Beech (Fagus sylvatica L.)

# General

I already served as a reviewer on the previous version of the manuscript. The authors have completely reworked the paper and have taken into account, wherever applicable, most of my comments. The manuscript is now considerably shorter and easier to read, focusing only on the radial patterns of three main wood properties that is ring width, specific gravity and specific modulus. I still have a couple of minor comments (see below), hope this is useful to the authors.

-          I feel the objectives still could be made clearer at the end of the introduction, with a couple of hypotheses based on the literature that would facilitate the reading and the analysis of the results regarding the main findings.

-          The authors used the present tense most of the times for the writing; I suggest using the preterit to stick to standards.

-          Regarding the decision rule to attribute UP, DOWN or FLAT radial patterns (Table 3 in the results section): it seems like the decision was made visually unless I am wrong. I think it would be better/less subjective if the decision was based on some kind of metric. Why not use the significance and the sign of spearman’s rank correlation coefficient to decide? Or something else?

-          Regarding the presentation of global mean radial patterns on L.242 to 247: I found it too vague and/or not exactly representative of what is drawn in figures. I would suggest rewording a bit these sentences so that it fits closer to the trends presented in the figures. For instance, what does mean “RW decreases regularly” L.242 while ring width clearly does not follow a linear trend and reaches a plateau after approx. 14/15 cm? Same comment for SG, it does not seem to “decrease thereafter” but it shows instead a more complex pattern. For SM, not sure the “then decreases regularly” is applicable, not sure the slight decrease can be considered actually significant.

-          I think Fig. 9 is still not up to standard quality. Maybe remove the symbols so that the trends are more visible. A better choice of contrasting colours might also help.

-          The discussion still might be a bit improved. I found the main findings directly gained from the data not clearly visible; as such, it rather feels like a sum of findings from the literature (with the exception of the last paragraph). This might be reworked without too much work.

# Misc.

-          The young modulus (MOE) appears in the Materials and methods section L.164 while it is not presented before in the introduction (only specific gravity and specific modulus are presented). Later on, MOE is sometimes omitted (e.g. Table 2, Table 3, Fig. 5, Fig. 6) but sometimes presented (e.g. Table 4 to 7, Fig. 9). A couple of words on the usefulness and/or relevance of MOE, besides SM, would be welcome.

-          Table 2: pls add the name of the parameters in the legend and indicate what the numbers from the ANOVA refer to.

-          Table 2: consider changing the symbol for non-significant effects by ns to stick to standards.

-          Table 4: it would be nice to have some kind of error (e.g. SE, confidence interval…) associated with the means presented.

-          Fig.9: why not indicate specific gravity here, instead of density, since every other figure presents specific gravity?

-          Table 6, 7 and 8: only two digits for correlation coefficients are enough. Pls also consider adding asterisks (corresponding to significance levels) next to coefficients instead of indicating the significance in the legend. Also, what kind of coefficients (Pearson’s, Spearman’s)? Pls indicate.

-          L.327, TRP: pls consider recalling the meaning of TRP here at the beginning of the discussion.

-          L.341-343: this seems out of context and should be incorporated somewhere in the discussion. 

Review: Radial variation of some wood properties in European beech

Authors stressed the topic of the paper on radial variations in beech wood properties that makes it more consistent and highlight better the originality of the study. However, some improvements in the presentation and analyses are to my opinion still necessary before publication. First, in the result section authors are jumping from the analysis of mean values and sources of its variation to analysis of radial variation patterns which results in redundancies. Result section should be streamlined and some figures removed. Figure style should be homogenized. Considering the analysis presented, some statistical assessments are missing, I do not think that visual assessments are sufficient. Further, discussion should concern all results or results that are not discussed should be removed from the paper.

Title: I think that the originality of the paper is to look at three different sources of the wood properties variability: stand site with different forest management, radial position and circumferential position. It would be nice if the title could better reflect these aspects.

Abstract

L15 First sentence is blurry, please rephrase.

L17 Objectives of the study should be introduced more clearly – analyze the variability of the radial pattern variation in function of the stand but also radius orientation?

L19 aspects of juvenility: Specify or rephrase

L23 more or less constant: I do no think we can accept this type of assessment, stick with statistical evaluation

L25 I am not sure we can assign the observed variability to differences in mechanical experienced by trees. It displays the plasticity of wood properties in relation with growth conditions in general and for sure mechanical environment of the tree is very important parameter. But to my opinion, data collected in the present paper do not allow mechanistic explanation of observed variations.

Introduction

L59 Authors present ring width as a mechanical wood trait which is not straightforward, please explicit the link, it would help the comprehension. However, can be really ring width introduced as a mechanical trait at the material level alongside with other wood properties?

L67 I am struggling with the definition of juvenility used in this paper: authors defined juvenility purely as an effect of cambial age while the origin of juvenile transition is widely discussed in the literature (presence/absence of living crown, age or diameter) and basically biological origin is not known. 

I think in this paper authors basically discuss two possible origins of the radial variations of wood properties: the effect of cambial age per se (that is one possible definition of juvenility among others) with a mechanical functional definition of juvenile transition related to the tree diameter and experienced mechanical environment. I think that wider definition of juvenility should be introduced and positioning of the paper clarified. Mechanical viewpoint is generic and may also explain the variations in properties around the tree circumference or between stands clearly pointing the limits of the juvenility based on the cambial age hypothesis.

L69 I think that reaction wood, its characteristics, location in the tree and impacts of sylviculture on its occurrence should be clearly introduced in this section. Further, forestry papers on this topic should be also cited alongside more theoretical studies (for example Dassot et al. 2015).

L94 I do not think the whole panels of figures is necessary for introduction, explanation of expected trends is sufficient. Figure presenting the relationship between the specific modulus and microfibril angle may be introduced in the material&methods section to justify the choice of measured parameters.  

L112 historical context – This wording may be misleading, please rephrase

L120 At the end of introduction, objectives and hypothesis tested in the study should be clearly introduced.

Material and Methods

L125 If available, it would be nice to have more information about the plots (rainfall, wind speed, soil type). Further, sampling of trees in the stand should be described – both, dominant and suppressed trees are considered or only dominant trees? Do they cover the whole range of presented diameters in the stand and if yes, what is the variability in diameters?

L131 20°C & 65RH in general yields equilibrium moisture content close to 12%?

L174 Does ‘core’ effect correspond to ‘rod’ effect? 

Results

L193 I do not feel comfortable with this part of analysis based on visual observations. I think that the significance and importance of the slope should be assessed for each South or North profile and a clear limit for asymmetry should be given, of course the limit will be arbitrary but at least it is quantified and identical for all profiles. Further, extension of the variability in term of distance to pith would be interesting to model. It would help for quantitative discussion of these original results and would also improve the analysis at the stand level (Figure 9).

L203 Table – Percentage of symmetry are lacking for specific modulus?

L203 It is a bit surprising that the stand 2 located on mountain slopes does not represent a high asymmetry contrary to stand 8, how do you explain that? 

L210 Authors are using both terms, specific gravity and density in the result section (Table 3 vs Table 4), please select one and use it throughout the text. 

L217 There were no systematic difference between the Northern and Southern samples: how did you test it? On mean values or on trendlines? Please specify. I think it should be evaluated as a parameter affecting the trendline (in function of radial position) and considering that the symmetry ratio is of 70%, it would be surprising that the effect is not significant?

L232 no noteworthy difference could be observed – this statement is based on visual assessment? If yes, I would recommend to use a statistical test instead. Further, I do not think that Fig. 7 is necessary (or you can move it in annex), test on mean values will be sufficient for this point. 

L251 Some strange characters in the figure legend.

L262 These plots are very interesting because they display variations of trendlines between stands however the figure is difficult to read. One option to improve could be to fit colors to the forest management type. 

L265 Global results – the term global is not very precise. Some results from this section are already present before, the others should be incorporated in the section dealing with the variability sources before the results about trendlines as both deals with variability of mean values at the stand, tree or rod level. 

L270: Data from Table 5 are already presented in Fig. 8, I don’t think that this Table is necessary. Results on the coefficient of variation may be mentioned in the description of Fig. 8. 

L284 tree dimension – number of trees?

L294 Table 8 deals with a relationship between measured properties with the idea about structure-function model which is not really introduced in the paper previously and not discussed neither. Explicit better or remove. 

L299 wood properties per tree – > per plot. These results are in the Figure 9 already, can be removed.

L310 Same comment as for Table 8. Correlation does not mean causality (L307). 0.1% level is not generally used for the significance of statistical tests, why to use this level? 

Discussion

Results about the variance components are not discussed. Results about relationship between measured parameters neither.

L318 For softwoods…Please rephrase the sentence, it is not very clear.

L325 Plourde et al. (2015) reported the relation between the inner wood density and variation expected (increase, decrease or no variation), what is expected for a medium density wood as beech? 

L333 I think the discussion about the interlocked grain trendline is out of scope or the relation should be explained.

L341 This is experimental choice and should be noted in material and methods section.

L346 (Germany, France) – Forgotten text?

L351 No information about thinning intensity is given in the study so it is difficult to related difference in the mean ring width at the plot level with the type of forest management. For example, soil properties or type of climate will be also important. Be less assertive. 

L357 This result is not at the right place and should not end the discussion section. 

Conclusion

I think that conclusion could be improved to fit better with results (please follow the same plan as for results and discussion above – mean values and partition of variation at first, radial variation patterns after) and discussion. 

L368 This is not discussed in the previous section. Typical for intraspecific variations of MOE.

Review of the paper ‘Some aspects of beech biomechanics: juvenile wood properties, sapwood pre-stress and growth forces’ by Jullien et al.

The manuscript investigates radial variation in a couple of wood traits mostly related to biomechanics (e.g. specific gravity, modulus of elasticity or specific modulus, maturation stress/strain) in trunks from 86 beech trees originating from 9 ‘natural’ forests across 5 European countries. Overall, the authors report radial patterns typical of those commonly observed for planted trees and show that pre-stress is independent of growth, whereas posture control is not.

As a side note, I am not a specialist in wood biomechanics. Since I cannot comment specifically on the theory behind, my comments remain probably very general and are mostly related to manuscript structure and readability from my understanding of the paper. Hope this is still useful to the authors.

The paper has merit with measurements performed on more than 1000 wood rods sampled from wood logs obtained on 86 adult beech trees varying from 70 to 200 years spread across 9 forest in Europe. However, in its present form, I think the paper suffers from several limits that the authors should consider. I also think the paper would benefit from language editing and a table of abbreviations would probably be a good idea considering the large number of traits and abbreviations used.

I found the structure of the paper awkward in several instances. For instance, the introduction is very short while followed by, as written by the authors, a state of the art related to spatial variation in wood properties. Besides, the choice was made to merge results and discussion (although results are not discussed in this section and there are no references) while another section is later called discussion and conclusion (and there is actually no conclusion in this section). I would recommend reshaping the whole ms structure as I do not see any reason why not following a typical layout (intro / MM / results / discussion and conclusion) and I think it actually does not require that much work.

I think the relevance of the paper is not sufficiently well presented and the paper suffers from a lack of clear objectives. The state of the art feels a bit disconnected from the work, at least it does not lead up to clear objectives and a clear rationale as it should do. There is one hypothesis stated at the end of this section (L. 149 – 151) but the rationale is not easy to catch.

In my opinion, the paper also suffers from a lack of precision and statistics. There are numerous examples of sentences like ‘There was a large majority of…’ (L. 204) or ‘Globally there were no noteworthy difference…’ (L. 211) or ‘Globally there is not significant difference…’ (L. 230) or ‘… no clear difference could be observed…’ (L. 242). The radial patterns observed between N/S directions, wood colour, trees or plots are not discussed in the light of statistical tests such that the affirmations above feel assertive without any proof. I guess there are tools available to compare patterns and/or trends that would strengthen the choices and comparisons made.

The number of figures is too high but this can be easily remedied by suppressing several unnecessary figures such as Fig. 2, 3 and 5. For Fig. 2 and 3, from my understanding these come from the literature; I think there is no need to keep them as long as patterns (and what they imply) are explained in the introduction. Fig. 5 serves as an example to illustrate N/S asymmetric patterns; I do not think this deserves necessarily a figure.

Still related to figure, I think Fig. 10 and 11 are not up to standard quality. I believe these figures can be improved through a different choice of symbols, lines, grids, etc…

The discussion is far too short considering the length of the results, with very few references. While the authors should streamline the results to be more to the point and aligned with statistics, I also think the authors should rework and strengthen the whole discussion so that findings are put in a broader literature context and the reader can see the relevance of the findings. Since the work seems to have benefited from a widespread network of forests across Europe, I actually expected the discussion to address underlying questions related to forest management, competition, climate, etc… Although this is noted at the beginning of the discussion, it is not discussed.

Specifics:

-          L.42: (bud role). Pls consider changing to ‘through bud activity’ or something similar.

-          L.43: (cambium role). Pls consider changing to ‘through cambial activity’ or something similar.

-          L.43-45: I did not get the sentence, pls reword.

-          L.56: unnecessary considering my comments above about reshaping the introduction section.

-          L.59: secondary growth performed by living wood cells. What is meant here by living wood cells? Meristematic cells? I think the sentence should be reworded, as you probably mean cell differentiation endured by cambium derivatives instead of secondary growth, which in my opinion does not refer to the same process. Need to be more specific. In addition, programmed cell death does not apply to all secondary xylem cells (e.g. parenchyma ray cells).

-          L.64: basic wood features are achieved. Such as what?

-          L.69: basic growth descriptors. Pls remove the word ‘basic’.

-          L.71: Δr. Definition is missing.

-          L.87: delete Savidge in the parentheses.

-          L.89: …. Pls change to etc…

-          L.101-107: I assume this whole paragraph would need one or several references.

-          Fig.2 inset on specific modulus: there is a typo to specific (specidic).

-          L.123: and so what?

-          L.141: see data for poplar. What data? Do you mean Fig. 3?

-          L.144: longitudinal shrinkage is a good parameter in case of reaction wood. A good parameter for what?

-          L.149-151: besides I did not get the rationale for the hypothesis, one-sentence paragraphs should be avoided.

-          L.154: such reference to projects should not be indicated in the text but kept for acknowledgements.

-          L.158: from the 9 plots. Stick to the same terms (the plots are referred to forests before).

-          L.219 and thereafter: the text switches to mainly present tense while past tense was generally used before.

-          L.231 and elsewhere: allowable. Pls change to possible.

-          Fig.7 inset on specific modulus: pls delete one parenthesis.

-          L.254 and elsewhere: such description falls short considering there are no stats or indexes that come in support.

-          L.266: I do not think comparing interindividual variation between traits based on standard deviation is incorrect since traits do not have the same units. In this case, relative variation such as coefficient of variation should be used.

-          L.277: why is that interesting?

-          L.341: GSI. Unless I am mistaken, the definition is missing.

-          L.347: were gathered in the same sheet (maturation). This is probably related to the fact that I am not a specialist, but I did not get the sentence.

-          L.351: Cirad 2015. Pls change to the authors of the report.

-          Table 10: typos (probably due to pdf export) for αm, σm and ΔF.

The sampling used in the present paper is the same as in the paper by Jullien et al. (2013) where the original biomechanical measurements (GSI distribution) in relation with the tree morphology were analysed. It seems to me that the present paper does not really bring new insights into the beech biomechanics, the hypothesis of the possible trade-off between the posture control and pre-stresses does not seem very relevant for such big trees with very low tilt angle (DBH of 60cm, 33m tall). The original data presented in the paper concern mainly the material properties and their variability. I would recommend to stress the paper to this focus and re-analyse the data as follows:

i) Partitioning of the variability (comparison of different sources of variability: plot vs tree vs radial position vs radial orientation) of different properties investigated ;

ii) Radial pattern analysis; relation with the tree morphology (data presented in the previous paper), growth speed (relation between figure 10 and 11) might be interesting as well to explain differences in the radial variation pattern in different plots;

Introduction and discussion should be revised to fit to this viewpoint and hypothesis clearly explicited. I would recommend an update of the literature refences concerning the radial variation patterns focusing more on hardwoods – works by Plourde et al., Bossu et al., Longuetaud et al., Purba et al. and likely others should be included.