Diversity Effects on Canopy Structure Change throughout a Growing Season in Experimental Grassland Communities

Round 1

Reviewer 1 Report

Please see the attached file:

Comments for author File: Comments.pdf

Author Response

#REVIEWERS COMMENTS#

To the Review 1:

 

General: This is a well-written paper that will make an important contribution to the literature. I would strongly recommend that the figures be produced at high resolution—they are not easy to read in the pdf copy that I have to read

• Thank you very much for reviewing our manuscript and for the positive overall evaluation. We are sorry that you could not see the high-resolution figures that we have uploaded to the system. we saved all figures with resolution 600 dpi and can only assume that these figures were not provided to you in the course of the review process. However, we agree that figure 4 had quite small x-axis fonts and so we decided to improve it. We have saved Figure 4 also in high resolution and we hope that it will be better visible in the editing format of the journal.

 

Equations are difficult to understand because an index of summation typically starts at, say, 1 and goes to n; your index (understood literally) starts at 5 and goes to n. It seems to me that line 265 should look like

• Thank you very much for noticing this mistake. Indeed, the mathematical formula was not correct. We consulted a mathematical reference (Bourchtein & Bourchtein, 2022; He, 2022) and checked how to better annotate in R-Markdown. Accordingly, we changed the formula in line 267, i.e. the evenness equation as you recommended. However, for the center of gravity equation, we decided to remove the term (please see the equation on the uploaded letter) since it referred only to the meaning of the center of gravity and not to a mathematical function. Now we show that the center of gravity is based on the proportion of filled voxels across strata of vegetation height and from this calculation, we identified the location with the highest volume allocated. Please see line 275 for our changes.

 

Lines 313ff: it is appropriate that blocks and plots are random effects; it is not at all clear, by your wording, what are blocks and plots nested in? A nested effect is an effect that is nested in another effect (e.g., “replications nested in treatments”); to say that “replications are nested” does not tell me what they are nested in. Please clarify

• This is an important point. We would like to clarify by spelling out our mixed effect model:

• Canopy structure metric ~ Evenness ~ -1 + time + log(Species):time, random = ~ 1| Block/Plot)
• Based on this model, we defined random effects as plots nested in blocks because we have repeated measurements in 92 plots within the 3 blocks. Now we added the following sentence: “The mixed-effects models treated block and plot as random factors in which plots are nested in blocks because we have repeated measurements (11 times) within each plot. Moreover, this parameterization allows responses to vary randomly between blocks and plots throughout the growing season. (lines 321 -322)

 

It is good that you modeled several candidate variance-covariance structures to describe the potential lack of independence. And selecting a “best” structure via AIC is routine. However, it is also clear that, if candidate structures are proposed and then some criterion is used to select a “best structure,” then (of course) the correct or true structure is unknown—this is the basis of model selection. Put this way, this is a “fishing expedition” of sorts: one proposes several structures and then uses some criterion to choose among them. And this is fine. But viewed this way, why use a package that is so limited in the candidate structures available to choose from? R’s nlme offers limited options compared to, say, SAS or SPSS—why not use a more powerful package so that you can expand the options available to best describe the potential auto-correlation in your data? It seems unfortunate to limit one’s analysis to a given package when other packages that are more powerful are available.

• Thank you very much for your comment and we appreciated a lot the suggestion about using SAS and SPSS for mixed-effects models with an auto-correlation structure. However, we considered the “nlme” package in r scientifically well recognized, as many publications have used it for mixed-effects models with repeated measurements and corrected by structure auto-correlation functions (Bolker et al., 2009; Pinheiro & Bates, 2000; Zuur et al., 2009) and for random structure in the Jena experiment with repeated measurements (Gottschall et al., 2022; Roscher et al., 2013). Therefore, we are confident that our analysis is in line with best scientific practices in our field of expertise and decided to stick with this form of analysis. Moreover, considering the manuscript was recommended with minor revisions with a due date of 5 days, we balanced in which way the potentially more powerful packages in SPSS could contribute to better interpreting the auto-correlation structure of these models, considering that we have tested several candidate models that best corrected for the temporal auto-correlation structure of our data. Finally, in the scientific community of ecologists, the free software R is the ‘lingua Franca’ and very few colleagues use and have access to the fee-based software SAS or SPSS. Using R thus strongly improves the reproducibility of our work. We hope you understand our reasoning for keeping the original analysis at this point.

 

 

With a model that tests richness effects, time, and their interaction, did you use df adjustment for simple main effects (say, based on Satterthwaite or Kenward-Rogers)? This is available in R as well as other packages.

• We used the “anova” function from the “nlme” package to test the main effects and the interactions terms taking care of degrees of freedom, however for this an F-test is used. Based on your recommendation, we now used an “anova” function from the car package which provides df adjustment for the main effects using Kenward-Rogers method. Therefore, we present a new Table 1 and 2 with a type II analysis of variance with Kenward-Roger´s method using χ² tests for terms of comparison. Please note that after using the new df adjustment test with KR, the results remained the same.

 

Results:

 

Line 325: Change “The volume calculated based on the voxel…” to “The volume calculated from the voxel…”

• - Line 333 now reads: “The volume calculated from the voxel grid approach”

 

Line 330: change “Our data shows…” to “Our data show...”

- Line 342 now reads: “Our data show that most of the volume”

 

The description of the interaction in lines 336 – 345 is important.

• - Thank you very much.

 

Lines 359ff: Change: “Moreover, the models testing the effects of diversity on the center of gravity showed a significant relationship with time…” to: “Moreover, the models testing the effects of diversity on the center of gravity showed a significant effect of time that was, however, independent of species richness.”

• - This line now reads in lines 441 to 413: “Moreover, the models testing the effects of diversity on the center of gravity showed a significant effect of time, that was, however, independent of species richness (Table 2)”

 

 

Lines 377 – 379 (“Even though…”) is a bit self-contradictory in this sense: you show no interaction (Table 2) but then you point out that “…diversity had a negative effect...during the recovery period after mowing”—and this implies to me that there may not have been a negative effect during periods other than recovery after mowing, and if this were the case, then there must also be an interaction. Further, the following text (“The strengths of the negative relationship…increased from June 17th to July 10th …but did not remain significant in consecutive dates…” also expresses clearly an interaction effect that is not supported by the results in Table 2. This needs to be thought through and presented consistently and cogently.

•  In lines 386 to 394, we presented two results. First, in Table 2, the output from the anova analysis shows the overall main effects considering the variable time and its interaction with species richness. Second, in Figure 4C, we evaluated the slopes between diversity and clumpiness of each individual time slice in the growing season separately. Figure 4C shows only three dates in which species richness negatively affected clumpiness (as you mentioned) the significant interactions. The specific dates in which clumpiness decreases with increasing species richness after the mowing pointed to a potentially crowded condition in more diverse communities, which has not happened in the other seven dates during the growing season. This might reflect the overall non-significant interactions between time and species richness in the anova analysis. To state this clearer we added in the sentence (line 388) that the non-significant interaction is an overall main effect result from the anova analysis.
• Now the sentence (lines 455 - 460) reads: “Even though the analysis of variance of the main effects showed no significant overall interaction between time and species richness, we did find significant effects within individual time slices during the growing season. Considering only the variation of the slopes, diversity had a negative effect on the clumpiness of vegetation patches during the recovery period after the mowing. The strengths of the negative relationship between clumpiness and species richness, only considering the variation of estimates, increased from June 17^th to July 10^th (p values varied between 0.01 and 0.007, respectively) but did not remain significant in the consecutive dates when the plant community again filled the available canopy space (Figure 4C).”

 

 

Discussion

1. The discussion is complete and the points that are made are generally supported by the data that are presented and the current literature. This is a good job.

• - Thank you very much.
•  
• Lines 523 – 525: the sentence “One potential reason…” does not seem to be a complete sentence…either a word is missing, or misspelled. Changing the last word from “strength” to “strengthen” might help?
• Thanks for your suggestion. Now lines 606 -608 read: “One potential reason for the strong seasonality in diversity effects on clumpiness could be that plant species show phenological asymmetry which might lead to prolonged periods of low plant-plant interaction.”

Author Response File: Author Response.docx

Reviewer 2 Report

The topic is of  interest esp in ecosystem where the measurements of traits are quite difficult. The use TSL method to determine the 2 vertical and 2 horizontal structural components were promising.

I found the results section weak. I was not sure if the data was missing species richness  8 as it is in legend but did not appear in the figures. The 11 time points covered well they cycles of all the species and the evenesss and canopy variation distributed agrees with the hypotheses mentioned b the authors.

The mathematical approach undertaken to extract the 3d point cloud data and convert it into the structural components was good. The experimental design for experiment set up was appropriate and it appears to be building on previous study already carried out at site. 

 

The authors was testing 4 hypothesis as quoted from manuscript

"We test the following hypotheses:

(1) Plant species richness increases vertical canopy evenness, most pronounced during the biomass peak due to more complementary filling of niche space and the height profile.

(2) Plant species richness increases the vertical center of gravity well before reaching and during the peak of biomass due to increased light competition and canopy height.

(3) Plant species richness will most strongly affect horizontal canopy clumpiness at phenological phases with lower interspecific interactions (at the start of the growing season and during recovery after mowing.

(4) Species richness will increase the horizontal canopy variation when height differences are most pronounced between component species, which may occur midway between the start of the growing season and the biomass peak

when using a new method, there is a need for validation. the authors did not present any data or figure that show any biological traits that have been measured. References of biomass and senescence have been made in the discussion but current findings of the study fail to corroborate this as biological traits were not mentioned but speculated that it could be that. This leads to a weakness of the study and thus the discussion was making statements without being supported by data.

Author Response

#REVIEWERS COMMENTS#

 

#To the Review 2:

The topic is of  interest esp in ecosystem where the measurements of traits are quite difficult. The use TSL method to determine the 2 vertical and 2 horizontal structural components were promising.

• Thank you very much for your revision of the manuscript. We are glad that you found our method to analyze the plant structural components using TLS technique promising.

I found the results section weak. I was not sure if the data was missing species richness  8 as it is in legend but did not appear in the figures. The 11 time points covered well they cycles of all the species and the evenesss and canopy variation distributed agree with the hypotheses mentioned b the authors.

• We are sorry to hear that you found the result section weak because of the uncertainty of the 8-species communities. The main reason is that in the very experiment we used (the trait-based experiment of the Jena Experiment) in pools 1 and 2 which our data is based on there are only two 8-species communities (please see (Ebeling et al., 2014, p. 20; Weisser et al., 2017)). Therefore, the mean lines in figures 2 and 3 represent only two communities with considerably lower variance than other community means along the species richness gradient. It is, for this reason, the box plots displaying the interquartile ranges are not shown for the 8-species mixtures. Please note, that despite the lower number of communities with greater diversity levels, the design of the experiment does provide sufficient statistical power since we have replicates in the design and also temporal repetition.
• To state this clearer, we also added a new sentence in the captions of  Figure 2. Now it reads: “Note that the eight-species communities represent two communities in our database, therefore the mean lines have lower variance than other communities and the interquartile ranges in the boxplot are not shown.”

The mathematical approach undertaken to extract the 3d point cloud data and convert it into the structural components was good. The experimental design for experiment set up was appropriate and it appears to be building on previous study already carried out at the site. 

• Thank you very much.

when using a new method, there is a need for validation. the authors did not present any data or figure that show any biological traits that have been measured. References of biomass and senescence have been made in the discussion but current findings of the study fail to corroborate this as biological traits were not mentioned but speculated that it could be that. This leads to a weakness of the study and thus the discussion was making statements without being supported by data.

• Thank you for pointing this out. We agree with your comment about validation when a new method is proposed. Therefore, I added a new sentence in the introduction section (lines 45-46) which now reads: “Standing biomass has been shown to be directly correlated to the volume of vegetation, which was recently successfully estimated using terrestrial laser scanning techniques (Cooper et al., 2017; Schulze-Brüninghoff et al., 2019; Wallace et al., 2017), as well as shown to be positively correlated with plant height (a basis for volume calculation) also using the same technique (Guimarães-Steinicke et al., 2019)".
• With this addition, we hope that it will be clear that this method was already validated in a previous study. Note that, in the Jena Experiment there are two harvests per year and in our previous studies we used two years of TLS and biomass data for the validation. Therefore, we feel confident to state that vegetation volume based on height is directly related to biomass production in these communities, and consequently, the canopy structural components calculated. Moreover, as based on this previous study, we could see from the temporal mean height distribution and from the literature of species-specific senescence periods, that at the end of the season the decreasing mean height parallels with the senescence of particular species in our communities as stated by the cited literature.

 

Author Response File: Author Response.docx