Functional Trait Responses of C4 Bunchgrasses to Fire Return Intervals in the Semi-Arid Savanna of South Africa

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This is a very interesting paper in elucidating some of the mechanisms underlying fire responses among savanna grasses. This is very interesting from the perspective of having an ecological understanding of savanna grass fire responses. These are relevant to not only an African audience, but also a global savanna audience given that some of these grass species are broadly distributed globally. 

I have no comments or amendments to suggest on this study. Its really good.

Author Response

Authors would like to thank the reviewer for the work done.

Reviewer 2 Report

Comments and Suggestions for Authors

General comments (diversity-2687272):

Fire is regarded as having an important ecological function in the semi-arid savanna of Africa because it can increase species diversity, plant quality and community productivity. Based on continuous field observation data from 1985 to 2022, the authors mainly analyzed the recover and growth of the selected four major perennial C4 grasses under four fire treatment scenarios. Furthermore they assessed the effects of varying fire return intervals (FRIs) on grass functional traits. As a result, they concluded that C4 bunchgrasses respond differently under recurrent fires, and among the four treatments biennial and quadrennial burns (i.e. 2-, 4-FRIs) can promote these grass’ vegetative and sexual regeneration. This study may help to understand post-fire recovery dynamics of C4 bunchgrasses and to provide scientific reference for management of semi-arid savannas in Africa.

 

This study is an interesting and informative. The authors express their results in a clear and coherent way. In addition, this paper is well balanced in structure, and very smooth in language expression. Therefore, I think that it can be acceptable for publication in this journal Diversity after a MINOR REVISION.

 

Additionally, I only have a few queries regarding certain aspects.

 

Specific comments:

In the section of Introduction:

The last sentence is “Additionally, a bivariate correlation analysis was conducted to ascertain the relationships between grass functional traits.” It would be much better, if the authors delete the specific research method like “a bivariate correlation analysis” in Introduction.

 

In the section of Materials and Methods:

 “2.3. Species and data collection”

The first line on Page 4: “but however,” superfluous? Please consider deleting the word “however,”.

 

In the section of Results:

I suggest to make each plant species’ scientific name in italics in Figure 2, 3, 4, and 5, respectively.

“RT” is missing in the Table 1.

What is the meaning for the abbreviation “HT” in Figure 5?

 

In the section of Discussion:

If possible, it is suggested that the authors consider the growth habits and community habitats (i.e. the physical and chemical properties of soils) of the four major grass species when they analyze their differences in tolerance to certain fire regimes.

In my opinion, the manuscript may be much more readable if adding subheadings in the part of discussion.

 

In the section of References:

A few references are inconsistent in format. Please keep the format of manuscript in accordance with the requirements of the journal Diversity. For example, the 15th reference is different from the 19th in DOI format.

 

Comments on the Quality of English Language

This study is an interesting and informative. The authors express their results in a clear and coherent way. In addition, this paper is well balanced in structure, and very smooth in language expression.

Author Response

General comments (diversity-2687272):

Fire is regarded as having an important ecological function in the semi-arid savanna of Africa because it can increase species diversity, plant quality and community productivity. Based on continuous field observation data from 1985 to 2022, the authors mainly analyzed the recover and growth of the selected four major perennial C4 grasses under four fire treatment scenarios. Furthermore they assessed the effects of varying fire return intervals (FRIs) on grass functional traits. As a result, they concluded that C4 bunchgrasses respond differently under recurrent fires, and among the four treatments biennial and quadrennial burns (i.e. 2-, 4-FRIs) can promote these grass’ vegetative and sexual regeneration. This study may help to understand post-fire recovery dynamics of C4 bunchgrasses and to provide scientific reference for management of semi-arid savannas in Africa.

 

This study is an interesting and informative. The authors express their results in a clear and coherent way. In addition, this paper is well balanced in structure, and very smooth in language expression. Therefore, I think that it can be acceptable for publication in this journal Diversity after a MINOR REVISION.

 

Additionally, I only have a few queries regarding certain aspects.

 

Specific comments:

In the section of Introduction:

1. The last sentence is “Additionally, a bivariate correlation analysis was conducted to ascertain the relationships between grass functional traits.” It would be much better, if the authors delete the specific research method like “a bivariate correlation analysis” in Introduction.

Response

As suggested by the reviewer, “a bivariate correlation analysis” was deleted and now the sentence reads:

Additionally, we assessed the relationships between grass functional traits.

 

2. In the section ofMaterials and Methods:

 “2.3. Species and data collection”

The first line on Page 4: “but however,” superfluous? Please consider deleting the word “however,”.

Response

Line 131; “however” was deleted after “but” and the sentence reads: The latter two species are regarded as low-quality grasses, but, in the absence of the highly nutritious species they form large part of ruminant diet in semi-arid savannas of South Africa.

 

3. In the section ofResults:

I suggest to make each plant species’ scientific name in italics in Figure 2, 3, 4, and 5, respectively.

Response

Authors would like to indicate that the software used for analysis is still under development, hence it is impossible to manipulate the graphs. We therefore humbly request the reviewer to consider these graphs with non-italicized species names.

4. “RT” is missing in the Table 1.

Response

Thanks to the reviewer for noticing this. RT and its associated statistic was added in the table.

 

5. What is the meaning for the abbreviation “HT” in Figure 5?

Response

Authors would like to clarify that HT stands for Height. This abbreviation is explained below Figure 5.

 

6. In the section ofDiscussion:

If possible, it is suggested that the authors consider the growth habits and community habitats (i.e. the physical and chemical properties of soils) of the four major grass species when they analyze their differences in tolerance to certain fire regimes.

Response

Thanks to the reviewer for the advice. However, authors would like to indicate that currently they do not have the data for soil properties. Furthermore, we like to indicate that all the species studied are bunchgrasses. So, we did not make changes in the manuscript regarding these suggestions.  

7. In my opinion, the manuscript may be much more readable if adding subheadings in the part of discussion.

The advice was taken, and the subheadings were added, and they read as follows:

Effect of varying fire return intervals on grass biomass production.

Effect of fire return intervals on tillering, crown size and flowering

 

8. In the section ofReferences:

A few references are inconsistent in format. Please keep the format of manuscript in accordance with the requirements of the journal Diversity. For example, the 15th reference is different from the 19th in DOI format.

Response

All references were reworked to conform with the journal style.

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

General comments

The manuscript of Mndela et al. entitled “Functional trait responses of C4 bunchgrasses to fire return intervals in the semi-arid savanna of South Africa” is a research article that studies the effect of fire recurrence on functional traits of 4 C4 bunchgrasses in South Africa. Overall, I think the paper addresses an interesting and topical issue, with a field work over many years, which is very valuable. However, the experimental design and the presentation of the results raise very important issues that the authors should resolve before the article can be published in a journal such as Diversity.

Specific comments

Note: It would be appreciated if the line number could be included in the next versions of the manuscript, so that the authors could know exactly which sentence or part of the specific manuscript the reviewers are referring to in their comments. This would greatly facilitate the review process.

Abstract

“in two +- 0.5-ha plots”: This is the size of the plots or the degree of error in the plots ¿? Please, review and clarify.

Introduction

Beest et al. 2022: I think this citation does not appear in the references section. Please, include and check that the same does not happen with other references in the article.

Materials and methods

2.1. Study area

“the respective minimum and maximum temperatures of −5 to 11°C and 26 to 41°C”: The sentence is not clear. Please, indicate the average of the maximum and the average of the minimum.

2.2. Experimental design and layout

When did the first burn of the experiment occur? In 1980? This is not totally clear in the manuscript and should be indicated.

Is it known when was the last time the vegetation in the experimental area was burned naturally before the experiment began? This is also relevant, since it is not the same to start the experiment in an area that had not been burned for 20-30 years as in an area that burned naturally 1-2 years ago, for example. This should be indicated in the manuscript.

When was the last year that the different FRI treatments were experimentally burned? This is of utmost importance in the experimental design, since the last fire in all of them should have occurred in the same year. Otherwise, the effects observed in the results could be due not only to fire recurrence (FRI), but also to the time since the last fire.

“Although burning was not applied in a 1-year FRI in 2020, the 2021 COVID restrictions due to new variant were imposed after burning in this treatment”: If in the 1-year FRI treatment the burning stopped during one year, we would no longer have an annual frequency, but a biannual frequency, and therefore the treatment 1-year FRI should be eliminated from the experimental design, since it would not comply with its own definition.

Do you have data on the intensity of burns in the different FRI treatments (thermocouple data or similar)? This is also very important, because if the intensity of the burns was different in the different treatments, then, again, the results obtained in the manuscript could be due not only to the recurrence of fires (FRI), but also to the intensity of the fires.

Figure 1: Please, mark in some way in the figure that the 3-year treatment has not been used in the experimental design.

2.3. Species and data collection

When was the field sampling done? I assume it was at the same time for all species, but this should be indicated in the manuscript, along with the sampling dates.

Perhaps it would be interesting to include a photograph of the various study species, either in the manuscript itself or in supplementary material, so that those readers who are not familiar with the savanna species of South Africa can get an idea.

“(n = 8 plants per species per treatment), totaling to 192 plants”: This total number of plants does not fit with the experimental design mentioned in the previous section. 5 FRI treatments (0, 1, 2, 4, and 6), 2 plots per treatment and 4 subplots per plot, with 4 plants (1 per species) selected in each subplot, results in: 5 x 2 x 4 x 4 = 160 plants…. not 192 plants ¿? Please, clarify and correct if necessary.

2.4. Statistical analysis

“we manually identified and removed extreme outliers using Q-Q and boxplots”: From my point of view, this practice is incompatible with a scientific article of a certain level. The problem is that the number of plants per treatment is very low (n=8) and that is why the authors obtain data with high errors. The solution would have been to increase the n during sampling, but not to eliminate the outliers manually now, because that is really manipulating the data. I think the authors should use their entire database if they want the article to be really valid.

Results

Table 1: Data for the reproductive tillers (RT) variable are missing. Please include it.

Figures 2, 3 and 4: In the format in which the graphs are currently presented, the reader can only see a bunch of bars of different colors that do not follow any defined pattern. I propose to the authors that on the x-axis, instead of including the FRI factor, the SPECIES factor should be included, so that the bars of the different FRIs of each species are shown together. In this way, the reader can more easily see how a species behaves in general with respect to the others and if the FRI really has an effect on it. An asterisk or symbol could even be included above those species on which the FRI has a significant effect. I believe that this is what is really relevant in the article, and the figures should be aimed at making it as easy as possible for the reader to understand the results.

Figure 5: According to the experimental design the n in all species should be 40 (5 treatments x 8 plants per treatment). Why then is the n different in each species? If it is due to the outlayers that have been eliminated, I think it is something unacceptable, because for example T. triandra has less than half of the n that would correspond to it (n=16), which means a massive elimination of data. I think the authors should include in the article their entire database, without excluding any outlayer, for the article to be valid.

Figure 5: “X = non-significant at p < 0.05”: If the correlation is not significant it is because the p is higher than 0.05, not lower. Please, correct it.

In general terms, since the authors present their data in the manuscript and figures in the order ATB, TB, LB, TP, RT, CA, and CD, I recommend that they include the variables in that same order in both Table 1 and Table 2 for easier understanding by the reader.

Discussion

“Generally, the storage reserves, root biomass and efficient use of growth resources vary by species type, hence, post-fire recovery potential of grasses differs by species”: I believe that the authors should go deeper in their discussion, characterizing how these types of functional and/or ecophysiological variables are in their study species, in order to try to find a pattern to explain the high variability observed among species.

Author Response

Response to reviewer 3

 

General comments

The manuscript of Mndela et al. entitled “Functional trait responses of C4 bunchgrasses to fire return intervals in the semi-arid savanna of South Africa” is a research article that studies the effect of fire recurrence on functional traits of 4 C4 bunchgrasses in South Africa. Overall, I think the paper addresses an interesting and topical issue, with a field work over many years, which is very valuable. However, the experimental design and the presentation of the results raise very important issues that the authors should resolve before the article can be published in a journal such as Diversity.

Specific comments

Note: It would be appreciated if the line number could be included in the next versions of the manuscript, so that the authors could know exactly which sentence or part of the specific manuscript the reviewers are referring to in their comments. This would greatly facilitate the review process.

Response

Thanks to the reviewer for indicating this. The line numbers could have been omitted when the manuscript was converted to the MDPI format by the editorial office. In the current version, the line numbers are indicated.

Abstract

“in two +- 0.5-ha plots”: This is the size of the plots or the degree of error in the plots ¿? Please, review and clarify.

Response

Line 14 of the abstract; “±0.5 ha” was changed to read “0.5±0.01 ha” to indicate that plot size varied around 0.5 ha, deviating by 0.01.

Introduction

Beest et al. 2022: I think this citation does not appear in the references section. Please, include and check that the same does not happen with other references in the article.

Response

The reference “Te Beest et al. 2022” is now added as reference 17 in the list of references.

Materials and methods

2.1. Study area

“the respective minimum and maximum temperatures of −5 to 11°C and 26 to 41°C”: The sentence is not clear. Please, indicate the average of the maximum and the average of the minimum.

Response

The sentence was rephrased in line 82-83 to read: The climate of the study area is semi-arid, with a long-term rainfall of 500 mm per annum [18] and the temperatures range from 4°C in July to 28°C in February [19].

2.2. Experimental design and layout

When did the first burn of the experiment occur? In 1980? This is not totally clear in the manuscript and should be indicated.

Response

Line 95; the time of first burning was added as follows: At the initiation of the experiment, trees were removed using cut-stump treatment [22] and all plots were burned for the first time in August 1980 to ensure uniformity of the vegetation and soils across the plots [18].

Is it known when was the last time the vegetation in the experimental area was burned naturally before the experiment began? This is also relevant, since it is not the same to start the experiment in an area that had not been burned for 20-30 years as in an area that burned naturally 1-2 years ago, for example. This should be indicated in the manuscript.

When was the last year that the different FRI treatments were experimentally burned? This is of utmost importance in the experimental design, since the last fire in all of them should have occurred in the same year. Otherwise, the effects observed in the results could be due not only to fire recurrence (FRI), but also to the time since the last fire.

Response

The authors would like to clarify that since fire return intervals are different it is impossible to burn all of them during the same year, as we must wait until the cycle is finished for each FRI. For example, 4-years FRI is burned two years before burning 6-years FRI. The FRIs that are burned at the same time are 1, 2 and 4-years FRIs.

 

“Although burning was not applied in a 1-year FRI in 2020, the 2021 COVID restrictions due to new variant were imposed after burning in this treatment”: If in the 1-year FRI treatment the burning stopped during one year, we would no longer have an annual frequency, but a biannual frequency, and therefore the treatment 1-year FRI should be eliminated from the experimental design, since it would not comply with its own definition.

Response

Authors would like to indicate that it is only one year that the 1-year FRI was not applied of which based on our field observations one year of no fire did not change vegetation. The sampling for this study also was conducted at least after 2 years of burning, hence, authors would like to humble themselves to ask the reviewer to consider that this treatment has been burned for 39 years and the omission of one year did not change vegetation, as the vegetation change is a long-term process.

Do you have data on the intensity of burns in the different FRI treatments (thermocouple data or similar)? This is also very important, because if the intensity of the burns was different in the different treatments, then, again, the results obtained in the manuscript could be due not only to the recurrence of fires (FRI), but also to the intensity of the fires.

Response

Authors would like to indicate that the fire intensity was not recorded during the last burn. However, we believe that the differences in fire intensity across FRIs that the reviewer perceives would be due to indirect effects of FRIs via influence on fuel load, as the longer FRIs allow for accumulation of more fuel load that increases fire intensity compared to short FRIs.

Authors would like to thank the reviewer for this concern, in fact we will add this as part of our future research in these long-term FRIs. 

Figure 1: Please, mark in some way in the figure that the 3-year treatment has not been used in the experimental design.

Response

In Figure 1, two plots of 3-years FRI were marked with a red cross (X) to denote that these plots were not used.

2.3. Species and data collection

When was the field sampling done? I assume it was at the same time for all species, but this should be indicated in the manuscript, along with the sampling dates.

Response

Line 118; The sentence was added to clarify the time of sampling and the sentence reads:

Sampling was conducted on 15 February 2023 during peak production and flowering.

Perhaps it would be interesting to include a photograph of the various study species, either in the manuscript itself or in supplementary material, so that those readers who are not familiar with the savanna species of South Africa can get an idea.

Response

Pictures of the studied species were added as a supplemental file.

“(n = 8 plants per species per treatment), totaling to 192 plants”: This total number of plants does not fit with the experimental design mentioned in the previous section. 5 FRI treatments (0, 1, 2, 4, and 6), 2 plots per treatment and 4 subplots per plot, with 4 plants (1 per species) selected in each subplot, results in: 5 x 2 x 4 x 4 = 160 plants…. not 192 plants ¿? Please, clarify and correct if necessary.

Response

Authors would like to acknowledge that they made a mistake on calculating the number of species, we mistakenly included triennial burn in our calculation. The corrections are made in line 128 where “190 plants” was changed to “160 plants” as suggested by the reviewer.

 2.4. Statistical analysis

“we manually identified and removed extreme outliers using Q-Q and boxplots”: From my point of view, this practice is incompatible with a scientific article of a certain level. The problem is that the number of plants per treatment is very low (n=8) and that is why the authors obtain data with high errors. The solution would have been to increase the n during sampling, but not to eliminate the outliers manually now, because that is really manipulating the data. I think the authors should use their entire database if they want the article to be really valid.

Response

The removed values are returned, and the changes can be seen in Figure 5, with n = 37 for T. triandra. After returning these values there was a slight change in the P-values of some parameters e.g., line 196.

We further corrected a mistake where we reported r² instead of r because we analyzed Pearson correlations not regression.

The results interpretation was reshuffled under subtopic “Canopy and crown size”. the last sentence was moved to the beginning of the paragraph in line 212.

Results

Table 1: Data for the reproductive tillers (RT) variable are missing. Please include it.

Response

As suggested by the reviewer, the RT and the associated statistics are now included in Table 1.

Figures 2, 3 and 4: In the format in which the graphs are currently presented, the reader can only see a bunch of bars of different colors that do not follow any defined pattern. I propose to the authors that on the x-axis, instead of including the FRI factor, the SPECIES factor should be included, so that the bars of the different FRIs of each species are shown together. In this way, the reader can more easily see how a species behaves in general with respect to the others and if the FRI really has an effect on it. An asterisk or symbol could even be included above those species on which the FRI has a significant effect. I believe that this is what is really relevant in the article, and the figures should be aimed at making it as easy as possible for the reader to understand the results.

Response

As suggested by the reviewer, the presentation of figures 2, 3 and 4 were changed and the species are presented on the x-axis with superscripts to indicate significant differences between means.

Figure 5: According to the experimental design the n in all species should be 40 (5 treatments x 8 plants per treatment). Why then is the n different in each species? If it is due to the outlayers that have been eliminated, I think it is something unacceptable, because for example T. triandra has less than half of the n that would correspond to it (n=16), which means a massive elimination of data. I think the authors should include in the article their entire database, without excluding any outlayer, for the article to be valid.

Response

Thanks to the reviewer for the suggestion, indeed the sample size for T. triandra is very small. The reviewer’s suggestion was taken and the values that we regarded as outliers were returned. The changes can be seen in figure 5. The reason we removed these values at the beginning was that they caused higher variation in the data such that even when we transform it, it could not achieve normality.

Figure 5: “X = non-significant at p < 0.05”: If the correlation is not significant it is because the p is higher than 0.05, not lower. Please, correct it.

Response

Thanks to the reviewer, it appears that the fault is in the software used as the statement “X = non-significant at p < 0.05” is automated in the software. The advice from the reviewer was taken and “<” was changed to “>” in figure 5.  

In general terms, since the authors present their data in the manuscript and figures in the order ATB, TB, LB, TP, RT, CA, and CD, I recommend that they include the variables in that same order in both Table 1 and Table 2 for easier understanding by the reader.

Response

Reviewer’s suggestion was accepted, and the changes are made in Table 1. Authors would like to indicate that the manuscript has one table.

Discussion

“Generally, the storage reserves, root biomass and efficient use of growth resources vary by species type, hence, post-fire recovery potential of grasses differs by species”: I believe that the authors should go deeper in their discussion, characterizing how these types of functional and/or ecophysiological variables are in their study species, in order to try to find a pattern to explain the high variability observed among species.

Response

Line 269 to 274; A paragraph was added to strengthen the discussion section and provide more clarity on different responses of the study species.

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

This article deals with very important questions about the impact of fire frequency on plants. The article is well written, but there are some shortcomings that need to be addressed. 

1. Keywords need to be adjusted and it is important not to repeat keywords that are already in the title of the article. 

2.  In the introductory part of the article, I very much missed an analysis of the functional traits of plants. I think that this should be addressed, as it is the basis for all the data analysis and discussion. At the very least, it should be stated which sources were used to select the functional traits.

3. The aim of the article and the objectives to be addressed must be stated. The hypothesis on which the study was based could also be formulated. 

4. The section on statistical methods in the Methods section should be clarified. Some of the methods referred to, if used to analyse the data, are not clearly identified in the Results section. For example, both two-way ANOVA and GLMM are indicated, but the results obtained by these methods are not clearly stated.

5. Clarification is needed on how the biomass study was actually conducted. Parts of the current description contradict each other. If leaves were separated from stems, then how can the biomass of stems be calculated from the total biomass subtracting the biomass of leaves? It is unfortunate that the lines of the article are not numbered, which makes it difficult to provide accurate references. 

6. What is the method used to obtain the results in Table 1? Are these ANOVA results? It is necessary to specify so that the reader is clear and does not have to guess. 

7. In all cases, the x-axis names should be given in the figures. A full stop is not needed when referring to parts of the figure in lower case. In Figure 5, the x-value (actually the crossed-out correlations) should not be indicated somewhere inside the figure, but in the figure caption.

8. In my opinion, the recommendations in the conclusions and the text as a whole are very relevant for the discussion section. I would also suggest expanding the recommendations to include in which cases fires would be appropriate for habitat management and under which conditions. The conclusions could provide a much shorter summary of the study results.

9. The text of the article needs to be edited and drafted as required. There are technical errors.

 

Comments on the Quality of English Language

Moderate editing is required.

Author Response

Response to reviewer 4

This article deals with very important questions about the impact of fire frequency on plants. The article is well written, but there are some shortcomings that need to be addressed. 

1. Keywords need to be adjusted and it is important not to repeat keywords that are already in the title of the article. 

Response

On the keywords, “tillering” was changed to “tiller production” and C4 in “C4 grasses” was deleted to avoid repeating word mentioned in the title.

2. In the introductory part of the article, I very much missed an analysis of the functional traits of plants. I think that this should be addressed, as it is the basis for all the data analysis and discussion. At the very least, it should be stated which sources were used to select the functional traits.

Response

Line 38 to 42; the paragraph was added to enhance the introduction and highlight the functional traits considered in this study.

3. The aim of the article and the objectives to be addressed must be stated. The hypothesis on which the study was based could also be formulated. 

Response

Line 76 to 79; authors would like to indicate that objectives of the were listed. As suggested by the reviewer, two research questions were formulated to account for hypothesis. The questions are found in line 80 to 82 and they read as follows:

How do different C4 bunchgrasses respond along a gradient of fire return intervals? And What are the relationships between functional traits of bunchgrasses subjected to recurrent fires?

 

4. The section on statistical methods in the Methods section should be clarified. Some of the methods referred to, if used to analyse the data, are not clearly identified in the Results section. For example, both two-way ANOVA and GLMM are indicated, but the results obtained by these methods are not clearly stated.

Authors would like to indicate that the two-way interactions obtained using ANOVA and GLMM are reported under each section of the results section. Hence, no change was made to the manuscript.

5. Clarification is needed on how the biomass study was actually conducted. Parts of the current description contradict each other. If leaves were separated from stems, then how can the biomass of stems be calculated from the total biomass subtracting the biomass of leaves? It is unfortunate that the lines of the article are not numbered, which makes it difficult to provide accurate references. 

Response

Authors would like to thank the reviewer for this concern. The sentence in line 147 to 149 was rephrased to read: Tiller and leaf biomass were weighed separately and later added to determine total biomass.

6. What is the method used to obtain the results in Table 1? Are these ANOVA results? It is necessary to specify so that the reader is clear and does not have to guess. 

Response

The title of Table 1 was changed to read: ANOVA results showing the effect of species, FRIs and their interactions on plant functional traits of bunchgrasses.

7. In all cases, the x-axis names should be given in the figures. A full stop is not needed when referring to parts of the figure in lower case. In Figure 5, the x-value (actually the crossed-out correlations) should not be indicated somewhere inside the figure, but in the figure caption.

Response

All the x-axis of the figures were changed according to the reviewer’s 3 recommendations. X-axis are now represented by species instead of fire return interval. The authors would like to indicate that they did not change the numbers of FRIs to letters because we want to indicate intervals as numbers.

Authors would like to indicate that the software used for the analysis is automated such that it is impossible to manipulate the graph. Thus, authors would like to humble themselves to ask the reviewer to consider the correlation figure as it is.

8. In my opinion, the recommendations in the conclusions and the text as a whole are very relevant for the discussion section. I would also suggest expanding the recommendations to include in which cases fires would be appropriate for habitat management and under which conditions. The conclusions could provide a much shorter summary of the study results.

Response

Authors would like to express that the conclusion section was combined with the recommendations which is a general practice. However, regarding the suggestion of the reviewer to indicate use of fire for habitat management and conditions under which fires can be applied, we added sentences under discussion section in line 315 to 317. The sentences read as follows:

Fire application every 2-4 years may be more significant for plant population maintenance, habitat management and diversity conservation. Generally, maintaining T. triandra populations through application of this fire regime could be more beneficial for game and livestock production due to its high grazing value.

9. The text of the article needs to be edited and drafted as required. There are technical errors.

Response

Authors went through the whole document and rephrasing of sentences was done to improve the manuscript (see lines 145 to 151, 155 to 157, 219 to 220, 224 to 225, and 236 to 238).

Author Response File: Author Response.docx

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

The authors have modified different parts of the manuscript following the indications of the reviewers. However, there are still several issues to which they have not provided answers. If the authors cannot address these important shortcomings of the manuscript, I am afraid the article should be rejected, although I hope they can address them in the next revision. These issues are discussed below:

The time since the last fire is not considered in the experimental design:

The experimental design presents an important deficiency, as at the time elapsed since the last fire in the different fire return interval treatments (FRI) was apparently not the same when the plant sampling was done, according to the authors' response to the previous review.

To solve this, the authors should clearly specify in the manuscript the exact year in which the last fire occurred in each FRI treatment and, in case the elapsed time is not the same in all treatments (as it appears to be), the factor time since the last fire (TLF) should be included in the experimental design and analyzed together with the factors FRI and species, since it is highly probable that it has an effect on the vegetation response. In fact, some studies have already shown recently that the time since the last fire (TLF) may be even more important in ecosystem response than the degree of fire recurrence (FRI):

Albert-Belda, E., Hinojosa, M. B., Laudicina, V. A., & Moreno, J. M. (2023). Soil biogeochemistry and microbial community dynamics in Pinus pinaster Ait. forests subjected to increased fire frequency. Science of The Total Environment, 858, 159912.

The authors could also choose a more radical alternative if they do not want to introduce the TLF factor in their study, which would be to eliminate those treatments that have not been burned for the last time at the same moment.

The intensity of the burns has not been recorded:

If the intensity of the burns was different in the different treatments, which was not recorded, the results obtained in the manuscript could be due not only to the recurrence of fires (FRI), but also to the intensity of the fires.

As the authors rightly point out in their response to the previous review, the differences in fire intensity across FRIs could be due to indirect effects of FRIs via influence on fuel load, as the longer FRIs allow for accumulation of more fuel load. This should be explicitly included in the manuscript. In addition, it would also be advisable for the authors to indicate the methodology used to carry out the fires (if any type of accelerant is used for ignition, etc.), as well as the time of year and the typical meteorological conditions that occur during burns, so that the reader is clear that if there are differences in fire intensity between treatments, it is only due to the fuel load of the different FRIs and not to other factors such as burning at different seasons of the year, different methods for starting the fire, etc.

Pictures of the various study species:

The authors have included as supplementary material photos of the study species, but these images are taken from google. The ideal is to provide photos of the plants in the study area, not generic photos from google. I understand that in 40 years of monitoring numerous photos of the study species have been taken in situ, so please include those images better than the current ones.

Please, also include the name of the different species in the figure caption, as right now they only appear as A, B, C and D.

Outliers and n in the experiment:

The authors indicate that they have reintroduced into the database the outliers that they had deleted in the first version of the manuscript.

However, in Figure 5 there are still several species with an n less than 40, which means that some samples are still missing according to the experimental design. Which is the reason of this?

In addition, the new Figure 5 shows the terms “nmin”, “nmode” and “nmax”, the meaning of which is not understood and not explained anywhere in the figure caption or in the manuscript itself. What is the exact n used in each statistical analysis? This is what the authors must be made clear in Figure 5 and in the manuscript, and if the n is lower than that described in the methodology, they should explain the reason for this discrepancy.

Figure 5 and variable HT:

The height (HT) variable is included in the figure 5, but this response variable has not been statistically analyzed or represented in any graph of the manuscript. If this variable appears in the correlations of Figure 5, its statistical analysis and graph should also appear in Table 1 and corresponding figure, as in the rest of plant response variables.

Author Response

Response to reviewer 3

 

The time since the last fire is not considered in the experimental design:

The experimental design presents an important deficiency, as at the time elapsed since the last fire in the different fire return interval treatments (FRI) was apparently not the same when the plant sampling was done, according to the authors' response to the previous review.

To solve this, the authors should clearly specify in the manuscript the exact year in which the last fire occurred in each FRI treatment and, in case the elapsed time is not the same in all treatments (as it appears to be), the factor time since the last fire (TLF) should be included in the experimental design and analyzed together with the factors FRI and species, since it is highly probable that it has an effect on the vegetation response. In fact, some studies have already shown recently that the time since the last fire (TLF) may be even more important in ecosystem response than the degree of fire recurrence (FRI):

Albert-Belda, E., Hinojosa, M. B., Laudicina, V. A., & Moreno, J. M. (2023). Soil biogeochemistry and microbial community dynamics in Pinus pinaster Ait. forests subjected to increased fire frequency. Science of The Total Environment, 858, 159912.

The authors could also choose a more radical alternative if they do not want to introduce the TLF factor in their study, which would be to eliminate those treatments that have not been burned for the last time at the same moment.

Response

Authors would like to indicate that we attempted to include the time since last fire as a covariate via analysis of covariate (ANCOVA). However, due to the fact that three FRIs (1, 2 and 4) had similar time since fire and 6-years FRI was the only treatment that had different time since fire, this gave us a difficulty during analysis due to large errors we got and we decided to take the second option suggested by the reviewer to eliminate 6-years FRI (see Figures 2, 3 and 4).

The intensity of the burns has not been recorded:

If the intensity of the burns was different in the different treatments, which was not recorded, the results obtained in the manuscript could be due not only to the recurrence of fires (FRI), but also to the intensity of the fires.

As the authors rightly point out in their response to the previous review, the differences in fire intensity across FRIs could be due to indirect effects of FRIs via influence on fuel load, as the longer FRIs allow for accumulation of more fuel load. This should be explicitly included in the manuscript. In addition, it would also be advisable for the authors to indicate the methodology used to carry out the fires (if any type of accelerant is used for ignition, etc.), as well as the time of year and the typical meteorological conditions that occur during burns, so that the reader is clear that if there are differences in fire intensity between treatments, it is only due to the fuel load of the different FRIs and not to other factors such as burning at different seasons of the year, different methods for starting the fire, etc.

Response

Due to the fact that we eliminated the 6-years FRI which was not burned during the same time with other treatments, we believe that we eliminated the effect that might arise from differences in fire intensity as a result of one FRI having longer time to produce fuel load compared to others. Furthermore, we would like to indicate that fires are applied in August (spring season) during the same day to avoid variation in weather conditions and seasonal variability in fuel moisture.

In lines 109 to 122, we added details of the weather conditions and the type and method of fire application to clarify further. The paragraph reads as follows:

However, the 3- and 6-years FRIs were excluded in this study because they were not burned in the same year as other FRIs, as their inclusion would potentially cause variation in time since last since (TSF) between these FRIs. The fire is normally ignited during spring around August just after first rains to manage herbaceous vegetation composition. The fire was applied when air temperatures and humidity were on average 23°C and 52%, respectively, to avoid high intensity fires that would cause a severe damage on herbaceous plants. The wind speed for the last fire application varied temporarily with time between 15 to 18 km/hr, with fires burning for 5 to 7 minutes in each plot. The fuel load was on average 4989.0, 4459.2, 3473.0 and 4560.7 kg/ha in 0, 1, 2, and 4-years FRIs, respectively, during the last fire application. The last fires were applied in August 2022 for all fire frequencies except for 3- and 6-years FRIs which were both excluded in this study (Figure 1). Fire ignition was conducted using drip torch filled with diesel and petrol at a 60:40% ratio. The head fires were applied in which fire was ignited in the direction of wind through quick pulling of drip torch along the periphery of the plot.

 

Pictures of the various study species:

The authors have included as supplementary material photos of the study species, but these images are taken from google. The ideal is to provide photos of the plants in the study area, not generic photos from google. I understand that in 40 years of monitoring numerous photos of the study species have been taken in situ, so please include those images better than the current ones.

Please, also include the name of the different species in the figure caption, as right now they only appear as A, B, C and D.

Response

The new pictures of species were added to replace the ones taken from Google. The names of species are also listed in the caption as suggested by the reviewer.

Outliers and n in the experiment:

The authors indicate that they have reintroduced into the database the outliers that they had deleted in the first version of the manuscript.

However, in Figure 5 there are still several species with an n less than 40, which means that some samples are still missing according to the experimental design. Which is the reason of this?

Response

We would like to indicate that Digitaria eriantha and Themeda triandra were not found in two points in the 6-years FRI. However, due to elimination of this FRI in the analysis, there is no more missing data.

In addition, the new Figure 5 shows the terms “nmin”, “nmode” and “nmax”, the meaning of which is not understood and not explained anywhere in the figure caption or in the manuscript itself. What is the exact n used in each statistical analysis? This is what the authors must be made clear in Figure 5 and in the manuscript, and if the n is lower than that described in the methodology, they should explain the reason for this discrepancy.

Response

The n_max was the one used to indicate the overall n. To avoid confusion, we remained with one n value which is now 32 for all species as we removed 6-years FRI in the analysis (See Figure 5).

Figure 5 and variable HT:

The height (HT) variable is included in the figure 5, but this response variable has not been statistically analyzed or represented in any graph of the manuscript. If this variable appears in the correlations of Figure 5, its statistical analysis and graph should also appear in Table 1 and corresponding figure, as in the rest of plant response variables.

Response

The plant height was analyzed and added in Table 1 and Figure 4 (as Figure 4c).  

 

Additional comments

The graphical abstract was added as suggested by the assistant editor of the manuscript.

Due to changes which occurred because of removing 6-years FRI, F and P values in Table 1 changed for some parameters.

The results interpretation also changed as follows:

Lines 192, 221, 222, and 238: sample size and degrees of freedom changed to 128 and 9 for interactions, respectively.

 

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

The article has been substantially revised after the reviews, but there are still some shortcomings:

1. The newly added sentence (lines 139-140) is completely incomprehensible. The sentence needs to be corrected. 

2. The main issue is the statistical methods. As I wrote in the previous review, the results in Table 1 and in the text are ANOVA results, but not two-way ANOVA results as indicated in lines 153-154. I did not find any results of two-way ANOVA in the article. Similarly, I did not find GLMM results in the results, even though the use of this method is described in lines 155-156. Why do the authors use Persons r and not Spearman r_s for correlation between small data samples?

Comments on the Quality of English Language

Moderate editing is required. 

Author Response

Response to reviewer 4

1. The newly added sentence (lines 139-140) is completely incomprehensible. The sentence needs to be corrected. 

Response

The sentence in lines 139-140 was rephrased and now it reads:

From each individual plant the total number of tillers were counted, and the reproductive tillers were determined.

 

2. The main issue is the statistical methods. As I wrote in the previous review, the results in Table 1 and in the text are ANOVA results, but not two-way ANOVA results as indicated in lines 153-154. I did not find any results of two-way ANOVA in the article. Similarly, I did not find GLMM results in the results, even though the use of this method is described in lines 155-156. Why do the authors use Persons r and not Spearman r_sfor correlation between small data samples?

Response

The authors would like to indicate that indeed they misunderstood this comment in the previous review and would like to thank the reviewer for making this clarity.

Line 153, “two-way ANOVA” was changed to “ANOVA”

Authors would also like to indicate that they made a mistake for calling the model “general linear mixed model”, the correct model used was general linear model (GLM). The changes are made in lines 153 to 155. The model was also added in line 158 to make clarity and it reads as follows:

 

Authors would like to indicate that the choice of correlation statistical measure was based on the fact that we were more interested on the linear relationships which are best depicted by Pearson r rather than monotonic relationships which are best calculated by Spearman. Authors also believe that their data with n = 40 for each species was large enough to allow use of Pearson r correlation.

 

***NB

The last paragraph under the subheading “Effect of varying fire return intervals on grass biomass production” of the discussion section was moved to lines 260 to 265. This was done to ensure a good flow of discussion.

 

Author Response File: Author Response.docx