Who Eats the Grass? Grazing Pressure and Interactions Between Wild Kangaroos, Feral Goats and Rabbits, and Domestic Sheep on an Arid Australian Rangeland

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This is an interesting and relevant study. The authors aimed to examine the spatio-temporal interactions of mammalian herbivores in relation to resource availability, particularly the quantity and quality of the pasture layer. They quantified the impact of different herbivore species (sheep, free-ranging and introduced feral rabbits and goats, and native kangaroos of four species) at multiple levels: paddock, landclass within paddock, and plot within landclass. In summary, destocking did not lead to increased grazing pressure from other herbivores, with pasture biomass primarily influenced by prior rainfall and temperature. Medium-term de-stocked paddocks exhibited the greatest pasture recovery, while short-term and stocked paddocks, impacted by past overgrazing, showed less recovery. Although kangaroos exerted temporary grazing pressure, the long-term effect was pasture accrual in the absence of sheep, albeit with reduced plant diversity.

The main idea of the study is compelling and suitable for Wild readers. However, my primary concern is the volume of content and its organization throughout the manuscript. Despite being well-illustrated, the 57 pages of content make the reading experience lengthy and, at times, exhausting. Readers may struggle to maintain focus and are unlikely to read the paper in one sitting, which compromises the flow and assimilation of the material. Another clear example of the lack of fluidity and clarity in the presentation of content is the conclusion section, where the authors continue to discuss results rather than succinctly summarizing the study's key findings.

I suggest the authors revise the manuscript to condense the content, focusing on the most significant and interesting results. These should be presented more clearly and objectively to enhance readability. Additionally, please review the text for proper punctuation and ensure that cited references are presented correctly (using only brackets, without parentheses). In my view, in its current form, the study does not present its findings clearly and concisely enough for publication.

Author Response

Reviewer’s advice in italics.

We thank the reviewer for the sage advice to improve our manuscript.

The main idea of the study is compelling and suitable for Wild readers. However, my primary concern is the volume of content and its organization throughout the manuscript. Despite being well-illustrated, the 57 pages of content make the reading experience lengthy and, at times, exhausting. Readers may struggle to maintain focus and are unlikely to read the paper in one sitting, which compromises the flow and assimilation of the material. Another clear example of the lack of fluidity and clarity in the presentation of content is the conclusion section, where the authors continue to discuss results rather than succinctly summarizing the study’s key findings.

We have revised the introduction to make it more succinct and eliminated some superfluous text and citations. We have added three lines of text to clarify the aims by stating the hypotheses tested as advised by another reviewer. We have removed superfluous text from the methods but have added a short paragraph about the statistical analysis as advised by another reviewer. The results now refer only to pasture and its components. This eliminates a figure and some rows of two tables and provides consistency across the results. We only reference shrubs and trees briefly as potential competition with pasture species, especially grasses. We have revised the discussion and removed superfluous information. We have created a sub-heading to incorporate our previous ‘conclusions’ into the discussion. The reviewer is correct that the text is appropriate to discussion. As the MDPI template allows ‘Conclusions’ as an option we have opted not the include such a section for brevity. The ‘simple summary’ encapsulates our conclusions and thus another section would be repetitious.

I suggest the authors revise the manuscript to condense the content, focusing on the most significant and interesting results. These should be presented more clearly and objectively to enhance readability.

In our defence we present a study that includes results on 7 mammalian herbivores at three levels of resolution in the landscape. A typical paper on rangeland grazing would include one species or pairs of species in competition or be a desktop study. Thus, our manuscript is longer than average, but the alternative would be to partition it into several publications. We contend that our approach gives the best overview of the grazing system and its heterogeneity rather than several fragments. This is supported by another reviewer.

Additionally, please review the text for proper punctuation and ensure that cited references are presented correctly (using only brackets, without parentheses).

We apologize for the wayward parentheses which were not intentional. We have addressed this throughout the manuscript. We have likewise addressed faults in our punctuation. In addition, we have broken long sentences into concise short ones to improve readability and remove excessive punctuation.

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript is a robust monograph on the differences of grazing pressure exerted by native and livestock species in Australia. I have to commend the authors for opting to maintain the manuscript in its monograph form, not abridging it to fit modern article conventions, preserving the original ideas in their fullest.

In spite of the "outdated" methods employed, given that the study was performed in the 1990s, the results and conclusions are based on a simple yet elegant design and well-defined hypotheses. The development and results are copiously documented and supported with visual elements.

While I have no issues to raise with the scientific soundness of the manuscript, there are a few minor adjustments that I would recommend, and a few questions I would like to raise for the authors' consideration.

One thing I am curious about is regarding the lack of consideration for potential effects of trampling pressure on the vegetation, in particular given the possibility of high-density sheep herding being practiced. This could be a mechanism affecting the results observed, as the grazing pressure parameters observed did not present a clear effect on the vegetation. Is trampling pressure a significant issue on sheep herding? Do you have some idea on how the particular soil of the region could respond to livestock trampling and compactation?

on the methods: were all plants surveyed considered palatable for all potential grazers? 

on the results section, I am curious about the negative relationships found between kangaroos and some of the vegetational outcomes observed. Given that kangaroos were among the animals that were free to leave the enclosures by whichever reason they might have, has this been taken into account on the analyses? To what extent are the relationships here observed na effect of kangaroo presence and grazing pressure, and to what extent those are simply because the conditions in the plot were unfavourable and stimulated migration? I am curious of this specially given the results regarding kangaroo usage of the plots

I would recommend to convert the y axis of the graphs tracking the biomass per herbivore of the different strata and foliage groups into log. a lot of values are very close, and this could aid visualisation.

 

paragraph 827-835: there are parts of this paragraph that should be moved to the discussion. citations are better suíte for the discussion, as the authors use them to discuss particular aspects of the results.

 

line 75, 101, 1519 you can remove the parentheses around the reference

line 150 missing parenthesis

line 491 there's a word missing in "from seed to..."

 

Author Response

The reviewer’s advice is shown in italics

The manuscript is a robust monograph on the differences of grazing pressure exerted by native and livestock species in Australia. I have to commend the authors for opting to maintain the manuscript in its monograph form, not abridging it to fit modern article conventions, preserving the original ideas in their fullest.

We thank the reviewer for supporting our comprehensive and lengthy report on our study. We believe this gives the best overview of the grazing system and its heterogeneity rather than fragments in multiple publications. We appreciate the time invested in reviewing our lengthy manuscript. In response to other reviewers’ advice, we have made some modest revisions to reduce the content.

One thing I am curious about is regarding the lack of consideration for potential effects of trampling pressure on the vegetation, in particular given the possibility of high-density sheep herding being practiced. This could be a mechanism affecting the results observed, as the grazing pressure parameters observed did not present a clear effect on the vegetation. Is trampling pressure a significant issue on sheep herding? Do you have some idea on how the particular soil of the region could respond to livestock trampling and compactation?

We thank the reviewer for raising these questions. We introduced the ‘piosphere’ which is a function of both water-focussed grazing and trampling at a point water source. We addressed this briefly in respect to open plots sampled in Stocked2 and Unstocked2 where heavy trampling occurred. Sheep are more responsible for trampling and compaction of soils by virtue of their hard hooves and aggregation in flocks than kangaroos. The latter do leave narrow linear trails into water points but do not show water-focussed grazing (Reference 174 has more detail on this). A previous study on our site used a penetrometer to estimate soil compaction on the open and ungrazed plots but results were inconsistent. Although we refer to overgrazing by sheep, the density in the arid rangelands is relatively low and we refer to evidence in the results/discussion that parts of our paddocks were little used. We do agree that offtake of pasture by trampling and lack of growth on compacted soils are relevant issues.

on the methods: were all plants surveyed considered palatable for all potential grazers? 

As part of our revision, we now reference only pasture and its components. Thus, all plants presented in the results are palatable to all seven herbivore species. We do note that palatability of some plants, especially copperburrs, varies through their life cycle. At maturity many species are thorny and woody. The shrub layer is palatable to sheep, goats and, where accessible, to rabbits. Kangaroos typically do not graze/browse shrubs. We only now reference shrubs and trees as a potential competitor to pasture plants, especially grasses.

on the results section, I am curious about the negative relationships found between kangaroos and some of the vegetational outcomes observed. Given that kangaroos were among the animals that were free to leave the enclosures by whichever reason they might have, has this been taken into account on the analyses? To what extent are the relationships here observed na effect of kangaroo presence and grazing pressure, and to what extent those are simply because the conditions in the plot were unfavourable and stimulated migration? I am curious of this specially given the results regarding kangaroo usage of the plots

We report differential selection of plots by both sheep and kangaroos. We argue that the vegetation on some landclasses is more responsive to light rainfalls than others. Some other landclasses are dominated by unpalatable plants. Kangaroos, in particular, track this heterogeneity but sheep are also capable. We did radio-track a cohort of kangaroos and sheep and have extensive data on mobility. This will be the subject of a forthcoming paper.

I would recommend to convert the y axis of the graphs tracking the biomass per herbivore of the different strata and foliage groups into log. a lot of values are very close, and this could aid visualisation.

 We thank the reviewer for this sage advice. We have some zeros in the datasets that cannot be resolved on a log scale and so have not transformed the dependent variables.

paragraph 827-835: there are parts of this paragraph that should be moved to the discussion. citations are better suíte for the discussion, as the authors use them to discuss particular aspects of the results.

 We agree and have removed this reference to another study which is addressed in the discussion.

line 75, 101, 1519 you can remove the parentheses around the reference

line 150 missing parenthesis

We apologize for the haphazard use of parentheses which was not intentional. We have corrected this misuse throughout the manuscript.

line 491 there's a word missing in "from seed to..."

Corrected.

Reviewer 3 Report

Comments and Suggestions for Authors

The authors assessed grazing pressure and interactions among seven herbivore species, including four wild kangaroos. To do so, among other methods, they removed livestock to see pasture production after grazing pressure had occurred. Pasture production was examined at three scales: (1) paddock, (2) land class within paddock, and (3) overlapping excluded and grazed plots within land class and paddock. Biomass variation was assessed as a function of rainfall, temperature, and herbivore density.

Below are some brief notes that I believe may help improve understanding of the manuscript, and I hope the authors find them useful:

1.      Title: I think the title is very well chosen.

2.      Introduction:

·       The discussion is generally clear and well-structured. However, starting from line 207 it starts to become a bit difficult to follow, and I think the authors could try to shorten it and get a bit more to the point. Although I personally like long introductions, I think this one is too long and at some point it is difficult to keep the focus without getting lost. This is not an essential action, but I encourage authors to considerably summarize the introduction to improve reading.

·       The authors said “The impact of each herbivore species on their vegetation resource base was quantified at the paddock, landclass within paddock and plot within landclass and paddock levels”. I encourage authors to use this phrase to formulate the specific objectives of the study.

·       The authors said: “The results are discussed in relation to the value of de-stocking for 308 restoration or other purposes and the value of ‘total grazing pressure’ for rangeland management”. I think this phrase is not relevant here, but more appropriate for a simple summary or abstract.

·       I miss the hypotheses or predictions. Based on the literature and the authors' knowledge, what results do the authors expect to obtain? Why? The objectives and hypothesis section are a critical part of the articles in which the objectives and expectations are clearly summarized.

3.      Material and methods:

·       I think there should be a specific section within the material and methods regarding statistics in which the authors explicitly mention which are the dependent and independent variables, as well as the statistical tests used in each case and the statistical program used. It would also be good if the authors indicated here how they are going to present the results, whether they are means and standard errors or medians and IQRs. The authors also did not specify how the data was distributed or whether they performed a prior analysis of homoscedasticity to choose the most appropriate statistical methods.

·       I wonder if the authors considered running general o generalized models rather than simple linear regressions to assess multiple predictor variables as well as their interactions. If so, why did they rule this out?

4.      Results:

·       Fig 24. It seems that the authors performed Wilcoxon rank tests, which I do not recall being mentioned anywhere in the methodology. Also, I do not know why the authors use the mean and standard error instead of the median and IQR if these are non-parametric tests. As far as I understand, to compare data groups with non-parametric distribution, medians and IQRs should be reported instead of means and standard errors, since this non-parametric test does not assume normality and evaluates differences based on data ranges, not average values. Could the authors indicate the reasons why they showed the results this way?

·       Throughout the results section, comparisons between means are presented, mentioning “t-tests”, which I assume are Student T tests, but which are not mentioned in material and methods. For a better understanding for the reader, I suggest that the authors specify in material and methods each of the tests used with the variables involved in each case. Given that the article is very long and contains many results, a clear and concise explanation on the subject helps the reader to have an outline of the statistical methodology used.

·       Tables 3 and 8 are informative, but I still think that a section that previously explains all the variables that come into play in the article would be appropriate so as not to distract the reader's attention.

5.      Discussion and conclusions:

·       I am a bit confused about the conclusions section. It seems that the authors discuss the results of the article by confirming or refuting results from other articles. I think that part should go to the end of the discussion. On the contrary, I have the feeling that the last paragraph of the discussion (line 1525) does correspond to a pure conclusion. I suggest moving the discussed part of the conclusions to the end of the discussion and combining the key conclusions with the paragraph on line 1525, which should be moved to the conclusions section.

Author Response

The reviewer’s advice is given in italics.

Below are some brief notes that I believe may help improve understanding of the manuscript, and I hope the authors find them useful:

We thank the reviewer for the comprehensive advice on improvements to our manuscript and have endeavoured to accommodate all.

1. Title: I think the title is very well chosen.

We thank the reviewer.

2. Introduction:

The discussion is generally clear and well-structured. However, starting from line 207 it starts to become a bit difficult to follow, and I think the authors could try to shorten it and get a bit more to the point. Although I personally like long introductions, I think this one is too long and at some point it is difficult to keep the focus without getting lost. This is not an essential action, but I encourage authors to considerably summarize the introduction to improve reading.

We have revised the introduction and removed superfluous lines of text to improve clarity. We have also revised long sentences to short concise ones to improve readability.

The authors said “The impact of each herbivore species on their vegetation resource base was quantified at the paddock, landclass within paddock and plot within landclass and paddock levels”. I encourage authors to use this phrase to formulate the specific objectives of the study.

We have clarified the objectives based on the reviewer’s advice.

The authors said: “The results are discussed in relation to the value of de-stocking for 308 restoration or other purposes and the value of ‘total grazing pressure’ for rangeland management”. I think this phrase is not relevant here, but more appropriate for a simple summary or abstract.

We have removed this text.

I miss the hypotheses or predictions. Based on the literature and the authors' knowledge, what results do the authors expect to obtain? Why? The objectives and hypothesis section are a critical part of the articles in which the objectives and expectations are clearly summarized.

We have followed this advice and framed three hypotheses to conclude the paragraph on the aims of the study.

3. Material and methods:

I think there should be a specific section within the material and methods regarding statistics in which the authors explicitly mention which are the dependent and independent variables, as well as the statistical tests used in each case and the statistical program used. It would also be good if the authors indicated here how they are going to present the results, whether they are means and standard errors or medians and IQRs. The authors also did not specify how the data was distributed or whether they performed a prior analysis of homoscedasticity to choose the most appropriate statistical methods.

We have added a section explaining statistical testing and the presentation of central tendencies. We agree with the reviewer that the central tendency and variability from non-parametric testing should be the median and IQR. We have therefore amended such results accordingly.

I wonder if the authors considered running general o generalized models rather than simple linear regressions to assess multiple predictor variables as well as their interactions. If so, why did they rule this out?

We recognize the greater efficiency of GLM and insight into interactions. We have used multiple regression to assess the outcome from multiple predictor variables in many instances but did not examine interaction terms in detail. We have used stepwise regression to refine models of the best explanatory predictor variables. We have used simple regression mostly to present trends across samples to assist the reader’s interpretation of several figures. Plants grow in a finite space and so interactions between pasture components are likely. We have more comprehensive data on plant dynamics and perhaps this will be the subject of a future publication. We contend that our statistical analysis was sufficient to meet the objectives of the study, and we present a relevant subset of all the analyses undertaken on our large datasets.

4. Results:

Fig 24. It seems that the authors performed Wilcoxon rank tests, which I do not recall being mentioned anywhere in the methodology. Also, I do not know why the authors use the mean and standard error instead of the median and IQR if these are non-parametric tests. As far as I understand, to compare data groups with non-parametric distribution, medians and IQRs should be reported instead of means and standard errors, since this non-parametric test does not assume normality and evaluates differences based on data ranges, not average values. Could the authors indicate the reasons why they showed the results this way?

We have added a section on statistical analysis in the methods. We agree that the median and IRQ should be presented and have amended all relevant tables and figures accordingly.

Throughout the results section, comparisons between means are presented, mentioning “t-tests”, which I assume are Student T tests, but which are not mentioned in material and methods. For a better understanding for the reader, I suggest that the authors specify in material and methods each of the tests used with the variables involved in each case. Given that the article is very long and contains many results, a clear and concise explanation on the subject helps the reader to have an outline of the statistical methodology used.

In one instance we used a Paired t-test (a variant of the Students T-test) where assumptions for parametric testing were met. In more cases, we used the non-parametric equivalent, Wilcoxon signed-ranks matched pairs test, because data were categorical and not continuous.

Tables 3 and 8 are informative, but I still think that a section that previously explains all the variables that come into play in the article would be appropriate so as not to distract the reader's attention.

We have condensed Table 3 to include only pasture and its components which may simplify understanding and provides consistency across the three levels of the study. We hope this is sufficient given the other reviewers’ advice for brevity.

5.Discussion and conclusions:

• I am a bit confused about the conclusions section. It seems that the authors discuss the results of the article by confirming or refuting results from other articles. I think that part should go to the end of the discussion. On the contrary, I have the feeling that the last paragraph of the discussion (line 1525) does correspond to a pure conclusion. I suggest moving the discussed part of the conclusions to the end of the discussion and combining the key conclusions with the paragraph on line 1525, which should be moved to the conclusions section.

We apologize for this confusion and agree our ‘conclusions’ are more appropriate to the ‘discussion’. We have therefore revised the discussion and removed superfluous information. We have created a sub-heading to incorporate our previous ‘conclusions’ into the discussion. As the MDPI template allows ‘Conclusions’ as an option we have opted not the include such a section for brevity. The ‘simple summary’ encapsulates our conclusions and thus another section would be repetitious.

 

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Although I still find the manuscript somewhat extensive and dense in content, I recognize the originality and level of detail it provides, which offer clear scientific merit. It has the potential to serve as a valuable theoretical reference for future studies. Therefore, in my view, the paper aligns well with the scope of Wild.