Landscape Impact on the Roadkill of Mammals in Brazil

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Dear Editor,

 

The authors of this manuscript have made an in-depth analysis of 1,418 cases of road collisions with animals. The detour of the highway was carried out daily every 3 hours, so there is no doubt about the objectivity of the collected data, which has been collected for 8 years. All types of mammals that live in the area of the highway are marked as victims of the traffic collision. Some of these species are forest dwellers, while others are inhabitants of open spaces, agricultural lands, and aquatic biotopes. Most of the noted mammals (24 species in total) are widespread animals and are not in danger of extinction, but 9 species are rare and each case of a raid can affect the state of the populations in the wild.

 

The main issue of the study is to find a link between the deaths of those species of mammals on the road that prefer a particular landscape located near the highway. The authors of the article showed that there is a direct relationship between the preferred landscape and the mammals that live there. The only exception is the capybara, which dies not only in the area of reservoirs, but also at a great distance from them.

This article is quite relevant, as more and more roads are being built everywhere and the death of animals on the roads is becoming an increasingly powerful negative factor for the conservation of species in the wild. This article clarifies that the landscape plays an important role in determining in advance which species will be most affected and, when building a new road, will allow measures to be taken to protect these particular species against high rates of roadkills on the road.

This article not only determines the frequency of road collisions, but also explains why certain mammal species are most often victims of accidents.

The only complaint about this article is that its title is too long. For example, it is better to replace a rather long local place name that few people know with the name of the country (Brazil). The topic of the article is quite general, and such cases occur not only in the south of Brazil, but in many countries of the world, so the title of the article can be significantly shortened. For example: “Landscape impact on the roadkill of mammals in Brazil”.

The methodology is not new and has been used in many articles, only this article examines the impact of the landscape factor on roadkills in great detail, which has rarely been studied so far.

The conclusions of the study are fully consistent with the collected material and the analysis of the data obtained and answer the main question posed at the beginning of the article.

The references are appropriate. The tables and diagrams illustrate the analysis very well and clearly show the patterns found.

 

Author Response

Dear Reviewer,
Thank you very much for your thorough and thoughtful evaluation of our manuscript. We sincerely appreciate your positive assessment of the study and your constructive suggestions.
We fully agree with your comment regarding the length of the title, and we have adopted your recommendation to shorten it. The revised title now reflects the broader relevance of the study, as suggested.
We are grateful for your careful review and valuable insights, which have helped us improve the clarity and impact of our work.
Kind regards,

Francisco

Reviewer 2 Report

Comments and Suggestions for Authors

Dear Authors,

I find your paper very interesting and informative, and it is well written in good English. However, prior to acceptance I would like you to address several small mistakes in Table 1.

In both column 1 and 2 there are missing parts, e.g. in column 1 rows 4, 7 and 8, years are missing. Please see also further down in the column 1.  

In column 2, in several rows the text is cut and missing. As an example, please see row 3 and 4, but also further down in the column 2.

Author Response

Dear Reviewer,
Thank you very much for your careful reading of our manuscript and for pointing out the issues in Table 1. We identified the problem, which was caused by a formatting error in the table, and we have now corrected all missing and truncated entries in both columns.
We appreciate your attention to detail and your helpful feedback.

Kind regards,

Francisco Alves

Reviewer 3 Report

Comments and Suggestions for Authors

This study uses statistical analysis to estimate the relative likelihood of roadkill along the Dom Pedro Hwy in the Cantereira-Mantiqueiria Ecological Corridor, Brazil, covering seven different landscapes. Background, context, and organization are evident. The methods model roadkill incident predictions using 1,410 recorded roadkill events over 7 years. It is unclear how the number of roadkill incidents correlates with species' migration patterns, particularly with the mean home range. Given that the most significant number of roadkills occurs with the Capybara, it would be helpful to understand the species' radius, typical terrain, and migratory patterns in relation to the modeled area. In the Model Selection, you select two models, one 3,000 m and one 6,000 m, for Group 3 - large home-range mammals. Please qualify this selection as both models show positive effects of pasture and agricultural types. It makes me wonder whether the study's grain, while showing a clear impact of human-managed landscapes, needs to clarify its scale and how that affected the results. 

 

 

Author Response

Dear Reviewer,
Thank you very much for your thoughtful comment. We would like to clarify that, to the best of our knowledge, the mammal species included in our study are not migratory, and their movements are restricted to their established home ranges. This point has been clarified in the manuscript.
The buffer distances used in the analysis were defined according to the typical home‑range sizes of each functional group, including the capybara, ensuring that the spatial scale matched the ecological scale at which species interact with the landscape. As explained in the Materials and Methods section, this rationale also applies to Group 3, for which both 3,000 m and 6,000 m radii were evaluated. The positive effects of pasture and agriculture at both scales are therefore consistent with the species’ spatial ecology and with our multiscale analytical framework. We also address the preferred habitat of capybaras in the Discussion section.

Best regards.

Francisco Alves

Reviewer 4 Report

Comments and Suggestions for Authors

This study looks into how roads affect medium- and large-sized mammal roadkill. Measuring the impact of various landscape features on roadkill occurrences and carrying out evaluations at different spatial scales are important research challenges. The paper tackles an interesting subject and is generally well-written. The following are specific revision recommendations:

1. Although "nine threatened species" are mentioned in the abstract, no representative species are listed, nor are the spatial scales that correspond to these species' conservation statuses specified. This makes it more difficult for readers to understand how serious the threat really is.
2. Lines 64–82 of the introduction say that "traditional mitigation measures focus on wildlife crossings in native vegetation and riparian zones," but they don't say how this study is different from traditional ones. This study creatively combines "functional species group classification + multi-spatial scale analysis," while traditional methods only look at places where animals are killed on the road and different types of landscapes. Readers may not understand how important this study is if it doesn't focus on this important new idea.
3. The only thing missing from lines 140–149 is the phrase "data were validated by an environmental consulting firm." For example, what percentage of samples were randomly selected to verify species identification, and what was the accuracy rate of the verification? Patrol staff often make mistakes when identifying species, which can skew roadkill data and make people doubt its accuracy.
4. Lines 187–193 say that making pseudo-absence points "generated a random point" without saying whether the generation includes the whole SP-065 highway or leaves out areas that are not good for species to live in. If inhospitable habitats are included in pseudo-absence points, the GLM model may either overestimate or underestimate the effects on the landscape.
5. In Lines 320–341 of the discussion, when explaining why capybara roadkill is unrelated to water bodies, the study only attributes this to "Jersey barriers installed on bridges, which maintain underwater connectivity." However, it provides no field survey data on the barriers and does not mention whether there is on-site observation of capybaras passing under the bridges. The argument relies solely on speculation, resulting in insufficient persuasiveness.
6. Lines 373–386 of the conclusion propose that "mitigation measures should shift from focusing on hotspots in native vegetation to agricultural/pasture areas" but fail to provide specific, actionable measures—such as how many wildlife crossings should be installed per kilometer in agricultural areas, or whether 10–20% of native vegetation patches should be retained in pastures to reduce species movement. The recommendations are overly vague and lack practical guidance value.

Author Response

Dear Reviewer,

Thank you for your careful reading and constructive feedback. The threatened species recorded in our study are listed in Table 1, along with their conservation statuses at local, national, and global scales. During revision, we also verified the total number of threatened species and corrected the abstract to reflect the accurate number (eight).

Regarding comment nº 2, we incorporated an explanation in the Introduction clarifying how our approach differs from traditional mitigation studies, strengthening the contextualization of our framework.

In response to comment nº 3, we expanded the description of the data‑validation process, clarifying the role of the environmental consulting firm and the procedures used by inspectors, whose records are reviewed by trained biologists and veterinarians.

For comment nº 4, we clarified that pseudo‑absence points were generated along the SP‑065 corridor, restricted to the same linear environment where roadkill occurs, avoiding the inclusion of unsuitable habitats and following standard practices in road‑ecology studies.

Regarding comment nº 5, we acknowledge that our initial explanation relied on inference rather than field evidence. The discussion has been revised to explicitly state this limitation and to present the mechanism as a plausible hypothesis.

Finally, in response to comment nº 6, we clarified that agricultural and pasture areas are not prioritized over forest or riparian habitats, but should also be considered in mitigation planning. As there is currently insufficient evidence to propose specific, actionable measures for these landscapes, we emphasize the need for further studies to support evidence‑based recommendations.

Kind regards,

Francisco Alves