Habitat Association of Key Wildlife Species in One of the Largest Lowland Evergreen Forests in Southeast Asia

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

In the methodology section, they mention that when they didn't find individuals of certain species along a transect, they waited a while before continuing to sample. Why do this, instead of simply stating that nothing was found along the transect?
Why weren't environmental variables such as temperature or relative humidity, among others, considered to relate species richness and abundance?
The authors mention using the Shannon index; however, it's clear that Shannon doesn't measure diversity, but rather the entropy of the system. Therefore, true diversity is now measured using Hill numbers (D0, D1, and D2), so they should remove the Shannon index.

Furthermore, I believe the authors should perform Chao extrapolation and interpolation curves to determine the significance of their sampling.

They don't mention how many times they retraced the transects. Regarding the habitat preference models they developed, they don't clearly specify which variables they considered.

The authors discuss species' habitat preference for well-preserved forests; however, they study different species within mammal and bird groups. They use only one sampling technique to measure abundance, without employing camera traps, mist nets, or nighttime sampling to draw these inferences. Therefore, I believe the authors should provide a more detailed explanation.

The study claims to use distance sampling; however:

They do not calculate detection functions.

They do not calculate true density.

They do not model detectability (p).

Without modeling detectability:

The abundance comparisons are biased.

Author Response

For research article

Response to Reviewer 1 Comments

We would like to express our sincere gratitude to the reviewer for the careful evaluation of our manuscript and for the valuable and constructive suggestions provided. The comments and recommendations have greatly contributed to improving the clarity, quality, and scientific rigor of the manuscript. Detailed responses to each comment are presented below, and all corresponding revisions have been highlighted using track changes in the revised manuscript.

 

Comments and Suggestions for Authors

• In the methodology section, they mention that when they didn't find individuals of certain species along a transect, they waited a while before continuing to sample. Why do this, instead of simply stating that nothing was found along the transect?
  Why weren't environmental variables such as temperature or relative humidity, among others, considered to relate species richness and abundance?
  The authors mention using the Shannon index; however, it's clear that Shannon doesn't measure diversity, but rather the entropy of the system. Therefore, true diversity is now measured using Hill numbers (D0, D1, and D2), so they should remove the Shannon index.

       We thank the reviewer for this valuable comment. We do this because some species are hard to detect because they are hidden or inactive, so pausing increases detection and helps avoid recording false absences. The environmental variables mentioned were not measured during the surveys; we acknowledge this gap, and they could therefore not be included in the analyses. Instead, we frame our focus on habitat characteristics and anthropogenic pressure so that we can provide a general landscape pattern. For the next survey phase, we will measure and include them in the analysis. Therefore, we have explained a bit more about our limitations in the discussion section (see lines 383-385).

In addition to the shanon diversity, we have calculated the Hill number as suggested. Please see line 305-307 in the revised manuscript.

• Furthermore, I believe the authors should perform Chao extrapolation and interpolation curves to determine the significance of their sampling. They don't mention how many times they retraced the transects. Regarding the habitat preference models they developed, they don't clearly specify which variables they considered.

       We thank the reviewer for these helpful suggestions. Chao extrapolation/ interpolation methods were not applied because our study focused on species–habitat associations of selected taxa rather than estimating total species richness. Moreover, each transect was surveyed twice (morning and evening within the same day), and this has now been explicitly stated in the Methods section (lines: 233-236).

      We have also revised the Methods to clearly specify the variables used in habitat models, including vegetation composition and anthropogenic proxies, and clarified the criteria for variable selection (lines: 240-255).

• The authors discuss species' habitat preference for well-preserved forests; however, they study different species within mammal and bird groups. They use only one sampling technique to measure abundance, without employing camera traps, mist nets, or nighttime sampling to draw these inferences. Therefore, I believe the authors should provide a more detailed explanation.

        We thank the reviewer for this important comment. We agree that our study includes two species groups (mammal and birds) and relies on a single survey method (daytime line transects). Our intent was not to comprehensively assess species’ habitat preference across all detection methods, but rather to examine patterns of habitat association based on observed occurrence under a standardized sampling approach applied consistently across the study area.

       We acknowledge that the absence of additional survey techniques (e.g., camera traps, mist nets, or nocturnal sampling) may limit detection of certain taxa and could influence inference for some species analyzed in our study. These limitations have now been clarified in the revised manuscript (lines: 435-440). Accordingly, we have revised our wording to avoid overgeneralization and emphasize habitat association rather than definitive habitat preference.

• The study claims to use distance sampling; however:
• They do not calculate detection functions.
• They do not calculate true density.
• They do not model detectability (p).
• Without modeling detectability:
• The abundance comparisons are biased.

We thank the reviewer for this important clarification. Although our field data were collected following distance sampling procedures, we did not perform formal distance sampling analyses (e.g., detection functions or density estimation). Our study focused on assessing species occurrence and relative abundance patterns rather than estimating absolute densities.

Given limitations in the dataset, particularly the lack of sufficient independent detections across distance intervals and the constraints of dense forest conditions, we considered that reliable estimation of detection functions and detectability (p) was not feasible. Applying distance sampling models under these conditions could therefore result in biased or unreliable estimates.

Accordingly, we adopted a more conservative analytical framework based on logistic regression and count-based models to examine relative patterns (lines: 236-239), while acknowledging that imperfect detection is not explicitly accounted for. We have clarified this distinction and revised the manuscript to avoid overstating the use of distance sampling and to more accurately describe the analytical approach and its limitations.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This study investigated the distribution, diversity, and habitat preferences of key wildlife species in the Prey Lang Wildlife Sanctuary, Cambodia, and assessed the impacts of forest types and anthropogenic pressures on their distribution. It confirms the ecological value of this sanctuary as a crucial lowland evergreen forest in Southeast Asia, with its intact forest blocks serving as key refuges for many threatened specialist species. The research provides scientific evidence for REDD+ projects to protect biodiversity by reducing deforestation and offers targeted management recommendations. After careful review, I suggest that the paper be accepted for publication after certain revisions. Specific comments are as follows:

1. Line 31: Delete the "and" before "great hornbill" in the keywords.
2. Table 1: The citation format for "Seed dispersal and keystone herbivore(Quin, Morgan,& Murphy, 2023)" in the "Sambar deer" row is inconsistent with other rows. It is recommended to verify if this reference is included in the bibliography. If it is, change it to the corresponding superscript number; if not, it needs to be added to the reference list and the corresponding number used here.
3. Line 190: The use of "distance to the nearest deforested areas" as a proxy for anthropogenic pressure is mentioned. However, only the analysis of "distance to the nearest village" is presented subsequently, with no mention of the results for "distance to the nearest deforested areas". If this was omitted, the results should be supplemented; if it was intentionally left out, the reason needs to be stated in the text to maintain consistency between the methods and results.
4. Section 3.1 describes gaur and sambar deer as "the least detected species, with only one sighting". However, the title of Appendix A1 is "Distribution pattern of the four least detected species (less than 20% of total transects)". The threshold "less than 20% of total transects" is a new, undefined criterion in the main text. It groups green peafowl (detected 13 times, ~17.6%), northern red muntjac (10 times, 13.5%), and Indochinese silvered langur (8 times, 10.8%) with gaur and sambar deer (detected only once, 1.4%), which is questionable. It is recommended to unify the criteria. Either explicitly define "least detected species" in the main text as "species detected in less than 20% of the total transects" and apply this definition in the results description of Section 3.1, or revise the appendix title to align more closely with the expression in the main text.
5. Lines 212-215: For comparing diversity metrics between habitat types, it states "use an unpaired t-test...; otherwise, use the Mann-Whitney U test". However, in the results section, it directly reports "based on the Mann-Whitney U test, species richness (p=0.008) and species diversity H (p=0.015)...", without explaining why the U test was chosen. The reason for selecting the non-parametric test (Mann-Whitney U) should be added in the results section.
6. Lines 388-390: The statement "One notable example of REDD+ conserving wildlife species is the Orangutan Conservation Project in Indonesia..." is a very important supporting case but is not followed by any references, weakening the persuasiveness of the argument. At least 1-2 high-quality references (e.g., academic papers, official project reports) must be provided for this case.

Author Response

Response to Reviewer 2 Comments
We sincerely thank the reviewer for taking the time to review our manuscript and for providing valuable and constructive comments. We greatly appreciate the insightful suggestions, which have helped improve the clarity and scientific quality of the manuscript. We have carefully revised the manuscript and addressed all comments in detail below. All corresponding revisions have been highlighted using track changes in the revised manuscript.

Comments and Suggestions for Authors

This study investigated the distribution, diversity, and habitat preferences of key wildlife species in the Prey Lang Wildlife Sanctuary, Cambodia, and assessed the impacts of forest types and anthropogenic pressures on their distribution. It confirms the ecological value of this sanctuary as a crucial lowland evergreen forest in Southeast Asia, with its intact forest blocks serving as key refuges for many threatened specialist species. The research provides scientific evidence for REDD+ projects to protect biodiversity by reducing deforestation and offers targeted management recommendations. After careful review, I suggest that the paper be accepted for publication after certain revisions. Specific comments are as follows:

We thank the reviewer for their constructive comments and suggestions to improve the manuscript. Please find the following for the point-by-point response to the comments.

 

1. Line 31: Delete the "and" before "great hornbill" in the keywords.

We have deleted “and” before “great hornbill”.

1. Table 1: The citation format for "Seed dispersal and keystone herbivore (Quin, Morgan,& Murphy, 2023)" in the "Sambar deer" row is inconsistent with other rows. It is recommended to verify if this reference is included in the bibliography. If it is, change it to the corresponding superscript number; if not, it needs to be added to the reference list and the corresponding number used here.

We thank the reviewer for this important suggestion. We have revised the reference and the corresponding in-text citation, which is now indicated as [22].

 

1. Line 190: The use of "distance to the nearest deforested areas" as a proxy for anthropogenic pressure is mentioned. However, only the analysis of "distance to the nearest village" is presented subsequently, with no mention of the results for "distance to the nearest deforested areas". If this was omitted, the results should be supplemented; if it was intentionally left out, the reason needs to be stated in the text to maintain consistency between the methods and results.

We thank the reviewer for this valuable comment. Actually, “Distance to the nearest deforested areas” was initially included in the full model but was not retained in the final model after AIC-based forward stepwise selection due to its low contribution to model performance. We also have clarified this in the Methods section as shown from lines: 242 – 255.

 

1. Section 3.1 describes gaur and sambar deer as "the least detected species, with only one sighting". However, the title of Appendix A1 is "Distribution pattern of the four least detected species (less than 20% of total transects)". The threshold "less than 20% of total transects" is a new, undefined criterion in the main text. It groups green peafowl (detected 13 times, ~17.6%), northern red muntjac (10 times, 13.5%), and Indochinese silvered langur (8 times, 10.8%) with gaur and sambar deer (detected only once, 1.4%), which is questionable. It is recommended to unify the criteria. Either explicitly define "least detected species" in the main text as "species detected in less than 20% of the total transects" and apply this definition in the results description of Section 3.1, or revise the appendix title to align more closely with the expression in the main text.

We thank the reviewer for this valuable comment, we have revised the manuscript and referred to only gaur and sambar as the only species with least detection. To avoid confusion and appending, we have moved all maps to the result section.

 

1. Lines 212-215: For comparing diversity metrics between habitat types, it states "use an unpaired t-test...; otherwise, use the Mann-Whitney U test". However, in the results section, it directly reports "based on the Mann-Whitney U test, species richness (p=0.008) and species diversity H (p=0.015)...", without explaining why the U test was chosen. The reason for selecting the non-parametric test (Mann-Whitney U) should be added in the results section.

As the data were not normally distributed (Shapiro–Wilk test), the Mann–Whitney U test was applied. We have added clarification in the Results section explaining that the Mann–Whitney U test was used (See line: 302- 303).

 

1. Lines 388-390: The statement "One notable example of REDD+ conserving wildlife species is the Orangutan Conservation Project in Indonesia..." is a very important supporting case but is not followed by any references, weakening the persuasiveness of the argument. At least 1-2 high-quality references (e.g., academic papers, official project reports) must be provided for this case.

We thank the reviewer for this valuable suggestion. We have revised the text to improve clarity by adding one references directly after “ in particular ecosystems” and two references “the Orangutan Conservation Project in Indonesia” to provide clearer and stronger supporting evidence (Please see line 455-457).

 

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

General comment:

The authors present an interesting study regarding key wildlife species of Cambodia’s Prey Lang Wildlife Sanctuary. Despite the substantial and particularly challenging field data collection, the analytical treatment of the data is somewhat limited. In addition, more consistent use of terminology is needed throughout the manuscript. Below I provide my comments, which I hope the authors will find useful in revising the manuscript. While the study has potential for publication, I strongly encourage a major revision.

Major issues:

• The field sampling effort conducted in this study is substantial and appears particularly challenging given the dense forest conditions and repeated transect surveys. However, the analytical treatment of the data may not fully exploit the information collected. By reducing repeated detections to simple presence/absence summaries, the analyses effectively ignore imperfect detection and part of the temporal replication inherent in the sampling design. Because each transect was surveyed twice (even if it was on the same day), the study design appears suitable for the application of single-season occupancy modeling, which would allow separate estimation of occurrence (occupancy probability, ψ) and detection probability (p). This could substantially strengthen ecological inference, particularly for rare or cryptic species where non-detection does not necessarily imply absence. In addition, while occupancy modeling may not ultimately be feasible for all species because of limited repeat surveys or sparse detections, the limitations associated with imperfect detection should be discussed more explicitly.

 

• Similarly, if count data were recorded during surveys, abundance could potentially be modelled using N-mixture models or related count-based frameworks, rather than reducing the data to a basic spatial representation (map), which does not allow robust inference about the observed patterns.

 

• The manuscript would also benefit from more careful and consistent use of ecological terminology. Terms such as “presence” and “occupancy”, as well as “habitat preference” and “habitat association”, are currently used somewhat interchangeably, although they represent distinct concepts with different inferential meanings, particularly with respect to imperfect detection and habitat selection versus statistical association.

 

Specific comments:

Lines 19-21: Please provide the scientific name of each species when first appeared in the text.

Line 36: The statement assumes that species occupy “suitable” habitats providing all essential resources. However, habitat use does not always correspond to optimal conditions, as animals may persist in suboptimal or constrained environments due to various reasons (e.g.: competition, disturbance, or landscape fragmentation). As such, it may be preferable to avoid or more clearly define the term “suitable”.

Lines 39-42: Same as before (please provide the scientific name of each species when first appeared in the text).

Line 45: Consider modifying to “species-habitat relationships”

Lines 49-50: Same as before (please provide the scientific name of each species when first appeared in the text).

Line 51: Replace “form” with “from”

Line 65: Consider modifying from “poorly understood” to “poorly studied”

Lines 72-73: The stated objective refers to assessing the “levels of impact of evergreen forest proportion and anthropogenic pressure on faunal distribution.” However, the analytical approach is limited to logistic regression models of presence/absence, which estimate variation in occurrence probability rather than spatial distribution per se. In addition, “distribution” is only illustrated using descriptive mapping without accompanying spatial or inferential analyses. As a result, the analyses appear better aligned with examining species-habitat relationships rather than formally assessing species distribution patterns.

Line 82: Please modify “km2” to “km²”

Line 83: Remove the dash “-“ from “deciduous”

Line 99: Remove the dash “-“ from “plantations”

Line 102: Figure 1. High- and moderate-intact forests should also be included in the legend of the map.

Lines 114-115: “For additional information on each species, we refer to Table 1 below.” I believe this sentence is unnecessary and could perhaps be removed entirely from the main text.

Line 125: Please revise “….selected 400 1 km² grids”

Lines 123-131: The authors state that the reduced sample of 74 grids represents a “representative” subset of the study area. While random selection can reduce sampling bias within a defined sampling frame, representativeness depends critically on whether the initial set of 149 grids adequately captures the full variability of habitat types and environmental gradients in the landscape. Given that grids adjacent to one another and previously deforested areas were excluded prior to random sampling, it is unclear whether some habitat types may be underrepresented or omitted entirely from the final analytical dataset. This limitation should be acknowledged, particularly with respect to inference about habitat-related effects.

Lines 151-159: Since each square transect was walked twice, the authors could perhaps consider using other analytical approaches when analyzing their data (e.g.: occupancy and N-mixture models). In addition, the effect of “observer” could be used as a variable influencing detection probability of each species.

Lines 158-159: The study relies exclusively on visual encounter data for estimating species presence and abundance. While this approach is common in ecological surveys, it may introduce substantial detectability bias, particularly in dense forest environments where many species are cryptic, nocturnal, or otherwise difficult to observe. The manuscript would benefit from a clearer justification for the exclusive use of visual encounters, as well as a discussion of how this may influence estimates of species richness, abundance, and habitat associations.

Lines 178-191: Could you please provide the format and spatial resolution of the forest cover, deforestation and administrative data?

In addition, how is “an area of 9 km² indicative of habitat preference for animals’ daily activities, especially foraging”? Could you perhaps be more specific, considering the species under consideration in this study? Is 9 km² indicative of all species analyzed?

Finally, did you check for multicollinearity between variables before conducting the analyses that followed?

Line 210: Consider using an alternative term to “project” when referring to map-based visualizations, as this wording may be interpreted as model-based spatial projections or extrapolations. If the intent is to present a descriptive spatial illustration of observed data, a more neutral term such as “visualize” would be more appropriate.

Section “2.6.2. Influences of Habitat Characteristics and Anthropogenic Pressure”: Please refer to my “Major issues” comments for this part.

Line 241: Please provide the citation for the “R statistical software”

Lines 247-250: The scientific names are not needed if they have previously appeared in the text.

Lines 250-253: The statement that “no distinct spatial pattern emerged” appears to be based solely on visual interpretation of mapped data. However, no formal spatial statistical analyses were conducted to test for clustering, gradients, or non-random spatial structure in species presence or abundance. As such, this conclusion should be interpreted cautiously or rephrased to reflect its descriptive nature.

Figure 3: Please provide a more explanatory description of the legend in each map. In addition, consider changing the caption of the figure as follows “Figure 3. The distribution pattern of (a) species richness and (b) abundance of the target species recorded from the transect survey.”

Line 263: Remove the additional full stop at the end of the heading. In addition, could you include the measurement unit for “cluster size” and “encounter rate”?

Line 276: Perhaps it would be useful if the authors could include overall model results for each species in a supplementary file.

Section “3.2. Habitat Preference and Anthropogenic Pressure”: Coefficient estimates should be reported using the regression coefficient (β). In addition, standard errors (SE), 95% confidence intervals (95% CI), and p-values (where appropriate) should be provided to allow proper assessment of estimate precision and statistical significance. This would improve interpretability and allow readers to evaluate the strength and uncertainty of the reported effects.

Line 289: The heading of Table 3 could be revised to better reflect the analysis presented, for example “Species-habitat associations based on logistic regression results” (or similar wording). The current phrasing may imply inference about habitat preference, whereas the analysis appears to assess habitat associations in terms of occurrence probability. The manuscript would benefit from consistently distinguishing between habitat preference, habitat use, and statistical habitat-occurrence associations throughout.

Line 299: The terminology used to describe Figure 6 and the corresponding text is unclear and potentially misleading. The figure appears to present frequency-based measures of presence/absence, whereas the text refers to “presence probability,” which typically implies a model-derived probability. The authors should clarify whether these values represent empirical frequencies or predicted probabilities, and ensure consistent terminology throughout the manuscript.

Line 304: The terms “occurrence frequency” and “presence probability” appear to be used interchangeably, although they represent different quantities. The authors should ensure consistent terminology throughout the manuscript.

Figure 6: The use of occurrence frequency across high-intact and medium-intact forest habitats is used to infer habitat “preference.” However, occurrence frequency alone reflects patterns of use within the sampled design and does not account for habitat availability or detectability differences. As such, it does not provide a robust basis for inferring habitat preference in the formal ecological sense, but rather describes relative occurrence across habitat categories. The manuscript should therefore rephrase these interpretations as habitat associations.

Line 318: Similar to the above, this finding does not necessarily imply “preference”.

Line 339: Perhaps another wording for “…exhibited greater adaptability” would be more appropriate. For example: “….exhibited less specialization”.

Lines 354-368: The authors acknowledge substantial evidence of imperfect and heterogeneous detectability for multiple species, including non-detection despite confirmed presence. This suggests that detection probability is < 1 and likely varies across species and habitats. However, this limitation is not accounted for in the analytical approach, and results should therefore be interpreted as conditional on detection rather than true occupancy.

Line 370: Use the words “detections” or “sightings” instead of “occurrences”

Lines 383-386: The authors report the presence of anthropogenic pressures such as hunting, habitat fragmentation, agricultural clearance, and illegal logging. They subsequently conclude that these factors “inevitably affect” wildlife distribution and populations. However, these relationships (at least not all of them) are not quantitatively tested in the manuscript. As such, this statement appears speculative and should be reframed to reflect observational evidence rather than inferred effects.

Lines 400-401: Could you please provide a relevant reference. In addition, “absence” does not necessarily mean that a species is not present in the area. Absence could arise from imperfect detection, i.e..: 9 species were detected in this study, while 54 were detected in the same area using a different methodological approach. Do you mean that these species have gone entirely extinct from the area?

Line 408: Do you mean strict protection against deforestation, illegal logging, hunting, etc?

Line 419: Author contributions should follow the journal’s format guidelines

Author Response

Response to Reviewer 3 Comments
We sincerely thank the reviewer for taking the time to carefully review our manuscript and for providing valuable and constructive comments. We greatly appreciate the insightful suggestions, which have helped improve the clarity, quality, and scientific rigor of the manuscript. Please find the following for the point-by-point response to the comments and suggestions. All corresponding revisions have been highlighted using track changes in the revised manuscript.   Comments and Suggestions for Authors General comment: The authors present an interesting study regarding key wildlife species of Cambodia’s Prey Lang Wildlife Sanctuary. Despite the substantial and particularly challenging field data collection, the analytical treatment of the data is somewhat limited. In addition, more consistent use of terminology is needed throughout the manuscript. Below I provide my comments, which I hope the authors will find useful in revising the manuscript. While the study has potential for publication, I strongly encourage a major revision. We thank the reviewer for the positive comments and constructive suggestions. We have carefully revised the manuscript to address the concerns raised, including improving the analysis, analytical limitations and ensuring consistent terminology throughout the manuscript. We believe these revisions have strengthened the study. Major issues: ·        The field sampling effort conducted in this study is substantial and appears particularly challenging given the dense forest conditions and repeated transect surveys. However, the analytical treatment of the data may not fully exploit the information collected. By reducing repeated detections to simple presence/absence summaries, the analyses effectively ignore imperfect detection and part of the temporal replication inherent in the sampling design. Because each transect was surveyed twice (even if it was on the same day), the study design appears suitable for the application of single-season occupancy modeling, which would allow separate estimation of occurrence (occupancy probability, ψ) and detection probability (p). This could substantially strengthen ecological inference, particularly for rare or cryptic species where non-detection does not necessarily imply absence. In addition, while occupancy modeling may not ultimately be feasible for all species because of limited repeat surveys or sparse detections, the limitations associated with imperfect detection should be discussed more explicitly.         We thank you very much the reviewer for this valuable suggestion. While our sampling design included two surveys per transect, both surveys were conducted within the same day and are therefore unlikely to represent independent detection occasions. In particular, repeated surveys may have influenced animal behavior (e.g., temporary avoidance), especially for disturbance-sensitive species, violating key assumptions required for occupancy modelling. As a result, the data do not provide sufficient independent temporal replication to reliably estimate detection probability (p) separately from occupancy (ψ). Applying occupancy models under these conditions could therefore also lead to biased or misleading inference. We have clarified this limitation in the revised manuscript and now explicitly acknowledge that our presence/absence approach does not account for imperfect detection (see lines: 233-239). Given the constraints of the dataset, we adopted a conservative analytical approach as described in our study to avoid violating model assumptions.   ·        Similarly, if count data were recorded during surveys, abundance could potentially be modelled using N-mixture models or related count-based frameworks, rather than reducing the data to a basic spatial representation (map), which does not allow robust inference about the observed patterns.         We thank the reviewer for this valuable suggestion regarding the use of N‑mixture models to better analyze count data. We agree that such approaches can provide robust inference on abundance while accounting for imperfect detection when their assumptions are met. However, we opted not to apply N‑mixture models in this study due to limitations in the sampling design. Although count data were recorded, repeated surveys at each transect were conducted within the same day and therefore cannot be considered independent detection occasions. For disturbance-sensitive species, particularly mammals, detections during subsequent surveys may have been influenced by behavioral responses (e.g., temporary avoidance or reduced activity), violating the assumption of independent counts required for reliable estimation of detection probability. Applying N‑mixture models under these conditions could therefore bias estimates of abundance and detection. Instead, to strengthen inference beyond simple spatial representation, we applied model-based analyses of observed counts using negative binomial generalized linear models (GLMs). This approach accounts for overdispersion in count data and allows examination of relationships between relative abundance (observed counts) and ecological covariates (e.g., habitat quality and proximity to human disturbance), while avoiding violations of key model assumptions. We also provided the description of the analysis (lines: 267-272) and the summary result (lines: 361-366) in the revised manuscript. ·        The manuscript would also benefit from more careful and consistent use of ecological terminology. Terms such as “presence” and “occupancy”, as well as “habitat preference” and “habitat association”, are currently used somewhat interchangeably, although they represent distinct concepts with different inferential meanings, particularly with respect to imperfect detection and habitat selection versus statistical association
We have modified the entire manuscript regarding the terms used. Specifically, we have replaced the term “preference” or “prefer” by the term “association” or “associated”, respectively.   Specific comments: ·        Lines 19-21: Please provide the scientific name of each species when first appeared in the text. We thank the reviewer for this valuable comment. We have revised the manuscript accordingly by providing the scientific names of all species at their first mention in the text to ensure clarity and consistency throughout the manuscript. ·        Line 36: The statement assumes that species occupy “suitable” habitats providing all essential resources. However, habitat use does not always correspond to optimal conditions, as animals may persist in suboptimal or constrained environments due to various reasons (e.g.: competition, disturbance, or landscape fragmentation). As such, it may be preferable to avoid or more clearly define the term “suitable”. We have revised the sentence by removing the term “suitable” and replaced by “depend on habitat that provides essential resources…” (line: 39)   ·        Lines 39-42: Same as before (please provide the scientific name of each species when first appeared in the text). This was also revised based on your comment.   ·        Line 45: Consider modifying to “species-habitat relationships” This has been modified in the text. ·        Lines 49-50: Same as before (please provide the scientific name of each species when first appeared in the text). This has been revised same as the previous ones. ·        Line 51: Replace “form” with “from” “form” has been replaced with “from”   ·        Line 65: Consider modifying from “poorly understood” to “poorly studied” “Poorly understood” has been modified to “poorly studied based on this comment.   ·        Lines 72-73: The stated objective refers to assessing the “levels of impact of evergreen forest proportion and anthropogenic pressure on faunal distribution.” However, the analytical approach is limited to logistic regression models of presence/absence, which estimate variation in occurrence probability rather than spatial distribution per se. In addition, “distribution” is only illustrated using descriptive mapping without accompanying spatial or inferential analyses. As a result, the analyses appear better aligned with examining species-habitat relationships rather than formally assessing species distribution patterns.         We thank the reviewer for this helpful observation. We agree that our analyses are more aligned with examining species–habitat relationships rather than formally modelling spatial distribution nor the habitat preference. Accordingly, we have revised the manuscript to emphasize “habitat association” rather than “preference”. The logistic regression models assess how habitat characteristics and anthropogenic factors influence species occurrence, which we interpret as habitat association. Distribution maps we used here are presented as a descriptive and conservative visualization of observed patterns, rather than as inferential spatial analyses. We have modified the objective #1 as “ to describe patterns of species occurrence, diversity, richness, and relative abundance of ecologically important mammal and bird species”.   ·        Line 82: Please modify “km2” to “km2” This has been modified in the manuscript.   ·        Line 83: Remove the dash “-“ from “deciduous” It has been removed in the manuscript as suggested.   ·        Line 99: Remove the dash “-“ from “plantations” It has been removed in the manuscript as suggested.   ·        Line 102: Figure 1. High- and moderate-intact forests should also be included in the legend of the map.         We thank the reviewer for this valuable suggestion. We clarified the legend by adding a separate “Quadrat-Transects” section showing the high-intact and moderate-intact forest categories, which represent localized sampling units based on evergreen forest cover thresholds (>80% and < 80%, respectively) rather than landscape-wide land cover classes.   ·        Lines 114-115: “For additional information on each species, we refer to Table 1 below.” I believe this sentence is unnecessary and could perhaps be removed entirely from the main text. With this suggestion, we have removed the sentence from the manuscript to improve conciseness and readability.   ·        Line 125: Please revise “….selected 400 1 km2 grids” It has been revised in the text.   ·        Lines 123-131: The authors state that the reduced sample of 74 grids represents a “representative” subset of the study area. While random selection can reduce sampling bias within a defined sampling frame, representativeness depends critically on whether the initial set of 149 grids adequately captures the full variability of habitat types and environmental gradients in the landscape. Given that grids adjacent to one another and previously deforested areas were excluded prior to random sampling, it is unclear whether some habitat types may be underrepresented or omitted entirely from the final analytical dataset. This limitation should be acknowledged, particularly with respect to inference about habitat-related effects. Following your suggestion, we have revised the text by removing the term “representative”. We revised the text as here “We removed grids that were previously deforested, adjacent or connected to each other, although may reduce the representation of habitat characteristics in this area, to allow to maintain spatial independence of the sampled grids” (please see line 130-132). Lines 151-159: Since each square transect was walked twice, the authors could perhaps consider using other analytical approaches when analyzing their data (e.g.: occupancy and N-mixture models). In addition, the effect of “observer” could be used as a variable influencing detection probability of each species.         We thank the reviewer for this suggestion. While count data were recorded, repeated surveys were conducted within the same day and therefore could not be considered independent sampling occasions. As noted above, this limits the application of occupancy and N‑mixture models, which require independent replicate counts to obtain reliable estimates. Applying these models under our conditions could lead to biased inference, and we therefore did not adopt this approach.         For the effect of observers, we acknowledged this aspect and we also discussed this matter in the discussion. Sampling techniques (e.g., camera trapping or nocturnal surveys), together with variability among observers in the field (e.g., differences in experience, skill level, attention, and familiarity with species identification), may have may have limited detection probability for our studied taxa and could influence estimates of their occurrence and abundance  (lines: 438-440). ·        Lines 158-159: The study relies exclusively on visual encounter data for estimating species presence and abundance. While this approach is common in ecological surveys, it may introduce substantial detectability bias, particularly in dense forest environments where many species are cryptic, nocturnal, or otherwise difficult to observe. The manuscript would benefit from a clearer justification for the exclusive use of visual encounters, as well as a discussion of how this may influence estimates of species richness, abundance, and habitat associations.         We thank the reviewer for this important comment. We have clarified in the Methods section that direct visual sightings along line transects were used to ensure standardized observations during the surveys (See line 161-165). We also discussed in the Discussion section that some cryptic or elusive species may have been difficult to detect in dense forest habitats using this method (See line: 415-424) ·        Lines 178-191: Could you please provide the format and spatial resolution of the forest cover, deforestation and administrative data?         We thank the reviewer for this valuable suggestion. We have revised the manuscript to explicitly include the format and spatial resolution of these datasets. The text now reads: " The forest cover and deforestation datasets were sourced from the Regional Land Cover Monitoring System (RLCMS) and GLAD laboratory databases and acquired as georeferenced raster datasets (GeoTIFF) with a spatial resolution of 30 m, while the administrative data were provided as vector shapefiles” (See line 201-205). In addition, how is “an area of 9 km2 indicative of habitat preference for animals’ daily activities, especially foraging”? Could you perhaps be more specific, considering the species under consideration in this study? Is 9 km2 indicative of all species analyzed?​         We thank the reviewer for this excellent point. We agree that a single spatial scale cannot perfectly mirror the exact home range of every species, but it serves as an optimized multi-species baseline for our target fauna. As suggested, we have updated the text in Section 2.5 (Methodology) to be specific to our study species, explaining how this scale encompasses local foraging patches for wide-ranging species (e.g., great hornbill, gaur, sambar deer) while capturing the landscape matrix and edge effects for primates and generalists. Finally, did you check for multicollinearity between variables before conducting the analyses that followed?         We thank the reviewer for this excellent question. All variables were initially included in the logistic regression analysis without assessing multicollinearity to ensure that no potential predictors were excluded.  We instead relied on the AIC-based forward stepwise selection procedure to iteratively eliminate less important or redundant variables, resulting in a parsimonious final model that retains predictors with meaningful contributions while preserving overall model performance. We have added a brief clarification to the methodology section of the revised manuscript to reflect this: Multicollinearity among the predictors was assessed using the Variance Inflation Fac-tor (VIF), with all values falling within acceptable thresholds. (See Lines 249–257)   ·        Line 210: Consider using an alternative term to “project” when referring to map-based visualizations, as this wording may be interpreted as model-based spatial projections or extrapolations. If the intent is to present a descriptive spatial illustration of observed data, a more neutral term such as “visualize” would be more appropriate. The word visualize has been replaced the word “project” in the text as suggested.   ·        Section “2.6.2. Influences of Habitat Characteristics and Anthropogenic Pressure”: Please refer to my “Major issues” comments for this part. We have modified the title and replaced the term “Habitat Preference” by the term “Habitat Association” across the whole manuscript.   ·        Line 241: Please provide the citation for the “R statistical software” The citation for this has been added.   ·        Lines 247-250: The scientific names are not needed if they have previously appeared in the text. The scientific name of each species in that section was deleted. ·        Lines 250-253: The statement that “no distinct spatial pattern emerged” appears to be based solely on visual interpretation of mapped data. However, no formal spatial statistical analyses were conducted to test for clustering, gradients, or non-random spatial structure in species presence or abundance. As such, this conclusion should be interpreted cautiously or rephrased to reflect its descriptive nature.       We thank the reviewer for this insightful comment. we have revised the text to clarify that the observation is based on visual inspection of the mapped data and does not reflect results from formal spatial statistical analyses (see line: 282-283).   ·        Figure 3: Please provide a more explanatory description of the legend in each map. In addition, consider changing the caption of the figure as follows “Figure 3. The distribution pattern of (a) species richness and (b) abundance of the target species recorded from the transect survey.” These have been modified and changed based on your suggestion (See the updated figure 3).   ·        Line 263: Remove the additional full stop at the end of the heading. In addition, could you include the measurement unit for “cluster size” and “encounter rate”?       We thank the reviewer for this important suggestion. We have now removed the additional full stop added the measurement units for cluster size and encounter rate in Table 2.   ·        Line 276: Perhaps it would be useful if the authors could include overall model results for each species in a supplementary file.       We thank the reviewer for this important suggestion. We have created a table that include overall model results for each species in a supplementary file (Tale S1) Section “3.2. Habitat Preference and Anthropogenic Pressure”: Coefficient estimates should be reported using the regression coefficient (β). In addition, standard errors (SE), 95% confidence intervals (95% CI), and p-values (where appropriate) should be provided to allow proper assessment of estimate precision and statistical significance. This would improve interpretability and allow readers to evaluate the strength and uncertainty of the reported effects.         We thank the reviewer for this valuable suggestion. Overall model results have been provided in a supplementary file (Table S1). We have also revised Section 3.2 to report all coefficient estimates using regression coefficients (β), along with their SE, 95% CI, and p-values where appropriate.   ·        Line 289: The heading of Table 3 could be revised to better reflect the analysis presented, for example “Species-habitat associations based on logistic regression results” (or similar wording). The current phrasing may imply inference about habitat preference, whereas the analysis appears to assess habitat associations in terms of occurrence probability. The manuscript would benefit from consistently distinguishing between habitat preference, habitat use, and statistical habitat-occurrence associations throughout.         We thank the reviewer for this constructive suggestion. We agree with the importance of distinguishing between ecological preference and statistical association. Accordingly, we have revised the title of Table 3 and ensured that the terminology regarding habitat preference, use, and occurrence associations is applied consistently throughout the manuscript. The revised caption now reads:       "Table 3. Species-habitat associations based on logistic regression results. The important variables associated with the statistical occurrence of wildlife species are based on the results of multiple logistic regression analysis and are retained (ordered from most to least important) in the models based on the minimum AIC value." ·        Line 299: The terminology used to describe Figure 6 and the corresponding text is unclear and potentially misleading. The figure appears to present frequency-based measures of presence/absence, whereas the text refers to “presence probability,” which typically implies a model-derived probability. The authors should clarify whether these values represent empirical frequencies or predicted probabilities, and ensure consistent terminology throughout the manuscript. To improve clarity and consistency, we have replaced “presence probability” with “occurrence frequency” (See line 349)   ·        Line 304: The terms “occurrence frequency” and “presence probability” appear to be used interchangeably, although they represent different quantities. The authors should ensure consistent terminology throughout the manuscript.       We thank the reviewer for pointing out this constructive comment. We agree that these terms represent distinct quantities, and we have carefully revised the manuscript to ensure consistent terminology throughout. Specifically, we have standardized our phrasing to use "occurrence frequency" in this section and throughout the text to maintain clarity and accuracy. (See line 354)   ·        Figure 6: The use of occurrence frequency across high-intact and medium-intact forest habitats is used to infer habitat “preference.” However, occurrence frequency alone reflects patterns of use within the sampled design and does not account for habitat availability or detectability differences. As such, it does not provide a robust basis for inferring habitat preference in the formal ecological sense, but rather describes relative occurrence across habitat categories. The manuscript should therefore rephrase these interpretations as habitat associations.        We thank the reviewer for this deep and insightful comment. We have now rephrased the title and as well as modified the whole text with the focus on “Habitat Association”   ·        Line 318: Similar to the above, this finding does not necessarily imply “preference”. We have changed it to “association”.   ·        Line 339: Perhaps another wording for “…exhibited greater adaptability” would be more appropriate. For example: “….exhibited less specialization”. Following your suggestion, we have revised the text to “exhibited broader habitat associations,” which may be more appropriate as suggested.   ·        Lines 354-368: The authors acknowledge substantial evidence of imperfect and heterogeneous detectability for multiple species, including non-detection despite confirmed presence. This suggests that detection probability is < 1 and likely varies across species and habitats. However, this limitation is not accounted for in the analytical approach, and results should therefore be interpreted as conditional on detection rather than true occupancy.        We thank the reviewer for this important comment. we have revised the text to clarify that the results are based on detection/non-detection data and to emphasize imperfect detection rather than true absence (See line: 414-418).   ·        Line 370: Use the words “detections” or “sightings” instead of “occurrences” Correction for this suggestion has been made in the text.   ·        Lines 383-386: The authors report the presence of anthropogenic pressures such as hunting, habitat fragmentation, agricultural clearance, and illegal logging. They subsequently conclude that these factors “inevitably affect” wildlife distribution and populations. However, these relationships (at least not all of them) are not quantitatively tested in the manuscript. As such, this statement appears speculative and should be reframed to reflect observational evidence rather than inferred effects. We thank for this constructive comment, we have reframed the statement by replacing ““inevitably affect” with “may affect” and changed “key wildlife biodiversity” to “key wildlife species in the area” (see line: 450-451).   ·        Lines 400-401: Could you please provide a relevant reference. In addition, “absence” does not necessarily mean that a species is not present in the area. Absence could arise from imperfect detection, i.e..: 9 species were detected in this study, while 54 were detected in the same area using a different methodological approach. Do you mean that these species have gone entirely extinct from the area? Following your suggestion, we have revised the text and added a reference for this statement (See line: 468).   ·        Line 408: Do you mean strict protection against deforestation, illegal logging, hunting, etc? Yes, we meant that. To improve clarity, we revised the sentence to specify that “strict protection” refers to protection against deforestation, illegal logging, hunting, and other human disturbances (Lines 474–475).   ·        Line 419: Author contributions should follow the journal’s format guidelines This section has been modified with this suggestion (see line: 487-492).

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

The authors did a great job revising their original manuscript. All comments and suggestions have been adequately addressed. I believe the paper can now be considered for publication, given some minor changes listed below.

Line 130: Remove the word “selected” after “generated”.

Line 130-132: Please check the phrasing and revise.

Line 166: Consider modifying to “although it may produce false-negatives”.

Figure 2: Figure indications (a) and (b)would be better fitted on the outer part of each figure.

Lines 194-199: Could you please provide a citation to support the argument?

Line 311: Please correct “3.2. Habitat Associations and Anthropogenic Pressure”.

Lines 312-330: Several confidence intervals seem to overlap zero, indicating that they are not very informative. On the other hand, those with no overlapping zero values are highly informative and should be highlighted in the text. The authors could also check the 85% confidence intervals; variables with no negative values of the 85% confidence intervals are usually considered moderately informative. This is described in: Arnold, T.W. Uninformative Parameters and Model Selection Using Akaike’s Information Criterion. J. Wildl. Manag. 2010, 74, 1175–1178.

Line 434: The opening of the sentence may benefit from some re-phrasing “Despite the results highlight”.

Author Response

Response to Reviewer 3 Comments

1. Summary

We sincerely thank the reviewer for the positive evaluation of our revised manuscript and for recognizing that the comments and suggestions from the previous review have been adequately addressed. We greatly appreciate the reviewer's continued careful assessment and the additional constructive suggestions provided in this round. These valuable comments have further improved the clarity, accuracy, and presentation of the manuscript. We have carefully considered each recommendation and revised the manuscript accordingly. A detailed point-by-point response to each comment is provided below, and all corresponding revisions have been highlighted using track changes in the revised manuscript.

2. Overall comments from the reviewer

The authors did a great job revising their original manuscript. All comments and suggestions have been adequately addressed. I believe the paper can now be considered for publication, given some minor changes listed below.

 

3. Point-by-point response to Comments and Suggestions for Authors

• Line 130: Remove the word “selected” after “generated”.

The word the word “selected” after “generated” has been removed as suggested.

 

• Line 130-132:Please check the phrasing and revise.

We thank the reviewer for this valuable comment. We have revised the text to improve clarity, flow, and grammatical correctness while retaining the methodological justification (See line: 130- 133).

 

• Line 166: Consider modifying to “although it may produce false-negatives”.

This has been modified in the text as suggested.

 

• Figure 2:Figure indications (a) and (b)would be better fitted on the outer part of each figure.

The figure indications (a) and (b) of figure 2 have been modified as suggested.

 

• Lines 194-199: Could you please provide a citation to support the argument?

We thank the reviewer for this valuable suggestion. We added citations to support the argument (See line: 195-200).

 

• Line 311: Please correct “2. Habitat Associations andAnthropogenic Pressure”.

This has been correct as suggested.

 

• Lines 312-330:Several confidence intervals seem to overlap zero, indicating that they are not very informative. On the other hand, those with no overlapping zero values are highly informative and should be highlighted in the text. The authors could also check the 85% confidence intervals; variables with no negative values of the 85% confidence intervals are usually considered moderately informative. This is described in: Arnold, T.W. Uninformative Parameters and Model Selection Using Akaike’s Information Criterion. J. Wildl. Manag. 2010, 74, 1175–1178.

We thank the reviewer for this valuable and constructive comment. In response to this suggestion, we added a sentence to Section 2.6.2 of the Methodology (see Lines 256–259). We also revised the Results and Discussion sections to make them more specific and appropriate (see Lines 318–337 and 400–401, respectively).

 

• Line 434:The opening of the sentence may benefit from some re-phrasing “Despite the results highlight”.

We thank the reviewer for this valuable comment. We have revised the text by replacing the word “Despite” with “Although” which is grammatically correct and understandable.