Prediction of the Potentially Suitable Area for Anoplophora glabripennis (Coleoptera: Cerambycidae) in China Based on MaxEnt

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This paper proposes to look at the suitable areas for Anoplophora glabripennis distribution now and in the future in China.  There are issues with the methods.  The introduction and discussion need to be better organized  and do not fully place this work within the present literature on this topic (e. g. some Max Ent work already done not referenced).  Some of the conclusions are not fully explained and others are not supported by the beetle’s ecology/biology. The application of the work is not clearly defined.  A major revision of the paper is needed.

Introduction

Line 36  Asia covers Japan, Korea and China so remove that word.  Also it is native to Korea but may be invasive to Japan.

Line 38 be more specific bores what? Why is this wood boring pest a problem?  Way should it be managed?  Give the reader more information.

Line 40  How is this insect a serious pest? Give more information on impacts.

Line 42 How does it threaten the ecology of the forest?

Line 45 these references are not particularly good for this statement.  There has been Max Ent work done on Anoplophora glabripennis before that the authors are not acknowledging.  Also, how can you say future predictions are not accurate?

Line 52 Max Ent does not predict the distribution, but the potential distribution of a species based on its presence and absence in particular locations.

Line 58 “so on” is not a correct way to end a list.  You could say among others if there are more options.

Lines 60-72 What does the potential distribution of all these plants have to do with this insect? Why not reference the other studies that have looked at Anoplophora  glabripennis using Max Ent? Or related species of beetles or its host plants?  This is a poor choice of examples.

Materials and Methods

Lines 83-85  Poor wording.  Should be changed: All occurrence data for Anoplophora glabripennis was obtained form GBIF and data points for China were selected for use.  What were the other data sources for data that were used?  I know CABI has some occurrence data for this insect and there are scientific articles (which ones were used?) with data.  You should have to specifically provide information on all the data sources used.

Lines 87-89 what was considered abnormal data? How was the data filtered to generate geographically unique occurrences and account for potential sampling bias? Generally, a buffer of 5 km around each record is used to identify unique data points.

Figure 1  The map is too small to clearly see the data points or the scale.  For the size of the country I would expect many more data points.

Lines 93-101  The resolution is 2.5 arc minutes not meters. 

Section 2.3  give more information about why this map was used?

Section 2.4  knife-cut method?  Do you mean jack-knife method?  You could have used R package “MaxentVariableSelection” (Jueterbock et al., 2016) to select a relevant set of variables because it  looks at contribution and accounted for autocorrelation.  Usually you get down to 10 or less variables that are the best. 

Section 2.5  This is a very brief description of what was done.  You have not provided the parameters used in the model or how sampling bias or correlation was accounted for.  Why only 10 iterations, that seems small.

Results

Section 3.1  Again the figure is too small to evaluate accurately.  The number of variables is too high.  You should be comparing the models progressively by adding in variables and find the model that has the the lowest Akaike’s Information Criterion for small sample size scores (AICc) (Warren & Seifert, 2011) to use.

You mention only a few variables in the paragraph.  Are these the most important ones?  Tell the reader why precipitation and temperature may be critical for survival of this insect.  Elevation is likely correlated with both temperature and precipitation so not a good variable. There is plenty of data on the affect of temperature on the insect in the literature to refer to.  Using biologically informed, uncorrelated variables has been shown to achieve better dispersal predictions, particularly across time.

Section 3.2  More reliable than what?  Just because you predict more of the data points by having more variable does not make the model better.

Discussion

Paragraph one lines 257-273   This is a start to describing why the parameters that did the best at predicting distribution might have biological significance, but it needs to be flushed out and more specifics given. For example, why is temperature in September more important that in the other months that this insect is active? Why is precipitation important to this insect in the Fall?  How does it affect survival?  Why is an alternation of precipitation important?

I don’t think elevation and flight are connected directly but there could be some effect of temerpture on flight. Distribution in higher elevation would be limited by temperature and precipitation would vary by which side of montians receive the precipitation and which ones are too dry.  Elevation is not a good variable anyway since it is correlated with both precipitation and temperature.

Lines 276-284  some of this should be incorporated earlier as reasons for the variables being the best.   Your discussion is poorly organized.  You should start with a general paragraph that highlights the major findings.  Then follow it with paragraphs that delve into more detail on why those findings, how they compare with already published work and why they are important to managing this insect.

Section 4.2  Warmer temperature (≥ 35°C) effects on this insect have been documented (poorer survival of all stages, fewer eggs, shorter adult life span, etc).  Why talk about male beetle fertility based on speculation when there are documented effects.

The increase in suitable area may be the opening up of more northern areas to use by the insect while the warmer areas have not yet reached high enough temperatures to cause population crash.  Then as the warm areas become unsuitable the range would decrease.

You could also explore adding human habitation layers and host layers to the variables since this beetle can move longer distances if aided by humans and establish if suitable hosts exist in the new areas.  The hosts are not able to migrate as fast as the insects so that too can limit the speed with which they move north and the timing of when warmer areas may become unsuitable (hosts can’t survive there any longer).   Many more factors involved that you have presented.

You do not compare these finding with previous distribution modeling for this insect.  Were the same variables found to be the best? Did the same pattern of present and future distribution appear?  You need to place this work within the body of literature already published much better.

You also need to talk about how these finding will help guide management.

Comments on the Quality of English Language

There are a few places that the wrong term or word is chosen.  Most of the English is acceptable.

Author Response

Comments 1:This paper proposes to look at the suitable areas for Anoplophora glabripennis distribution now and in the future in China.  There are issues with the methods.  The introduction and discussion need to be better organized  and do not fully place this work within the present literature on this topic (e. g. some Max Ent work already done not referenced).  Some of the conclusions are not fully explained and others are not supported by the beetle’s ecology/biology. The application of the work is not clearly defined.  A major revision of the paper is needed.

Response 1:Thank you for the expert's advice. Here are our explanations and revision notes. We appreciate your review.

Introduction

Comments 2:Line 36 Asia covers Japan, Korea and China so remove that word.  Also it is native to Korea but may be invasive to Japan.

Response 2:Thank you for your suggestions. We have made the revisions.

Comments 3:Line 38 be more specific bores what? Why is this wood boring pest a problem?  Way should it be managed?  Give the reader more information.

Response 3:Thank you for your suggestions. We have already made the revisions.

Comments 4:Line 40  How is this insect a serious pest? Give more information on impacts.

Response 4:Thank you for your suggestions. We have already made the revisions.

Comments 5:Line 42 How does it threaten the ecology of the forest?

Response 5:Thank you for your suggestions. We have added relevant examples in the following sections.

Comments 6:Line 45 these references are not particularly good for this statement.  There has been Max Ent work done on Anoplophora glabripennis before that the authors are not acknowledging.  Also, how can you say future predictions are not accurate?

Response 6:Thank you for your suggestions. We have also recognized our mistakes and have made corrections to the language expression and replaced the references. Thank you again for the reviewer's suggestions.

Comments 7:Line 52 Max Ent does not predict the distribution, but the potential distribution of a species based on its presence and absence in particular locations.

Response 7:Thank you very much for your professional suggestions. We have made the changes accordingly.

Comments 8:Line 58 “so on” is not a correct way to end a list.  You could say among others if there are more options.

Response 8:Thank you for your suggestions. We have made the revisions.

Comments 9:Lines 60-72 What does the potential distribution of all these plants have to do with this insect? Why not reference the other studies that have looked at Anoplophora  glabripennis using Max Ent? Or related species of beetles or its host plants?  This is a poor choice of examples.

Response 9:Thank you for your suggestions. We have made the revisions.

Materials and Methods

Comments 10:Lines 83-85  Poor wording.  Should be changed: All occurrence data for Anoplophora glabripennis was obtained form GBIF and data points for China were selected for use.  What were the other data sources for data that were used?  I know CABI has some occurrence data for this insect and there are scientific articles (which ones were used?) with data.  You should have to specifically provide information on all the data sources used.

Response 10:Thank you for your suggestions on the data sources. We have made the corrections and provided the data sources for this study.

Comments 11:Lines 87-89 what was considered abnormal data? How was the data filtered to generate geographically unique occurrences and account for potential sampling bias? Generally, a buffer of 5 km around each record is used to identify unique data points.

Response 11:Thank you for your suggestions. During the processing, points falling outside the Chinese regions of this study were considered as outliers. We used GIS clipping tools (such as ArcGIS) to remove points outside the habitat areas (e.g., water bodies, extreme climate zones). We also filtered the data based on the predicted suitability probabilities from the MaxEnt model, removing points with probabilities less than 0.05 to generate geographically unique occurrence points and account for potential sampling biases.

Comments 12:Figure 1  The map is too small to clearly see the data points or the scale.  For the size of the country I would expect many more data points.

Response 12:Thank you for your suggestions. We have revised the figures. For the size of this country, we used GIS clipping tools (such as ArcGIS) to remove points outside the habitat areas (e.g., water bodies, extreme climate zones) and filtered the data based on the predicted suitability probabilities from the MaxEnt model, removing points with probabilities less than 0.05 to generate geographically unique occurrence points and account for potential sampling biases. Therefore, the data points are shown in Figure 1.

Comments 13:Lines 93-101 The resolution is 2.5 arc minutes not meters. 

Response 13:Thank you for your suggestions. We have made the revisions.

Comments 14:Section 2.3 give more information about why this map was used?

Response 14:Thank you for your suggestions. We have added the information as suggested.

Comments 15:Section 2.4 knife-cut method?  Do you mean jack-knife method?  You could have used R package “Maxent Variable Selection” (Jueter bock et al., 2016) to select a relevant set of variables because it looks at contribution and accounted for autocorrelation.  Usually you get down to 10 or less variables that are the best. 

Response 15:Thank you for your suggestions. We have taken autocorrelation into account. Given the large volume of data on autocorrelation, we have placed the results of the autocorrelation analysis in a supplementary file named “Pearson Correlation Analysis of the 56 Environmental Variables.”

Comments 16:Section 2.5  This is a very brief description of what was done.  You have not provided the parameters used in the model or how sampling bias or correlation was accounted for.  Why only 10 iterations, that seems small.

Response 16:Thank you for your suggestions. We have provided the parameters used in the model. We also conducted a correlation study, which is detailed in Section 2.4. The reason for performing 10 iterations is that we referred to some literature, and after 10 iterations, the difference in AUC met the model's requirements. And we have conducted 50 replicates. The results show that the AUC value of the 50 replicates is 0.848, which is not as good as that of the 10 replicates.

Results

Comments 17:Section 3.1  Again the figure is too small to evaluate accurately.  The number of variables is too high.  You should be comparing the models progressively by adding in variables and find the model that has the the lowest Akaike’s Information Criterion for small sample size scores (AICc) (Warren & Seifert, 2011) to use.

You mention only a few variables in the paragraph.  Are these the most important ones?  Tell the reader why precipitation and temperature may be critical for survival of this insect.  Elevation is likely correlated with both temperature and precipitation so not a good variable. There is plenty of data on the affect of temperature on the insect in the literature to refer to.  Using biologically informed, uncorrelated variables has been shown to achieve better dispersal predictions, particularly across time.

Response 17:Thank you very much for your valuable suggestions. We have modified the graphs to facilitate the readers' comprehension. We are grateful for your meticulous review of this paper and the important suggestions regarding model optimization. Since we are not clear about the application of the Akaike Information Criterion for small sample sizes (AICc), there are technical difficulties in directly applying AICc. The variables mentioned in the paragraph are based on the contribution rate results of the jackknife method. These variables have higher contribution rates and are thus the most important ones. The statement "the minimum temperature and precipitation during the pupal and oviposition periods are the dominant factors affecting the suitable distribution of Anoplophora glabripennis" is a result. We have explained in the discussion why temperature and precipitation may be crucial for the survival of this insect. We agree with your suggestion that altitude may be correlated with temperature and precipitation. However, our autocorrelation results (see the attached Pearson correlation analysis) show that altitude has low correlation with temperature and precipitation. Therefore, we did not exclude this variable. We sincerely appreciate your suggestions for improving this paper.

Comments 18:Section 3.2  More reliable than what?  Just because you predict more of the data points by having more variable does not make the model better.

Response 18:Thank you for your meticulous review and valuable comments on the model evaluation section of this paper. We fully understand your concern that an increase in the number of variables does not necessarily mean an improvement in model performance. To clarify, the core basis for model optimization in this study is not the number of variables, but a comprehensive assessment based on objective statistical indicators (such as the AUC value). Moreover, Section 3.2 is intended to explain the rationality and reliability of the maximum entropy model in predicting the future suitable distribution range of Anoplophora glabripennis, and that is what we mean by reliability. Thank you once again for your suggestions.

Discussion

Comments 19:Paragraph one lines 257-273   This is a start to describing why the parameters that did the best at predicting distribution might have biological significance, but it needs to be flushed out and more specifics given. For example, why is temperature in September more important that in the other months that this insect is active? Why is precipitation important to this insect in the Fall?  How does it affect survival?  Why is an alternation of precipitation important? I don’t think elevation and flight are connected directly but there could be some effect of temerpture on flight. Distribution in higher elevation would be limited by temperature and precipitation would vary by which side of montians receive the precipitation and which ones are too dry.  Elevation is not a good variable anyway since it is correlated with both precipitation and temperature.

Lines 276-284  some of this should be incorporated earlier as reasons for the variables being the best.   Your discussion is poorly organized.  You should start with a general paragraph that highlights the major findings.  Then follow it with paragraphs that delve into more detail on why those findings, how they compare with already published work and why they are important to managing this insect.

Response 19:Thank you very much for your valuable suggestions. We fully agree with your advice on the structure of our discussion. Therefore, we have revised it according to your suggestions and have also included the content of the issues you raised. Thank you once again for your help in improving the quality of this paper. Regarding the issue of altitude being correlated with both precipitation and temperature, we have already provided an explanation in the section where you previously mentioned the altitude issue. Thank you again for your professional suggestions.

Comments 20:Section 4.2  Warmer temperature (≥ 35°C) effects on this insect have been documented (poorer survival of all stages, fewer eggs, shorter adult life span, etc).  Why talk about male beetle fertility based on speculation when there are documented effects.

The increase in suitable area may be the opening up of more northern areas to use by the insect while the warmer areas have not yet reached high enough temperatures to cause population crash.  Then as the warm areas become unsuitable the range would decrease.

You could also explore adding human habitation layers and host layers to the variables since this beetle can move longer distances if aided by humans and establish if suitable hosts exist in the new areas.  The hosts are not able to migrate as fast as the insects so that too can limit the speed with which they move north and the timing of when warmer areas may become unsuitable (hosts can’t survive there any longer).   Many more factors involved that you have presented.

You do not compare these finding with previous distribution modeling for this insect.  Were the same variables found to be the best? Did the same pattern of present and future distribution appear?  You need to place this work within the body of literature already published much better.

You also need to talk about how these finding will help guide management.

Response 20:Thank you very much for your insightful and constructive comments on our paper. Your suggestions about the impact of human activities and the availability of host plants on the distribution of Anoplophora glabripennis are extremely valuable. Although the conditions are not ripe at present, we will further provide relevant information for the academic community in the future.

We fully agree with your point about the potential impact of the temperature threshold (≥35℃) on population collapse and how these findings can guide management. These points have been added to the revised manuscript. We have also added content regarding the impact of past distribution models on the distribution of Anoplophora glabripennis.

Once again, we appreciate your help in situating our research within a more rigorous academic framework. Your comments have significantly enhanced the scientific value and practical significance of our paper. Should there be any further adjustments needed, we are more than willing to make the necessary improvements.

 

 

Reviewer 2 Report

Comments and Suggestions for Authors

The order and family should be included in the title.
The keywords could be improved.
In the Introduction, the wording regarding the distribution of Cerambycidae and Anoplophora glabripennis should be corrected as it is confusing.
Explain in more detail about its introduction in North America, from where? and how or why?
Suddenly, they start writing about many species of trees (litchi, Alsophila, hawthorn, Tilia, bamboo)... It seems to come out of nowhere. It's confusing; the authors have to organize their thoughts.
Here they only give examples of plant distribution, it would be much better to give examples of predictions of potential insect distribution.
The map (Figure 1) is of poor quality, looks tiny and cannot be read at all.
The figures really need to be enlarged.
The novelty is that it predicts a distribution in northwest and southwest China, because the species is already present in other places today.

Comments for author File: Comments.pdf

Comments on the Quality of English Language

English needs to be improved a lot.

Author Response

Comments 1:The order and family should be included in the title.

Response 1:Thank you very much for your suggestions. We have modified the title according to your suggestions.
Comments 2:The keywords could be improved.

Response 2:Thank you for your suggestions. We have adopted your suggestions and improved the keywords.
Comments 3:In the Introduction, the wording regarding the distribution of Cerambycidae and Anoplophora glabripennis should be corrected as it is confusing.

Response 3:Thank you very much for your suggestions. We have corrected the wording about the distribution of the species in the introduction.
Comments 4:Explain in more detail about its introduction in North America, from where? and how or why?

Response 4:Thank you very much for your suggestions. We have already added the relevant information to the text.
Comments 5:Suddenly, they start writing about many species of trees (litchi, Alsophila, hawthorn, Tilia, bamboo)... It seems to come out of nowhere. It's confusing; the authors have to organize their thoughts.

Response 5:Thank you very much for your suggestions. We have revised the introduction section of the text in response to this suggestion.
Comments 6:Here they only give examples of plant distribution, it would be much better to give examples of predictions of potential insect distribution.

Response 6:Thank you very much for your suggestions. We have revised the introduction section of the text in response to this suggestion.
Comments 7:The map (Figure 1) is of poor quality, looks tiny and cannot be read at all.

Response 7:Thank you very much for your suggestions. We have adjusted the size and fill color of the distribution points in Figure 1.
Comments 8:The figures really need to be enlarged.

Response 8:Thank you very much for your suggestions. We have adjusted the figures in this paper.
Comments 9:The novelty is that it predicts a distribution in northwest and southwest China, because the species is already present in other places today.

Response 9:Thank you very much for your suggestions. We have made revisions to the results section in light of the suggestions you have put forward.

Reviewer 3 Report

Comments and Suggestions for Authors

The authors predict the distribution of parasitic tree species under present and future climate conditions using a local circulation model. While the approach is appropriate, the analysis of the results could be more comprehensive. The methodology requires further explanation, and the interpretation of the findings should be strengthened.

Keywords should include terms that are not already used in the title to enhance discoverability

Introduction

Line 62. The species’ name should be reported in italics.

The authors should justify why they expect their predictions to be more accurate or insightful than those of previous studies. Additionally, the introduction lacks a clear statement about the future distribution predictions. Specifically, it would be helpful to explain what outcomes they anticipate based on the characteristics of the BCC-CSM2-MR model under the SSP245 scenario for the periods 2021–2040 and 2041–2060.

Clearly state what you expect to find based on the information presented. What is your hypothesis?

Methods

Improve the quality of Figure 1

Line 88. Specify the minimum spatial distance between records.

Line 98: Please specify the characteristics of the Beijing Climate Center Climate System Model (BCC-CSM2-MR) under the SSP245 shared socioeconomic pathway scenario

The WorldClim layers are available at different spatial resolutions; the authors should specify which resolution was used and describe the procedure applied to match or resample them to the resolution of the BCC-CSM2-MR model (2.5 m).

Line 130. Since the authors modeled the distribution of a single species, using only 10 replicates is insufficient. It is recommended to use at least 50 replicates to ensure greater robustness and reliability of predictions.

Since the main objective of the study is to predict the distribution of a tree-pest species, I recommend estimating the forest areas currently at risk of being affected by its presence, as well as projecting the areas that could be impacted under future scenarios.

Results

Figure Size: Increase the size of the figures for better clarity.

Unspecified Results: The section reports results that were not described in the methods, such as model contributions, curves of major environmental variables, AUC, and the classification of suitable distribution areas.

Line 142: The species name should be italicized.

Lines 148-168: Report the presence-absence threshold used to identify the intervals where the distribution is most suitable.

Lines 174-177: This information should be included in the Methods section.

Include a description of the criteria used to define the four grades via the natural discontinuous point classification method.

Figure 5: In the legend, include the percentage occupied by each class.

Table 7 presents the percentage of the distribution across the four categories, but this does not allow for an assessment of the species' spatial persistence and changes among the three models. I suggest that the authors perform a Markov chain analysis to estimate the probability of persistence and change between the four classes across consecutive periods (present-2021–2040 and 2021–2040–2041–2060). This could be achieved by vectorizing the predictions, unioning the layers, calculating the polygon areas in ArcMap, and analyzing the data in Excel (using a dynamic table).

Table 3. Specify at the bottom of the table what SSP means.

Figures 6 and 7: The ability to identify changes in species distribution between the present-2021–2040 and 2021–2040–2041–2060 periods are limited. I suggest that the authors vectorize their layers (using QGIS), union them (via Geoprocessing in ArcMap), and reclassify the changes and persistences.

Discussion

The results presented limit the ability to assess the extent of permanence and expansion of species' distribution. The interpretation of the results is insufficient. The authors should consider discussing factors that may undermine their findings, such as the number of records or the selection of variables. The models generated solely with climatic variables assume that the species is in equilibrium with the environment, and this assumption may not hold in the case of the species studied.

The authors could interpret their results clearly within the context of the study’s objectives and hypothesis. They should also explain the implications of their findings for forest conservation and management. Furthermore, the study’s limitations should be acknowledged without diminishing its overall value. Finally, future research directions could be suggested based on the current findings and any identified knowledge gaps.

Author Response

Comments 1:The authors predict the distribution of parasitic tree species under present and future climate conditions using a local circulation model. While the approach is appropriate, the analysis of the results could be more comprehensive. The methodology requires further explanation, and the interpretation of the findings should be strengthened.

Keywords should include terms that are not already used in the title to enhance discoverability

Response 1:Thank you very much for your suggestions. The keywords section of this paper has been revised in accordance with your opinions.

Introduction

Comments 2:Line 62. The species’ name should be reported in italics.

Response 2:Thank you very much for your suggestions. We have already made the species’ name italicized in the introduction.

Comments 3:The authors should justify why they expect their predictions to be more accurate or insightful than those of previous studies. Additionally, the introduction lacks a clear statement about the future distribution predictions. Specifically, it would be helpful to explain what outcomes they anticipate based on the characteristics of the BCC-CSM2-MR model under the SSP245 scenario for the periods 2021–2040 and 2041–2060.

Clearly state what you expect to find based on the information presented. What is your hypothesis?

Response 3:Thank you for your professional advice. Based on your suggestions, we have rewritten the introduction, incorporating a clear statement about our prediction of future distribution, which may be in the northwest and southwest regions. We have also included an example of using the CLIMEX model to predict the potential global distribution of the starry sky buprestid beetle (Agrilus ater) based on historical (1987–2016) and future (2021–2050) climate conditions. Our research predicts the potential distribution range of the starry sky buprestid beetle under the SSP2-4.5 climate scenario for 1970–2000, 2021–2040, and 2041–2060. It is helpful to expand the time range of research on this species to explain our expected results.

Methods

Comments 4:Improve the quality of Figure 1

Response 4:Thank you for your suggestions. We have already improved Figure 1.

Comments 5:Line 88. Specify the minimum spatial distance between records.

Response 5:Thank you for your suggestions. We have already specified the minimum spatial distance between records as 2.5 arc minutes.

Comments 6:Line 98: Please specify the characteristics of the Beijing Climate Center Climate System Model (BCC-CSM2-MR) under the SSP245 shared socioeconomic pathway scenario

Response 6:Reply：Thank you for your suggestions. We have already added the explanation.

Comments 7:The WorldClim layers are available at different spatial resolutions; the authors should specify which resolution was used and describe the procedure applied to match or resample them to the resolution of the BCC-CSM2-MR model (2.5 m).

Response 7:Thank you for your suggestions. Our spatial resolution has already been specified as 2.5 arc minutes. After our reprocessing of the data, we found that when matching or resampling it to the resolution of the BCC-CSM2-MR model (2.5 arc minutes), the ArcGIS software directly carried out the matching process, so there are no detailed records. Thank you again for your guidance and help.

Comments 8:Line 130. Since the authors modeled the distribution of a single species, using only 10 replicates is insufficient. It is recommended to use at least 50 replicates to ensure greater robustness and reliability of predictions.

Response 8:Thank you for your suggestions. Following your advice, we have conducted 50 replicates. The results show that the AUC value of the 50 replicates is 0.848, which is not as good as that of the 10 replicates. Moreover, we based our decision to conduct 10 replicates on the existing literature, specifically the paper titled "Prediction of suitable area of Solidago canadensis based on Maxent and ArcGIS (in Chinese)." Once again, thank you for your professional advice.

Comments 9:Since the main objective of the study is to predict the distribution of a tree-pest species, I recommend estimating the forest areas currently at risk of being affected by its presence, as well as projecting the areas that could be impacted under future scenarios.

Response 9:Thank you for your important suggestions. Quantifying the forest area currently and potentially affected by the Anoplophora glabripennis is crucial for assessing its ecological and economic risks. In response to your comments, we have added further explanations to our study and planned follow-up research. The specific reply is as follows: Due to the limited availability of fine-scale distribution maps of host tree species and continuous outbreak verification data at the national level, there is considerable uncertainty in directly estimating the affected forest area. Future work will collaborate with forestry departments to obtain specialized monitoring data to improve this analysis. Thank you once again for your professional suggestions to enhance this paper.

Results

Comments 10:Figure Size: Increase the size of the figures for better clarity.

Response 10:Thank you for your suggestions. We have made the corrections.

Comments 11:Unspecified Results: The section reports results that were not described in the methods, such as model contributions, curves of major environmental variables, AUC, and the classification of suitable distribution areas.

Response 11:Yes, this section mainly describes our research results. As you said, it includes research results related to model contributions, response curves of the main environmental variables, AUC, and classification of suitable distribution areas. Thank you for your suggestions.

Comments 12:Line 142: The species name should be italicized.

Response 12:Thank you for your suggestions. We have made the corrections.

Comments 13:Lines 148-168: Report the presence-absence threshold used to identify the intervals where the distribution is most suitable.

Response 13:Thank you for your suggestions. Yes. This section reports the presence–absence threshold used to identify the most suitable distribution range. We have already added the explanation according to your suggestions.

Comments 14:Lines 174-177: This information should be included in the Methods section.Include a description of the criteria used to define the four grades via the natural discontinuous point classification method.

Response 14:Thank you for your suggestions. This part of the information is an assessment of the accuracy of the Maxent model's predictions for this study and is part of the results. The results show that the Maxent model can accurately predict the potential future distribution areas of the starry sky buprestid beetle in this study. The description of the criteria for defining four levels using the natural discontinuity classification method was written by us to better describe our results, so that readers can better understand this part of the results. Therefore, it is appropriate to place this section in the results. Thank you again for your suggestions.

Comments 15:Figure 5: In the legend, include the percentage occupied by each class.
Response 15:Thank you for your suggestions. We have already made the modifications in the figure 5.

Comments 16:Table 7 presents the percentage of the distribution across the four categories, but this does not allow for an assessment of the species' spatial persistence and changes among the three models. I suggest that the authors perform a Markov chain analysis to estimate the probability of persistence and change between the four classes across consecutive periods (present-2021–2040 and 2021–2040–2041–2060). This could be achieved by vectorizing the predictions, unioning the layers, calculating the polygon areas in ArcMap, and analyzing the data in Excel (using a dynamic table).

Response 16:Thank you for your suggestions. We have employed Markov chain analysis to estimate the probabilities of persistence and change among the four categories across consecutive periods (current-2021-2040 and 2021-2040-2041-2060). Given that the data spans two periods, the probabilities of persistence and change are not particularly high. The specific results are provided in Attachment 2.

Comments 17:Table 3. Specify at the bottom of the table what SSP means.

Response 17:Thank you for your suggestions. We have already made the addition.

Comments 18:Figures 6 and 7: The ability to identify changes in species distribution between the present-2021–2040 and 2021–2040–2041–2060 periods are limited. I suggest that the authors vectorize their layers (using QGIS), union them (via Geoprocessing in ArcMap), and reclassify the changes and persistences.

Response 18:Thank you for your suggestions. Following your advice, we have conducted vectorization and Markov chain analysis through geoprocessing in ArcMap. Since the study involves two consecutive periods with a relatively small amount of data, the results obtained do not significantly affect our existing distribution situation. The detailed results are provided in Attachment 2.

Discussion

Comments 19:The results presented limit the ability to assess the extent of permanence and expansion of species' distribution. The interpretation of the results is insufficient. The authors should consider discussing factors that may undermine their findings, such as the number of records or the selection of variables. The models generated solely with climatic variables assume that the species is in equilibrium with the environment, and this assumption may not hold in the case of the species studied.

The authors could interpret their results clearly within the context of the study’s objectives and hypothesis. They should also explain the implications of their findings for forest conservation and management. Furthermore, the study’s limitations should be acknowledged without diminishing its overall value. Finally, future research directions could be suggested based on the current findings and any identified knowledge gaps.

Response 19:Thank you for your suggestions. Following your advice, we have added an example of the assessment of species distribution persistence and expansion range. At the same time, we have acknowledged the limitations of our study and added explanations regarding future research directions. We have also explained the significance of our findings for forest conservation and management. Thank you once again for your efforts in improving the quality of this paper.

Reviewer 4 Report

Comments and Suggestions for Authors

1. Authors are kindly requested to check all references to the species name Anoplophora glabripennis for typos.
2. In the introduction, to justify the choice of Maxent, it is worth citing examples of studies related to Maxent prediction of the spatial distribution of insects within China (such works exist), rather than only the spatial distribution of plant organisms (references: 13-18, 23, 24). This would lend more weight to the argument in favour of using the Maxent model for this species.
3. The list of references includes 30 sources, of which 13-18, 23, 24 are devoted to studies of the distribution of plant organisms, source 25 does not relate to the topic of the study at all.
4. The authors write ‘Moreover, the distribution of this species in China was supple-86 mented by a literature review’, but unfortunately there is no reference to literature, so it is not clear from which sources the distribution data are taken. GBIF contains 166 points of occurrence of the species in China. Unfortunately, there is no data on field collection by the authors themselves, which means that the analysis is based solely on data obtained from GBIF and literature data.
5. Please explain the discrepancy between the number of reported variables (56) and what is presented in Table 1. ‘In this study, a total of 56 environmental variables (Table 1) were selected from the World Meteorological Data website (WorldClim, https://www.worldclim.org)»
6. Explain the choice of script SSP245 and the reasons why, for example, script SSP285 was not used. Provide reasoning as to why this particular scenario is appropriate for your study.
7. To obtain an optimal model, a series of models with different settings must be analysed. Good modelling results in Maxent can be obtained by using multi-stage analysis, which involves successive model calibration based on the development of multiple models with different settings.
8. To obtain correct models, careful selection of all modelling components is necessary: species distribution points, environmental variables, detail of analysis settings.

 

Author Response

Comments 1:Authors are kindly requested to check all references to the species name Anoplophora glabripennis for typos.

Response 1:Thank you for your suggestion. We have rechecked the references to ensure that the spelling of the species name Anoplophora glabripennis is correct in all of them.

Comments 2:In the introduction, to justify the choice of Maxent, it is worth citing examples of studies related to Maxent prediction of the spatial distribution of insects within China (such works exist), rather than only the spatial distribution of plant organisms (references: 13-18, 23, 24). This would lend more weight to the argument in favour of using the Maxent model for this species.

Response 2:Thank you for your suggestion. The corrections have already been made in the corresponding places (highlighted in red).

Comments 3:The list of references includes 30 sources, of which 13-18, 23, 24 are devoted to studies of the distribution of plant organisms, source 25 does not relate to the topic of the study at all.

Response 3:Thank you for your suggestion. I have already revised it to references related to the content of this study. Thank you again for your excellent suggestions.

Comments 4:The authors write ‘Moreover, the distribution of this species in China was supple-86 mented by a literature review’, but unfortunately there is no reference to literature, so it is not clear from which sources the distribution data are taken. GBIF contains 166 points of occurrence of the species in China. Unfortunately, there is no data on field collection by the authors themselves, which means that the analysis is based solely on data obtained from GBIF and literature data.

Response 4:Thank you for your meticulous review of this manuscript and the valuable comments you have provided! Regarding the data source issue you pointed out, we would like to provide an explanation here: This study focuses on integrating existing public data to explore the distribution patterns of species, and therefore we chose the GBIF database as the primary data source (a total of 166 records). As an international authoritative biodiversity platform, GBIF’s data openness, standardization, and geographical coverage provide a reliable foundation for this study. At the same time, we have conducted supplementary searches for relevant literature records (and have added the relevant references) to ensure the comprehensiveness of the data. We fully agree with your suggestion of adding field data. Due to the limitations of the research cycle and field sampling conditions, we were unable to include our own data this time. In the future, we will carry out targeted field validation and are willing to share the new data to the GBIF platform for reference by the academic community.

Comments 5:Please explain the discrepancy between the number of reported variables (56) and what is presented in Table 1. ‘In this study, a total of 56 environmental variables (Table 1) were selected from the World Meteorological Data website (WorldClim, https://www.worldclim.org)»

Response 5:Thank you for your comments. In Table 1, the last three variables represent the mean precipitation, mean maximum temperature, and mean minimum temperature, respectively. Each variable is measured on a monthly basis, resulting in a total of 36 variables for these three categories. Combined with the previous 20 variables, there are a total of 56 variables.

Comments 6:Explain the choice of script SSP245 and the reasons why, for example, script SSP285 was not used. Provide reasoning as to why this particular scenario is appropriate for your study.

Response 6:Thank you for your suggestion. The SSP245 scenario was selected for this study primarily because it balances policy realism and scientific robustness. It aligns with the mid-term emission reduction pathways of the Paris Agreement and is consistent with the current global trend in climate action. Its moderate warming trend is more suitable for assessing the adaptive responses of the target species (or ecosystems) in the medium to long term (e.g., the mid-to-late 21st century). The SSP285 scenario, which may represent a higher level of climate forcing, is considered less appropriate for several reasons. It significantly deviates from actual policy progress (e.g., by overestimating emissions in the absence of mitigation) and may only show extreme effects towards the end of the study period. Its predicted spatial heterogeneity and ecological operability are relatively low. Moreover, the SSP245 scenario's CMIP6 data offer greater advantages in terms of multi-model consistency and resolution. We have validated its applicability through sensitivity analysis and comparison with other studies. Future research will incorporate multi-scenario comparisons to expand the scope of our conclusions. Thank you again for your suggestion; we will further refine our analysis!

Comments 7:To obtain an optimal model, a series of models with different settings must be analysed. Good modelling results in Maxent can be obtained by using multi-stage analysis, which involves successive model calibration based on the development of multiple models with different settings.

Response 7:Thank you for your suggestion. We have optimized the model through a series of settings, increasing the AUC value in the Receiver Operating Characteristic (ROC) curve of the model prediction results from an initial 0.832 to 0.948. According to the relevant literature, we have achieved model optimization to the best of our ability. We truly appreciate the valuable suggestions provided by the expert.

Comments 8:To obtain correct models, careful selection of all modelling components is necessary: species distribution points, environmental variables, detail of analysis settings.

Response 8:Thank you for your suggestion. We have once again carefully screened the details of species distribution points, environmental variables, and analysis settings. The species distribution points were collected based on websites and relevant references (which have been supplemented in the text). The environmental variables were selected considering the biological characteristics of the starry beetle. The analysis settings were determined after carefully considering the modeling components, including species distribution points and environmental variables. We truly appreciate the valuable suggestions provided by the professor.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The paper has been improved but still has issues. 

Still have not described fully how the geographic data points obtained were filtered to determine unique occurances and avoid sampling bias.  If data points are from very close sources they should not be counted as unique points.

The number of variables is too high.  You should be comparing the models progressively by adding in variables and find the model that has the the lowest Akaike’s Information Criterion for small sample size scores (AICc) (Warren & Seifert, 2011) to use.  The contribution to the model is not a sufficient methods for selection and peason's correlation also does not fully deal with ruling out correlated variable.  For example, elevation is correlated with both temperature and precipitation and should not be used with them.  There are MaxEnt papers that delve into how to select the best variables and why.

There is a paper that used MaxENt to look at niche overlap between the native and invaded ranges for A. glabripennis and it found that bio 1, 4, 6, 11, and 13 were the best variables to use.  The authors of that paper used some of the same data you used so why are your variables different? Why have you not compared your results with their results?  

 

Comments on the Quality of English Language

The English is worse now following the revisions. There are many poor word choices in the sentences too. 

Author Response

Comments 1: Still have not described fully how the geographic data points obtained were filtered to determine unique occurances and avoid sampling bias.  If data points are from very close sources they should not be counted as unique points.

Response 1: Thank you very much for the reviewer’s suggestions. We are very grateful for your professional advice. By using ENMTools (https://github.com/danlwarren/ENMTools) to filter the obtained geographic data points to determine unique occurrences and avoid sampling bias, our research methods have been greatly improved. Thank you very much for your suggestions.

Comments 2: The number of variables is too high.  You should be comparing the models progressively by adding in variables and find the model that has the the lowest Akaike’s Information Criterion for small sample size scores (AICc) (Warren & Seifert, 2011) to use.  The contribution to the model is not a sufficient methods for selection and peason's correlation also does not fully deal with ruling out correlated variable.  For example, elevation is correlated with both temperature and precipitation and should not be used with them.  There are MaxEnt papers that delve into how to select the best variables and why.

Response 2: Thank you for your suggestions. Autocorrelation does indeed have some limitations. Following your advice, we used the ENMeval package to calculate AICc and then screened for the optimal model by regularizing parameters and features step by step to obtain the best variables. We also agree that elevation is correlated with temperature and precipitation. Therefore, based on the AICc values, we removed the variable of elevation. This has greatly helped us improve the quality of our article and gain a clearer understanding of the results. Thank you once again for your help with this paper.

Comments 3: There is a paper that used MaxENt to look at niche overlap between the native and invaded ranges for A. glabripennis and it found that bio 1, 4, 6, 11, and 13 were the best variables to use.  The authors of that paper used some of the same data you used so why are your variables different? Why have you not compared your results with their results?  

Response 3: Thank you for your suggestions. We have compared our study with the one you mentioned, and the specific comparisons are marked in the Discussion section. The differences in variables between their study and ours are due to the inclusion of predator variables in their analysis. Additionally, the model optimization parameters they used yielded different results from ours. However, their findings are partially consistent with ours, which also serves to validate our research. Thank you once again for your valuable suggestions.

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript is not of sufficient scope to be published in an MDPI journal.

Comments for author File: Comments.pdf

Author Response

Comments 1: The manuscript is not of sufficient scope to be published in an MDPI journal.

Response 1: Thank you very much for taking the time to review our manuscript and for providing your valuable feedback. We sincerely appreciate your insightful comments regarding the scope of our study. We understand your concerns about the current scope of the manuscript and its suitability for publication in an MDPI journal. If possible, we would be very grateful for any specific suggestions on how to expand or strengthen the scope of the study. We are fully committed to revising the manuscript to meet the journal’s expectations and are happy to incorporate constructive recommendations.Once again, thank you for your thoughtful critique. We look forward to your guidance on improving this work.

Reviewer 3 Report

Comments and Suggestions for Authors

The authors have addressed all suggested edits, but I have two critical observations:

1. Number of replicates per period
   The authors randomly allocated 75% of occurrence records for model training and 25% for validation, producing a new model with each replicate. With any sampling scheme, increasing the number of replicates improves confidence that model performance reflects the true estimation rather than chance. Conversely, a low number of replicates is effectively a small sample size.

However, the authors report that AUC decreases as the number of replicates increases. This counterintuitive trend likely arises because the environmental variables at the collection sites are highly heterogeneous—early, small-n runs may have by chance sampled more homogeneous conditions, inflating AUC. I therefore recommend that the Methods include a formal statistical justification for choosing only 10 replicates (e.g. bootstrap convergence rates or a plot of AUC mean ± SE vs. replicate count), since this choice runs contrary to expectations under the Law of Large Numbers.

2. Transition matrices

The current 4 × 4 matrices contain multiple errors, obscuring any spatially explicit analysis of suitability change. Specifically:

• Time intervals. Each matrix should be clearly labeled (“Current [1950–2000] → Future [2021-2040]”), matching the four suitability classes used in Figs. 5–7.
• Row sums. Every row of each transition matrix must sum to 1.0, yet in the supplementary archive sums up to 2.0 or even 0.0.
• Diagonal values. The first matrix implies no change (all diagonals = 1.0), which contradicts the mapped results.
• Other intervals. The remaining matrices likewise contain impossible totals given the figures 5-7.

I’ve attached an Excel file with corrected categories-by-categories Markov calculations and a suggested figure to incluye in transition map. Incorporating these will allow you both to (a) quantify the proportion of categories that remain in or transition between suitability classes from time t to t + 1, and (b) present a spatially explicit map of gains, reduction, and persistent areas.

Comments for author File: Comments.zip

Author Response

Comments 1: 1. Number of replicates per period
The authors randomly allocated 75% of occurrence records for model training and 25% for validation, producing a new model with each replicate. With any sampling scheme, increasing the number of replicates improves confidence that model performance reflects the true estimation rather than chance. Conversely, a low number of replicates is effectively a small sample size.

However, the authors report that AUC decreases as the number of replicates increases. This counterintuitive trend likely arises because the environmental variables at the collection sites are highly heterogeneous—early, small-n runs may have by chance sampled more homogeneous conditions, inflating AUC. I therefore recommend that the Methods include a formal statistical justification for choosing only 10 replicates (e.g. bootstrap convergence rates or a plot of AUC mean ± SE vs. replicate count), since this choice runs contrary to expectations under the Law of Large Numbers.

Response 1: Thank you for your suggestions. We have filtered out sampling bias according to your advice to improve the accuracy of the model, and we have also added the specific steps in the Methods section. Through this method, our AUC values have also been well validated, which can be seen in the attachment. Thank you once again for your contribution to improving the quality of this study.

Comments 2: 2. Transition matrices

The current 4 × 4 matrices contain multiple errors, obscuring any spatially explicit analysis of suitability change. Specifically:

• Time intervals. Each matrix should be clearly labeled (“Current [1950–2000] → Future [2021-2040]”), matching the four suitability classes used in Figs. 5–7.
• Row sums. Every row of each transition matrix must sum to 1.0, yet in the supplementary archive sums up to 2.0 or even 0.0.
• Diagonal values. The first matrix implies no change (all diagonals = 1.0), which contradicts the mapped results.
• Other intervals. The remaining matrices likewise contain impossible totals given the figures 5-7.

I’ve attached an Excel file with corrected categories-by-categories Markov calculations and a suggested figure to incluye in transition map. Incorporating these will allow you both to (a) quantify the proportion of categories that remain in or transition between suitability classes from time t to t + 1, and (b) present a spatially explicit map of gains, reduction, and persistent areas.

Response 2: Thank you for your suggestions. We have reconstructed the relevant 4×4 matrix according to your advice and added it to the paper. Once again, we appreciate your professional suggestions, which have greatly enhanced the quality of this paper.

Reviewer 4 Report

Comments and Suggestions for Authors The authors made adjustments based on the comments provided The authors made adjustments based on the comments provided.

Author Response

Comments 1:The authors made adjustments based on the comments provided The authors made adjustments based on the comments provided.

Response 1:We sincerely appreciate your time and effort in reviewing our work. We believe that the revisions have significantly enhanced our manuscript and addressed the concerns raised. Wish you a successful career and a pleasant life!

Round 3

Reviewer 1 Report

Comments and Suggestions for Authors They have responded appropriately to the comments I had on the last revision. The paper is much improved.

Author Response

Comments 1:They have responded appropriately to the comments I had on the last revision. The paper is much improved.

Response 1:Thank you very much for your positive feedback on our revised manuscript. We are very pleased to hear that you found our responses to your previous comments appropriate and that the revisions have resulted in a much-improved paper. We sincerely appreciate the time and expertise you dedicated to reviewing our work and are grateful for your invaluable insights which significantly enhanced the quality of the manuscript.

Reviewer 2 Report

Comments and Suggestions for Authors

The scientific novelty is low.
The use of MaxEnt to model the potential distribution of A. glabripennis is not new.
Multiple recent studies have addressed this same objective in China and on a global scale.
The manuscript does not incorporate new biological variables.
The main contribution suggested as novel is the more accurate prediction for northwest and southwest China, but there is no clear justification for why these results are more reliable or useful than existing ones.
The potential impact is limited.
The topic is of regional and applied interest, but the methodology used does not generate important findings or solid bases for forest policies beyond what is already known.
The scope of the manuscript in Forests is moderately adequate. 
However, Forests requires a clear and novel contribution to forest science.
The technical quality is poor.
The use of MaxEnt is well described, but the choice of the SSP245 scenario is not justified.
The limitations of the MaxEnt model are not discussed beyond a few general lines.
External validation or comparison with field data is lacking, which weakens the model's applicability.
The figures have improved substantially; they now meet the minimum standard, are larger, and are easier to read.

Suggestions for authors:

Consider other journals with a lower impact factor or a more regional focus.
Clarify the novel contribution of the study in relation to previously published work on A. glabripennis.
Incorporate non-climatic ecological variables (e.g., vegetation cover, host distribution, human population density).
Improve the English language and paragraph structure.

Comments on the Quality of English Language

The writing contains multiple errors in style, grammar, and punctuation.

There are redundant sentences and poor logical structuring in several sections. 

 

Author Response

Comment 1：Consider other journals with a lower impact factor or a more regional focus.

Response 1： Thank you very much for your suggestions. We believe that this journal is highly relevant to our research topic.

Comment 2：Clarify the novel contribution of the study in relation to previously published work on A. glabripennis.

Response 2：Thank you very much for your suggestions. We have added a detailed discussion of previous research related to A. glabripennis in the introduction section.

Comment 3：Incorporate non-climatic ecological variables (e.g., vegetation cover, host distribution, human population density).

Response 3：Thank you very much for your suggestions. Regarding the incorporation of non-climatic ecological variables, due to the current insufficiency of data, we will consider conducting research in future work. Thank you once again for your professional advice.

Comment 4：Improve the English language and paragraph structure.

Response 4：Thank you very much for your suggestions. We have optimized the language and structure, which has significantly enhanced the quality of our paper. Thank you once again for your valuable advice.

Reviewer 3 Report

Comments and Suggestions for Authors

The authors made the changes I suggested. In this regard, the incorporation of the transition matrices makes it possible to identify the habitability classes that will tend to remain or disappear. A final observation is that the authors can generate a map with the information presented in Tables 3 and 4, which will allow the identification of the areas where the habitability of the species will persist and change. 

 

Author Response

Comments 1:The authors made the changes I suggested. In this regard, the incorporation of the transition matrices makes it possible to identify the habitability classes that will tend to remain or disappear. A final observation is that the authors can generate a map with the information presented in Tables 3 and 4, which will allow the identification of the areas where the habitability of the species will persist and change. 

Response 1:We sincerely thank the reviewer for recognizing the value of incorporating transition matrices to analyze habitat class persistence. As suggested, we have now generated a comprehensive map (Figure 8,9) based on Tables 3 and 4, visually identifying areas where A. glabripennis habitat is projected to persist, decline, or transition under future scenarios. This enhancement directly addresses your insightful observation and significantly strengthens the spatial applicability of our findings. We deeply appreciate your expert guidance, which has greatly improved the clarity and utility of this study.