Review Reports

Round 1

Reviewer 1 Report

Impression:

Manuscript (ID forests-1109740) with the title “Ecological niches and suitability areas of three host pine species of bark beetle Dendroctonus mexicanus Hopkins” by Fátima M. Méndez-Encina et al. represents an average case study that correlates geographical distribution of parasitic insect and its host tree species under the control of climatic variables in the region of Central America. It is based on sufficient data records, combining own fieldwork and available digital databases. Statistical analyses are built on meticulous and successive procedures of ecological modelling, where although I am not a specialist, they seem to lack serious flaws. The text is relatively well readable and understandable, but in places its fluency and clarity is worse, especially several sentences are hard to comprehend. As to my feeling, too much stress is given to the methodological issues throughout the discussion, although in the type of case study I would prefer stronger consideration of concrete results of similar studies based on real data. This will provide better the usefulness and real potential of the submitted manuscript for both forestry sector and for nature conservationists.

Main comments:

L. 125: Species occurrence records were primarily obtained as point data, or were they treated primarily as grid cells? In the latter case, define geographical grid used (in the recent text this fundamental information occurs at first on l. 363).

L. 157: According to which key have you defined the three different sets of environmental variables?

L. 225: Why you cannot use precisely 70% and 30% of the data for modelling and validation, respectively? It seems you made some type of arbitrary selection from the data, which is hidden to the reader (i.e., why there is a portion of 30.39%, why not e.g. 28.5% or 31%?).

Table 2: Two numbers of parameters used for the construction of final distribution model look strange – models for P. devoniana and especially for P. teocote have extraordinarily large sizes, in contrast to three parameters only used in the model for P. leiophylla. You wrote about the problems when a statistical model contains many parameters (l. 452–460), so your model containing 81 parameters seems to be extremely overparametrized and I wonder how it can do a good job? If we assume each model parameter posses its own random error, then the additive effect of 81 random errors can make the whole model far from reality.

L. 281: You wrote about minimum and maximum intra-stratum variance, but on l. 187 you state minimizing the variance of the estimate only. Probably correct is to write “assuring minimum intra-stratum variance and maximum inter-stratum variance”.

L. 282: Suitability model for D. mexicanus was constructed using the lowest number of original occurrence records (N=86), but it yielded the largest area of high suitability (~ 234 thousands sq. km). Can you assure that the whole high suitability area is composed of highly reliable grids according to the quality of suitability estimation? In other words, how the model treated the geographic patches lacking occurrence records that enter the modelling algorithm?

Table 4: Thermal bioclimatic variables should be transformed to their ambient values (not as multipliers of 10). Also delete authorities from the species scientific names. In the legend, change Mad -> MAD. In the footnote, change Confidence interval of mean -> Confidence interval of the mean; and Mad -> MAD.

L. 324–325: It does not sound logical, why the regions with high climatic suitability for the bark beetle have to be classified as SAD-free SAP areas! Does it mean that currently these are the regions lacking records of the bark beetle incidence, yet highly suitable for its potential occurrence? Reformulate this sentence.

L. 343–346: I would expect the construction of an environmental space using variables BIO1, BIO10 and BIO11 with the highest contribution to the models as based on the presented results in the Table 4. Why there is a different set of bioclimatic variables here? Even more, variable BIO5 is missing in the Table 4.

L. 413: You mention “systematic sampling” here for the first time, but the method of systematic sampling has to be given already in the Methods section (chapter 2.3).

Chapter 4.2 is not appropriate for the presented case study. Rather than pure methodological and statistical notes, I would expect a comparison of achieved results with case studies using the same procedures, if available. This part is better to move to Supplementary Material as e.g. “Notes to the adopted methodology”. The only part worth to mention in the main text is on l. 432–434.

L. 481: This fact is antagonistic to the reasoning given on l. 452–460, so giving examples of models containing many parameters adds inconsistency in your discussion.

Chapter 4.3: Actually, I miss the explicit implications for future threat/spread/shrinkage of the bark beetle from your results – there are many less important details among which the most important message is lost. Also, this fundamental outcome cannot be detected from the further text (l. 507–525).

L. 497–500: I cannot fully understand what is delivered in this sentence, it lacks clear meaning for me – please streamline the sentence.

L. 508–514: I wonder to which extent do these reasonings affect the reliability of devising SAD-free SAPs from your model?

Supplementary material is not referred within the main text – it has to be added to proper places.

Minor comments: Recommendations for the text changes

L. 24: Please be more specific in the future distribution of D. mexicanus (“new climates and geographic areas” -> e.g. “colder climates and new northern areas”).

L. 25: Give full words of keywords, it is not a good practice to give abbreviations.

L. 28: Give more appropriate reference to some standard work (e.g. Kuennecke 2008, Temperate Forest Biomes).

L. 31: 120 species -> 120 pine species.

L. 32: “…occupying one of the first three places worldwide” sounds strange, better to write “…reaching one of the highest diversities worldwide.”

L. 37: …effects on species physiology, phenology and adaptation. -> …effects on adaptations of living organisms.

L. 45: term “endophyte” is confusing, it is used in a standard ecological terminology to describe parasitic fungi (phyte = plant).

L. 62: the region of geography -> the geographical region.

L. 77–79: This is a complicated sentence, hard to understand.

L. 80: Although much as been gained… -> Although much has been gained…

L. 92: There is missing a heading of a subchapter 2.1.

L. 98: …they are coniferous forests… -> …there are coniferous forests…

L. 103: altitudes are 1600 m -> altitudes are from 1600 m.

L. 105: Reference nr. 22 should be in fact reference nr. 32 (CHELSA climatic data by Karger et al.).

L. 165: The model was selected according to: -> The best model, or the parsimonious model, or another model type?

L. 265: between parentheses -> in parentheses.

L. 278: 3.728 -> 3.728, respectively.

Figure 2: Delete authorities from the species scientific names (also in other figure captions).

L. 294: Table 1 -> Table 4.

L. 296: “from three (variables)” – but in Table 4 there are four bioclimatic variables used in the model for D. mexicanus.

L. 301: “93.3%” – in Table 4 the value is 93.9%.

L. 305: had high variability -> experienced high variability.

L. 307: Relative contribution of the suitability model variables… -> Relative contribution (in percents) of the suitability model bioclimatic variables…

L. 317: 55,964.8 km2 -> 55,964.8 km2, respectively.

L. 322: first pine species -> P. leiophylla; for the second -> for P. teocote.

L. 324: area of distribution -> area of its distribution.

Figure 3: Swap and amend the caption text like this: “Suitable areas free of the bark beetle Dendroctonus mexicanus for: a) Pinus leiophylla (orange colour), b) Pinus teocote (blue colour) and c) Pinus devoniana (green colour)”. In line 339, change “of each species d), e) and f)” -> “of each pine species d), e) and f), respectively”. Also add explanation what is behind the variation coefficient (i.e., variation coefficient of what feature?).

Figure 4: Add explanation of “UC and “PN” into the caption.

L. 366: I=0.69 differs from the value I<0.66 on Figure 5i.

L. 378: “conservatism” rather than “conservationism”.

L. 381: delete “de”.

Figure 5: Add explanation of “Up, E, D” in the caption; edit “devoniana” in italics (l. 395).

L. 399–400: Void sentence, delete it (The criteria used…).

L. 414–416: Move first sentence (“The records used…”) to the chapter 2.3; delete the next sentence (it just repeats the point of Methods).

L. 440: Better to write “large/extensive” instead of “ample”.

L. 443–444: It seems more correct to finish the sentence with “in an unstable environment.” rather than “which does not occur with static variables (soil, …)”.

L. 461: Delete “(2002)”.

L. 464: “in accord with these authors” -> “in accord with theory”.

L. 473: Delete authorities from the species scientific names.

L. 481: “…plant species distribution.” -> “…plant species distribution very well.” Delete “(2011)”.

L. 504: “Pinus-Dendroctonus diversity” – what does it mean? Do you mean “host diversity”?

L. 520: “this figure” – which one?

L. 532: The variables representative of extreme temperatures… -> The variables representing extreme temperatures…

L. 538: more ample ecological niche -> broader ecological niche.

Author Response

 

Article

Ecological niches and suitability areas of three host pine species of bark beetle Dendroctonus mexicanus Hopkins

Main comments: REVIEWER 1.

1. 125: Species occurrence records were primarily obtained as point data, or were they treated primarily as grid cells? In the latter case, define geographical grid used (in the recent text this fundamental information occurs at first on l. 363). Yes, the species occurrence data correspond to point data.
2. 157: According to which key have you defined the three different sets of environmental variables? Setting the number of sets is optional, this procedure offers the opportunity to select the variable set that best predict the climatic suitability of the species. However, the variables are chosen once they have passed all the criteria established in section 2.2.
3. 225: Why you cannot use precisely 70% and 30% of the data for modelling and validation, respectively? It seems you made some type of arbitrary selection from the data, which is hidden to the reader (i.e., why there is a portion of 30.39%, why not e.g. 28.5% or 31%?). The distribution of the species occurrence data for training and validation is very important in ENM, especially the validation data set. The validation data set must be composed of independent records, so it is not possible to distribute precisely 70% and 30% since this depends on the availability of the records and the data source.

Table 2: Two numbers of parameters used for the construction of final distribution model look strange – models for P. devoniana and especially for P. teocote have extraordinarily large sizes, in contrast to three parameters only used in the model for P. leiophylla. You wrote about the problems when a statistical model contains many parameters (l. 452–460), so your model containing 81 parameters seems to be extremely overparametrized and I wonder how it can do a good job? If we assume each model parameter posses its own random error, then the additive effect of 81 random errors can make the whole model far from reality. The Maxent algorithm, implemented in kuenm, offers the possibility of modifying some parameters, e.g.: 1) "regularization multiplier" (β) and 2) "feature classes" (section 2.5) and thus make combinations to generate n candidate models, of which the same algorithm (kuenm) validates each of them with the set of independent records; what cannot be done with the traditional method, even with this last procedure, a single model is generated, which in many cases is random and not significant (Loiselle et al., [60]; Shcheglovitova & Anderson [60]). So the number of parameters in the final model depends on the number of "feature classes" that made up the model, as you can see, it is a combination of entity classes, it is a single model (there are no 81 models for the number of parameters), details on this can be consulted in Cobos et al., [16]. Overparameterization occurs when modeling the distribution of the species with multiple predictor variables. In order not to confuse the reader, this information was eliminated from the table.

1. 281: You wrote about minimum and maximum intra-stratum variance, but on l. 187 you state minimizing the variance of the estimate only. Probably correct is to write “assuring minimum intra-stratum variance and maximum inter-stratum variance”. The reviewer's observation is correct, the correction was made
2. 282: Suitability model for D. mexicanus was constructed using the lowest number of original occurrence records (N=86), but it yielded the largest area of high suitability (~ 234 thousands sq. km). Can you assure that the whole high suitability area is composed of highly reliable grids according to the quality of suitability estimation? In other words, how the model treated the geographic patches lacking occurrence records that enter the modelling algorithm?. The suitability of a species is not dependent on the number of records of presence of the species. Maxent, like most algorithms, uses presence records to model the climatic suitability, during this process, the algorithm generates “pseudopresences” of the species where there are no records, generating a class variable (presence and absence), which is modeled through a logistic model. The predictions (models) are validated according to different criteria (Section 2.5) using independent records and as considered by each researcher. In a study in this same species [27], they used 1500 presence records, there are no details of the cleanliness of the records, nor of the criteria for choosing the variables, more than 20 predictors are included in the modeling, the resulting areas are lower than those in our study. On the other hand, in this same species [Salinas et al., 2010], they used another modeling algorithm (Bioclim), with a different number of records of presence of the species; their results are very similar to ours. Unequivocally, the differences in the predictions are due to the modeling methodology. Therefore, the surface area of ​​the estimated climatic suitability is dependent on: the quality of the records, the geographical distribution of the records, the bias of the records, the type of variables, the number of variables included in the model, the criterion choice of predictor variables, number of candidate models for validating, algorithm, etc. Likewise, some variables overestimate the predictions [32], in our case they were discarded. Due to the criteria considered in our study as a rigorous selection of predictors (Section 2.2), an exhaustive cleaning of the records of occurrence of the species (2.3), generation of more than a thousand candidate models (rasters) and calibration, creation and evaluation from the candidate models (1392) for choosing the best model (section 2.5), we consider our results to be very robust.

Table 4: Thermal bioclimatic variables should be transformed to their ambient values (not as multipliers of 10). The variables are presented in the units as they were originally generated, as they are used for modeling, and as used in all the analyzes of this study; these have been expressed in their original form by most of the authors for comparison purposes. We believe that these should remain as they are. Also delete authorities from the species scientific names. Done. In the legend, change Mad -> MAD. Done. In the footnote, change Confidence interval of mean -> Confidence interval of the mean; and Mad -> MAD. Done.

2. 324–325: It does not sound logical, why the regions with high climatic suitability for the bark beetle have to be classified as SAD-free SAP areas! Does it mean that currently these are the regions lacking records of the bark beetle incidence, yet highly suitable for its potential occurrence? Reformulate this sentence. It is not SAD-free SAP; it is SAP free of SAD. Which, is where the climatic suitability (from a set threshold) of the Pinus species does not overlap in G space with the climatic suitability of the bark beetle species (established in section 2.7). According to the results, the SAD-free SAP areas do not occur exclusively in the regions of high climatic suitability of the species (Figures 2 and 3a-3b and 3c). The high climatic suitability of one species is not dependent on the high climatic suitability of another species even though they coexist in the same area. It may happen that the high suitability of the species is present in different regions (as observed in the results) presented or in the same area. The overlap in the fundamental niche (space E) between the pine and bark beetle species is very high, but this does not imply that the high climatic suitability of the pine species occurs in the same geographic space (G). Does it mean that currently these are the regions lacking records of the bark beetle incidence, yet highly suitable for its potential occurrence? Reformulate this sentence.
3. 343–346: I would expect the construction of an environmental space using variables BIO1, BIO10 and BIO11 with the highest contribution to the models as based on the presented results in the Table 4. Why there is a different set of bioclimatic variables here? Even more, variable BIO5 is missing in the Table 4. The environmental space was built with Bio5, Bio6 and Bio12 for the purpose of comparing the niche between the three species. In addition to verifying the correct cleaning of the records and the adequate classification of the species, in space E, with the same Bios it is possible to identify the overlap of the niche: existing and fundamental, the shape of the ellipsoid in three-dimensional space and the location of the species in different climates. Performing E with the Bios that represent the bioclimatic profile of each species, it would not be possible to make the previous comparisons.
4. 413: You mention “systematic sampling” here for the first time, but the method of systematic sampling has to be given already in the Methods section (chapter 2.3).

The mention of “systematic sampling” in this section (Discussion) is to support that it is better for modeling purposes in ENM. These cannot be in the Materials and Methods section because we did not do the sampling. The species records were downloaded from the databases available on the internet and others were requested from official institutions that had them (CONAFOR) and they were distributed in this way.

Chapter 4.2 is not appropriate for the presented case study. Rather than pure methodological and statistical notes, I would expect a comparison of achieved results with case studies using the same procedures, if available. This part is better to move to Supplementary Material as e.g. “Notes to the adopted methodology”. The only part worth to mention in the main text is on l. 432–434. By agreement of the authors, section 4.2 was not modified. We consider that this information should not be a complement to the article, here it is emphasized that a large part of the ENM studies are probably not robust, even not significant. On the other hand, according to ENM experts, the comparison between studies is valid only if it is the same model, which is complicated by methodological differences.

1. 481: This fact is antagonistic to the reasoning given on l. 452–460, so giving examples of models containing many parameters adds inconsistency in your discussion. Text removal changes were made as suggested by the Reviewer.

Chapter 4.3: Actually, I miss the explicit implications for future threat/spread/shrinkage of the bark beetle from your results – there are many less important details among which the most important message is lost. Also, this fundamental outcome cannot be detected from the further text (l. 507–525). Like this and all the suggestions made by the reviewers, each of them was agreed between the authors. The agreement reached by the authors ..., “it is considered important to leave the text as it is”. The fact is based on the fact that showing that the number of Bios considered in a model is fundamental to determine the vulnerability of the species, as long as the best modeling parameters ("regularization multiplier" and "feature classes") have been chosen, especially the selection of the best predictors. Also, this fundamental outcome cannot be detected from the further text (l. 507–525). They are two different things, the first refers to the fact that the number of variables that predict the climatic suitability of a species is important, especially if they are undergoing changes (for more details see the text). The last one refers to niche overlap, in which its vulnerability cannot be assessed. The overlap between the two species may be very high or very low, and despite this, the specie may or may not be susceptible to climate change.

1. 497–500: I cannot fully understand what is delivered in this sentence, it lacks clear meaning for me – please streamline the sentence. The paragraph was reformulated
2. 508–514: I wonder to which extent do these reasonings affect the reliability of devising SAD-free SAPs from your model?. Broennimann et al., [55] precursors of distributional ecology, mention that evaluating the niche overlap (climatic suitability) in G is complicated; since the models that predict the climatic suitability of each species are different, the bioclimatic variables that determine its profile are also different, as well as its extension. Therefore, the interpretation of the overlap metrics in the G space would be inconsistent and must be carefully evaluated. In our study, we obtained SAD-free SAPs through raster algebra. The overlap of the climatic suitability of the species of this study (evaluated in G), was carried out as suggested for [48].

Minor comments: Recommendations for the text changes

1. 24: Please be more specific in the future distribution of D. mexicanus (“new climates and geographic areas” -> e.g. “colder climates and new northern areas”). The paragraph was modified.
2. 25: Give full words of keywords, it is not a good practice to give abbreviations. Corrected.
3. 28: Give more appropriate reference to some standard work (e.g. Kuennecke 2008, Temperate Forest Biomes). Done.
4. 31: 120 species -> 120 pine species. Done
5. 32: “…occupying one of the first three places worldwide” sounds strange, better to write “…reaching one of the highest diversities worldwide.”. Done
6. 37: …effects on species physiology, phenology and adaptation. -> …effects on adaptations of living organisms. Done
7. 45: term “endophyte” is confusing, it is used in a standard ecological terminology to describe parasitic fungi (phyte = plant). Bark beetles are considered endophytes because they live inside of the bark of the tree. This term has been coined and published by a specialist in bark beetles [1, 12, 13]
8. 62: the region of geography -> the geographical region. Done.
9. 77–79: This is a complicated sentence, hard to understand. Corrected.
10. 80: Although much as been gained… -> Although much has been gained… Done
11. 92: There is missing a heading of a subchapter 2.1. We consider that it is not necessary because they are generalities.
12. 98: …they are coniferous forests… -> …there are coniferous forests… Done
13. 103: altitudes are 1600 m -> altitudes are from 1600 m. Done
14. 105: Reference nr. 22 should be in fact reference nr. 32 (CHELSA climatic data by Karger et al.).

References are correctly cited.

1. 165: The model was selected according to: -> The best model, or the parsimonious model, or another model type?. Corrected.
2. 265: between parentheses -> in parentheses. Done
3. 278: 3.728 -> 3.728, respectively. Done

Figure 2: Delete authorities from the species scientific names (also in other figure captions). Done.

4. 294: Table 1 -> Table 4. Corrected
5. 296: “from three (variables)” – but in Table 4 there are four bioclimatic variables used in the model for D. mexicanus. Corrected
6. 301: “93.3%” – in Table 4 the value is 93.9%. Corrected
7. 305: had high variability -> experienced high variability. Corrected
8. 307: Relative contribution of the suitability model variables… -> Relative contribution (in percents) of the suitability model bioclimatic variables… Done
9. 317: 55,964.8 km2 -> 55,964.8 km2, respectively. Done
10. 322: first pine species -> P. leiophylla; for the second -> for P. teocote. Done
11. 324: area of distribution -> area of its distribution. Done

Figure 3: Swap and amend the caption text like this: “Suitable areas free of the bark beetle Dendroctonus mexicanus for: a) Pinus leiophylla (orange colour), b) Pinus teocote (blue colour) and c) Pinus devoniana (green colour)”. Done. In line 339, change “of each species d), e) and f)” -> “of each pine species d), e) and f), respectively”. Done. Also add explanation what is behind the variation coefficient (i.e., variation coefficient of what feature?). Done

Figure 4: Add explanation of “UC and “PN” into the caption. Done

1. 366: I=0.69 differs from the value I<0.66 on Figure 5i. The value showed in the figure represents p-value, which is different from value of I.  
2. 378: “conservatism” rather than “conservationism”. Done
3. 381: delete “de” Done

Figure 5: Add explanation of “Up, E, D” in the caption; edit “devoniana” in italics (l. 395). Done

1. 399–400: Void sentence, delete it (The criteria used…). Corrected.
2. 414–416: Move first sentence (“The records used…”) to the chapter 2.3; delete the next sentence (it just repeats the point of Methods). It is not possible to move it, it is discussion, It is not part of the methodology of this study.
3. 440: Better to write “large/extensive” instead of “ample”. Done
4. 443–444: It seems more correct to finish the sentence with “in an unstable environment.” rather than “which does not occur with static variables (soil, …)”. Done
5. 461: Delete “(2002)”. Done
6. 464: “in accord with these authors” -> “in accord with theory”. Done
7. 473: Delete authorities from the species scientific names. It is the first time that the scientific name of this species has been cited. The authorities of this species should be included.
8. 481: “…plant species distribution.” -> “…plant species distribution very well.” Delete “(2011)”. Done.
9. 504: “Pinus-Dendroctonus diversity” – what does it mean? Do you mean “host diversity”?. Corrected.
10. 520: “this figure” – which one?. It was a translation error, it was corrected.
11. 532: The variables representative of extreme temperatures… -> The variables representing extreme temperatures… Corrected
12. 538: more ample ecological niche -> broader ecological niche. Corrected

 

Author Response File: Author Response.docx

Reviewer 2 Report

Review of forests-1109740

 

This manuscipt titeld "Ecological niches and suitability areas of three host pine species of bark beetle Dendroctonus mexicanus Hopkins" is dealing with modelling of a bark  beetle species and his host trees. Although I find the paper interesting, the methodological approach is somewhat hard to follow. The results, explaining that variables related to temperature were the most important, are in agreement with other scientific studies.

I suggest a minor revision.

Intro: Is it bad under current cicumstanec? Is there large tree mortality caused by bark beetles now?

 

l.14-15: author mention expansion of bark beetles, but what about the expansion/shrinking of the habitat of the host tree species. This should be mentioned in the Intro or dicussion (or both).

23-24: what new areas and climates?

47: twelwe of the 19 know species of bark beetles? I think there are more than 19 species of bark beetles.

l.67-68: the reference is 14 years old so I would not claim that it is still valid

l.98: map or a figure would be good

l.105: what is the scale of the study area?

111: the PCA was used for each species? I do not understand this part. Variables were analysed with the current range of species?

147-189: this part was hard to follow. Authors refer to Fig.1, but I think they should use numbers in figures and refer to each step, because I was not able to understand the workflow from the text.

206: " distance called  I  using ordination methods (PCA)" ?

223-226: occurence records were from the GBIF (l.126)

Table 1: I do not understand how this PCA was done? Species occurence? How have you calculated PCs for each species?

Fig.5: the red colour is what?

l.407: 86 records of the D. mexicanus were from which database, source?

l.457: have you mentionded the problem of collinearity in the methods section? For the PCA or for your modelling?

535-536: I would not use "G" or "E" in conlcusions. Be more precise what you mean.

542-543: what do you mean by "priority for conservation"? In case bark beetles will spread, how do you want to conserve the current state? Is this possible? More info should be added on the current state of bark beetle management in the study area. Is the area mostly unmanaged (mountains etc.).

The conlcusion could be writtent more clearly - do the authors expect a large danger of spread of beetles into Pinus species? What do they propose?

Author Response

Article

Ecological niches and suitability areas of three host pine species of bark beetle Dendroctonus mexicanus Hopkins

Main comments: REVIEWER 2.

Intro: Is it bad under current cicumstanec? Is there large tree mortality caused by bark beetles now? Yes, tree mortality due to bark beetles species is increasing, and not only in Mexico, throughout the world, due in part to the increase in global temperature. In the discussion section there are quotes that demonstrate this.

 l.14-15: author mention expansion of bark beetles, but what about the expansion/shrinking of the habitat of the host tree species. This should be mentioned in the Intro or dicussion (or both).  The requested information was added in the introduction, already under discussion. However, this information is about niches, there are no studies on bark beetles. Niches is another concept.

23-24: what new areas and climates? The sentence was reformulated.

47: twelwe of the 19 know species of bark beetles? I think there are more than 19 species of bark beetles. According to the literature consulted, this is the number of species that are registered in Mexico.

l.67-68: the reference is 14 years old so I would not claim that it is still valid. Yes, it is valid, Elith (2006) is one of the greatest exponents in ENM. This document emphasizes the performance of various modeling algorithms, evaluation metrics, etc. This reading should not be lacking in any ENM study.

l.98: map or a figure would be good. We do not think that it is necessary to show a map of the study area due that in our results we showed it,

l.105: what is the scale of the study area?. It is indicated in the text, it is justified.

111: the PCA was used for each species? I do not understand this part. Variables were analysed with the current range of species?. These procedures were carried out in each species.

147-189: this part was hard to follow. Authors refer to Fig.1, but I think they should use numbers in figures and refer to each step, because I was not able to understand the workflow from the text. An attempt was made to modify the figure as suggested by the referees, but by adding numbers or in the section that each action belongs to, the figure does not improve, on the contrary.

206: " distance called  I  using ordination methods (PCA)" ? Yes, distance I, calculated using PCA ordination methods, is indicated in the text.

223-226: occurence records were from the GBIF (l.126). No, the species records were obtained from various sources, those indicated in section 2.3.

Table 1: I do not understand how this PCA was done? Species occurence? How have you calculated PCs for each species?. The PCA performed is explained in section 2.2. A PCA was made for each species. Table 1 shows only the contributions of the variables determined by the PCA. Variables that were preselected to do the modeling.

Fig.5: the red colour is what?. Symbols and text were added to the title of Fig 5 to explain missing details.

l.407: 86 records of the D. mexicanus were from which database, source?. The records of the presence of D. mexicanus and each of the species were obtained from different sources, it is mentioned in section 2.3.

l.457: have you mentionded the problem of collinearity in the methods section? For the PCA or for your modelling?. Yes. In the methods section (2.2), it is mentioned that to avoid collinearity in the modeling, non-correlated variables (r <0.8) were selected, determined through parametric correlation (Pearson), on the variables transformed to logarithms. Collinearity is avoided by modeling with few variables, as it was done in this work, it is also explained in this section. For the PCA, performed to obtain the contribution of the variables, it is necessary that the variables are correlated, otherwise this analysis is not justified. Likewise, this issue and how it was avoided in this study is addressed in discussion.

535-536: I would not use "G" or "E" in conlcusions. Be more precise what you mean. The paragraph was modified

542-543: what do you mean by "priority for conservation"? In case bark beetles will spread, how do you want to conserve the current state? Is this possible? More info should be added on the current state of bark beetle management in the study area. Is the area mostly unmanaged (mountains etc.). The sentence refers to the priority of conservation of suitable areas (free of suitable areas of the bark beetle) of the pine species, not of the conservation of the species, and no of the bark beetle specie.

The conlcusion could be writtent more clearly - do the authors expect a large danger of spread of beetles into Pinus species? What do they propose?. The conclusions were enriched

 

Author Response File: Author Response.docx

Reviewer 3 Report

Review for Forests 1109740 by Mendez-Encina et al.

This is a very good paper. Most papers addressing the effect of climate change on the future distribution of insects do not include the effect of climate change on the host. This work is quite innovative in this sense.

The paper does rely heavily in the models with scant information on forest ecology and the interactions of the insect and their hosts. I would like to see more ecological and entomological discussion in the papers.

In the discussion, the authors should also discuss their findings and how they fit within the current literature. I did not see much of this.

I am not very familiar with the modeling part of the work so hopefully other reviewers can comment on that.

While revising the paper check the gramma for verb and subject agreement – noticed a few errors on this.

Some additional comments and suggestions follow:

Line 19, sentence indicating “trend to precipitation was positive and negative to temperature, the latter determining climatic suitability of the species.” – Which species is this referring to? Please clarify.

Line 21, indicates: “Indeed, a single variable (Bio 1) contributed 93.9% to the model (Pinus leiophylla Schl. & Cham). – Why is the pine species in parenthesis? I suppose that is the subject of the sentence, if so include in sentence.

Line 22, same comment of following sentence: “The overlap of suitable areas for Dendroctonus–Pinus is 74.95 % (P. leiophylla)

Line 31, don’t think 40% is needed

Line 49, suggest delete “irreversible damage” – human disturbances can cause this level of damage but not bark beetles in their natural environment.

Line 51 – is this because these species are more susceptible of more abundant.

Line 68, states “Maxent performs better than standard methods”. Suggest briefly indicate why this may be the case.

Lines 82-84 indicate: “Thus, for the SDM objective in Pinus to be possible, it is necessary to have precise, reliable predictions for suitable areas of Pinus (SAP), free of suitable areas of Dendroctonus (SAD) in order to implement management and conservation strategies efficiently. – I would disagree with this idea. Conservation, management of pine forests does not need to occur under the absence of Dendroctonus. The interaction has been occurring for thousands of years and they still co-exist in an equilibrium.

Line 86, should be “its” instead of “their”

Lines 93-95, need to distinguish whether these are the most susceptible species or the most geographically abundant.

Line 105 – please refer to Table 1 here so the readers understand this.

Line 125, should be “were” instead of “was”

Line 129, “1) outside the geographic range (latitude and longitude)” – is this of the species or the study areas.

Line 140, has this figure been referred to earlier?

Lines 159-160, sentence reads kind of awkward. Suggest edit.        

Line 176, consider referring to this figure earlier.

Line 225, suggest delete “only”

Line 225, why were so many records eliminated for leiophylla? Suggest explain.

Line 228, are the numbers listed here in Table 1?

Line 236, suggest mention what criteria were not met.

Figure 2, would be good to include the species in each figure. Would be easier to follow.

Line 284-286, Very confusing sentence. Please edit for clarity.

Line 365, I would disagree that D. mexicanus should be considered an invasive species

Line 398, Instead of “Cleansing” suggest something like “The used of reliable records”

Lines 402-403, I would also delete this statement – it is basically stating that the model is indicating where the insects are. Moreover even in this case would be biased since those records were eliminated.

Line 407, Suggest delete “according to the authors, this is a valid model” – of course, otherwise you would not be writing this paper.

Line 428, delete “severely”

Line 434 and 441, should be “are” instead of “is”

Line 445, is “Fourcade” like a citation as in Fourcade et al. – if so needs to be corrected. Also suggest delete “and other researchers surprised the scientific community when they”

Line 446, what does “paints” refers to here

Author Response

 

Article

Ecological niches and suitability areas of three host pine species of bark beetle Dendroctonus mexicanus Hopkins

REVISOR 3

Line 19, sentence indicating “trend to precipitation was positive and negative to temperature, the latter determining climatic suitability of the species.” – Which species is this referring to? Please clarify.  The paragraph was modified.

Line 21, indicates: “Indeed, a single variable (Bio 1) contributed 93.9% to the model (Pinus leiophylla Schl. & Cham). – Why is the pine species in parenthesis? I suppose that is the subject of the sentence, if so include in sentence. Yes, Pinus leiophylla is in parenthesis because is the referred specie.

Line 22, same comment of following sentence: “The overlap of suitable areas for Dendroctonus–Pinus is 74.95 % (P. leiophylla). Same answer.

Line 31, don’t think 40% is needed. Corrected.

Line 49, suggest delete “irreversible damage” – human disturbances can cause this level of damage but not bark beetles in their natural environment. It was left as it was. “irreversible damage” according to the literature because even bark beetles are natural part of the ecosystems are one of the agents most destructive killing larges areas of coniferous forest.

Line 51 – is this because these species are more susceptible of more abundant. A small change was made in the paragraph.

Line 68, states “Maxent performs better than standard methods”. Suggest briefly indicate why this may be the case. It is not the aim of the study to compare algorithms. Details on the performance of various modeling algorithms can be found in the attached quote.

Lines 82-84 indicate: “Thus, for the SDM objective in Pinus to be possible, it is necessary to have precise, reliable predictions for suitable areas of Pinus (SAP), free of suitable areas of Dendroctonus (SAD) in order to implement management and conservation strategies efficiently. – I would disagree with this idea. Conservation, management of pine forests does not need to occur under the absence of Dendroctonus. The interaction has been occurring for thousands of years and they still co-exist in an equilibrium. Corrected.

Line 86, should be “its” instead of “their”.  Done

Lines 93-95, need to distinguish whether these are the most susceptible species or the most geographically abundant. It was indicated in the paragraph

Line 105 – please refer to Table 1 here so the readers understand this. Done

Line 125, should be “were” instead of “was”. Done

Line 129, “1) outside the geographic range (latitude and longitude)” – is this of the species or the study areas. It refers to eliminating data outside the geographic range of the species. The paragraph was modified.

Line 140, has this figure been referred to earlier?. No. The paragraph was corrected because this refers to the authors' BAM diagram [28]. This figure is not included.

Lines 159-160, sentence reads kind of awkward. Suggest edit.   This observation was corrected at the request of another reviewer. The confusion is because the paragraph says “with approximately 70% of the records; with independent data (∼30%). In results, this percentage is specified, making this confusion clear. In response to the other reviewer, it is explained why these values.

Line 176, consider referring to this figure earlier. We believe that referring to Fig 1 in this place is fine, because it is presented after each one of the modeling process has been explained.

Line 225, suggest delete “only”. Done.

Line 225, why were so many records eliminated for leiophylla? Suggest explain. It's good reviewer observation. In section 2.3 the criteria used in the cleaning of the species occurrence records are specified. The elimination of a large percentage of records in this species was due to the fact that they did not meet the established criteria, and is dependent on the species and its ease of identification. The cleaning procedure is crucial in ENM and the most important, on this depends the selection of the variables and the results of the modeling. It all depends on the cleanliness of the records. The paragraph was modified at the request of the reviewer.

Line 228, are the numbers listed here in Table 1?. Yes. The cumulative percentage of the variance explained by the first two main components of PCA in each species is shown in Table 1.

Line 236, suggest mention what criteria were not met. The paragraph was modified to explain why these variables in question were not selected.

Figure 2, would be good to include the species in each figure. Would be easier to follow. Each species is referred by: a), b), c) and d).

Line 284-286, Very confusing sentence. Please edit for clarity. A section of the paragraph was removed to better understand the results.

Line 365, I would disagree that D. mexicanus should be considered an invasive species. Under the hypothesis of niche conservatism of the genus Pinus; it is known that it was established ∼145 million years ago in the lower Cretaceous [57]. We did not find information on when D. mexicanus was established, but it was considered an invasive species due to its mobility capacity in the face of changes in climate, especially temperature.

Line 398, Instead of “Cleansing” suggest something like “The used of reliable records”. Done

Lines 402-403, I would also delete this statement – it is basically stating that the model is indicating where the insects are. Moreover even in this case would be biased since those records were eliminated. The authors' idea is to emphasize that including altitude with the bioclimatic variables in the PCA analysis (statistical procedure) was crucial in the cleanliness of the records. This procedure has not been documented in NMD studies. With the rest of the cleaning criteria of the records, it was not possible to detect those that do not correspond to the species.

Line 407, Suggest delete “according to the authors, this is a valid model” – of course, otherwise you would not be writing this paper. Done

Line 428, delete “severely”. Done

Line 434 and 441, should be “are” instead of “is”. Done

Line 445, is “Fourcade” like a citation as in Fourcade et al. – if so needs to be corrected. Also suggest delete “and other researchers surprised the scientific community when they”. The text was not modified. We consider this statement is important to support the correct selection of predictors.

Line 446, what does “paints” refers to here. Information was added in the paragraph to clarify the reviewer's doubt.

Author Response File: Author Response.docx