All reviews of published articles are made public. This includes manuscript files, peer review comments, author rebuttals and revised materials. Note: This was optional for articles submitted before 13 February 2023.
Peer reviewers are encouraged (but not required) to provide their names to the authors when submitting their peer review. If they agree to provide their name, then their personal profile page will reflect a public acknowledgment that they performed a review (even if the article is rejected). If the article is accepted, then reviewers who provided their name will be associated with the article itself.
Thank you for choosing PeerJ for publishing your work.
[# PeerJ Staff Note - this decision was reviewed and approved by David Roberts, a PeerJ Section Editor covering this Section #]
Please revise the points raised by the reviewers.
In general, the new version of the article has improved considerably compared to the first. The introduction is easier to follow, the results are better structured and the manuscript is improved and has a clearer distribution. Therefore, the manuscript falls within the scope of the Peer J Journal.
no comment
I suggest that authors would improve some other minor points.
Introduction:
L 61-69: I suggest that is not necessary that explain all the infection process…of the huanglongbing (HLB). Please, resume this part and rewrite more general.
L126: “Asian citrus psyllid” to “D.citri”
M&M:
L 136-143: This is not M&M. I think that the authors could use this information for the discussion part.
L 168: Remove link and introduced this reference (EPPO, 2019) that appear the data acces to this link...
Discussion:
L 292: Change: true skill statistic (TSS)
Introduction
L-61: Check for duplicated under dash
L-87, try to avoid using etc. in the manuscript
Materials and Methods
Model selection. The justification to use MaxEnt is extense, reduce this section to a few lines or move to Introduction.
L-159, Table 1 refers to correlation coefficients, not variable contribution
L-197, there are two Wang et al. 2018 in the references, use letters to differentiate them.
Results
L-263 : L-285, move criterion to obtain suitability values for environmental variables to M&M and leave only the ranges for each variable.
Discussion & Conclusions
Some suggestions and corrections are highlighted in the manuscript.
Last paragraph of conclusion is more general and is not directly related to objectives and results.
References
Check upper and lower case words in paper titles and journal names.
Figure 2. Rewrite the legend to avoid repeating D. citri
Figure 3. Check that x-axis title of each plot only contains uppercase letter at the beginning.
Figures 2-4. Write legend of figures in a continuous way, that is, join the two separated statements of the caption.
Table 1 just provide one value above 0.8 to fulfill the exclusion criterion, thus, this particular result is better presented in the text and Table 1 removed.
Other comments are annotated in the manuscript in yellow and sticky notes.
Authors should describe in more detail how they extracted the values to compute the correlation coefficient between the bioclimatic variables: Extracting from the observed presence data of D. citri? From a random sample? Using all the cell values in the raster layers? They should also annotate the sample size. Usually, a significant correlation (abs(r) > 0) is found with large sample sizes even for numerically low r values, as presented for most of the variable pairs in Table 1 at p < 0.01. Explain.
Usually, altitude is not used in modeling species distributions because it is highly correlated with environmental variables and it is strange is not included in Table 1. Explain.
Response curves in Figure 4 should not be the marginal response curves as they are difficult to interpret if variables are correlated, as presented in Table 1. Instead, use the response curves for each variable.
Last paragraph of conclusions is more speculative and beyond the purpose of the research.
The manuscript is well written with few redaction errors, some of them annotated in the document. Most of the issues are minor corrections/ suggestions. The only point that need to be explained or better described is the computations of the correlation coefficient. Usually, only for large sample sizes small r values are significantly higher than zero, as presented in Table 1. Thus, the results in Table 1 suggest the analysis was carried on with all the values of the raster layers. Usually, correlation among variables is performed by extracting the values at each of the presence point data and not using the whole raster layer. While it may not affect the general conclusions and results, the marginal response curves generated by MaxEnt are difficult to interpret if variables are correlated. Thus, response curves should be computed for a single variable and, if that was the case, the authors should make it clear.
The study has potential but the manuscript is poorly prepared and it needs major revisions before it can move forward.
The manuscript analyzed the potential spread of D. citri in china using the MaxEnt model to map the potential spread of D. citri in China under current climate scenarios, and the relationship between the spread of D. citri and the environmental factors. Therefore, the manuscript falls within the scope of the Peer J Journal. However, this manuscript is a bit confusing and it is hard, at least for this reviewer, to follow the introduction, the hypotheses, methodology and results obtained. In this senses, the manuscript would improve if it had a clearer distribution. Otherwise, it is very difficult to evaluate the findings. My comments attachments try to improve some of the shortcuts I found during the reading.
It is knowing that D. citri is distributed in China (Yang et al. 2006; Ke et al., 1988; Tsai et al., 1988; CABI 2011 and EPPO 2014). However, there are four areas (Chongqing and Hubei) where this psyllid is not localized and the model confirm there are” high susceptibility” areas as a potential D. citri distribution. Maybe the authors could be focus the study on these areas where the psyllid is not present, it could be more interesting.
Furthermore, for the development of the potential distribution model the authors should be considered three main factors: a) the availability of susceptible hosts, b) data about the presence or absence of the D. citri in the areas and c) the suitability of climatic conditions for the development of D. citri. This might have affected the results obtained and therefore, it should be, at least, explained in the discussion.
The manuscript should be checked against misspelling to better understand the paper. Some examples are marked in yellow in the manuscript (for example lines 49, 83, 100, 314 et seq.). Also, the authors should verify if they refer to the vector D. citri or host (lines 170, 336), kiwi is not known to be a host of this species, only Rutaceae. Some directions to paper sections are confuse, for example, line 166 “…evaluation criteria in 2.2” does not match any part of the manuscript because it is not numbered in their sections.
Some key papers are missing in the introduction, which address the potential distribution of D citri in other parts of the world, for example in Mexico (Lopez-Collado et al. 2013. J Insect Sc. 13, Art. 114), Mediterranean basin (Gutierrez & Ponti. 2013. Fla. Entomol. 96, 1375-1391), Australia (Aurambout et al. 2009. Ecol. Model. 220, 2512-1514), worldwide (Narouei-Khandan et al. 2016. Eur. J. Plant Pathol. 144, 655-670). These papers involve a diverse array of methodologies dealing with the potential distribution of D. citri and may also serve to compare with variables selected in this paper.
In several places of the manuscript appears “contribution rate” referring to the proportion or fraction of the given variable to the total model fitting, therefore rate in the sense of change is not an accurate description of the value.
Figures also contain some misspelled words (Figs. 2, 3: Sensitive vs Sensitivity, prediciton vs prediction). Figure 1 is recommended to combine with Fig. 4 as it will show observed versus predicted presence. Caption in Figure 5 refers to a single model not several models.
Table 1 refers to categories of suitability in distribution of the target species not habitat type.
Table 2 need to use period not comma to separate decimals.
The manuscript need to be checked such that citations in the text matched the references at the end of the manuscript (for example lines 62, 111, 121).
The list of references should be check to write the appropriate names, for example lines 360, 366, 418; removing or adding blank spaces (lines 395, 408). Also, some references are unordered (line 449), and duplicated (lines 466 and 469).
It is recommended to upload the point presence data of D. citri in China for future users.
The paper addresses a valid question about the distribution of D citri in China and uses an accepted approach to estimate the potential distribution of this species. However, some issues make the paper confusing and contradicting. For example, the methodology in the paper of Worthington et al. (2016), cited in lines 111-112, is used to select from the 19 bioclimatic variables. However, Worthington et al. used different predictor layers and applied rank correlation, which usually is a non parametric index better suited to analyze ordinal variables. However, in the current paper, all the bioclimatic variables are continuous, therefore the Pearson coefficient (r) is more appropriate, yet it is confused with the Spearman coefficient (line 182). Moreover, they correctly indicates that if r > 0.8, the correlation is high, but table 2 shows low correlation values AND a significant departure from zero, which would indicate the variables used in the final model are correlated and should be dismissed from this model! Usually, the selection of predictor variables is a step before building the final model but table 2 presents correlation results after the variables have been selected which does not make sense.
Also, lines 111-112 indicates using the methodology in the paper of Worthington et al. (2016) to select predictor variables as mentioned before but in later lines (141-143), they indicate using jackknife within MaxEnt to select environmental variables. But if they were selected previously by using correlation VIF or PCA (as described by Worthington op. cit.) then how is that other variables escaped the selection process?
Regarding the presence point data presented in Fig. 1, it is not recommended to use presence data from a distant region to model occurrence in another zone, instead project the model to a novel zone, also known as trasnferability (Peterson et al. 2007. Ecography, 30, 550-560; Phillips. 2017. A brief tutorial on MaxEnt. Available from url: http://biodiversityinformatics.amnh.org/open_source/maxent/). Therefore, it is recommended to drop those presence data that do not occur within the target region, that is, China.
On the other hand, lines 148-150 present a misunderstanding in the wording of the distribution layer nature. The output file from MaxEnt, ASCII by default, is a text file but the content is in raster format and so the other data types (grd, bil, mxe). It is strange the ArcGis program could not read the file because the structure of the ASCII file is proprietary by ESRI, the creators of ArcGis, thus some steps might be skipped that did not allowed visualizing the output file. Probably the best way to deal with this issue is simply remove this step, which is not fundamental to the analysis.
In general, results are clear but some words do not represent the content, for example, lines 186, 243, Fig. 3 refers to a reconstructed model but it would be more appropriate to use final model. This is because the process of model building does not consider assembling and disassembling model components, rather, there is process of variable selection to make the final model parsimonious.
The first paragraph of the discussion (lines 231-243) does not relate directly with the results, rather, is information that justify using MaxEnt. Therefore, this paragraph may be better moved to the introduction or removed at all.
Lines 244-249 refers to a common problem with model overfitting and again, is not the objective of this paper.
The discussion about environmental variables may be enriched by contrasting the findings in this papers with those variables reported in previous papers, as referenced before.
Regarding the conclusions, some parts (lines 334-336) does not derive from the analysis but are more speculative and out of context.
In summary, the paper address an important aspect of knowing the potential distribution in China because both HLB and D citri are native from Southeast Asia. However, the paper would improve by addressing language and grammar issues in both the manuscript and figures; clarify some methodological steps in the analysis, comparing results with previous work and limiting the conclusions to the scope of the research. Other specific comments are annotated within the manuscript
All text and materials provided via this peer-review history page are made available under a Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.