Forecasting Dendrolimus sibiricus Outbreaks: Data Analysis and Genetic Programming-Based Predictive Modeling

Round 1

Reviewer 1 Report

Reviewer Comments

The topic of this paper was interesting; however, the methods are not fully described, which makes it difficult to assess merit.  

MAJOR COMMENTS:

All figures: I am unable to interpret these figures due to the small font size. Additionally, the lettering has been cut off in some of the figures.

Data: What was the size of the dataset? How many presences versus absences were recorded? Was this a presence-only dataset where background points were used? If so, how many background points were used, and how were these selected? Given the spatial nature of the data, these points should be randomly selected spatially with a method that assesses model performance at different numbers of absences. This information should be included in the methods section. The results section discusses data imbalance, but this issue seems to be introduced abruptly and without context.

Analysis: This paper presents only one accuracy metric. Were the data separated into training and test/validation datasets? If so, what were the metrics for both? If the test dataset has an accuracy (0.9941) much higher than the training/validation set, the model may be overfitting the data. Regardless, splitting the data is crucial for proper model evaluation. Moreover, this should be done spatially (using spatial cross-validation) to adhere to best practices. Spatial cross-validation is vital because it accounts for the possibility that accuracy in one location could vary significantly from another. If all points in the test/validation dataset are close together and located in an area with uniformly high or low accuracy, the reported accuracy could be misleading. You mention spatial cross-validation briefly in the discussion regarding another paper, so perhaps this was performed but not included in the methods section.

Analysis: You present feature importance plots but do not discuss this aspect in your methods. Given how decision trees are constructed, highly correlated features can influence feature importance plots, which sometimes do not affect predictive capabilities, but can influence feature importance. How was this addressed? If it was not addressed, your model predictions may be reliable (although potential overfitting cannot be assessed without additional details), but caution should be exercised when stating that these are the variables most important for management, etc. They were important for the predictions and this model, but it is not justifiable to claim they are the most significant for the species or to assign biological meaning without further validation. Figure 5 suggests that you may have employed a technique where you left out important features to see how the model responds, thus assessing if the absence of a variable influences the importance of other features (this would allow you to discern whether the variables you are identifying as most important are indeed the most important, or if other highly correlated features are equally important). However, as this is not discussed in the methods, its implementation remains unclear.

Discussion: The discussion is primarily descriptive of the models rather than discussing the relevance of the study to forestry as a whole and connecting it to other published work. I would suggest further developing this section.

MINOR COMMENTS:

Line 1: Species names should be italicized.

Line 5: Use the specific model (random forest), as opposed to machine learning, which is very broad.

Line 8: Words that are in the title do not need to be included as keywords.

Line 11: Species names should be italicized.

Line 16: I would recommend clarifying what you mean by “advancing”.

Line 17: While numbers under 10 are typically spelled out, as written, units should be displayed as numbers. I recommend changing to 8 cm.

Line 44 and 58: Species names should be italicized.

Line 184: September is missing the r.

Line 198: Add an “a” between synthesizing and dataset.

Line 260: Species names should be italicized.

References should be double checked, as there are typos in some. 

There were errors throughout. I've detailed some in my minor comments to authors, but the manuscript should be given a more thorough editorial review. 

Author Response

Dear Reviewer 1,

We would like to extend our heartfelt gratitude for taking the time to review our manuscript.

Your expertise and attention to detail have been invaluable in identifying areas for enhancement and suggesting thoughtful recommendations for further development. Your dedication to the peer review process is commendable and greatly appreciated.

We genuinely value the time and effort you have invested in assessing our manuscript. Your expertise has undoubtedly enriched the quality and rigor of our research, ultimately contributing to its scholarly integrity.

Once again, thank you for your invaluable contribution to our manuscript. We are grateful for your commitment to advancing scientific knowledge and for your support of our work.

Sincerely, Ivan.

Author Response File: Author Response.pdf

Reviewer 2 Report

The article is well designed and prepared. But few small correction required before  final processing. My comments are given in attached file.

Thanks 

Comments for author File: Comments.pdf

Author Response

Dear Reviewer 2,

We wanted to express our sincere appreciation for your thoughtful review of our manuscript.

We are truly grateful for the time and effort you dedicated to assessing our manuscript. Your expertise and attention to detail have provided us with valuable insights and suggestions for improvement that we will carefully consider in our revisions.

Your commitment to the peer review process and your dedication to advancing scientific knowledge are truly commendable. We feel fortunate to have benefitted from your expertise and guidance.

Thank you once again for your invaluable contribution to our manuscript. Your feedback has been instrumental in enhancing the quality and impact of our research.

Warm regards, Ivan.

Author Response File: Author Response.pdf

Reviewer 3 Report

 The paper "Forecasting DS outbreaks: data analysis and genetic programming based predictive modeling" is undoubtedly a very interesting and innovative work with a broad range of applications in agronomy and the forest industry. It is well presented and explained from the introduction, through the objectives, coherent methodology, description of results, and discussion. However, I have some general comments that I believe could help improve the quality of the writing.

General aspects:

I am concerned about the number of times "Dendrolimus sibiricus" is written without italics. It should be italicized. Please correct this in the abstract, keywords, introduction, and throughout the text. Additionally, in the references, species names should also be italicized, just like "Dendrolimus sibiricus."

Introduction:

The introduction is generally well laid out. However, I would suggest that the authors reread it and try to reorganize the structure of the text to improve the flow and linearity of the ideas presented. Furthermore, there are some sentences that could be clarified as they repeat ideas unnecessarily. Please reread and revise accordingly. Finally, it might be an alternative to use acronyms to avoid mentioning concepts or species throughout the text, for example, "Siberian Moth" could be abbreviated as "SM". This could provide an alternative approach.

In lines 58-59, it is not necessary to put the scientific name "Dendrolimus sibiricus" in parentheses.

The figure 5 is not referenced in the text.

The description of the confusion matrices in Figure 5 could be more detailed to provide a better understanding of the performance of the classifier under different conditions.

The significance of the identified climatic variables could be emphasized further, highlighting their role in predicting forest conditions and pest outbreaks.

Author Response

Dear Reviewer 3,

I wanted to extend my heartfelt gratitude for taking the time to review our manuscript.

Your expertise and careful consideration have provided us with valuable perspectives that we may not have otherwise considered. Your dedication to ensuring the quality and rigor of scientific research is truly commendable, and we are grateful for your commitment to the peer review process.

Your insights have undoubtedly contributed to the refinement of our manuscript, and we are confident that your suggestions will enhance its overall quality and impact.

Thank you once again for your invaluable contribution to our research. Your efforts are deeply appreciated.

Warm regards, Ivan.

Author Response File: Author Response.pdf