Modeling Fire Boundary Formation Based on Machine Learning in Liangshan, China

Round 1

Reviewer 1 Report

In the present study, LightGBM and RF models have been utilized to estimate the fire boundary formation. For this purpose, a dateset has been gathered to develop the models

Generally the topic and methodology are interesting, however some critical issues and concerns exist and it is needed to in-depth revise the present manuscript. The main problem is that, nowadays and with considerable progress of the optimizing algorithms, single models cannot reach the highest accuracy in comparison to the optimized models. Another major problem is the poor and sometimes incomprehensible writing of the text of the article. All in all, I decided to invite the authors to revise their manuscript. Without a major revision, this manuscript cannot be accepted to publish. Special comments are as below that the authors should answer, apply and address them carefully in the manuscript.

 

Comments:

1.     The title of the article is incomprehensible, and it should be expressed in a more expressive way, relying on being a case study.

2.     The table of abbreviations should be presented.

3.     The keywords should be placed based on high importance and better order from higher importance to lower one. Some non-important ones must be removed.

4.     The English of the paper is poor making the paper difficult and even impossible to understand in some sections. It is strongly recommended that the authors rewrite the entire paper.

5.     The overall objectives, novelty and contribution are not well presented.

6.     Poor justification and what is missing is the novelty of the present work compared to previous work conducted along the same line of research. The authors should be more critical on what is lacked and how this work will bring the readers to increase their knowledge in the related field.

7.     The introduction of the paper is poor and generally not organized. There is not a reasonable flow between different paragraphs and sections in the introduction. To show the value of using the proposed approach, a review and comparison are necessary. Please complete the background of this research and update the references until 2023 with details and results. The authors should clearly explain this research's originality aspects that make this paper valuable for publication. The literature review for the AI-based studies is needed to consider in the introduction. Following references are recommended to consider in the introduction or methodology descriptions:

·        "Ensemble deep learning-based models to predict the resilient modulus of modified base materials subjected to wet-dry cycles." Geomechanics and Engineering 32.6 (2023): 583-600.

·        "A novel ensemble modeling approach for the spatial prediction of tropical forest fire susceptibility using LogitBoost machine learning classifier and multi-source geospatial data." Theoretical and Applied Climatology 137 (2019): 637-653.

8.      Why are optimized models not used in this study? In other word, how are the authors sure that the single-tuned models are more accurate than the models optimized by meta-heuristic algorithms?

9.     Statistical specifications of the dataset should be presented for training and testing datasets, separately (such as min, max, mean, median etc.). 

10.  Statistical analysis of the dataset using the schematic and visualization tool (such as boxplot, violin plot, frequency distribution, histogram, etc.) should be provided for training and testing datasets, separately.

11.  The explanations given about data dividing are often general. Specific and quantitative explanations are required.

Extensive editing of English language is required.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Dear Editor,

All comments are included in the attached file. I suggest the authors to rethink the methodology and answer the question of how many fires were analyzed to create the model. The article needs to be refined in some respects.

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

The article is very well written and at a high scientific level. Two issues need to be modified.

1. In the Conclusions, Abstract and Introduction, there must be noted and bolded that the conclusions are related to local conditions and the models were not checked for many different and sometimes unexpected locations.

2. The Authors write: "These factors include topography, vegetation, climate, and human activity". Is the weight of these factors the same? The authors should explain this. In addition, human activity is unpredictable, especially in terms of deliberate ignitions. It should be also discussed.

 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The revision made is appropriate and the paper can be accepted for Publication. 

 Minor editing of English language is required

Author Response

The manuscript has been revised using an English editing service.

Reviewer 2 Report

Dear Editor,

I am wondering about the correctness of the keywords in the article, but I think that the way they are written (abbreviations) depends on the requirements of the Editorial Board.

Most comments have been addressed.

In the sentence "N. Ryzhkova et al. presented the idea that climate drives the fire cycle and that humans influence fire occurrence in the East European boreal forest using an analytical fire model", it follows that people influence fires. The results that people respond to fires. The cause of fires, which is humans, affects not only the boreal forests, but it is a problem in various climatic zones. I propose to change that sentence. Please let me know how the authors of the article understand the definition of a boreal forest.

Author Response

The relevant content has been revised, as seen in lines 56-61 of the manuscript.