Prediction of Forest Fire Occurrence in Southwestern China
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsDear authors!
Your article is in the scope of Forests journal.
But at present time artcile has some disadvantages.
Introduction
I study your Introduction section and I found only 11 references taking into account that Introduction is combined with Background section. There is no separate Background Section. It is totally unacceptable situation.
You must completely rework Introduction with published international literature on fire occurrence and prediction topic.
You must provide references on published works around the world taking into account USA, Canada, Russian Federation, Australia and South Europe at least, so the Forests is international journal.
I estimate the needed quantity of additional cited works abou 50 new original references cited in Introduction to discover current situation in forest fire occurrence and prediction field of science.
Resources and Methods
Table 1. Why did you not use information on Lightning activity and human activity? Prediction of forest fires can not be processed without this information on reasons of ignition. Please, rework this section according suggested modification of Table 1.
3 Results
Please, rename this section to Results and Discussion taking into account information on forest fire ignition reasons like Lightning activity and human activity over controlled territory.
At least, you must provide extensive explanation why did you not taking into account such significant information.
Maybe, you should change the title of the article to "Prediction of forest fire occurrence on weather and vegetation conditions: Case Study".
Maybe, you can link road density, residential areas and holidays information with human activity reasons of forest fuels ignition.
Conclusion
There is no Conclusion section.
Please provide this section in the article.
Please, mark out 3-5 numbered key findings with corresponding conclusions. Briefly summarize potential impact of your article in the scientific field like forest fire prediction.
References
Please, add approximately 50 new cited works used in reworked Introduction.
Comments on the Quality of English LanguageNo comments
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for Authorsforests-2575459 “Prediction of Forest Fire Occurrence in Southwest China” Xiaodong Jing, Donghui Zhang, Xusheng Li, Wanchang Zhang, Zhijie Zhang
This article describes the application of the LightGBM decision tree method on predicting fire occurrence with a dataset from China. It does a reasonable job of describing the methods and testing, but does not do as well differentiating between the model results and just describing the observed fire data. For example, describing where the fires are, or their seasonality, can be done by summarising the nearly 20,000 data points they have, whereas predicting the occurrence of fires in areas without data requires the model. There are numerous minor editing corrections needed, most obviously the inclusion or omission of spaces, before and after, commas “,” and periods “.”.
Several places have a discussion of a list of points (page 2 lines 50-60; pages 3+4 lines 121-147; page 7 lines 225-245), but they are all presented differently. The best way to do it is the list on page 7 with short paragraphs for each separate point.
On Page 6 line 201 the phrase “θ represents the orientation angle of the ellipse” is repeated twice.
In §2.3.4 please state specifically that TP, TN, FP and FN are the true positive, true negative, etc as this is the first time they are seen.
Figure 3 has two y-axes. Are all lines aligned with one, or are some with the other? Please indicate which axis each is measured on, or remove one if it is incorrect, or put them all on the same y-axis. If one or more of them has a different range then this is worth seeing visually.
Figure 4 lacks clarity as all the lines are very thin and have similar colours – several browns and two blue-greens. Please use colours that are more easily differentiated, with regard to colour-blind readers too, or use different line widths and/or dash patterns for different years.
It is a trivial observation that the authors have a scale that is 980km long in Figures 3, 4 and 7. For the sake of round numbers could this be changed to 1000km?
It may be an optical illusion, but it appears that the seasonal maps in Figure 7, which should be indicated as panels (b) through (e), are compressed laterally so the shape no longer matches the first panel. The options are to simply reduce the size of each image with constant aspect ratio, or to cut off part of the western provinces where there is little data and almost no fire risk. Given that Figure 7 is never actually referenced in the article, it is difficult to know what it is showing. Is this a model result that describes the predicted occurrence of fires, is it a probabilistic result showing how often each datum is a fire point in each season, is it another product? Please describe the figure in the text and figure caption.
There are far too many decimal places used to describe the accuracy statistics of the method (in Abstract and §3.3). Even as integers, the training, test and full sets are differentiable, but one decimal place is all that is required to adequately illustrate the differences.
§3.4 needs a complete title, such as “Results of Fire Zone mapping”. The wall of text from lines 322 to 347 is completely unnecessary, and useful only to someone with a very good grasp of the provinces, prefectures and municipalities of southern China (not marked on any maps in the article). The maps of modelled fire hazard tell the spatial story, while the previous standard deviation ellipses show changes over time. The rest of the text of §3.4 and the ellipses do a better job of talking around the fire hazard and occurrence.
Comments on the Quality of English Language
It is mainly punctuation that requires attention, rather than English language expression.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsDear Authors,
The submitted manuscript titled ”Prediction of Forest Fire Occurrence in Southwest China” contains interesting and valuable results which might interest an international audience. Nevertheless, I have found some imperfections-which in my opinion-should be improved or at east clarified before an eventual publication. I have listed them below:
1. In my opinion at the end of section Introduction, the aims of investigations should be listed. Moreover, the information about other investigations focused on prediction of forest fires in China (if they exist) would be very desired.
2. Lines 322-347 the description with recalling place names is difficult to read. Perhaps this part maight be written in more concise way.
3. Table 2-Please add the explanation of abbreviations in capation.
4. Figure 4 is a bit illegible. Perhaps change of colours into more flashy would be helpful.
5. I suggest to divide part Discussion and Conclusions into two separate chapters.
6. Please, look into below-listed publications. Perhaps, some of them would be help ful in manuscript corrections:
· Gao C, Lin HL, Hu HQ, Song H. A review of models of forest fire occurrence prediction in China. The Journal of Applied Ecology. 2020, 31(9):3227-3240. DOI https://doi.org/10.13287/j.1001-9332.202009.014.
· Pang, Y.; Li, Y.; Feng, Z.; Feng, Z.; Zhao, Z.; Chen, S.; Zhang, H. Forest Fire Occurrence Prediction in China Based on Machine Learning Methods. Remote Sens. 2022, 14, 5546. https://doi.org/10.3390/rs14215546
· Zhang, G., Wang, M. & Liu, K. Forest Fire Susceptibility Modeling Using a Convolutional Neural Network for Yunnan Province of China. Int J Disaster Risk Sci 10, 386–403 (2019). https://doi.org/10.1007/s13753-019-00233-1
· Yudong Li, Zhongke Feng, Shilin Chen, Ziyu Zhao, Fengge Wang, "Application of the Artificial Neural Network and Support Vector Machines in Forest Fire Prediction in the Guangxi Autonomous Region, China", Discrete Dynamics in Nature and Society, vol. 2020, Article ID 5612650, 2020. https://doi.org/10.1155/2020/5612650
· Paulo A. Martins Fernandes. Fire spread prediction in shrub fuels in Portugal, Forest Ecology and Management, 144 (1–3,), 2001, 67-74. https://doi.org/10.1016/S0378-1127(00)00363-7.
· Kajol R Singh, K.P. Neethu, K Madhurekaa, A Harita, Pushpa Mohan. Parallel SVM model for forest fire prediction. Soft Computing Letters,3, 2021, 100014. https://doi.org/10.1016/j.socl.2021.100014.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Comments and Suggestions for AuthorsDear authors!
Thank you for your responses!
I can accept all your reasons and explanation, but I strongly suggest to revise Discussion and Conclusion section.
At present time Conclusions is a first part of this section. It is totally unacceptable structure of this section.
Please, organize separate Conclusion section at the end of the article. Please move from the Discussion first paragraph with numbered conclusions to the Conclusion section. Also add brief description of impact of your research to forest fire prediction topic.
Comments on the Quality of English LanguageNo comments
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsDear Authors,
I do not have any further remarks for corrections of manuscript.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
