Comparative Analysis of Machine Learning-Based Predictive Models for Fine Dead Fuel Moisture of Subtropical Forest in China

Round 1

Reviewer 1 Report

This paper investigates a noteworthy issue that may be helpful in forest fire management and forest destruction prevention. This paper analyzes five machine learning techniques for forecasting the moisture content of fine dead surface. However, I have some comments that I would want the authors to consider, as well as some questions that I am awaiting responses to. They are as follows:

- Lines 13 and 14: How may the moisture content of a fine dead surface help in the early detection of forest fires?

- Include the name of the study area in the abstract.

- If the referring format in MDPI journals has not been changed, it seems that the references in the text do not follow the journal format.

- Please give a detailed description of the automated moisture content monitor. What steps were taken to prepare this device? Do you have a reference for this?

- In the conclusion part, try to address all the research questions.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

The manuscript by Hou et al. devoted to modelling of fine dead fuel moisture using models trained on an observations data. Authors compare prediction accuracy for five model and four combinations of input factors to reveal the best fuel moisture content prediction model. 

The structure of the article is clear and straight. The introduction is sufficient, explains the manuscript idea and describe existing knowledge gaps. The experimental part and approaches to model training and testing are well described and can be reproduced.

The result section should be expanded by adding figures of obtained time series of meteorological variables and moisture of dead fuel. As far I understand, only five observation plots were used for model training. Please characterize these sites according to meteorological conditions. Alco describe the weather conditions for the study year. Was it typical or dry according to long-term weather stations data record?

Atmospheric precipitations are the most important factor controlling soil moisture. Please discuss the impact of precipitation and mention why it was not taking into account into the model.

Investigation of the forest fires was one of the motivation of the present study. Please discuss the fire dangerous conditions according to the observation and modelling results. Does the obtained prediction accuracy was enough for assessment of dangerous fire conditions? What are the possible future application of the obtained model? What are the restrictions of the model?

Specific comments

Table 3 – replace Temperature by Air temperature, Humidity by Relative air humidity.

Table 5 – I recommend to move this table to the Supplementary.

Table 6 – I recommend to move this table to the Supplementary.

Author Response

Thank you very much for reviewing this manuscript on your busy schedule. Please find the detailed responses below and the corresponding corrections in track changes in the re-submitted files.

We are grateful to you for your approbation and constructive suggestions. Together with your comments, we have completed the revision accordingly, with changes tracked and responded on the one-to-one basis as bellow.

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

After reviewing the changes made to the manuscript, I confirm that all my suggested corrections have been included. The introduced modifications are adequate and meet the requirements for publication. On this basis, I recommend to accept the revised manuscript.