Predicting Fine Dead Fuel Load of Forest Floors Based on Image Euler Numbers

Round 1

Reviewer 1 Report

It is proposed to expand the discussion by comparing the results obtained by different authors from the issue in question in the European Mediterranean area, where the importance of fine dead forest fuel is very high.

Author Response

Yunlin Zhang

School of Biological Sciences, Guizhou Education University

Gaoxin St.115

Guiyang, Guizhou Province 550018

PR China

• mail: [email protected]

Dear Reviewer:

Thank you for reviewing the manuscript. All revisions have been made as required, the response are as follows:

Comments:

It is proposed to expand the discussion by comparing the results obtained by different authors from the issue in question in the European Mediterranean area, where the importance of fine dead forest fuel is very high.

Answer: Thank you for your advice, further discussion of fine dead forest fuel in the European Mediterranean area has been added.

Reviewer 2 Report

The work is well-prepared, but the introduction could be enhanced by using more current references. Furthermore, it would be interesting to add a location map of the study area to the methodology.

The main question addressed in the research is the possibility of incorporating image Euler numbers into predictive models of fine dead fuel load via the taking photo method. The work is quite interesting and has important relevance for this segment of study and as I said, the work is very interesting, but it is not original in the area of study.

The work presents an interesting possibility for studying predictive models of fine fuel load using photographs. The paper is well-written and the text is clear and easy to read. The conclusions are consistent with the evidence and arguments presented and the authors posed a main question.

Author Response

Yunlin Zhang

School of Biological Sciences, Guizhou Education University

Gaoxin St.115

Guiyang, Guizhou Province 550018

PR China

• mail: [email protected]

Dear Reviewer:

Thank you for reviewing the manuscript. All revisions have been made as required, the response are as follows:

Comments:

1. The work is well-prepared, but the introduction could be enhanced by using more current references. Furthermore, it would be interesting to add a location map of the study area to the methodology.

 Answer：References have been added to the introduction and location map of the study area has been added to the Material and Methods.

2. The main question addressed in the research is the possibility of incorporating image Euler numbers into predictive models of fine dead fuel load via the taking photo method. The work is quite interesting and has important relevance for this segment of study and as I said, the work is very interesting, but it is not original in the area of study.

Answer: Thank you for your advice, the inappropriate description has been revised.

3. The work presents an interesting possibility for studying predictive models of fine fuel load using photographs. The paper is well-written and the text is clear and easy to read. The conclusions are consistent with the evidence and arguments presented and the authors posed a main question.

Answer: Thank you for your comment. All modifications have been made as requested.

Reviewer 3 Report

See attachment.

Comments for author File: Comments.pdf

Author Response

Yunlin Zhang

School of Biological Sciences, Guizhou Education University

Gaoxin St.115

Guiyang, Guizhou Province 550018

PR China

• mail: [email protected]

Dear Reviewer:

Thank you for reviewing the manuscript. All revisions have been made as required, the response are as follows:

Comments:

The main concept of this paper, that is to use image analysis to automate the measurement of wildland surface fuels, is a good one. And because of that, I think this paper should be published after some improvements. In general, some sections and concepts in the paper need more description. Also, there are some errors in the writing that must be fixed.

Answer: Thank you for your comment. All revisions have been made as requested.

General comments:

1. I suggest changing all mention of the term “forest fires” to “wildland fires”, as I think these methods could apply to fires burning in other types of vegetation such as grass and shrubs.

Answer: Thank you for your advice, all “forest fires” have been converted to “wildland fires”.

2. Can the Euler number be used to determine other fuel metrics such as number and size of fuel elements? If so, this should be mentioned because it could be of use to future fire behavior models.

Answer: Thank you for your question. Euler number are currently mainly used for number statistics. Further analysis is needed on whether it can be used for fuel size discrimination. Perhaps other image features may also be used to determine the size of fuel materials.

The relevant description has been added to the next research topic under discussion.

3. An enhanced description of the Euler number calculation and interpretation should be given toward the beginning of the paper. This is needed to give the reader a background of this calculation to judge whether it might be useful and help identify advantages/disadvantages of its use.

Answer: Thank you for your advice, the description of the Euler number calculation and interpretation has been added to the manuscript.

4. It should be made clear that fuel was obtained from the field site and then brought to the lab for the analysis. Initially I thought fuel loading analysis was being done in the field and lab (meaning pictures were taken in the field and lab and then analyzed).

Answer: Thank you for your advice, relevant content has been added in the manuscript.

5. What if the fuel depth is large enough that a significant portion of the fuel elements are obscured in the image? At some point, as fuel load increases there would be no change to the Euler number (the image would only show fuel elements on the top surface) but loading is increasing. So the Euler number calculation along must eventually fail for high loadings. Some discussion of this should be included.

Answer: Thank you for your question. As you mentioned, if the loading is infinitely increased, using this method may fail or the performance may not be good. But for this study, the loading gradient is 3 t/ha~20.5 t/ha, fully covering the range of surface fuel load in the study area. In addition, forest stands with a surface fuel loading exceeding 20 t/ha are a minority, and are still suitable for most types of fuel load prediction.

Regarding the issue you mentioned, it will also be we next research focus, which is not only analyzing the relationship between fuel load and image features, but also including the relationship between depth and features, attempting to predict loading through depth and image features.

Relevant content has been added in the manuscript.

Specific comments:

6. Line 34: It would probably be more appropriate to describe fine dead fuels as the fuel most important to fire spread, not ignition (“igniters” in your paper could be interpreted as cause for the ignition of the fire, from a match or other source for example). Consider changing this first sentence to “Fine dead surface fuels are generally considered the most important fuel class in wildland fires. They are surface fuels with a diameter of less than 6 mm (Rothermel, 1972; Gould et al., 2011) that are flammable, can spread fast, have a high combustion efficiency, and possess a large surface-area-to-volume ratio.”

Answer: Thank you for your advice, it has been modified as required.

7. Line 27: “securing scientific forest fire management” is a strange sounding statement. Consider changing to something like “advancing science-based wildland fire management”.

Answer: Thank you for your advice, it has been modified as required.

8. Lines 115-123: Provide more description on how load was measured. What size were the 20 sampling points and what method was used to measure the load? Was all the fuel in a certain size area gathered and weighed? Something else?

Answer: Thank you for you advice, it has been added.

9. Line 128: Be consistent with naming conventions in the paper. Here you use the term “thickness” of the fuel bed, but in Table 1 you have a column labelled “Range of height”. Are these the same quantities? Also, I think the best and most used term for this is the “fuel bed depth”, and suggest this term be used.

Answer: Thank you for your advice, change the entire manuscript to fuel bed depth.

10. Line 131-132: I suggest changing “the grading of artificial 131 coniferous and broad-leaved bed layers was set at 0.5 t/ha increments” to something like “loading increments of 0.5 t/ha were investigated”.

Answer: Thank you for you advice, it has been revised as required.

11. Lines 136-137: What was the purpose of soaking the fuel in water?

Answer: Very sorry, there was a writing error here and it has been deleted.

12. Lines 146-147: Change “for each load along 146 the gradient.” to “for each loading.”

Answer: Thank you for you advice, it has been revised as required. 

13. Line 151: What is meant by “for the best storage effect”? Do you mean to reduce storage used? Are these JPEG images compressed in a lossy way (ie. is information lost during compression) and does this effect the analysis?

Answer: Thank you for your question, here is only an introduction to the experimental process and how the images are saved. It had no impact on the analysis.

14. Figure 1: It appears that some images are not uniformly covered by the fuel bed. For example the second image on the top row corresponding to 8 t/ha of Piinus massoniana has a very significant change in loading along the border compared to the center area. Will this negatively affect the analysis of this paper?

Answer: Thank you for you question. In order to better simulate the edge state of fuel bed in the field, this situation may occur. This will not affect the analysis, as when extracting the Euler number, we first binarized the image to black and white, which has no effect on the final extraction of Euler numbers with different loads.

15. Lines 159-162: There appears to inadvertently be a comment from possibly an earlier review of the paper here. The comment is a good one and should be incorporated into the paper properly.

Answer: Thank you for your advice, it has been added in the introduction.

16. Line 170: Change “a gradient” to “increments”.

Answer: Thank you for your advice, it has been revised as required.

17. Lines 177-179: Were these statistics computed for all the different images for each loading (so 3 images X 4 conversion thresholds = 12 images per loading)? Or just the 3 different images per loading per conversion threshold? Specify this in the paper.

Answer: Thank you for your advice. The description here is not detailed, which confuses you. It has been revised.

To analyze the impact of Euler numbers obtained from different thresholds on the results, this study a range of thresholds from 0.05 to 0.95, with increments of 0.05. Select different thresholds for each image to binarize and extract Euler numbers, as there are 19 threshold gradients, each image has 19 Euler numbers. Euler numbers were obtained from images binarized at different conversion thresholds via the ‘bweuler’ function in MATLAB R2023b.

Fig.2 is only an example given to demonstrate the effect of binarization, with thresholds of 0.05, 0.3, 0.6, and 0.9 for a loading of 9 t/ha, respectively. Only partial display, explanation has been added in the manuscript.

18. Line 214: There are missing variable names in this sentence.

Answer: Thank you for your advice, it has been added.

19. Figure 3: In the legend, probably “medium” should be “median”.

Answer: Thank you for your advice, it has been revised.

20. Figure 4: The caption needs much more description of what is being shown in this figure. I don’t understand how to read this. What do the elliptical shapes represent? How/where is loading read from? What are the values in some of the cells, and why are they only in some? There is also a Chinese word/symbols in the upper right that should be removed.

Answer: Thank you for your advice, the elliptical represents the correlation between load and Euler number under different thresholds, with blue indicating a negative correlation and red indicating a positive correlation. The darker the color, the stronger the correlation. The value in the cells represents the correlation coefficient.

It has been added in the manuscript and the Chinese word/symbols has been removed.

21. Lines 255-261: Why were only certain image threshold values used in the models, and why different ones for each fuel type? This should be described here.

Answer: Thank you for your question, because the correlation between Euler number and load varies at different thresholds, only certain image threshold values used in the models. The shape and structure of fine dead fuels on different surfaces are different, so the threshold corresponding to the Euler number selected by the model is also different.

Relevant content has been added to the discussion.

22. Figure 5: The functional forms used here seem to be very arbitrary and the data seems to have lots of scatter. Why were these functional forms used? What justifies this? Why not some other form? And why was a piecewise function used, with a transition at 11 t/ha? It is not clear from the data that this is valid.

Answer: Thank you for your question. It can be seen from the Fig.5 that when the load was 11 t/ha, it exhibited different trends with respect to the logarithm of the Euler number for the two fuel types, so with a transition at 11 t/ha. The model form is selected based on the variation of bed load with Euler number, selected an appropriate model form (with R² maximum as the selection criterion), and least squares fitting parameters were selected.

The amount of data in this study is sufficient, and the R², MAE, MRE, and RMSE in Table 3 also demonstrate that the predictive performance of the model is feasible.

Relevant content has been added to the manuscript.

23. Figure 7: The last sentence in the caption is incomplete

Answer: Thank you for your advice, it has been added.

Reviewer 4 Report

Dear Authors,

The subject of the paper is a current and very important topic since it is a contribution to the study of forest fires. In my opinion, the subject fits the scope of the Forests Journal.

The study presented investigated the viability of incorporating image Euler numbers into predictive models of fine dead fuel load via taking photo method.

The document is well structured and organized.

Main Remarks:

- Lines 134-136: “We constructed 30 × 30 cm fuel beds in the laboratory using the structurally intact leaves that were collected in the field, ensuring a thickness of 3.2 cm for coniferous and 4.0 cm for broad-leaved beds.”

In my opinion, the thickness of natural fuel bed varies with the fuel load. It is not clear why the thickness of the fuel bed was constant for each fuel type.

- Line 248: Figure 4. In this figure, load scale is not presented.

- Lines 365-367: “Both random forest and nonlinear fitting methods based on the Euler number produced a better prediction performance for P. massoniana leaf fuel than for that of Q. fabri, which can mainly be ascribed to the fuel characteristics.”

This section should be further developed, clarifying the limitations of this technique in some types of forest fuel beds or for high fuel loads. In the particular case of broad leaves, the leaves tend to overlap (as shown in Figure 1), hiding each other and this technique can be expected to produce very erroneous results.

Author Response

Yunlin Zhang

School of Biological Sciences, Guizhou Education University

Gaoxin St.115

Guiyang, Guizhou Province 550018

PR China

• mail: [email protected]

Dear Reviewer:

Thank you for reviewing the manuscript. All revisions have been made as required, the response are as follows:

Comments:

The subject of the paper is a current and very important topic since it is a contribution to the study of forest fires. In my opinion, the subject fits the scope of the Forests Journal.

The study presented investigated the viability of incorporating image Euler numbers into predictive models of fine dead fuel load via taking photo method.

The document is well structured and organized.

Main Remarks:

1. - Lines 134-136: “We constructed 30 × 30 cm fuel beds in the laboratory using the structurally intact leaves that were collected in the field, ensuring a thickness of 3.2 cm for coniferous and 4.0 cm for broad-leaved beds.”

In my opinion, the thickness of natural fuel bed varies with the fuel load. It is not clear why the thickness of the fuel bed was constant for each fuel type.

Answer: Thank you for your question. As you mentioned, the depth of the fuel bed in the field does indeed vary and is not stable. However, the main purpose of this study is to analyze the applicability of using image Euler numbers to estimate loading, and different depths will affect the analysis results. Therefore, this study only sets one bed depth (selecting the most representative depth: the average depth of the fuel bed in the study area). In furthermore research, the analysis of fuel bed depth gradient will also be added.

Relevant explanations also added to the manuscript.

2. - Line 248: Figure 4. In this figure, load scale is not presented.

Answer: Thank you for your advice. The meaning of load in Fig.4 is a variable, and the main purpose of Fig.4 is to analyze which threshold the extracted Euler number is more relevant to the load within the scope of this study, so the load range is not reflected

3. - Lines 365-367: “Both random forest and nonlinear fitting methods based on the Euler number produced a better prediction performance for P. massoniana leaf fuel than for that of Q. fabri, which can mainly be ascribed to the fuel characteristics.”

This section should be further developed, clarifying the limitations of this technique in some types of forest fuel beds or for high fuel loads. In the particular case of broad leaves, the leaves tend to overlap (as shown in Figure 1), hiding each other and this technique can be expected to produce very erroneous results.

Answer: Thank you for your advice, the limitations of different types of fuels have been further discussed.