Global Patterns and Dynamics of Burned Area and Burn Severity
Round 1
Reviewer 1 Report
Dear authors,
I have carefully reviewed your manuscript titled "Global patterns and dynamics of burned area and burn severity" submitted to Remote Sensing. I appreciate the valuable contribution of your study in characterizing the spatiotemporal patterns of burned area and severity and their relationships with climate variables. Overall, the study is informative and presents important findings in the field. However, I have identified some areas that require further attention and improvement. I have listed my comments and suggestions below:
1. Definition of burn severity and more references: The manuscript would benefit from a clear and concise definition of burn severity. Additionally, I recommend providing more references to support the understanding and background of burn severity as it relates to your study.
2. English: The language used in the manuscript needs improvement. I suggest a thorough proofreading to correct grammatical errors, sentence structure, and clarity of expression. It is important to ensure that the text is written in clear and concise English to facilitate understanding for readers.
3. Acronyms: Please ensure that all acronyms used in the manuscript are defined upon first mention, even if they are commonly known in the field. This will enhance the clarity and readability of the manuscript.
4. Figure clarity: The figures presented in the manuscript should be revised to improve clarity. Ensure that labels, legends, and axis titles are legible and properly explained. Consider using larger font sizes and high-resolution images to enhance readability.
5. Conclusion section: The manuscript lacks a dedicated conclusion section. I suggest including a concise summary of the main findings and their implications in a separate section to provide closure to the study.
6. Database validation: It is important to provide more evidence regarding the validation of the database used in your study, particularly in terms of its accuracy and representativeness of global forest fires. Consider discussing the strengths and limitations of the database, as well as any quality control measures implemented.
7. Concise writing: The manuscript would benefit from more concise writing. I recommend revisiting the text to eliminate redundancies and unnecessary repetitions, ensuring that the information is presented in a clear and succinct manner.
8. Provide a brief overview of the methodology used for analyzing burned area and severity trends to help readers better understand the approach.
9. Discuss the implications of your findings in the context of fire management and climate change mitigation strategies.
10. Consider including a discussion on the potential future research directions that could build upon your study and address any remaining gaps in understanding burned area and severity dynamics.
Sincerely,
Fair
Author Response
REVIEWER 1
General comments
Comment: Dear authors, I have carefully reviewed your manuscript titled "Global patterns and dynamics of burned area and burn severity" submitted to Remote Sensing. I appreciate the valuable contribution of your study in characterizing the spatiotemporal patterns of burned area and severity and their relationships with climate variables. Overall, the study is informative and presents important findings in the field. However, I have identified some areas that require further attention and improvement. I have listed my comments and suggestions below:
Answer: We thank Reviewer 1 for her/his careful review and positive general comment. We hope the improvements made to our manuscript are to your satisfaction as well as to the editor’s and Reviewers 2,3 and 4 satisfactions.
Specific comments
Comment: Definition of burn severity and more references: The manuscript would benefit from a clear and concise definition of burn severity. Additionally, I recommend providing more references to support the understanding and background of burn severity as it relates to your study.
Answer: a definition of burn severity is provided in L48-49 (tracked changes). We have also included more information on BS in methods (L113-116), stating that we have followed an initial assessment approach. We have also added 9 new references to the document.
Comment: English: The language used in the manuscript needs improvement. I suggest a thorough proofreading to correct grammatical errors, sentence structure, and clarity of expression. It is important to ensure that the text is written in clear and concise English to facilitate understanding for readers.
Answer: The English has been triple checked by the two authors and by a native speaker, we have corrected some minor issues that have been corrected. We hope the English is fully correct after the review and of high quality as Reviewer 4. We kindly ask you for a more specific comment if some specific issue was detected.
Comment: Please ensure that all acronyms used in the manuscript are defined upon first mention, even if they are commonly known in the field. This will enhance the clarity and readability of the manuscript.
Answer: The acronyms had been defined upon first mention (L38 and L37).
Comment: Figure clarity: The figures presented in the manuscript should be revised to improve clarity. Ensure that labels, legends, and axis titles are legible and properly explained. Consider using larger font sizes and high-resolution images to enhance readability.
Answer: we have re-sized and increased the resolution of the figures to improve legibility.
Comment: Conclusion section: The manuscript lacks a dedicated conclusion section. I suggest including a concise summary of the main findings and their implications in a separate section to provide closure to the study.
Answer: Good point! We have included a Conclusion section following your recommendation (L475-499).
Comment: Database validation: It is important to provide more evidence regarding the validation of the database used in your study, particularly in terms of its accuracy and representativeness of global forest fires. Consider discussing the strengths and limitations of the database, as well as any quality control measures implemented.
Answer: the database has been published and tested in one of the most prestigious journals for global databases (Earth System Science Data). The database has a proven relationship with dNBR calculated using Landsat, which is the imagery traditionally used to develop spatial models of burn severity. According to the reviewer’s concern, we have provided the goodness-of-fit between MOSEV and Landsat-derived dNBR in the methods section (L117-119).
Comment: Concise writing: The manuscript would benefit from more concise writing. I recommend revisiting the text to eliminate redundancies and unnecessary repetitions, ensuring that the information is presented in a clear and succinct manner.
Answer: we have revised the document to improve conciseness. We hope the main body of the current version (5870 words including figs and tables) is concise enough.
Comment: Provide a brief overview of the methodology used for analyzing burned area and severity trends to help readers better understand the approach.
Answer: according to this comment, and to Reviewer’s3 comment, we have provided details on the methodology and robustness of the method used to test BA and BS trends (L173-176).
Comment: Discuss the implications of your findings in the context of fire management and climate change mitigation strategies.
Answer: According to the reviewer’s comment we have included a statement in the conclusions (L496-498). We have not expanded the discussion to conciseness and to not distract from our objectives. We think that a meaningful and accurate discussion on these two aspects should be done region-by-region and could be object of a different work.
Comment: Consider including a discussion on the potential future research directions that could build upon your study and address any remaining gaps in understanding burned area and severity dynamics.
Answer: we have expanded the discussion according to this comment. We consider that future research might address the analysis of BA and BS at finer spatial units, which could provide further insights on fire-climate relationships. Moreover, as new satellite imagery and computational capacities are available, we recommend future analysis to analyze BA and BS trends using higher spatial resolution imagery or statistically refined data, as BA is largely underestimated by MODIS, particularly in areas where small fires dominate. In addition, new methods for burn severity quantification could be employed and contrasted with our findings.
These potential future research has been highlighted in the discussion (L449-471) and concisely in the conclusions (L493-496).
Reviewer 2 Report
1. The burned area and burn severity data come from the MOSEV database, which provide monthly global data at 500 m spatial resolution from 2000 to 2019. However, all analysis in the manuscript was done based on regional scale. This is a bit of waste of the spatial resolution of this dataset. I strongly suggest to add some results at pixel scale.
2. On the basis of BA and BS analysis at the reginal scale, the BA and BS seem to be nonbusiness with the type of land cover. Therefore, all results in BA and BS were done in different climatic zones. It is recommended to supplement the climatic zoning map.
3. Figures in this manuscript are too blurry to see clearly. Please add high resolution figures.
Author Response
Specific comments
Comment: The burned area and burn severity data come from the MOSEV database, which provide monthly global data at 500 m spatial resolution from 2000 to 2019. However, all analysis in the manuscript was done based on regional scale. This is a bit of waste of the spatial resolution of this dataset. I strongly suggest to add some results at pixel scale.
Answer: In view of this comment and on the coarse spatial scale analysis of our study, we have recommended in the discussion and in the conclusions the use finer analytical units in future regional analysis as well as in analysis more focused on disentangling fire-climate relationships (L470, L494 – Tracked changes).
We acknowledge that the analysis might be done at finer scales, but it is not possible to do the analysis at the pixel scale. The reason is that on most of the planet, one single pixel has never burned, burned only once, or twice between 2003 and 2019, making impossible any trend analysis at the pixel level. By this reason the analysis must be aggregated by analytical units larger than pixels or by regions.
In this first global analysis of burn severity, we have decided to provide global trends as well as trends aggregated by regions. We consider that our 64 regions are still a manageable number to present and visualize the results on a global scale, but more regions would make results difficult to follow particularly if presented in Figs. (Fig.1) or tables as (Tables A1, A2).
Comment: On the basis of BA and BS analysis at the reginal scale, the BA and BS seem to be nonbusiness with the type of land cover. Therefore, all results in BA and BS were done in different climatic zones. It is recommended to supplement the climatic zoning map.
Answer: We have provided the climatic characteristics (mean values and trends) of each region in Fig. A3. The climatic values can be also seen in the main document in Figs. 5 and 7 (X axis).
Comment: Figures in this manuscript are too blurry to see clearly. Please add high resolution figures.
Answer: We have replaced the figures in the previous version with high resolution figures.
Reviewer 3 Report
The manuscript entitled “Global patterns and dynamics of burned area and burn severity” submitted to Remote Sensing offers a global characterization of the spatiotemporal patterns of burned area and burn severity, analyzing the trends in different biomes and considering the influence of climate. It offers an interesting view of these important variables, with results that question some of the assumed ideas about forest fires like the global increase of burned area and burn severity or the strong influence attributed to climate change as the main driver of changes in fire regime.
The manuscript is well written, the objectives are clear, the methods are correct and well explained and the results are adequately described and discussed, so I strongly recommend the publication of this research. I have just two suggestions to make to this research, that could be easily introduced. First, in section 2 the authors should state clearly if they are analyzing data of short- or long-term burn severity. The second issue is related to dNBR intervals. The official threshold for high severity is a dNBR of 660 and they are using the threshold value of the moderate-high burn severity. They should indicate the reason for these three and not four intervals and for considering the value of 440 as high severity, since it is significantly lower than the value usually used in the research focused on analyzing trends in high severity. Maybe the authors should also include some considerations related to the problems of these “official” dNBR thresholds, since many researchers have demonstrated that these thresholds are not adequate for all the different ecosystems.
Author Response
General comment
Comment: The manuscript entitled “Global patterns and dynamics of burned area and burn severity” submitted to Remote Sensing offers a global characterization of the spatiotemporal patterns of burned area and burn severity, analyzing the trends in different biomes and considering the influence of climate. It offers an interesting view of these important variables, with results that question some of the assumed ideas about forest fires like the global increase of burned area and burn severity or the strong influence attributed to climate change as the main driver of changes in fire regime.
The manuscript is well written, the objectives are clear, the methods are correct and well explained and the results are adequately described and discussed, so I strongly recommend the publication of this research.
Answer: Thank you very much for your effort in reviewing our manuscript, constructive comments and for your recommendation of acceptance. We hope the improvements made are to your satisfaction.
Specific comments
Comment: I have just two suggestions to make to this research, that could be easily introduced. First, in section 2 the authors should state clearly if they are analyzing data of short- or long-term burn severity.
Answer: Good point! We have provided the requested information (L113-117) to clarify our approach, as the two approaches (initial and extended assessments) are common in the literature. In our study we focus on initial (short-term) assessments of burn severity based on the MOSEV algorithm.
Comment: The second issue is related to dNBR intervals. The official threshold for high severity is a dNBR of 660 and they are using the threshold value of the moderate-high burn severity. They should indicate the reason for these three and not four intervals and for considering the value of 440 as high severity, since it is significantly lower than the value usually used in the research focused on analyzing trends in high severity. Maybe the authors should also include some considerations related to the problems of these “official” dNBR thresholds, since many researchers have demonstrated that these thresholds are not adequate for all the different ecosystems.
Answer: We further explained the thresholding possibilities in the methods section (L138-149).
We have indicated that the number of BS levels and threshold values depend on the user's objectives as stated by Key & Benson and other authors afterwards. Moreover, we have explained the two main reasons for selecting the threshold of 440 instead of 660 (the value of 660 has been found excessive in some studies, and more important, values above 660 are scarce in MOSEV. See for instance L179-180)
We have also explained the rationale of selecting the same thresholds in L147-150. It is not only because the lack of adapted thresholds for every region, but also for consistency as in that way they indicate the same overall spectral change. We hope this explanation for categorical BS data among with our analysis using continuous BS are of your satisfaction.
Reviewer 4 Report
Thanks to the authors for very interesting work. It is of interest in connection with the global assessment of flammability around the world, not only by single region. Still I have a few questions and comments.
The work seems to use data for the entire territory of ecoregions, including infrastructure, urban areas, crop fields and unvegetated areas. This can introduce a significant error in the calculation of the ratio BA and BS – many spring fires are caused by fields ignited by farmers before sowing. The same is for fires in urban area – usually they are not connected with climate.
Given the many works showing an increase in the burned area and number of fires, it is somewhat strange to see a statement about a trend towards a decrease by 1.5% in the total area.
For example, in Alaska an increase of 9% is stated
(Calef, M.P., A. Varvak, A.D. McGuire, F.S. Chapin III., and K.B. Reinhold. 2015. Recent changes in annual area burned in interior Alaska: The impact of fire management. Earth Interactions 19:1–16. https://doi.org/10.1175/EI-D-14-0025.1.)
and in Asian Arctic burned area increased almost 2.5 times for last decade comparing previous decade. (Kharuk, V.I.; Dvinskaya, M.L.; Im, S.T.; Golyukov, A.S.; Smith, K.T. Wildfires in the Siberian Arctic. Fire 2022, 5, 106. https://doi.org/10.3390/fire5040106 )
Figure 1 is hard to read. Too small font-size and upside-down letters are not good for understanding.
Figures 5,6,7 are out of focus
In addition:
Have you checked the rows for "outliers"? Single anomalous year can significantly affect the analysis, up to the complete removal the trend. For example, in the above mentioned article Kharuk et al., there may be no trend in BA due to values in 2000 but an increase is significant.
Author Response
General comment
Comment: thanks to the authors for very interesting work. It is of interest in connection with the global assessment of flammability around the world, not only by single region. Still I have a few questions and comments.
Answer: thank you for your effort and constructive comments. We hope our response and the improvements made to our manuscript are of your satisfaction.
Specific comments
Comment: The work seems to use data for the entire territory of ecoregions, including infrastructure, urban areas, crop fields and unvegetated areas. This can introduce a significant error in the calculation of the ratio BA and BS – many spring fires are caused by fields ignited by farmers before sowing. The same is for fires in urban area – usually they are not connected with climate.
Answer: We agree with the reviewer comment, that land cover as well as many other factors (density of roads, cattle density, population density etc.) have a demonstrated effect on burned area and eventually on BS. To point out the reviewer concern, and considering that the analysis of the drivers of BA and BS falls out of the scope of our work, we have encouraged future works to further disentangle the drivers of BA and BS (L462-463, tracked changes).
Comment: Given the many works showing an increase in the burned area and number of fires, it is somewhat strange to see a statement about a trend towards a decrease by 1.5% in the total area. For example, in Alaska an increase of 9% is stated (Calef, M.P., A. Varvak, A.D. McGuire, F.S. Chapin III., and K.B. Reinhold. 2015. Recent changes in annual area burned in interior Alaska: The impact of fire management. Earth Interactions 19:1–16. https://doi.org/10.1175/EI-D-14-0025.1.)
and in Asian Arctic burned area increased almost 2.5 times for last decade comparing previous decade. (Kharuk, V.I.; Dvinskaya, M.L.; Im, S.T.; Golyukov, A.S.; Smith, K.T. Wildfires in the Siberian Arctic. Fire 2022, 5, 106. https://doi.org/10.3390/fire5040106 )
Answer: As the reviewer states, some regions might have experienced increases in burned area. Our results also show differences among regions (Fig. 3), and for Siberia (Asian part) our results are in the line with those found by Kharuk et al. (2022). As our discussion is concise and mainly focused on significant trends, we have included the trends in BA found BY Kharuk et al. (2022) in the introduction.
In spite of this, the significant decrease in global BA last decades is solid evidence (See some of the references provided such as Andela et al., 2017 published in Science). Our work complements this evidence by expanding the period of analysis. We think this is very important as many works wrongly state the opposite as an easy way to give importance to their studies generating confusion in the scientific community– See also Doerr & Santin, 2016 published in Proc. Royal Soc. B).
Comment: Figure 1 is hard to read. Too small font-size and upside-down letters are not good for understanding. Figures 5,6,7 are out of focus
Answer: We have increased the resolution and size of the figures to increase readability.
Comment: In addition: Have you checked the rows for "outliers"? Single anomalous year can significantly affect the analysis, up to the complete removal the trend. For example, in the above mentioned article Kharuk et al., there may be no trend in BA due to values in 2000 but an increase is significant.
Answer: We think it is not appropriate to remove outliers in our case (even would be tricky) all the data has been checked and values are not affected by wrong observations/wrong data. Also, the existence of exceptional fire years it is normal, and those must be also included in analyses.
In any case, we have used a robust statistical analysis (Theil-Sen and Mann-Kendall) which is insensitive to outliers. You can check the statistical basis of Theil-Sen and Kendall here https://en.wikipedia.org/wiki/Theil%E2%80%93Sen_estimator or in any other source.
Round 2
Reviewer 1 Report
Accept in present form
Fair
Reviewer 2 Report
Authors have revised the manuscript according to reviewers' comments.
