Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Severity-driven assessment of greenhouse gas emissions from large Mediterranean wildfires using remote sensing and vegetation mosaics

 

by Helena van den Berg Sesma, Edgar Lorenzo Sáez, Victoria Lerma-Arce, Jose-Vicente Oliver-Villanueva and Mauricio Acuna

 

In this study, the authors estimated greenhouse gas emissions from the large Bejís wildfire in eastern Spain using remote data on burn-severity. The authors use Sentinel-2-derived NBR to estimate burn severity and consumed biomass in Mediterranean landscapes. The authors found that the total emission estimate from the Bejís fire was 766165 t CO₂eq with conifers and shrubland–conifer mixtures being responsible for the highest CO₂eq emissions. I believe the manuscript is suitable for the Fire journal, and I have only minor suggestions.

 

Lines 69 – 70: “Most carbon is emitted as CO₂, CO, N₂O, and CH₄.” The N₂O molecule does not contain carbon.

Lines 165 – 166: How the composite was created? Was the average value, maximum value composite, or some other method used?

Line 194: “The values were adapted…”. Perhaps it would be worth clarifying how exactly the coefficients were adapted. Were they interpolated or taken from another source?

Table 3. The table header states that it contains percentage values, but it appears that actual values are fractions.

Table 4. Emission factors for CO₂ and CH₄ are almost the same (1.377 and 1.4). Maybe EF should be 1377 for CO₂?

Table 7. Perhaps a heading should be added to the table explicitly stating that the columns correspond to severity classes.

Figure 4. Please, consider increasing legend font.

Tables 9 and 10: It is not clear what the color highlighting means.

Table 10. Possible rounding error in Table 10. According to Table 5 and Eq. 3, the CO₂eq value in Table 10 should be 728597 + 28 *739.416 + 265 * 60.446 = 765318.838 instead of 766165. Perhaps the authors should double-check the data in the table.

 

Author Response

We sincerely thank Reviewer 1 for the careful and detailed evaluation of our manuscript. The reviewer’s comments have been extremely helpful in improving the precision of terminology, correcting minor inconsistencies, and enhancing the clarity of methodological explanations and tables.

All suggested corrections have been implemented, including:

• Clarification of emission species terminology

• Specification of the Sentinel-2 compositing procedure

• Clarification of the biomass coefficient adaptation

• Corrections to tables and rounding inconsistencies

• Improvements in figure readability and formatting

We are grateful for these observations, which have significantly improved the technical accuracy of the manuscript. Our detailed responses to each individual comment are provided below.

Reviewer 1 – Comment #1

Lines 69–70:
“Most carbon is emitted as CO₂, CO, N₂O, and CH₄.” The N₂O molecule does not contain carbon.

Authors’ response:

We thank the reviewer for pointing out this imprecision. The sentence has been corrected accordingly. The revised text now reads:

“Forests that once acted as carbon sinks can become substantial sources of greenhouse gases (GHGs) during large fires, releasing gases such as CO₂, CH₄, and N₂O, which contribute to global warming and air pollution.”

Reviewer 1 – Comment #2

Lines 165–166:
How was the composite created? Was the average value, maximum value composite, or some other method used?

Authors’ response:

We thank the reviewer for this important clarification request. The compositing procedure has now been explicitly described in the Methods section.

The fire severity methodology section has been updated following the work of Quintero et al. (2025), which provides a clear framework for designing a compositing approach. The revised Methods section now includes a detailed description of the procedure used. Specifically, a mean value composite was generated from the selected Sentinel-2 images.

Reviewer 1 – Comment #3

Line 194:
“The values were adapted…”. Perhaps it would be worth clarifying how exactly the coefficients were adapted. Were they interpolated or taken from another source?

Authors’ response:

We appreciate the reviewer’s suggestion. The manuscript now clarifies how biomass consumption coefficients were adapted. The following sentence has been added:

“Biomass consumption coefficients (Table 5) were aligned with the European Forest Fire Information System burn severity classification. For intermediate severity classes, coefficients for each vegetation type were derived through linear interpolation between adjacent severity levels.”

Reviewer 1 – Comment #4

Table 3:
The table header states that it contains percentage values, but it appears that the values are fractions.

Authors’ response:

Thank you for this observation. The header of Table 3 (now Table 5) has been corrected and now reads:

“BEadj fractions (0–1) by burn severity class.”

Reviewer 1 – Comment #5

Table 4:
Emission factors for CO₂ and CH₄ are almost the same (1.377 and 1.4). Maybe EF should be 1377 for CO₂?

Authors’ response:

We thank the reviewer for identifying this inconsistency. The emission factor has been carefully checked and corrected in Table 4. The corrected value now reads:

“CO₂: 1,377 g·kg⁻¹ dry biomass.”

Reviewer 1 – Comment #6

Table 7:
Perhaps a heading should be added to the table explicitly stating that the columns correspond to severity classes.

Authors’ response:

We agree with the reviewer’s suggestion. The heading of Table 7 (now Table 9) has been revised so that each column explicitly indicates the corresponding severity class. Additionally, the values have been converted from tons per hectare to total consumed biomass (tons) to improve clarity and interpretability.

Reviewer 1 – Comment #7

Figure 4:
Please consider increasing the legend font.

Authors’ response:

Thank you for this helpful suggestion. The legend font size has been increased to improve readability. In addition, Figure 4 has been revised to better align with the recommendations provided by the other reviewers.

Reviewer 1 – Comment #8

Tables 9 and 10:
It is not clear what the color highlighting means.

Authors’ response:

We appreciate this comment. To improve clarity and interpretability, these tables have been replaced with Figures 5, 6, and 7, which more clearly illustrate the distribution of emissions by vegetation type and burn severity level.

Reviewer 1 – Comment #9

Table 10:
Possible rounding error. According to Table 5 and Eq. 3, the CO₂eq value should be 765318.838 instead of 766165.

Authors’ response:

We sincerely thank the reviewer for carefully verifying the calculation. Upon rechecking the computations, we identified a calculation error. The values have now been recalculated and corrected in the revised manuscript. All totals have been carefully verified to ensure consistency across the tables and equations.

Reviewer 2 Report

Comments and Suggestions for Authors

The paper by van den Berg Sesma et al. “Severity-driven assessment of greenhouse gas emissions ...” proposes an approach to derive greenhouse gases emissions carried out in a pilot case study of a large wildfire occurred during summer 2022 in the Valencian Region, Spain.

Study area interested by this extreme fire event represents a typical Mediterranean-type ecosystem, where ecosystem landscape is characterized by a mosaic of vegetation cover with heterogeneous composition of vegetation species and fuel loads.

Combining remotely sensed derived fire severity with biomass and land-cover data allows to better understand how heterogeneous vegetation mosaics and fire severity influence greenhouse gases emissions supporting more informed forest management and carbon accounting.

The proposed approach of deriving greenhouse gases emissions at the appropriated ecosystems scale is interesting, and the manuscript deserves to be published.

However, it is recommended to insert in the Abstract a sentence highlighting uncertainties and limitations still present in the proposed framework concerning with the use of a set of pre-defined parameters, as the Authors recognize in the “Discussions” and  “Conclusions”.

Minor remarks

Line 56.   (MTEs): this acronym is mentioned first time, please make explicit

Line 158  “a composite approach was applied” please clearly indicate which rule was used for compositing multiple Sentinel-2 images.

Moreover could you specify which Sentinel-2 data were used (Level-2A i.e. orthorectified atmospherically corrected surface reflectance) ?

How many Sentinel-2 images were used for the pre- and post-fire composite

- Pre-fire composite 15 June – 15 Sept 2021 and 15 June – 15 Aug 2022

- Post-fire composite 16 Aug – 15 Sept 2022

For pre-fire composite two years data (2021 and 2022) were used. Please briefly justify this choice

Line 517- 518  and Line 537: In these two References, please use the right format (not all Capital letters)

[6] R. MADONDO, N. MUTINGWENDE, S. SHWABABA, R. J. BAYNE, Á. RESTÁS, and R. TANDLICH, “ANALYSES OF TRENDS IN THE FIRE LOSSES AND THE FIRE-BRIGADE CALL-OUTS IN SOUTH AFRICA BETWEEN 2004 AND 2017,”

[15] C. STRECK and S. M. SCHOLZ 

Author Response

We would like to sincerely thank Reviewer 2 for the positive assessment of our work and for recognizing the relevance of the proposed framework for wildfire emission assessment in Mediterranean-type ecosystems.

We appreciate the reviewer’s constructive suggestions aimed at improving clarity and transparency. Following these recommendations, we have:

• Added a sentence in the Abstract explicitly acknowledging methodological uncertainties;

• Clarified the Sentinel-2 compositing methodology and data characteristics;

• Expanded acronyms at their first occurrence;

• Corrected formatting inconsistencies in the reference list.

These modifications have strengthened the methodological transparency and improved the overall coherence of the manuscript. Below we provide our point-by-point responses and describe the corresponding revisions introduced in the manuscript.

Reviewer 2 – Comment #1

The paper by van den Berg Sesma et al., “Severity-driven assessment of greenhouse gas emissions ...”, proposes an approach to derive greenhouse gas emissions applied in a pilot case study of a large wildfire that occurred during summer 2022 in the Valencian Region, Spain.
The study area affected by this extreme fire event represents a typical Mediterranean-type ecosystem, where the landscape is characterized by a mosaic of vegetation cover with heterogeneous species composition and fuel loads.

Combining remotely sensed fire severity with biomass and land-cover data allows a better understanding of how heterogeneous vegetation mosaics and fire severity influence greenhouse gas emissions, supporting more informed forest management and carbon accounting.

The proposed approach for deriving greenhouse gas emissions at the appropriate ecosystem scale is interesting, and the manuscript deserves to be published.

However, it is recommended to insert in the Abstract a sentence highlighting uncertainties and limitations still present in the proposed framework concerning the use of a set of predefined parameters, as the Authors recognize in the Discussion and Conclusions.

Authors’ response:

We sincerely thank the reviewer for the positive evaluation of our work and for this valuable suggestion. In response, we have incorporated an explicit statement in the final sentence of the Abstract acknowledging the methodological uncertainties associated with the use of predefined parameters.

The Abstract now concludes with the following sentence:

“This severity-driven, vegetation-explicit framework demonstrates robust potential for quantifying wildfire emissions across heterogeneous Mediterranean landscapes; however, uncertainties remain due to predefined biomass estimates, burning efficiency parameters, emission factors, assumptions in fire severity mapping, and limited field validation.”

Minor Remarks

Reviewer 2 – Comment #2

Line 56: (MTEs) – this acronym is mentioned for the first time; please make it explicit.

Authors’ response:

We thank the reviewer for this helpful remark. The acronym is now defined at its first occurrence in the manuscript (line 36).

Reviewer 2 – Comment #3

Line 158: “a composite approach was applied” – please clearly indicate which rule was used for compositing multiple Sentinel-2 images.

Moreover, could you specify which Sentinel-2 data were used (Level-2A, i.e., orthorectified atmospherically corrected surface reflectance)?

How many Sentinel-2 images were used for the pre- and post-fire composites?

• Pre-fire composite: 15 June – 15 Sept 2021 and 15 June – 15 Aug 2022

• Post-fire composite: 16 Aug – 15 Sept 2022

For the pre-fire composite, two years of data (2021 and 2022) were used. Please briefly justify this choice.

Authors’ response:

We appreciate the reviewer’s request for further clarification. The compositing methodology has now been described in detail in the Methods section, including the Sentinel-2 product level, compositing rule, number of images used, and temporal selection criteria.

The fire severity methodology section has been revised following the framework proposed by Quintero et al. (2025), which provides a clear approach for designing compositing strategies in fire severity analysis. The revised text now specifies that a mean-value composite was used.

Additionally, the manuscript now clarifies the Sentinel-2 data characteristics and explains the rationale for combining pre-fire imagery from 2021 and 2022, which was done to increase data availability and reduce the influence of cloud cover and seasonal variability.

Reviewer 2 – Comment #4

Lines 517–518 and Line 537:
In these two references, please use the correct format (not all capital letters).

Example:
[6] R. MADONDO, N. MUTINGWENDE, S. SHWABABA, R. J. BAYNE, Á. RESTÁS, and R. TANDLICH, “ANALYSES OF TRENDS IN THE FIRE LOSSES AND THE FIRE-BRIGADE CALL-OUTS IN SOUTH AFRICA BETWEEN 2004 AND 2017,”
[15] C. STRECK and S. M. SCHOLZ

Authors’ response:

We thank the reviewer for identifying this formatting issue. The referenced citations have been corrected and are now presented using the appropriate capitalization format in accordance with the journal’s reference style guidelines.

Reviewer 3 Report

Comments and Suggestions for Authors

The manuscript explained the technology of direct fire emission estimating, which integrates burn severity, vegetation composition, and standardized emission factors. It could be used as provides a methodological basis for integrating wildfire emissions into regional and national inventories.
In my opinion, the subject matter is of considerable interest, as it contributes to the investigation of fire emissions monitoring and to the technique of emission evaluating.
The manuscript will be suitable for publication in Fire after minor revisions.

A few issues need to be clarified by the authors, such as the following:
1. An important feature in the context of this work is detail of the vegetation distribution. Nevertheless, the authors provide a map that is exceedingly coarse (Figure 4a) in comparison to the fire severity detail (Figure 3). How much does this fact affect the accuracy of the other parameter estimates and lower the method's performance?
2. Should the impact of weather conditions be considered? Would the weather during a specific fire season also influence the parameters used, such as biomass consumption rates (Table 3)?
3. Table 6 and Line 220: “Pre-fire biomass: Values were derived from the SIOSE 2014 dataset”. The methodology by which the biomass loads data for the fire site was acquired is not entirely clear to me, as it contradicts the data in Appendix 1. Please, clarify. In particular, why did the load estimate per hectare go down in Table 6?
4. How can we interpret the extreme values in Tables 7 (86.73, 102.76, etc.) and 8 (162.56, etc.), which surpass the pre-fire biomass reserve values? Please, clarify in the text.
5. Line 421–242: “Fire severity, quantified using Sentinel-2 dNBR composites, proved a decisive predictor: both mean and median CO₂eq increased progressively from low to high severity classes, with high-severity areas concentrating extreme emission values.” Nonetheless, this is a consequence of the calculation method (see Table 3) and is not a conclusion? Please, clarify this issue.
6. Line 394, 438–444: “However, some uncertainty is inherent in the estimation of wildfire emissions.” It would be crucial to not only compile a list of potential uncertainties, but also to offer a qualitative evaluation, in my opinion. Compared to other parameters, fire severity estimates are probably the most accurate data for these kind of issues. Please, discus this issue.

Comments for author File: Comments.pdf

Author Response

We sincerely thank Reviewer 3 for the constructive evaluation of our manuscript. We particularly appreciate the recognition of the methodological contribution of integrating burn severity, vegetation composition, and emission factors for regional-scale wildfire emission assessment.

The reviewer raised several important conceptual and methodological points that have allowed us to substantially refine the Discussion section. In response, we have:

• Clarified the implications of vegetation map resolution and explicitly acknowledged this as a methodological limitation;

• Expanded the discussion on the potential influence of meteorological conditions on biomass consumption rates;

• Clarified biomass data derivation and interpretation of extreme values;

• Refined the interpretation of severity-dependent emission patterns to avoid presenting structurally embedded relationships as independent findings;

• Introduced a quantitative uncertainty through a sensitivity analysis and discussed their relative influence on emission estimates;

• We strengthened the conceptual discussion of uncertainty propagation and identified directions for future methodological improvements.

We are grateful for these insightful comments, which have significantly improved the scientific depth and clarity of the manuscript. A detailed response to each comment raised by the reviewer is presented below, together with the corresponding revisions implemented in the manuscript.

Reviewer 3 – Comment #1

Issue: The vegetation map provided (Figure 4a) is coarser than the fire severity map (Figure 3). How does this difference affect the accuracy of emission estimates and the method’s performance?

Authors’ response:

We thank the reviewer for this insightful comment. We fully agree that differences in spatial resolution between the vegetation dataset and the Sentinel-2 burn severity map may influence local-scale emission estimates. The SIOSE dataset was selected due to its regional consistency, institutional validation, and compatibility with official land cover inventories, making it appropriate for landscape-level analysis.

We acknowledge that a formal sensitivity analysis would be required to quantify the magnitude of this effect. While such analysis falls beyond the scope of the present study, we recognize it as a valuable direction for future research. Accordingly, the following text has been added to the Discussion:

“However, the spatial resolution of SIOSE remains coarser than that of the Sentinel-2 burn severity mapping, which may introduce local-scale uncertainty in emission estimates. Evaluating this uncertainty would require field measurements or higher-resolution biomass data, and technologies such as LiDAR could help address this limitation.”

Reviewer 3 – Comment #2

Issue: Should meteorological conditions be considered? Would the weather during a specific fire season influence parameters such as biomass consumption rates?

Authors’ response:

We thank the reviewer for this highly relevant comment. Meteorological conditions such as drought intensity, heatwaves, and fuel moisture can influence combustion completeness and biomass consumption rates. Operational wildfire emission assessments often rely on standardized coefficients rather than meteorology-dependent factors. However, emerging research indicates that extreme fire weather can affect combustion efficiency.

We have added the following text to the Discussion section:

“Combustion completeness may vary under extreme meteorological conditions, such as severe drought, local heatwaves, or low fuel moisture levels, which can influence fire behavior and fuel consumption dynamics [46]. Incorporating meteorology-dependent consumption coefficients could improve event-specific emission estimates. This requires detailed empirical datasets or dynamic fire behavior models linking meteorological variables to combustion efficiency. Such refinements fall beyond the scope of this study but constitute an important direction for future research.”

Reviewer 3 – Comment #3

Issue: Clarify biomass data derivation. Why do per-hectare biomass estimates in Table 6 differ from Appendix 1?

Authors’ response:

We thank the reviewer for this request. Initially, per-hectare biomass values were calculated as a weighted average based on vegetation type and the proportion of each polygon occupied by that type in the SIOSE dataset. Upon further consideration, we recognized that this approach could obscure the true spatial distribution of biomass across the landscape.

We have therefore revised the methodology: now biomass values are kept as absolutes and the areas are weighted by the percentage-cover of each vegetation type, and the analysis focuses on total biomass and total emissions rather than per-hectare values, which are less meaningful for this purpose. For transparency, the per-hectare biomass values for each vegetation type (former Table 6) have been recalculated and added in Appendix 1.

Reviewer 3 – Comment #4

Issue: How should extreme values in Tables 7 (86.73, 102.76) and 8 (162.56, etc.) be interpreted, as they surpass pre-fire biomass?

Authors’ response:

We appreciate this request for clarification. To prevent misinterpretation, we added the following sentence in the Results section. Extreme values reported in Tables 7 and 8 refered to CO₂eq emissions per hectare and do not imply biomass consumption exceeding pre-fire biomass. However, now the results focus on totals instead of per hectare values. 

Reviewer 3 – Comment #5

Issue: The progressive increase in CO₂eq with severity is structurally embedded in the methodology. Should it be presented as a conclusion?

Authors’ response:

We thank the reviewer for this important clarification. We agree that the increase in emissions with burn severity is methodologically embedded, as biomass consumption coefficients are severity-dependent. We revised the Discussion to emphasize that the study’s main contribution lies in the spatial quantification and magnitude of emissions across heterogeneous vegetation mosaics.

The Conclusions section was updated to focus on the interaction of vegetation types and burn severity on emissions. Mentions of simple conclusions, such as conifers being the main source or the increase of emissions by severity, were either removed or explicitly noted as methodological consequences.

Reviewer 3 – Comment #6

Issue: Discuss uncertainties qualitatively and compare their relative influence on emission estimates.

Authors’ response:

We thank the reviewer for this valuable suggestion. The Discussion now provides a sensitivity analysis of uncertainty sources, emphasizing the relatively high reliability of fire severity mapping and the comparatively greater uncertainty associated with biomass, consumption coefficients and emission factors.

While a formal sensitivity or Monte Carlo analysis was not conducted, we included three exploratory assessments to illustrate uncertainty propagation:

1. A -10% perturbation of input variables to evaluate the effect on resulting emissions;

2. Analysis under extreme theoretical fire severity scenarios;

3. Comparison with global products such as GFAS and GFED.

These analyses have been incorporated in the Methods, Results, and Discussion sections (lines 508–532) to provide a clearer conceptual understanding of uncertainty and guide future methodological improvements.

Reviewer 4 Report

Comments and Suggestions for Authors

Review paper fire-4150842

 

This study focuses on the analysis of one large fire in eastern Spain. It applies a methodology to estimate wildfire emissions based on the Seiler and Crutzen equation, even though the equation was not included in the manuscript. They applied the approach developed by De Santis et al. (2010), which assigned a burning efficiency value to each vegetation class and severity level. The study only provides emissions estimates for that one large fire. It doesn’t offer any methodological change that could be considered novel. Considerable gaps have been found in the literature review, and the methods are not thoroughly described, which makes the results section confusing. This paper has significant room for improvement. I have included my comments and suggestions below.

 

Major comments:

 

The introduction has many subsections, some with only one paragraph. I suggest removing the subsections and presenting the introduction as a continuous section. In addition, the review of previous methods for estimating biomass loss and emissions is truly lacking many references and a proper description of wildfire emission estimation methods. Without that description, the research needs are not justified. If the authors don’t explain the characteristics of previous products or methodologies, they can not argue that they are covering a research gap. 

 

Materials and methods:

• The materials and methods section is missing the Data subsection, where the input data used in the analysis should be described. That is a very important information gap in this manuscript. Later in this section, the authors discuss limitations and mention the SIOSE land cover map for the first time in the manuscript. The authors should describe the data used properly.
• Why is the description of the biomass estimation in an Annex? It is a crucial part of the methodology. The authors should include a map of the biomass spatial distribution.
• The authors explained they chose to compute composites to estimate fire severity, but the composite rules and criteria are not specified in the manuscript. Why do you use images from the previous year to create the pre-fire image? What would have happened if there had been a fire in the area the previous year? Pre-fire vegetation conditions can vary significantly from one year to another. In my opinion, including images from the previous year would introduce unnecessary variability to the composite data. You should read Quintero et al. (2025), as they explain the use of composites for fire severity mapping very well.
• In addition, the authors stated that using composites improves reliability, but they did not offer any analysis to prove the reliability of the approach. 
• Despite the limitations, the authors stated that their approach provides a robust framework, but in my opinion, it has some gaps. Besides, it does not offer anything new from the work presented by De Santis et al (2010).
• Table 3. The values contained in the table are not biomass consumption rates; they are burning efficiency or combustion completeness coefficients.

 

Results

• Table 7. The caption should include which variable it refers to. In addition, I suggest you present these results as a boxplot. Why are a range of values within the severity classes of the vegetation classes if you used an average biomass value per vegetation class and a burning efficiency coefficient per severity and vegetation class? It is not clear to me how you are computing those values.
• Table 8, What is the meaning of this table? And what is the meaning of the acronyms? These results can not be understood because the methodology is not clearly described.
• I don’t understand how the authors transition from showing results of mixed vegetation classes to showing results of general vegetation classes. It is very confusing, and it happens the same in the discussion section.

 

Discussion:

• The statement about the scalability of the methodology is repeated in this section. However, the methodology relies on biomass estimates calculated at the regional level from airborne data, which are not available for all European countries, making the approach inapplicable to other regions. 
• In addition, the approach was not described for the first time by the authors of this paper, as this is the same approach used by De Santis et al (2010) in California, which was also replicated by Oliva et al (2020) in Chile.
• To show the reliability of the approach, the results should have been compared with other global wildfire emission products that are freely available, which would be a way to showcase the need for higher spatial resolution.

 

 

Minor comments:

Line 39, citation number 9, is not the most appropriate as it is outdated and it doesn't represent the "recent decades" as stated at the beginning of the sentence. A more recent citation should be used.

Line 45, citation 14. It is not appropriate for a research article to use that website as a reference. The authors should refer to the studies mentioned on the website.

Line 60, citation 17. It is not appropriate to use the WWF report as a reference here, as it is not a peer-reviewed scientific document. There are several research papers that focus on Extreme Wildfires events, the authors should use those as a reference.

 

Figure 2 could show the land cover classification within the burned area. The image resolution should be improved, and the size of the coordinate numbers should be increased because they are not legible.

 

Line 146. The citation used is not appropriate, as the EFFIS recommendation refers to Key and Benson’s work. The proper reference should be Key and Benson.

 

Line 306. It should read “the highest per-hectare CO2eq emissions”. 

Line 330. emission rate or emission intensity?

Author Response

We sincerely thank Reviewer 4 for the thorough and constructive evaluation of our manuscript. The reviewer’s comments were extremely valuable in identifying sections that required clearer methodological explanations, a stronger literature context, and improved presentation of the results.

Following these suggestions, we have substantially revised the manuscript. The main improvements include:

• Inclusion of the Seiler and Crutzen reference to clarify that the study builds upon this conceptual framework.

• Expansion of the literature review, including existing wildfire emission products and alternative methodological approaches.

• Addition of a Data subsection in the Materials and Methods section describing all datasets used, including the SIOSE land cover database, its vegetation detail, and spatial resolution.

• Integration of the biomass estimation methodology into the main Methods section instead of the Annex.

• Clarification of the Sentinel-2 compositing procedure, including criteria for pre-fire and post-fire composites and references to recent methodological studies (e.g., Quintero et al., 2025).

• Correction of terminology related to burning efficiency / combustion completeness coefficients.

• Recalculation and clarification of Table 7, now using consistent mean biomass values per hectare.

• Improvements in tables, figures, captions, and acronyms to enhance readability and interpretability of results.

• Expansion of the uncertainty and discussion sections, including clarification of mixed vegetation class contributions and methodological limitations.

We believe these revisions significantly improve the clarity, transparency, and scientific robustness of the manuscript. Our detailed responses to each comment are provided below.

Major Comments

Comment 1 – Introduction:
The introduction has many subsections with only one paragraph; the literature review lacks many references and does not describe wildfire emission estimation methods.

Authors’ response:
We thank the reviewer for this suggestion. The Introduction has been restructured into a continuous section to improve readability.
The literature review has been expanded to include:

• The Seiler and Crutzen emission equation, explicitly framing the study’s conceptual basis.

• A description of existing wildfire emission products and methodologies.

These additions provide a clearer overview of the current state of the field and better justify the research motivation.

Comment 2 – Data subsection missing:
The Materials and Methods section lacks a description of the input datasets.

Authors’ response:
We agree with the reviewer. A new Data subsection has been added, describing:

• Sentinel-2 imagery used for fire severity mapping;

• The SIOSE land cover database, including spatial resolution and vegetation classification detail;

• The wildfire delimitation dataset.

These additions clarify data sources and improve methodological transparency.

Comment 3 – Biomass estimation in Annex:
Why is biomass estimation in the Annex, and a map of spatial distribution is requested.

Authors’ response:
We fully agree that biomass estimation is a fundamental part of the methodology. This section has been moved to the main Methods section, and a new figure illustrating the spatial distribution of biomass has been added to highlight variations across vegetation types.

Comment 4 – Sentinel-2 compositing procedure unclear:
The criteria and use of previous-year images for pre-fire composites are not explained.

Authors’ response:
We are very thankful for this comment. The recommendation to read Quintero et al (2025) was very helpful to describe the approach.The compositing methodology has been detailed in the Methods section following Quintero et al. (2025). The revised text specifies:

• The compositing method (mean-value composite);

• Temporal windows for pre-fire and post-fire imagery;

• Criteria for image selection;

• Rationale for including prior-year images to account for cloud cover and vegetation phenology variability (Parks et al., 2018).

Comment 5 – Reliability of composites:
No analysis was provided to support claims of improved reliability.

Authors’ response:
We appreciate this observation. The Discussion section now explicitly evaluates the limitations of the compositing approach. A formal claim of “reliability” has been removed, and sources of uncertainty are clearly stated. Exploratory analyses include a -10% perturbation and extreme fire severity scenarios to illustrate potential variability.

Comment 6 – Novelty relative to De Santis et al. (2010):
The methodology does not appear novel.

Authors’ response:
While the conceptual framework follows De Santis et al. (2010), our study adds value by:

• Integrating high-resolution SIOSE land cover information to represent mixed vegetation classes;

• Combining these with biomass estimates for detailed vegetation-specific distributions;

• Applying this methodology in a heterogeneous Mediterranean landscape, where mosaic vegetation is critical.

Additionally, model robustness was assessed through comparisons with global products, a sensitivity analysis and extreme theoric severity scenarios.

Comment 7 – Table 3 terminology:
Values in Table 3 are burning efficiency, not biomass consumption rates.

Authors’ response:
Terminology has been corrected. Table 3 (now Table 5) is updated to:

“BEadj fractions (0–1) by burn severity class.”

Comment 8 – Table 7 caption and calculation:
Clarify the variable and consider using a boxplot; explain within-class variability.

Authors’ response:
Table 7 has been recalculated using consistent mean biomass values per hectare, and the caption has been updated to specify the variable. The calculation now applies biomass proportionally to the area covered by each vegetation type. Total emissions are the primary focus, and per-hectare values are reported in Appendix 1 (former Table 6). Presentation has been revised for clarity. 

Comment 9 – Table 8 clarity:
Acronyms and transition from mixed to general vegetation classes are confusing.

Authors’ response:
All acronyms are now clearly defined (Table 2), and the methodology description has been expanded. Results presentation has been revised to improve interpretability, particularly regarding mixed vegetation classes.

Comment 11 – Scalability limitation:

The statement about the scalability of the methodology is repeated in this section. However, the methodology relies on biomass estimates calculated at the regional level from airborne data, which are not available for all European countries, making the approach inapplicable to other regions. 

Authors’ response:
We acknowledge that high-resolution biomass datasets are not universally available. The Discussion section now clarifies that while the methodology is conceptually transferable, practical application requires detailed biomass data with cover fraction for each vegetation type.

Comment 12 – Prior work:
Methodology was previously described by De Santis et al. (2010) and Oliva et al. (2020).

Authors’ response:
Now the literature sources are clearly stated to contextualize the framework.

Comment 13 – Comparison with global products:
Results should be compared with other wildfire emission datasets.

Authors’ response:
We have now included comparisons with GFAS and GFED in the Methods, Results, and Discussion sections. The approach highlights the advantages of high-resolution biomass-based estimates and quantifies differences in total CO₂ emissions relative to global products.

Minor Comments

All minor comments have been addressed:

• Line 39: updated reference;

• Line 45: replaced with peer-reviewed sources;

• Line 60: replaced with scientific literature;

• Line 146: citation corrected;

• Lines 306 and 330: results updated to reflect recalculations.