Sentinel-2 Reference Fire Perimeters for the Assessment of Burned Area Products over Latin America and the Caribbean for the Year 2019

Round 1

Reviewer 1 Report

General Comment:

The manuscript entitled “Sentinel-2 reference fire perimeters for the assessment of Burned Area products over Latin America and Caribbean for the year 2019” aims to describe and present the methodology used for the creation of the product “RP_LAC_2019_S2”, a database for the year 2019 containing the fire perimeters obtained using the images of Sentinel-2 imagery in Latin America and the Caribbean. The manuscript is overall well written, but I find it a little bit too long, particularly in the Results and Discussion section, where too many tables were added. For this reason, I would suggest adding some of them to Supplementary Information. The Introduction section provides an exhaustive overview of the state of the art and is, in my opinion, satisfactory. The methodology combines a number of previous approaches and data and presents an exhaustive description of the phases of the work.

The approach proposed is overall complete and replicable and can contribute to further similar studies aiming at mapping fire perimeters using satellite imagery in every biome. The results are clear, the discussion section is appropriate and compares the results and methodology of this study with previous similar works. I would recommend the publication of this work, after a few revisions. Specific comments are described below.

Specific Comments:

L11-please check the phrase

Fig.1, Fig.2, and Fig.3 please consider replacing the schemes of the methodology with more schematic-diagram flows with few words.

L188 and L195 please add the number of references

L202 please consider describing here the choice of using two criteria of biome and fire activity

Table 6. Please specify the year of the PS period for sampling, if there is no space inside the table, please refer to 2019 in the caption.

Table 8. Please specify the format of values (percentage) also inside the table, not just in the caption

In the discussion section, most of the comparisons are made by referring to the work of Stroppiana et al 2022. I would suggest the authors consider other similar works and discuss the agreement and disagreement between their results with them.

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 2 Report

The research “Sentinel-2 reference fire perimeters for the assessment of Burned Area products over Latin America and Caribbean for the year 2019”, presents a database of burn scars for use as validation data for burned area products using remote sensing.

The methodology is rich in detail, covering the various steps for creating the fire perimeter using Sentinel images and the reference base using Planet images. Sample selection follows the parameters indicated in the literature from Boschetti et al and other authors.

Below I provide some minor points for clarification or suggestions.

The Dice Coefficient values ​​are too low for a product to be used as validation. Was manual refinement performed on all scenes? Was the comparison of Planet with Sentinel carried out on close dates, avoiding scars from different dates?

L40-41: Detail more about the importance of burned area products for environmental management.

L292: Standardize citation for consultation in references

Figure 4. I suggest putting a small note in the caption about the discarded units that were left blank.

I suggest including in the discussions the advantages of using this product for sustainable fire management.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

Title: Sentinel-2 reference fire perimeters for the assessment of Burned Area products over Latin America and Caribbean for the year 2019

Overview: This paper follows similar methods to another study conducted over Sub Saharan Africa to map burned areas over South America, Central America, and Mexico for 2019. It is a very interesting study, but the methods are not clearly explained and the heavy use of abbreviations makes the writing even more difficult to follow (even by someone familiar with most of the shorthand already!)

Other comments:

Some uncommon word choices like “acquirement” (line 39) “stablished” (line 51), and “stablishes” (line 55) 

Line 56 I would use “image pairs” rather than “pairs images”

Line 187 Why are short temporal sampling units prone to more errors? 

Line 263-264 I would use “image interpretation” rather than “photo interpretation” since the reference images are not photos.

Were training data split so that some were used for testing/accuracy assessment? - after reading further it seems that PlanetScope and Sentinel based burned areas were compared to determine accuracy- is this correct? 

The use of “long unit” and “short unit” (e.g., line 229) is not clearly explained as meaning time intervals. Are these widely used terms rather than long internal and short interval? The distinction and purpose behind the two is also not clear. 

Line 322- Do you mean “extent” rather than “extension”?

There are a lot of acronyms and abbreviations used in the paper- and even though I’m familiar with almost all of them already, I still found myself having to double check some of them. For example, PlanetScope and Sentinel are not long words to begin with.

It’s also confusing that there is “RPS2” (which I’m not sure was even defined) and “S2RP” used in the paper.

Section 2.4.4 Again too many abbreviations.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 4 Report

A team created reference fire perimeters for Latin America and the Caribbean (LAC) in 2019 using Sentinel-2 (S2) imagery. Overall, the passage effectively describes the creation and application of reference fire perimeters.  With the suggested additions, it can be even more informative and valuable to users. Here is my comments to improve the manuscript: 

How does the data scarcity mentioned in the passage affect the generalizability of the accuracy metrics obtained from the PS image comparison?

1. Are there plans to collect more data in the future to improve the generalizability of the findings?

2. While the passage mentions efforts to maximize temporal coverage, can you quantify the average achieved temporal length for the reference perimeters?

3. How does this compare to the temporal coverage achieved in studies like Franquesa et al. (2022) mentioned in the table?

4. The passage highlights successful validation of FireCCI51 and MCD64A1. Can you elaborate on whether S2RP is suitable for validating other types of BA products besides these two examples?

5. The passage suggests a link between temporal length and validation accuracy. Are there plans for further research to explore this relationship in more detail?

The quality of the English language in the manuscript needs improvement, as I identified and corrected several grammar errors.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 4 Report

The author addressed all of my comments.