Impacts of Fire Frequency on Net CO₂ Emissions in the Cerrado Savanna Vegetation

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This paper analyzed burn area in Cerrado's savanna using Mapbiomas fire, improving the GHG emission estimation. It also shows the results of a re-growth simulation under different fire scenarios and this leads to the conclusion of Cerrado's savanna being a source of CO2. I think this paper is well written and scientifically sound.

This paper states the lack of accurate estimates of GHG emissions due to uncertainties in the burned area (line 30), but it seems to take Mapbiomas Fire at face value. I suggest adding a paragraph mentioning (or even better, analyzing) the uncertainties of Mapbiomas Fire (lines 46-47), which is still a relatively new product (collection 1).

I find the Fire Return Index interval index interesting. However, its reclassification intervals seem arbitrary (lines 202-207). Can you elaborate on the number of fires and number of years intervals? Do they match some vegetation or fire property? Perhaps a histogram would help to understand the data distribution and contextualize these intervals.

Author Response

We include 3 reviewer information. Please see the attachment!

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The paper “Impacts of fire regime on net CO2 emissions in the Cerrado savanna vegetation” (Gomes et al.) presents the results of a work developed to quantify emissions of CO2 and other greenhouse gases using 30 meter burned area products over 35 years (1985–2020). Moreover based on long-term field inventories of post fire vegetation regrowth the work simulated under different fire regimes atmospheric carbon removal to estimate net CO2 emissions in the Brazilian Cerrado savanna.

The contribution is interesting and worthy of being published.

 

Comments and remarks:

2. Materials and Methods

Before entering in more specific details such as “Cerrado Rupestre etc. ”, it would be interesting to show a map of the six Brazilian biomes, as defined by the Brazilian Institute of Geography and Statistics (IBGE) and give summary information on the Cerrado biome such as dimensions (area and %), main land cover an land use (Souza, C.M., Jr.; Z. Shimbo, J.; Rosa, M.R. et al., Remote Sens. 2020, 12, 2735. https://doi.org/10.3390/rs12172735).

For the equations presented in the manuscript (line 152; lines 168-169 and line 171), please utilize numbering and put them in the centre of the row, so that they appear immediately to the reader.

In equation and formula (lines 168-169 and line 171) time interval in years is indicated with different symbol ( T and t): you should use the same symbol to indicate the same quantity.

Figure 4: The diagrams in Figure 4 (b–e) concerning with other GHGs as CO, CH4, NOx and N2O are all with same shape throughout the time series as depicted by of CO2 and Burned Areas diagrams.

Removing these 4 diagrams from Figure 4, you don’t loose information which in turn can be gained by adding a Table reporting yearly values of different quantities.

 

Line 276 (Figure 6 (f)) should be (Figure 5 (f))

 

Line 284-285 “In 2021, the area corresponding to the moderate scenario was 4,247,981 Mg of biomass, …”: Mg is not a measure of “area”; please rephrase the sentence.

Author Response

We include 3 reviewer information. Please see the attachment!

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

Please see the attachments.

Comments for author File: Comments.zip

Comments on the Quality of English Language

-

Author Response

We include 3 reviewer information. Please see the attachment!

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

The authors have done a fine job making changes and refining the argument made. The article is worthy of publication with a minor change.

My one concern remains that the authors downplay (or ignore) the limitations of their study which uses only fire return interval as a critical variable in a study of vegetation recovery following fire. The literature on this subject is vast and while much of it emanates from African and Australian savannas, there are certainly studies from the cerrado region that document how fire intensity and seasonality can impact small trees, cause die-back and thus influence tree recovery rates. Hoffman's work is key in this regard as noted previously: Fires of higher intensity (flame length >2 m) caused greater mortality and topkill than fires of lower intensity (flame length <2 m).

This does not take away from the good work done in this paper, however, whose familiar with fire research--the readers of Fire journal (such as this reviewer) will be aware of this critical limitation.  The authors brush this off "due to the lack of specific field studies for Cerrado savannas to calibrate this model."

An improved article would include a limitation section (certainly this is true for any modelling work) that discusses the many limitations of the variables used (see previous review).  I suggest the authors include a 1 paragraph Limitations section that explains the limitations of not including fire intensity as well as the limits of the other variables (the EFs appear to be for savanna means and known to err on the side of grass dominated not woody ones). 

I think the limitation of the study could be mitigated with a simple statement that the data used to build the recovery model includes a variety of different fire intensity and seasonality values and, as such, if averaged over time, fire frequency or return can be isolated. This is not entirely true, of course as longer return intervals do increase fuel and, thus by definition, fire intensity.

Finally, an improved conclusion would suggest future work to improve the model.

Comments for author File: Comments.pdf

Author Response

Thank you for highlighting these relevant points. Please see the changes in the attached file.

Author Response File: Author Response.pdf