The authors study the fire-precipitation interactions and their feedback on the fires variability in Mexico and central America. The work is interesting suitable for the publication in EGU sphere. I recommended a major revision after addressing following comments.

• The present study period is 2003-2019. According to my knowledge, the GFED4s data is available during 1997-2022. QFED is also available during 2000-2022. I wonder if the authors could extend the analysis using the full dataset available in both GFED4s and QFED. If the authors do not plan to use the full dataset, the authors need to explain the reason why you choose to focus on the period during 2003-2019.
• The authors used the LAI as the surrogate of the fuel load and concluded that there is insignificant relationship between fuel load (inferred from the LAI) and fire consumption. I am concerned about this proxy choice since the LAI is a dimensionless quantity that characterize the canopy coverage and not biomass available to burn. I wonder if the authors could try some other metrics such as primary productivity dataset (e.g., MODIS GPP or NPP) to see if the similar results could be reached.
• Precipitation might be one reason to explain this biennial variability. I wonder if the authors get a chance to look at the vapor pressure deficit or relative humidity interannual variation. How are they related to the fire emission variations in Mexico and Central America?
• Please carefully check the figure caption. For example, Fig. 8 figure caption is confusing. The panel labels are labeled from a) to f). But only label a) and b) are mentioned in the figure caption.
• Some panel labels are missing in the figures (e.g., Figure 5 missing label a) and b)).

Overall, this work presents a compelling study of the quasi-biennial variability of fire characteristics over southern Mexico and Central America (SMCA) and demonstrates the role of fire-precipitation interactions at both interannual and subseasonal timescales in shaping the observed patterns. The manuscipt is well-written and the main results are clearly highlighted throughout the text. All the figures are appropriately labeled and capitoned; it's evident that the authors have devoted significant effort to effectively communicating their results. 

I'm flagging this manuscript as a major revision because there are a couple of important areas (see major comments below) that deserve a more careful examination. However, once these are addressed, I'll be happy to review the revised manuscript's suitability for publication.

Major comments:

Although in L200-201 the authors acknowledge that "fuel availability might play a role in the interannual variation of fires," they do not explore this further since LAI, their surrogate fore fuel load, does not show any correlation with fire characteristics. In the model world, this would be fine. However, in general, it has been conclusively demonstrated (see https://journals.ametsoc.org/view/journals/bams/84/5/bams-84-5-595.xml and https://www.publish.csiro.au/wf/WF19087) that, for arid and semi-arid regions, antecedent precipitation in 1-2 years prior to a major fire season shows significant correlations with burned area by promoting the growth of highly flammable fine fuels. Ignoring the effect of precipitation variability on fuel availability might artificially enhance the estimated amplification of the quasibiennial cycle by the short timescale feedback. I recommend at least discussing this potential source of bias.

Moreover, besides good physiological reasons for not using LAI, specifically that it does not account for variability of vegetation density on the surface, the AVHRR record used in the analysis also suffers from measurement issues due to orbital change and sensor degradation (see Section 3.4.2 in https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2018RG000608).I recommend that the authors explore other remote-sensed predictors such as: 

-- maximum GPP instead as illustrated in this paper https://iopscience.iop.org/article/10.1088/1748-9326/ac8be4 or, 

-- fractional land cover as used in https://gmd.copernicus.org/articles/16/3407/2023/gmd-16-3407-2023.html ; fractional land cover for the SMCA study region may be found here: https://2018mexicolandcover10m.esa.int 

The authors should also emphasize the fact that meteorological conditions, such as mean temperature and precipitation during a fire season, are a clear confounder to any estimated amplification from the short timescale feedback. That is, although they clearly illustrate using model experiments that there is a fire enhancing precipitation pattern contrast between strong and weak fire years (Fig. 9c), it's not quite clear what the magnitude of this effect is relative to the average difference in expected burned area due to meteorological variability between these years. One potential way to explore the magnitude of the short term feedback could be through artificially suppressing aerosol-radiation interactions (as in Huang et. al. 2023) and comparing burned areas among similar strong fire years. A careful analysis of this point in the Results or Discussion section would suffice if running new model experiments is cumbersome.

Minor comments:

L78: Change "characteristic" to characteristics

L79-80: Use present tense to maintain consistency throughout the paragraph; specifically change "explored" and "provided" to explore and provide respectively

L90: Is "fire consumption" correct here? It might be fuel consumption or fire-consumed fuel instead. Please verify.

L103: Change "previous" to prior

L135-136: What is the specific way in which the fire inventory is modified but anthropogenic emissions are kept unchanged? Consider adding a 1-2 sentence clarification

L162: Clarify if the difference in fire-consumed dry matter mentioned here is for one particular year or the whole study period.

L165 and L173: Omit "basically" -- the sentences read fine without it

L202 and L210: Rephrase "fire consumption" as suggested above

Fig. 2: I'm not sure if it's easily feasible with the current analysis set up, but including an additional plot of fire counts in the study region between 2003 and 2019 will help in visually emphasizing fire prone areas

Fig. 2: make the stippling bolder/bigger as it's hard to see the pattern in presence of colors

In Fig. 3 are the temperature and precipitation anomalies calculated with respect to the 2003-2019 mean? Please clarify in the caption

L383: Typo in "quai-"