General comments

This is an interesting and timely study. There has been a recent increase in studies linking health impacts from climate induced disasters and the manuscript touches on a highly relevant topic. However, I do have some questions about the data and the method, on which I have elaborated below.

Detailed comments

Why did the authors decide to use Global Forest Watch instead of, for example, CORINE?

Why did the authors select those four months? Would it have been an option to look at large events (>1000ha burnt) any time of the year? Can the authors explain this choice? Especially in light of air pollution (e.g., from PM) being worse in winter as air gets trapped under cloud cover (see for example Pey et al., 2010).

L.121: it may be helpful for the reader to refer to Fig 2 for a map of the locations of the monitoring stations.

L.121-125: could the authors expand on their method to correct PM10 data from wildfires from “normal” non-wildfire causes? I am a bit unclear as to what the authors mean by “background stations”.

L.133. What is the impact of the uneven distribution of the background stations on the outcomes of the study? Especially since some of the NUTS-regions don’t have any stations, while the urbanized NUTS regions have the highest number of stations but the lowest wildfire risk. Moreover, what did the authors do with the PM10 data in NUTs regions with more than 1 station? Did they average the PM10 values of all stations in one NUTS region?

L.140-142: I am unclear as to why a reduced number of deaths (over the selected time period?) prevented the authors from including a correlation for COPD and asthma. Then the authors would just find a smaller correlation between PM10 population from wildfires and these particular mortality causes? Can the authors elaborate?

L.145-147: if the authors use Pearson, then why explain Poisson as well? I found this a bit confusing. (Later, in section 3, I read that the authors use the Poisson for the RR. I recommend explaining this more carefully in section 2.3.)

L.154. From an earlier sentence (L. 111-113), I got the impression the authors only selected large fires in the time period of June-Sept, but from L.154 it appears that the authors correlated with both types of wildfire sizes (<1000ha and >1000ha). Maybe clarify this both in section 2.2.2 as well as in 2.3.2.

L.189-192: this paragraph seems out of place. I think it would fit better in the introduction, or – alternatively – link it better to the findings from Table 2.

Table 3: what is the difference between the 3 instances of columns headed with “All” and “Months w./LF”? From the text I gather it is the three types of death causes but it would be good to make that clear in the table.

Table 3 mentions 2016 population numbers while figure 2 shows 2011 population numbers. It would be good to match these. Also, were annual (monthly) mortality numbers corrected for annual population numbers? Would it be better to use mortality rates instead?

I think the paper could be restructured a bit by moving the content of section 3.1 up to section 2 (each of the different data sections). In my understanding, the sub-sections of 3.1 present (a discussion of) the input data while the actual analysis (and aim of the manuscript) of the relationship between wildfires and mortality are the core of the analysis and should indeed be presented in the results section.

L.261: how do the authors explain the lack of correlation between wildfires and PM10 for these regions?

L.268: “in some areas”: why only in some and not in others?

How do the findings compare to similar studies? How does the paper add to existing work such as the cited Faustini et al., (2015) paper?

In the conclusions, the authors remark that many other aspects could have influenced their findings, such as aged population, lower socio-economic status, etc. Why did they not correct their input data for this?

Minor comments:

• L13-14: “the 48% of wildfires occurred were large fires”. This doesn’t flow well, there appears to be a word missing.
• L16: “on the future” should be “in the future”
• 23: “an increase on” should be “an increase of”
• 26. I would replace “fustigated” with “hit” or something else. Fustigated is a bit lyrical in this context.
• 85 (and L.90): “is focused in” should be “focuses on”
• 94: the urbanized areas are depicted in red in Fig. 1, but it may be helpful to the reader to add the place names referred to in this sentence.
• 99: the use of “allied” and “verified” is incorrect in this sentence.
• 105-108: check referencing style and sentence structure
• 120: “network” is this a network of measuring points/stations?
• 128: this sentence seems to belong in the previous section on wildfire data?
• I think sentence L.169-170 and L.171-173 can be merged? Also, sentence L.171-173 doesn’t read well.
• 189: check referencing style. When the authors say “area most affected” do they mean globally, or in Europe?
• 194: “unequal spatial” -> “unequally spatially”
• 195 “in the in” -> “in the”
• 203: “represented by black dots” -> dotted or shaded area
• 216: “cause death then” -> causes of deaths than
• 235 – 237: sentence doesn’t read well.
• L 282 – 284: sentence doesn’t read well.

General comments:

The manuscript presents a valuable study of the impacts of wildfires on human health. While other studies of this kind have been more rigorous, the work presented here is for an understudied region, and therefore provides guidance for this region regarding the concern of wildfire smoke on human health. The work is generally OK, but the presentation needs a reasonable amount of revision. Additionally, the authors should work on providing a more complete discussion regarding the limitations and benefits of their study. See below.

Specific comments:

1- The paper needs some attention to the use of English. While not too bad, some attention to editing the text to correct wording, syntax, and general use of English will improve the manuscript and help readers to better understand the study. See below for specific location and suggestions. Some specific edits are required, while others are suggestions.

2-  The figures can be consolidated. Many of the figures are maps of Portugal – all good to have. I suggest the maps be combined for all “input” data (Figs 1 - 3) and a separate figure for results (Figs 4 - 6). (see note below regarding Fig 3) A layout similar to Fig 1 – a 3-up side-by-side – will work well.

3- Abstract – Has too much introduction and info on the “purpose”. Sentences on lines 2-7 can be condensed.

4- Introduction – This can be consolidated. The first part of the Intro through line 34 – should be consolidated into one short paragraph. Throughout the manuscript text, many sentences can be trimmed to make a better presentation. For example, the opening sentence can be changed to:

“Wildfires have a considerable impact on the environment and humans worldwide.” (I clipped out several phrases).”

(Section 2.1 is very good – very helpful to have this description for the purposes of this paper)

5- Datasets – It is unclear if the study is considering only forest fires or all wildland fires. Please clarify. It would be helpful to have a map of fires across the study region – dots on a map or polygons of areas burned, if appropriate. Total burned area – Figure 3 – is not a project result, it is input data. So, section 3.1.1 should be integrated into section 2.2.2. In 3.1.1 you can keep info that is derived from the distribution of fires, but the data itself (shown in Fig 3) is not results.

6- The manuscript lacks a proper “discussion” section. Much of the results and the text starting on line 214 (page 13) should be turned into Discussion. Also, some of the material in the “Conclusions” can be integrated into “Discussion”. The Conclusions should be a summary of the primary findings and relevance, but should not have anything new in it.

7- A Discussion section should be formulated (comment 6) and should include more from the authors reviewing the shortcomings, assumptions, and limitations of the study and the primary outcomes, and how they are relevant. Bring out the fact that this is new information for the region that can help with decision-making. In the limitations, there should be some acknowledgment of the coarse spatial (level III NUTS – why not better) and temporal (Monthly – why not better) scales, and what it means to interpretation of the results. These coarse-scale approaches will “dilute” the results meaning there may be health effects that are not found because wildfire smoke is so “episodic” and “local”. You may not be able to show this here, so say what is needed for a more complete understanding? Why is your study still useful?

Guide the reader to help them understand why this work is “good” or “helpful”. Otherwise, it may be seen as not as good as studies that use finer resolutions, smoke transport models, etc. I think it is worthwhile, but other readers may find it too simplistic. Discuss the value of AQ monitors – why they are helpful and why they are not for assessment of fire exposure.

Detailed comments: (I have many more, but will wait for the revision to re-read)

Line 9; What does “these” refer to?

Line 78: “… the effects of short-term pollutants exposure …” It is unclear if this is short-term exposure or short-term pollutants. Re-vise this sentence to be clear.

Lines 161-164: Can be deleted.

Line 202: Many sentences start with “Regarding …”. This is poor sentence structure. Please revise.

Line 282- 284: It is unclear why this is relevant to your study. Remove.